Donovan Posted January 14, 2009 Report Share Posted January 14, 2009 Hi Pete and MRJ. I have spent the last week or so building drivers and an amp circuit, testing, and plundering on. I do have a few questions which I am hoping you might clear up, but first let me bring you up to date... I have one driver built from old strat pickup> I removed old windings, blocked up the lower portion of the bobbin with epoxied on fiberglass strips. This left approximately 3mm at the top of the bobbin for the driver. I hand wound using only a mechanical tensioner. I had the tension set quite high which I now think was a mistake. Since, I have another driver with windings under moderate tension in the potting curing stage as we speak (or I write). The resistance of the original was spot on 8 ohms and very solid as I used urethane varnish to pot throughout the winding process and followed with 32 in of vacuum to degas. The amp circuit is functional. I have opted not to use the fetzer valve, but rather built a preamp based on the TL082 op amp, with variable gain from 2X to roughly 75X. The amp circuit is the "modded" LM386 Ruby found on page 280 of this thread. I attempted to optimize the amp for max clean output in the lab at my employer. We sent 1kHz sine wave at roughly 100mV through the input and proceeded to "tune" the pots from preamp gain, then to volume, then to amp gain. All looked good on the scope. The circuit was processing 1 kHz clean up to about 1.2W. However, there was about a 1/4 wave phase shift in the circuit, which is something I did not expect. Then, in real life application, the amp is very touchy, picking up a motorboat sounding low-mid frequency hum if gain is set too high. For this reason I am investigating other preamp/amp chip combinations, all looking for totals under 2W. I am also building a signal bypass/passthrough box for simplicity and ease of hookup for testing purposes with capability of switching phase at different points throught the cicuits to see just what the effects will be. I am building this box very modularly to facilitate ease of swapping out different preamp and amp circuits as hand slicing and even gator clips are a royal PIA. So actual sustain... I get it on the low E, the A, and the D, with a hint on the G if I really try, but nothing on B or E strings. This is all with the driver handheld at the 12th fret. The sustain is not exactly the fundamental or a nice harmonic. If I strum an open 6th string E, the driver fights with the string and gradually takes over with a harmonic that simulates the 7th fret on that string. The oscillation during the "takeover phase" is not all that pleasing. If you have ever tuned using the beginner 5th-fret-to-next-open method, the oscillation reminds me of this... as you get close to in tune, it disappears, but nothing at all like the nice sound bytes I have listen to which were posted via this thread. I am guessing this may be the embodiment of the 1/4 wave phase shift I mentioned earlier? Next step is to build a better circuit to try with the new driver, which will be ready tomorrow. Question 1: Is my 1/4 wave phase shift normal? I think it is more of a time constant xxx ms delay than a true phase shift and I am wondering... if I can not get rid of it, can I add a simple phase delay or phase shifter circuit with a knob control to "tune" such a delay out of the equation. Question 2: What if any are the consequences (other than EMI) of taking signal from the output jack of the guitar with pickup selector at bridge only, rather than the pickup leads themselves? I'd rather not mod my guitar (even adding leads) if I do not have to! My test box would take signal from guitar output, split to bypass jack (for out to Fender amp) and to sustainer circuit, then feed driver which is hand held until I get more positive results. Thoughts or suggestions? One comment I have from the beginner's perspective... not knowing if your driver or amp circuit is good is troublesome as there is no way to eliminate a variable! Link to comment Share on other sites More sharing options...
psw Posted January 14, 2009 Author Report Share Posted January 14, 2009 It's tricky to know exactly what is going on from my perspective as a lot of the conclusions are suppositions...I don't have the training or the equipment to test for them...so these are interesting results. However...you don't want anything like 2 watts (1 or less) and there is an important factor that may not be able to test for. Firstly...are you using all the caps and stuff that I added to the LM386 to stabilize it from internal oscillations? Also...are you using the suggested 100uF output cap? Anyway...you may be way overpowering or oscillating in the preamp circuit...that seems like a good basis for a circuit though...just may be putting too much power in. Even though it is giving a 1/4 wave phase shift...phase shift alters with frequency also...the legit sustainers all have phase compensation circuitry that auto adjusts with frequency. Obviously my appraoch is simpler and where the driver design comes in. Even though your circuit is showing this kind of discrepency, the phase of the vibrationof the string in the location of the driver is kind of unknown or at least variable if you think about it. ANd...the phase differences once it has been through the driver may have other phase effects that you might not be taking into account or measuring. Generally the idea is to get a reasonably close phase relationship and a driver that will operate at sufficient "speed" to keep up with the faster moving and less magnetic high strings...although simple, the wire gauge and thin coil design is where these phase differences are made to a close enough fit. As for winding by machine or under tension...I know I have said to hand wind them...hmmm. SOmetimes oscillation and weird effects can be caused if there is any internal vibration...sometimes you can feel or even hear this vibration in the driver or heat maybe generated...all lack of efficiency, but also creating unwanted resonances in the coil that can transmit into weird harmonic effects. Everything gets worse the more power you put into it...increasing power tends to accentuate unwanted effects, not the sustain effect. But it sounds like it is coming along...maybe some details of the circuit and pics of the driver might help a little to give advice to calm it down... hopefully, I too will calm down a little...have lost a lot of my patience lately...tough times I'm afraid can not always be contained... hope that helps for now...way past bed time at 1am here...so...i'm off! pete Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 14, 2009 Report Share Posted January 14, 2009 (edited) Just a thought...has anyone constructed a one string 'plank guitar' as a test bed? I'm thinking here a plank of wood with a hole drilled at one end to thread say a high E string through from underneath - the string then goes over a pseudo bridge (a small piece of wood lying on the aforementioned plank!) - at the other end of the plank -equi-distant to a typical guitar neck length - would be another piece of wood similar to the one used in the pseudo-bridge.....& then into a guitar tuner (machine head). An old pickup as the...erhmm - pickup...to feed the driver circuit Voila the makings of a guitar! The benefit here being... 1 you can get easy access under the string (to place the driver - asure I know you can hold the driver from above...but not steady & you have to remove the neck pickup first etc)) 2. Being able to quickly hold the driver securely in place ...at the optimum/preferred height from the string (nobody minds screwing test drivers into a plank of wood!) 3. The pickup electrics are stripped to a bare minimum - simplifies things. 4. No damage to your own guitar (I know from experience how testing stuff on your guitar ...eg hand held driver over the guitar can end in marks, scratches etc. Of course this only works if you have a spare pickup & tuner lying around (I do)...but thought it worth mentioning anyway, as I'm sure the amount of in & out at the R& D stage would get a little tiring on a restricted access normal guitar. Likewise for the circuit...I'd imagine a breadboard has got to be the way to go for fast prototyping of the circuit (http://www.tinyurl.com/86aezg - components can get whipped in & out very quickly) & how about this for brilliance - http://uk.youtube.com/watch?v=ZfNGYTT1IM0 (yes, I've read the "best wound by hand" comments...but I'm pretty confident that Fernandes don't!) Edited January 14, 2009 by Hank McSpank Link to comment Share on other sites More sharing options...
Donovan Posted January 14, 2009 Report Share Posted January 14, 2009 (edited) It's tricky to know exactly what is going on from my perspective as a lot of the conclusions are suppositions...I don't have the training or the equipment to test for them...so these are interesting results. Yes, exactly. However...you don't want anything like 2 watts (1 or less) and there is an important factor that may not be able to test for. Firstly...are you using all the caps and stuff that I added to the LM386 to stabilize it from internal oscillations? Also...are you using the suggested 100uF output cap? I am not planning to use all 2 Watts, just looking to maximize the cleanliness at 1W output through more headroom. Anyway...you may be way overpowering or oscillating in the preamp circuit...that seems like a good basis for a circuit though...just may be putting too much power in. I understand. I am guessing to further make things difficult, there is no direct correlation between what will make an audible 8 ohm speaker show oscillation and what will make an 8 ohm driver oscillate. Ughhh. Even though it is giving a 1/4 wave phase shift...phase shift alters with frequency also...the legit sustainers all have phase compensation circuitry that auto adjusts with frequency. Obviously my appraoch is simpler and where the driver design comes in. Even though your circuit is showing this kind of discrepency, the phase of the vibrationof the string in the location of the driver is kind of unknown or at least variable if you think about it. ANd...the phase differences once it has been through the driver may have other phase effects that you might not be taking into account or measuring. Once again the difficulties of having no natural human sense to pick up on. With the sustain your system is getting now, if you do have a similar phase shift in your circuit, I would imagine the performance would increase somewhat if the shift were eliminated. I am thinking about an op amp voltage follower followed by a push-pull transistor circuit to simplify this even more. As for winding by machine or under tension...I know I have said to hand wind them...hmmm. SOmetimes oscillation and weird effects can be caused if there is any internal vibration...sometimes you can feel or even hear this vibration in the driver or heat maybe generated...all lack of efficiency, but also creating unwanted resonances in the coil that can transmit into weird harmonic effects. I did not wind by machine. I wound by hand and the traverse was all but completely random. I auto-tensioned the wire as to not have it slack. I must be mistakn, but thought I had read somewhere in here where you (at least I thought it was you) said something to the effect of... it takes a few tries to get the tension right, so I assumed there was some tensioner being employed. Everything gets worse the more power you put into it...increasing power tends to accentuate unwanted effects, not the sustain effect. I will have another go with the same circuit, but with the preamp at unity gain. But it sounds like it is coming along...maybe some details of the circuit and pics of the driver might help a little to give advice to calm it down... Sure, the amp is the modded LM386 exacxtly as is on page 280 of this thread, as posted by MRJ. However, the preamp is not there. I have clipped off the fetzer portion and left only everything from the volume control and over to the right. The preamp (left) portion of the circuit has been replaced with this. The only difference is I have added a potentiometer and a resistor in place of the 100k and 1k resistor negative feedback loop to bring the gain down to an adjustable 2 to 70' sometyhing' ish. To make this a "unity gain" voltage follower, I will try something entirely different, a simple something like this to feed the LM386. Then, if the output is too low, I can add gain resistors back into the feedback loop. hopefully, I too will calm down a little...have lost a lot of my patience lately...tough times I'm afraid can not always be contained... I've been there and wish you the best. You have taken and continue to take on a lot with this project... answering the same questions over and over... must get very annoying. Maybe something to mitigate this is creation of a very organized thread locked to all but you and maybe an "assistant" or something, where current best practices, circuits, driver design, etc. can be in a "read-only" format and updated as progress occurs. One last thing... you didn't answer about the temporary test box setup, more specifically I am wondering if it is futile because there is some key something I am overlooking. Does the setup I described sound OK? To recap it is: I am also building a signal bypass/passthrough box for simplicity and ease of hookup for testing purposes with capability of switching phase at different points throught the cicuits to see just what the effects will be. I am building this box very modularly to facilitate ease of swapping out different preamp and amp circuits as hand slicing and even gator clips are a royal PIA. What if any are the consequences (other than EMI) of taking signal from the output jack of the guitar with pickup selector at bridge only, rather than the pickup leads themselves? I'd rather not mod my guitar (even adding leads) if I do not have to! My test box would take signal from guitar output, split to bypass jack (for out to Fender amp) and to sustainer circuit, then feed driver which is hand held until I get more positive results. Edited January 14, 2009 by Donovan Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 15, 2009 Report Share Posted January 15, 2009 (edited) Everyone talks of the Fetzer Ruby & how common it is on the net...therefore I googled it & found the basic schematic. looking at the board layout on p280 (which has the requisite mods?), I reckon the modified FR schematic might look like this? Have I missed anything (Harmonic switch aside!) or made any errors? re the wire...I'm hearing 0.2mm is paramount, but by the looks of it here in the UK at least, it's also rare. Best I can find is 0.19mm ...does 0.01mm matter that much? re going the opamp route...if stability is paramount, how about a dedicated single supply opamp (eg http://focus.ti.com/lit/ds/symlink/opa340.pdf) with a regulated VCC @5V? Sure, you're not saving much in the way of components, but biasing an opamp with a resistor network is a bit rough & ready...therefore you could dispense with that aspect & use the component space savings for a simple VCC regulator? As the battery life wanes, you'll still have a solid 5V ...this ought to be better where constant dependable results are the end game. Edited April 10, 2009 by Hank McSpank Link to comment Share on other sites More sharing options...
psw Posted January 15, 2009 Author Report Share Posted January 15, 2009 no, no...my surlyness is related to real life...we all know and have discussed at length how this forum needs to be more organized...I know I even tried to index it at one stage...and I know others volunteered. There has even been talk of starting an actual forum for the thing with different sections and such...and secretly, behind the scenes, there have been occasions where I have been asked to close it and start elsewhere...it only just came to mind. Anyway...interesting stuff... I kind of liked the idea of the opamp preamp myself and had done some like this. The early sustainers I had used a two transistor push pull preamp with filtering of frequencies outside of the audio range along with the data sheet type circuit for the LM386 and worked ok...but larger most likely. Strib had a really interesting opamp circuit with an alternative to the LM386 as well. The 386 is providing the power and I think that aiming for more headroom...which is a good idea...is more in the poweramp than the preamp stage. More power in the preamp only serves to distort the input of the LM386. All in all though, a lot of the headroom is restricted by the power...a 9 volt battery can only supply so much power and therefore headroom... Anyway...lots of discussion about this stuff. One interesting function is a thread view or something (can't see it while posting) that will give you a kind of one sentence overview of the thread...may make it easier to see where the eras kind of fall amongst the many pages. ... ok...so... looking at your circuit with the TL082 (might be worth exploring some of the other family like the more usual TL072 or others that have different specs but are interchangable ( a benefit of using an IC socket). I tried using plug boards and such, but never had much success...I tend to solder things up on an expaned boad so I can add components and shrink it all later...I know col did work this way though. The connections are never as good as solder joins and the wires and long connections in the board always seem to attract noise. There is a big problem with things like alligator clips though...the metal can generate a fair amount of EMI and also...attract a fair amount of RFI and other noise. Anyway...what I did notice is that your preamp has a 10uF cap between it and the LM386 which is pretty large...this may be causing the large phase difference and something like the little 22nF in the ruby might be more appropriate...I don't think my own circuit has a decoupling cap in it at all...but then that's different again. Speaking of which...check out the floyd patent for a bit of a look at the way phase correction circuitry may be organized...a series of different caps are switched in depending on frequencies to create the necessary "lead and lag" differences...page 7 from memory has a bunch of links there. As for the hand winding...well, I know that the commercial units are wound by machine and such...but then they have compensation circuits and things that I have tried to do away in my design. The problem with machine winding is that with DIY and so small a coil, I find that winding by hand gives a lot more control and actually takes less time. I do have a machine and do use it for my epoxy stuff...but that's a bit different again. A big difference in winding such a long coil is that the ends are always tight and often stretch, and the sides loose...a machine needs to be pretty sophisticated in the tensioning to provide variable tension between the sides and the ends...invariably you will need to push the sides in and maybe clamp them as in the pictorial. Plus...winding with glue can be pretty messy with a machine...spray you know! One last thing... you didn't answer about the temporary test box setup, more specifically I am wondering if it is futile because there is some key something I am overlooking. Does the setup I described sound OK? I must have missed that...well, it kind of worked...definitely not a performance thing...there is also no way of disconnecting the common ground to selector and neck pickup that is vital...it gives a bit of an indication though for experiments such as sticking a driver to an HB in my LP. I did also have a one string guitar for a bit...but not the best platform for testing...as you have found, getting one string going does not mean they will all work...plus, it is only hearing one signal, not a chord...so it has the advantages that the ebow has and may give some false results...I abandoned it for sacrificial guitars. I needed a new one last year and have a strat here that has a bathtub route...I removed everything and made a little control plate for components and use doublesided tape and spaces to mount any pickup as I choose...and all can be reached without detuning or anything. I use this for replicating systems intended for other guitars and experiments like the sustainer. OK...hmmm...I seem to be answering both donovan and hank...sorry, got to run out...so typing at speed... The circuit looks ok...I have used a 10uF from pin 7 to ground (+ towards the chip)...can't really speak to the fetzer part but it looks ok. The 100k volume control is often a 10k pot. I did do some experiments with single supply opamps but generally they are not suitable for audio...i even have a 340 I'm sure. I found on further research it wasn't really necessary...I did look into regulators too...this could be a good idea. A voltage divider is likely to be just as good...however I would suggest 1M-2M resistors at least there to stop loading...not sure if I can identify the values for the TL082 circuit...there are simpler circuits than that I suspect. 0.19mm should be fine...unless it is really 0.15 with a really thick coating...then not so good. There has been a consistent attitude that this project is simply an amplifier with any old coil...this is not the case...it really is simple, but a very specific coil and a balancing act. Good for you for perservering looking for it...it used to be as common here till very recently as just walking into an electronics shop and buying a small reall for $8 that would make several coils...alas, no more for some reason...but you can get other gauges...go figure! OK...so, got to run...ideally, a complete elimination of the phase shift would be idea...unfortunately not possible in the real world. Potentially, but with some sophistication, with a single string signal. Consider the ebow...this only needs to tackle one string input and exerts all it's energy in a concise point so needs less power. However, it also has the ability to move along the string length and get the perfect phase relation by positioning. There is no one universal phase correction, everything is a compromise...so even if you have correction circuitry which is kind of tricky...how does it handle a chord? Clearly...there are a lot of frequencies...similarly with harmonics, different string gauges, the type and position of the pickup in relation to the driver, the scale length and the length of string from the fretted note to these parts. My approach was to try and dispense with this by making a driver sufficiently fast enough (ie little phase delay) to work on the average guitar frequencies and with as simple a circuit as possible. This was the influence of the late Lovekraft who supported this thread in the very early days and encouraged me to stop frtteing about how others have done it and the patent mojo stuff...and try and do it my own way and concentrate on driver designs. I know the "solution" seems to easy but it did take a bit to get to that design and formula...there are many other driver styles and proposals and working solutions amongst the thread. For quite a bit there was a strong movement for dual coil and bilateral designs for instance...so there are other ways and different formulas I suppose...everyone who has come here tends to want to put their own twist on things I guess... anyway...again I must run...maybe I have answered something...touch typing and thinking is a bit of a juggling act... pete Link to comment Share on other sites More sharing options...
Donovan Posted January 15, 2009 Report Share Posted January 15, 2009 Hello Pete and Hank. Hank- I am very new to the worl of electronic design. I am in electronic manufacturing, but inner workings of semiconductors and their integration into circuits is something I am getting a crash course in as we speak. I hope we might be able to assist each other through this adventure. I am an instant op amp fan and it sounds like you have some experience with them, so I would be interested to hear more of your regulated idea. Are you suggesting a zener setup for the Vcc? Pete- Yes, you did answer some things there, Pete, thanks! A colleague also suggested one of the caps was causing the phase delay. I will remove it and see what happens. I am also going to create an entirely new preamp circuit, as bare bones as possible, and see what happens. I have a feeling that you are right about me overdriving the LM386 and a simple op amp voltage follower (no gain) might be the key to something good for a very basic preamp. I am also contemplating dual power supplies to get a bigger window. Any experience with this? I am wondering how a contrived ground at the -9v of one battery and the +9v at the 2nd battery might affect the guitar ground and any signal processed from the guitar to a regular piece of sound equipment with ground NOT at that same potential. I have 5 pieces of LM386N-4 which will take up to a 22v supply voltage. Back to op amps though, from what I read recently, most op amps have very high impedance inputs and as such should not load the guitar's output by nature. If my learning is correct, then the simple voltage follower pictured in my last post is all that is necessary for a preamp (sans gain if desired. With the TL082, which is dual op amps in one 8 pin DIP package, one could conceivably run a follower on one and gain on the other. The TL072 you mentioned might even be better as it is a lower noise version, but for now, I have very easy access to the 082's. Also, from what I have read the TL082 is a better choice than Strib's LM741 op amp, which is a very antiquated and slow design. I have a few of those as well and will attempt to confirm this at some point with a socket as you mentioned. I believe the zest of Strib's circuit is the TBA820M. I have samples en route of the TDA7267 2W amp chip from STMicro which I am eager to try out as well. This has a "depop" and a standby feature included and is a much more modern design. than the 386. I thought the "depop" might interest you. I know you are of the experienced mindset that driver construction is most of the puzzle, but I can't help think that a really clean circuit is also key. I tried my new driver and I have sustain on 5 strings now when held by hand above the neck pickup. I can also get sustain on these strings with bridge and middle running, as well as middle alone, with no EMI noise! I am getting excited now for sure. To add another twist just to further confuse from the norm, the sustain is harmonic sustain, not fundamental. I find this strange, since everything I have read in this thread has led me to believe fundamental sustain was easiest to get. When I reverse my leads, I get nothing except the oscillation (string not EMI) that sounds like the driver is trying, but no dice. OK, one last question (for now)... input to the circuit has a hot and a ground. While I understand the AC signal coming in should not really matter much which lead off the pickup is considered ground, isn't there a difference in some way with respect to the DC, since the guitar's ground is a much more massive body of metal and after all, the bridge pickup is connected to this ground when switched on. The reason I ask is my circuit, when hooked up to drive a speaker, will only work in one polarity when taking signal directly from the guitar jack. The mass of the guitar's ground is the only line of reasoning I could come up with to explain this phenomena on my own. I know it is real life that is turned upside down for you and was commenting that it can easily manifest here after answering the same questions over and over for us newcomers. I have been there, meaning low and with seemingly little hope, and lost everything due to my own mistakes. Counseling turned out to be the best thing I could have ever done to help get me back on track with my family and myself. Knowing nothing about your problems except what I have caught glimpses of in your posts, I can safely say you seem to me to be a good person. Keep your head up and you will find your way through it. Link to comment Share on other sites More sharing options...
psw Posted January 15, 2009 Author Report Share Posted January 15, 2009 Thanks donovan.... I am getting excited now for sure. To add another twist just to further confuse from the norm, the sustain is harmonic sustain, not fundamental. this is what keeps it going... beware though...I suspect you may be experiencing something I forget the name we made up for it if you are able to run it that close and with the middle pickup on. Anyway...one of the symptoms is as you describe...the neck pickup and middle can transfer the signal through to the ground wires and into the signal and so back into the feedback loop causing a kind of internal oscillation. A mild form of this is the symptom of distortion described as "fizz". The magnets and the grounded coils will be having a parasitic effect adding extra magnetic power or cancelling out the field as you can imagine if the pickup is held over it. Anyway...it certainly sounds as if you are getting closer. Preamp gain is tricky...the fetzer is fairly low but designed for it's pleasant distortion characteristics...this is of course not what we would want ideally, so designing it in seems like a bad idea. People also have trouble with the biasing of the transistor, including me...and is a part of the design I particularly don't like for this application...for only a few more parts you can get a decent op-amp running with a bit more versatility. If you have fairly hot pickups, a simple buffer probably will do...if a bit underpowered then maybe a little gain might be required...any more than that and you will slam the front end of the LM386 like putting super hot pickups into a guitar amp and the signal out will be more like a fuzz box! SOme have argued that such a squarewave might be advantageous but enough practical work has been done to suggest that the cleaner the better. Older sustainers did use dual battery power supplies...but while advantageous to get head room...perhaps a little impractical for installing and getting to the batteries easily in most guitars. My approach is more like the simple ebow really which also uses an LM386...keeping it as simple as possible. I would suspect that the F/R would work fine too. Anyway...it sounds like progress...was backing off the preamp power a key to this improvement? I have the parts but yet to get around to it...but last year ago I completely redesigned and configured a circuit for this project and my commercial side of things called the fireDrive. I has been tested a fair bit and seems to work well, very compact and has some basic AGC and improved control and stuff. I am looking at building them soon, the general size can be seen in the telecaster thread and will cost about A$60 I suspect. AGC is automatic gain control...once the thing is driving the strings it will continue to add more power...the AGC in mine turns it off at a particular threshold of drive saving power and adding control and reduced distortion. The "Drive" control is a little different in that it will adjust to override the AGC effectively. My circuit would directly replace anything that has been used and will make a decent comparison perhaps. Try not to get too carried away with improvements till you have the thing running as you will get a more intuitive feel for it...this whole thread is about taking the project further so of course it is welcome. At the present time there is far less activity, but there was a time last year where everyone was making dual coil and bilateral drivers to lessen EMI. At other times people have fiddled around with circuits and there is a lot of info in this regard. IN early stages there were some really interesting eras where I really did hope to get a design that could be top loaded with minimum or no modifications to the instrument (onelastgoodbye did some wonderful realizations and I thought we were getting close...but I had to rethink things). My telecaster project this year is probably as close to that as it's come (though it might not be so apparent) as well as my 1mm ultra thin coil for strat pickups and this new circuit. Life goes on...yes, some tough times...yes I have a councelor...I honestly do appreciate the concern as long as you guys appreciate I can get a bit grumpy due to ill health as much as anything. This thread has survived quite a few upheavals of monumental proportions over the years...but it makes a welcome distraction and the many people who have contributed or simply successfully got away with making a sustainer makes it worthwhile...or at least a welcome distraction... pete ps...frustrating I know...but there are some things I keep to myself like the new circuit design for commercial reasons...the exact workings of the now aborted hex things have also remained secret...however, I still maintain the "secret" is in the driver and I have provided adequate information for would be designers to come up with something of their own...or perhaps buy mine! Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 15, 2009 Report Share Posted January 15, 2009 (edited) Hello Pete and Hank. Hank- I am very new to the worl of electronic design. I am in electronic manufacturing, but inner workings of semiconductors and their integration into circuits is something I am getting a crash course in as we speak. I hope we might be able to assist each other through this adventure. I am an instant op amp fan and it sounds like you have some experience with them, so I would be interested to hear more of your regulated idea. Are you suggesting a zener setup for the Vcc? Actually at this early stage I'm just spewing out thoughts for others to comment on....the simplest form of regulator is a zener, but I'd be inclined to use a proper regulator transistor. I'm more of a tinkerer than designer, though as I mentioned a couple of posts back I was trained in Electronic fault finding by the RAF....so I can handle the usual associated test kit fine (eg I have a USB Digital storage scope, which might be handy for comparing phase differences between the original & driver signal) Having slept on it...I'm inclined to dispense with a single rail opamp solution altogether - this will obviously mean I'll walk a solitary path on the circuit aspect, but my reasoning being is that I'm into midi guitars...they already provide dual rail to the guitar (+7V & -7V) & I've already done a fair bit of experimentation with the TL series (to get my own DIY Midi guitar interface working with piezos) - my guitar will never be used on stage, so I've no need for a portable solution - & batteries suck! Therefore at this stage I'm thinking of a TL072 front end preamp (dual rail power supplied externally) driving a Fetzer Ruby amp variant. I'll also possibly head off on a path of individual driver per string (hex) as there may be some synergies with other ideas I have ...I'll therefore probably focus on just getting a small driver working for the top E string initally (as it goes I doubt I'd ever need a driver for the A & low E - for my needs, I only really need the sustainer for the high end). Since 8 ohms is not a variable for the fetzer Ruby & I need to condense the driver (becuase of one driver per string), then I'll need to go down much lower in wire gauge...becuase less cross section = more resistance, which means less coil turns, which means the single string driver can be made smaller (which is very necessary if you're going to have six drivers!) My problem here is my workshop is in my attic - & there's no heating up there ....it's freezing here in the UK right now...so this is just a fact finding/canvassing mission right now until the spring, when I can start this in earnest. Edited January 15, 2009 by Hank McSpank Link to comment Share on other sites More sharing options...
psw Posted January 15, 2009 Author Report Share Posted January 15, 2009 OK...well, I can see a lot more where you are perhaps heading...and maybe we can talk about that. There will require a fair amount of research and experimentation required however, although I have done a bit of work (about a solid year) on various miniture Hex designs as I used to hole a lot of promise for them. They by necessity have to be based on a completely different kind of formula. There can be problems with separation between drivers as any magnetic or electromagnetic devices interact with each other through magnetic coupling. Remote power is a possibility I believe...there is some information discussing these things too. It should be noted that it is not the voltage that is the issue for things like headroom...but the current...so an external constant power supply, even at a low voltage, can have significant advantages. One of the secrets to the driver design is the thin and compact coil very close to the strings...this is significantly different to pickup or the commercial driver designs and transformers...albeit simple, it would seem to be unique. It could well be worth emulating this with hex coils and in fact the design came directly from my hex designs which were even smaller than these small coils and contained an array of 12 miniture neodymium magnets. There can also be some unusual string bending and alignment effects with hex designs. See the high gain LM386 data sheet from national for circuits to keep the LM386 happy...that is the origin of my suggested mods that are designed to help prevent internal oscillations when driving this kind of load. The LM386 can also drive loads as low as 4 ohms and may need to be toned down if a number of separate amplifiers are used to drive the six coils...remember more power is not the solution...it is about efficiency and minimum power to minimise the EMI effects. Be aware that driving a single string is generally not that hard...all the energy on one string and with a monophonic source...it is in fact a stationary ebow and it may well be worth seeing the patents or the discussions on that device here also...but it should not be construed that if you can get one to work that they can be simply chained together or will not interact with each other or be able to differentiate the sound from all six strings to drive a single string. It is odd that you wish to drive the high strings...because they move faster the phase things have more of a problem, and they have less metal for the driver to work with so are always harder to move...a lot of people feel the desire to be able to drone bass notes and such. There is also an interesting effect with bass notes of chords in which in harmonic mode, the low E can sound higher than the high E and make interesting extensions to chords. Obviously when you start getting into hex things you are multiplying the whole project by six and it is still worth starting experiments with conventional proven designs. Perhaps between you and donovan we will finally see a good opamp based preamp/lm386 circuit. I did make some things like this myself and intended this to be the preferred standby circuit but instead took the next step into the AGC thing and so was disinclined to step back. If you want to get even more sophisticated there is also cols opamp and lm386 agc circuit with full details here and some different features again...a slightly different and milder kind of effect than mine, more controlled...but also some very valid ideas and a proven circuit with a couple more features as well as I recall. Someone did a bit of experiments with piezos as have I and it should work. You may need to filter the preaps on them to get a smoother less harmonically rich signal for the sustainer to work with. Actually, it is probably similar to that required by the midi out and if you already have separate preamps for the piezos they are probably good enough to run an LM386 circuit with no preamp (eg fetzer) at all. Mainly, the preamp or buffers are there to prevent loading with passive high impedance guitar pickups which would not apply if they are already preamped...similarly I suspect EMG pickups may be able to drive an LM386 directly saving a lot of hassle perhaps. ... Ahhh...cold weather...I think I would like that...although today was ok...yesterday and recently the heat is coming on, so got to 40C and does not bode well for the coming heat of february. Of course, I am in melbourne which is pretty low down and temperate and we don't have the humidity of the tropics so I can be thankful for that and the occassional cool breeze up from the antarctic when the wind changes direction. Today was about 23C so I guess that sounds like a heat wave for the UK at any time of year. Anyway...tinker when you can...a soldering iron is a source of heat! pete Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 18, 2009 Report Share Posted January 18, 2009 (edited) Ok, so I've made a start reading the thread at page 1! (a bit less onerous, once I realised it's possible to download the whole thread into a Word Doc)...it's going to take a while though! Funny how all the stuff in my head is making an appearance as I progress throughout the thread (phase lock looping etc!) Last night I ventured into my loft - still *way* too cold to make a start, but while up there I was able to measure the DC resitance for each individual string coil of an old Roland GK2 pickup I had lying around. Alas, each coil measures 54 ohms (I'd hoped it would be 8 ohms....voila instant hex driver to start working with!). Still, I guess all six coils could be wired in parallel, which would yield a resistance of about 9 Ohms. Anyway, I thought this little bit of info worth posting up as without a doubt the biggest stumbling block for most will be making up a driver .....if we can use Roland GKs (which some - like me - may have lying around, or there's always Ebay!), then so much the better! Edited January 18, 2009 by Hank McSpank Link to comment Share on other sites More sharing options...
psw Posted January 18, 2009 Author Report Share Posted January 18, 2009 Cool...I don't envy you reading the whole thing. Be aware that a lot of it is and was exploratory...a lot of ideas that went nowhere, a lot of ideas that were explored but failed...a lot of good ideas that were superseded...and a lot of bad ideas. Beware of circuits and such for instance, just because someone posts something didn't mean it would or did work or was a good idea at all. There are parts of the thread that mark some truly difficult parts of my life, a divorce for instance and posting occasionally from the road in internet cafe's occasionally...and the loss of LK...and a lot of others who have contributed along the way and I hope have gone on to live fulfilling lives somewhere out there. Anyway...the hex thing... Obviously you didn't get to the part where I point out strongly that... A Pickup is not a Driver Although the principle is the same (in reverse) not just any coil of wire will do the trick and the technology is completely different. The Hex thing has debatable merits...when you are ready we can discuss them, or you will find out for yourself. As for using something designed for another purpose like a hex synth driver...there are lots of things you have yet to really consider. First...the early synth drivers were made out of tape heads you know...basically a low impedance pickup just to get the signal, they are not designed to emit a signal, certainly not one strong enough to move the string. Also...you won't be able to mount it like a synth pickup near the bridge as it would be putting out so much EMI that will go directly into the other pickups for starters. A driver has far more in common with a speaker...think about it. My analogy has always been that the driver coil is like a speaker coil (and replaces a speaker on the circuit obviously) but the strings are like the cone of the speaker. Now a pickup is a very sensitive device with very fine wires...not like a speaker though the principle is the same. It is unlikely that any conventional pickup, certainly a tiny synth pickup...or a normal guitar pickup...is going to be built with wire to handle the kind of AC voltages being put through it ven with our little LM386 circuits. Also, there are many things about resonances, the reason for the 0.2mm thin coil design working from simple circuits, that you still need to come to terms with. That is not to suggest that this formula is the only one that worked...my hex designs were very radical...but, for instance they required considerable heat sinking or they would self destruct. I did try inductors as were suggested earlier, generated a fair bit of heat before fizzing out...lol! Anyway...there are a lot of ideas, and more that I didn't go into like my work with piezos and vibration "model c" types...much of which were failures. It is often 'felt' that winding a coil is an onerous task better replaced with some premade thing...but by the time you wire all these things up and add appropriate magnet arrays to ensure separation...you may as well have built a single effective and efficient coil...certainly easier and cheaper. Making the driver with sufficient planning, should only take about ten minutes to wind by hand as in my pictorial for instance...the circuit is going to take you longer to build than that! But, many many people, including myself have felt some anxiety about this part of the project. Personally I find the installation part the most difficult and there is no getting around that, even with the store bought off the shelf systems. ... Anyway...clearly this post is only making the thread longer for you...but feel free to ask things that haven't been fully explained as you went on or whatever. Also...may have to give me a heads up on downloading the thread sometime...someone should archive it I guess! pete Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 18, 2009 Report Share Posted January 18, 2009 The bit about a pickup not being a driver is coming through loud & clear! Since I'm no expert on driving speakers (though I'm fine with the concepts), i guess I need to do a lot of reading (& prototypes!). Re downloading the whole thread - it's far easier than you'd imagine... At the top of each page for the thread, there's a drop down menu entitled 'Options' - from there select 'Download this topic'...you'll then have two actual downloadable formats to choose from - I chose MS Word. It's all pretty self explanatory from there. After opening in Word, I opted to select 'all text' & then made all the fonts therein the same size, type & colour - makes it a whole lot easier on the eye! Link to comment Share on other sites More sharing options...
psw Posted January 18, 2009 Author Report Share Posted January 18, 2009 Thanks Hank There were some revisions a while back on the forum that added functions and this must be one of them that wasn't there years back but probably been there a while just not explored so well...it's a question that a few people have asked over the years...and as I say, potentially it could get lost. You know it was only a year back perhaps that someone decided to pin the thing to the top of the electronics forum...it was almost always on the front page there for years anyway...but ever now and and again it would slip into the back pages...it appears to be a thread that will not die. It was pointed out though that it is not the most subscribed to thread ever at PG...the old slam thread which was a place for forum members to abuse each other was much larger...of course not a lot of guitar building actually took place and eventually it had to be closed down...and the life thread seems to be catching up now I guess. ... Yes...it is important to consider not the principle at work here, that is very very simple...it is the actual technical details that make it work. I mean, there have been people who might expect that if you were to just stick sufficient power through a pickup it would become a driver. Things like frequency and phase relationships, resonance and things like that play a huge role...plus the amount of electromagnetic force you can generate in an isolated area without interfering with sensitive like pickup coils through magnetic coupling...all important aspects. But, it can be interesting and pose a lot of interesting thoughts and ideas that reflect on the nature of things like pickups to the physical vibration of strings...so good to develop a bit of a feel for these things...but it also has some massive lessons in problem solving and creative thinking if you engage with the project in your own approach or in collaboration with others. In many ways the question of why does my design work when many others obviously haven't is one that has yet to really be definitively answered really...but the question of why so many slight variations almost work is the way to such an answer which leads to what also might work! There are also some heavy things to face if you get really into it and I think you may well find this hank as I know I did...it is easy to be really enamored in an idea like hex drivers or more complex solutions, without ever really knowing why or what more such a scheme would achieve. pete Link to comment Share on other sites More sharing options...
Donovan Posted January 18, 2009 Report Share Posted January 18, 2009 Hi PSW and all. Here's an update of my progress toward a working sustainer with a focus on the amplifier circuit. I am on driver number 3 at this point after a few fumbles. It takes up the top 3 mm of a Fender Tex Mex single coil pickup, with staggered magnets manually moved level with the top of the bobbin surface. It measures 7.9 Ohms of 0.20 mm Polysol wire (32 AWG), was hand-wound under moderate machine-provided tension and potted in urethane varnish simultaneous to winding, then placed in 32 in of vacuum for 30 min to degas. I believe this driver is good due to test results of getting sustain on each string at least once during tests with various amp circuits, which is where my current focus is. I have wanted to create an amp circuit that is simple, very clean, and having sufficient headroom to guarantee the minimum necessary output to drive the sustainer coil. My current circuit is based on the TL082 op am as a preamp. This is followed by the TDA7267 2W power amplifier. Note, the plan is Not to utilize the full output potential of the 7267 chip. I will be utilizing about 1W and with supply is still one 9V battery. TL082 op amp preamp: The idea here was to create as simple a preamp as possible, using as few components as possible. As you will see, problems I encountered brought the component count up substantially. I first attempted to create a simple non-inverting voltage follower, using no external components beyond the 0.1uF DC-blocking capacitors on the input and output. I wanted to use the non-inverting input as I have read it is a higher impedance input than the inverting input. However, this failed with the TL082 as the chip goes to positive supply voltage on the output if it does not see a minimum voltage around 1v. I read this is a common problem with many op amp designs. To correct this behavior, I added to 50k resistors to bias the input to half of supply voltage following the input capacitor. The results were a very nice and clean 1 to 1 upon testing, but power consumptiom goes up due to the resistor network. Next, I attempted to add a 5X gain, but the TL082 did not seem to accept the standard divided negative feedback to ground technique as described in non-inverting mode on this page at allaboutcircuits.com. I still do not understand why this dod not work as expected, but as a work-around, I used the remaining half of the TL082 as a gain stage. This uses the inverting input, biases the non-inverting input to half of supply voltage again with 2 50k resistors, and produced a very clean, inverted output up to 2V AC. I knew at this point, the gain would need some tweaking, but have not installed a potentiometer for this purpose as of yet. I would rather see real world results with known values beforehand. TDA7267 power amp: I chose this chip as it is a relatively new design in contrast with the familiar LM386 and includes a 2W rating, de-pop circuitry, a standby mode, and good supply voltage options. This was relatively easy to implement, having only a supply + and -, input, output, standby, and 4 grounds to connect, utilizing only a cap for the standby line (47uF) and one on the output (470uF). My first attempt included a 10k potentiometer for colum, but the chip did not accept this between the 0.1uF input capacitor and the chip input. For now, the potentiometer is removed, leaving the preamp gain as the only circuit control, but I plan on reinstalling the potentiometer before the 0.1uF capacitor to see if this corrects the problem. Results: This circuit takes signal from the bridge pickup (Seymour Duncan JB full-size HB) from my Fender strat and will successfully drive 4 strings, the A, D, G, and B, but struggles on the low and high E. I believe preamp gain may be too high. I reduced it to 3X replacing a 20k resistor in the negative feedback loop of the TL082 second stage with a 33k resistor. This somewhat improved results. I also suspect the problem is due to my choice of capacitors and welcome suggestions on these values. The system seems to be biased toward mid-boost and the strongest sustain, which is very strong, is on the G string. I think I need to equalize this a bit, but do not yet understand how capacitor values affect frquency response. Also note my strings are sized .011" to .050". All other observations seemed normal, ie. I did observe the expected EMI squeals when the mid position pickup was selected or when I moved the driver too close to the bridge pickup. As I am a novice at circuit design/implementation, comments welcome and graciously solicited. Thanks. Link to comment Share on other sites More sharing options...
psw Posted January 19, 2009 Author Report Share Posted January 19, 2009 Well...that seems to be some progress... what I would suggest doing is installing the driver in the neck position and completely disconnecting the middle pickup and grounding which can pickup some of the driver EMI through magnetic coupling even with only the ground connected. You may be underestimating this effect and blaming it on other things. I am not sure if the 7267 is one of the chips previously suggested. I do know that a lot of these things seemed to be difficult to source though. The anti pop thing is cool, although the pop thing appears to be driver backlash and only effect the sustainer on turning the thing off, possibly when the grounds are reconnected at this point. The commercial units appear to have a slight delay on reconnecting the grounds and other switches and remain active when turning off through some electronic switching. There are room for improvements on the primitive DIY sustainer for sure...but as I say, I have found them mostly in the installation and bypassing things than in circuitry or driver solutions. It is kind of a shame that you skipped an LM386 circuit successfully in order to compare with this new design. There is probably a fair few other tweaks (like filters for anything above or below the guitar's frequency range, or at least above audio range) in the preamp as these things (hiss and stuff) can cause problems like isolation and certainly don't aid in drive! But, it is a start on a new circuit design. I didn't really study Strib's circuit, but this does look very similar in many respects. Of course, my motivation for new circuits waned when I did my own...in developing such a thing further you might wnhat to look at col's circuit or the reason for the desirability of AGC on the circuit. My circuit features this kind of thing which helps reduce power and scale back drive once the string is in motion. The driving of the G string or sometimes the B is not uncommon...my tele tends to do that. It is affected by the circuit bias (the 100uF cap which kind of compromises the bias for a good general all string response)...you may consider considerably scaling back that 470uF cap to alter the response to the effect you'd prefer. AGC can help even the response out a little more, it would here that the g string is driving hard and so stop driving or lowering drive till the signal is reduced below a threshold designed into the circuit (col, I an others had a lot of debate about this kind of thing) The other thing that is a little tricky is the resonant frequency of the driving coil...you may have some access to ways in which to test for this. It is a part of how this design works that the thin design appears to be keeping a fairly even impedance in it's intended range and so not as affected by phase changes and such...and appears to have a decent resonant frequency for the driving of strings. Making a deeper coil seems to severely alter this...as does the more obvious use of a different wire gauge. Once you get a "feel" for some of these things...it may be of interest to reflect on these affects when considering pickup designs for instance...or anything with coils like speakers and transformers. There are also some interesting lessons to observe about magnetic fields...if interested in that many of us have played with the FEMM program. One of the interesting things about this project is that we are playing with forces that are used and around us all the time, yet they are invisible...but the effects can be experienced in this unique project. good to see such enthusiasim... pete Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 19, 2009 Report Share Posted January 19, 2009 (edited) Donovan- thanks for taking the time to draw up your circuit. Re signal polarity, I see with your preamp, that your signal feeds into a non-inverting input then onwards to an inverting input. This means the signal driving your amp chip will be 180 degress out of phase to the original input. Does your final amp chip right that? psw (or anyone). Is the LM380 phase inverting or not? (or somewhere in between!) Re the orientation of the coils & subsequent electromagnetic field...I've still to read a lot more, but I'm thinking here that the electomagnetic field from the driver will be pulling down & up on each string - is this correct? (ie this force being 90 degrees removed from direction the string moves when plucked by the player) Edited January 19, 2009 by Hank McSpank Link to comment Share on other sites More sharing options...
mrjstudios Posted January 19, 2009 Report Share Posted January 19, 2009 Donovan: Great job in drawing up a new circuit. I had considered a TLA082 based preamp in one of my designs, but with it running into the good old 386 circuit you can see in the latest layout I posted. I ultimately didn't ever finish the design, because the more I read about the TLA082, the more it sounded like it would distort just about as much as a J201 or similar JFET. But there was a few things that did seem to be promising about the TLA082. Some of the driving problems you have sound like overpowering the driver (take it from me, I have tried that and as much as you want more power to = more sustain, it doesnt.) Any reason why you can't use your preamp running into a 386? I beleive that with a good driver, the 386 power stage has more than enough power and it is a familiar and easily manipulated power amp. Also, they are very easy to get about anywhere around the world. I think maybe some of the odd problems you are experiencing could be caused by the fact that your new power chip has extra features in it... but I can't be sure. It might be worth trying to run your new preamp into a 386 power section and see if that solves some of the problems. Link to comment Share on other sites More sharing options...
psw Posted January 19, 2009 Author Report Share Posted January 19, 2009 OK...some interesting ideas here...and you guys are starting to get a bit more of a feel for it and what is going on...keep your mind open and don't make too many assumptions...as for my conclusions, feel free to prove my experiences and suppositions wrong, but this thread has been going a fair while and a lot of the results and failures have been generally proven. Anyway...Donovan is on the right track in many respects. The square wave (distorted) signal vs clean driving signal was well debated and pretty conclusively feel into the clean camp. By using a higher powered amp at half the power you are hopefully going to get more headroom...but the swing of the signal is always going to be detirmined by the available power and with that the distortion characteristics. The TL082 is not necessarily the best choise, TL072 is perhaps more standard for audio but there are plenty of other more exotic and expensive audio opamps...the cool thing about most of them is that they are interchangeable... Some of the driving problems you have sound like overpowering the driver (take it from me, I have tried that and as much as you want more power to = more sustain, it doesnt.) Yes...a very important often repeated point...I am not really seeing on the 7267 how this is being run at half power though...hmmm...where is the gain control. Good to see some of the bigger caps that may have caused some signal phase problems greatly reduced...the decoupling cap a far more sensible 0.1uF I see. ... Hank, you will have gotten to the agonizing about these up and down vs side by side driving characteristics by now I suspect. My hex designs specifically tried to address this...but it is a little naive and not as much of an issue as one might think. The driver is continuing the strings sustain, not analogous to plucking it. Generally a string will vibrate in a circulatory motion ver soon after it is plucked. A common misconception is that the sustainer driver is continupusly plucking the strings...clearly this is not so. There will always be phase differences though, the idea is to have the strings physical vibration at the point in which the driver is mounted similar enough to the signal being picked up by the vibrations at the point along the string where the bridge pickup is. Clearly if the signal is not read and driven at the same point there will be unpredictable phase differences that will be different with frequency and with the ever changing string length caused by fretting. There are so many phase problems really...the lead and lag inherrient in the driver for instance...but the idea is to get close enough! As for the up and down thing...for risk of going on too long...have you considered that the pickup is generally hearing the string in much the same way...of that a guitar amp speaker reproduces the sound in a similar in and out movement? Re signal polarity, I see with your preamp, that your signal feeds into a non-inverting input then onwards to an inverting input. This means the signal driving your amp chip will be 180 degress out of phase to the original input. Does your final amp chip right that? It really does not matter if the preamp is inverting (usual use of an opamp) or non-inverting. If the signal is 180 out, reverse the driver leads will invert it back...or the harmonic switch. What might be cool and an attraction of the dual op amp would be to switch harmonics by switching in the second half of the op-amp to not only invert the signal (creating the harmonic effect) but perhaps add more drive or filters to best drive the strings in that mode. However...a quick warning...it is great that people are so interested in developing things further...many have done this before however. There is an important principle being ignored (for reasons not being said) and that is...this project has some generally "proven" basic principles that have taken a long time to evolve to this, most of it being in the driver and overall concept than any circuit you might be proposing. The principle ignored is that you are trying to develop something further that you have failed to get working in it's basic well established form to the best of it's potential. In fact, many of the problems that donovan has experienced are probably directly as a result of deviations from the design by going his own way. So...with what are you going to compare the results of a new circuit. Also...however you redesign the circuit, whether you use an opamp or jfet or the old LM386 or some other chip the end result is the same...don't mistake it for "innovation". An analogy could be designing a "better" radio...it won't change the some the radio is playing! Ideally, you don't hear the sustainer at all other than it, well, sustains...so whatever circuit achieves this result is a decent one...the easier the better. More innovative would be a driver design of a different kind focusing on better efficiency perhaps...or something more sophisticated in the area of signal and phase control, power conservation, AGC etc. See col's work and others at that time or my own more recent circuit development. So...beware trying to be too clever to achieve the same result...something I have often fallen too and my cations about things like hex designs. Be very clear about what benefits you are trying to achieve with such designs...make sure that you understand enough to see if there is any validity to these concerns you are trying to resolve with it or you aren't actually creating more problems for yourself in the process and trying to reinvint the wheel...as long as you catch a mouse, the mousetrap is a good one! Also...when considering designs...chip circuits are attractive in a lego kind of way...but there are other approaches that have kind of been ignored. One that could hold promise for instance would be a design with say an opamp preamp and a push pull transistor power amp stage...but we do like out poweramp chips don't we! Orientation of the coils is a tricky one to answer...and something that has a lot of problems if you go down that road. Most of my work in that kind of thing was not reported, the results interesting in some respects...but not the best ideas for actually making the effect that we require. Anyway...here I am making the thread ever longer...getting hot down here...but I guesss there could be worse things to do in the winter than read and think about sustainers....hehhe pete Link to comment Share on other sites More sharing options...
Donovan Posted January 20, 2009 Report Share Posted January 20, 2009 Donovan- thanks for taking the time to draw up your circuit. Re signal polarity, I see with your preamp, that your signal feeds into a non-inverting input then onwards to an inverting input. This means the signal driving your amp chip will be 180 degress out of phase to the original input. Does your final amp chip right that? psw (or anyone). Is the LM380 phase inverting or not? (or somewhere in between!) Re the orientation of the coils & subsequent electromagnetic field...I've still to read a lot more, but I'm thinking here that the electomagnetic field from the driver will be pulling down & up on each string - is this correct? (ie this force being 90 degrees removed from direction the string moves when plucked by the player) Hi Hank. The circuit is nowhere near ready. There are some issues regarding frequency response to figure out. I found today that C7 in that schematic is one of the keys and I am working on that now. The signal is 180 out. The amp does not correct that. But for the reasons Pete mentioned, this does not concern me. Link to comment Share on other sites More sharing options...
Donovan Posted January 20, 2009 Report Share Posted January 20, 2009 Pete- I do understand the fetzer/ruby circuit is proven as well as other things from which I may have deviated already. As you mentioned, everyone wants to put their own spin on things myself included. I am looking toward newer design for what improvements I might be able to achieve, as well as the overall learning experience of it. I do appreciate all the time others have spent on this project and I hope my attempts at improvement are not taken as some sort of slam on anyone's efforts. Indeed, I never could have got this far without this 4+ year, 280-page thread and everyone here's perseverance (especially yours). I may come back to the LM386 as I have a good stock of them on hand, but the fact that it is being run so close to its max potential bothers me to some extent, hence my wanting to go up to a bigger chip. Regarding the TL082 vs 072 choice... I agree the 072 is a much better choice, being lower noise and higher slew rate, but I have easy access to the 082 right now and will be switching it out with an 072 when I am more in the finishing stages. Regarding the TDA7267, it does not have variable gain like the LM386 as far as I know. As such, its behavior needs to be controlled in the preamp stage or with a volume control between stages. I will be opting for using preamp gain only, which I have added into my circuit today... a simple 2X to 5X variable gain. So I think the TDA7267's consumption is based on what, how much, and the manner in which you feed it. The quiescent current of the TDA7267 is very low. Combined with my dual op amp preamp, it runs now at about 35 mA where it is most efficient (guitar mid-range 500Hz to 750 Hz), up to .120 mA where it needs A LOT of work (frequency shaping). It is almost as though I am starving it and it is trying to compensate? I have ideas which I know are possible (just have to work them out properly) through which I plan on making this thing very efficient. I want to combine a simple AGC, even more simple than anything Col worked on and then couple this idea with some serious frequency shaping, which I just began educating myself on today. This will be space-consuming, but if I finish this I will move to an SMT circuit. SMT is what I teach and do in a production environment everyday since 1999, so it will not be all that difficult to make that move. Update... I did some testing on the scope yesterday and plotted frequency response, gain, and current. I don't have a chart, just some raw data. What is showed me was that my circuit attenuates the lower frequencies in the preamp stage, causing the amp to work hard below 500Hz, hence my low E-string problems. The gain increases with frequency. I believe this is due to my C7 choice or its mere existence. I believe I can remove it with no ill effects. I attempted raising it today to 1 uF and this exacerbated the problem, basically killing all signal below 10kHz! Amazing! I switched the TL082 out several times and even rebuilt my circuit from scratch thinking that I blew something or had an incorrect connection somewhere... until I realized the amp was working perfectly at high frequencies, above 10kHz. I do not understand filtering enough to comment on why this effect was so dramatic or in this direction, but if anyone else does, have at it. I am going to experiment with lower values (lower than 0.1 uF) next and with it removed. I know it may seem as though I am shooting for the stars and I am, but I am fairly confident in my abilities to get there with the help of those here and at my employer. Once I get this working efficiently, I am going to attack Moog's damping! How's that for star-shooting? Link to comment Share on other sites More sharing options...
Donovan Posted January 20, 2009 Report Share Posted January 20, 2009 Donovan: Great job in drawing up a new circuit. I had considered a TLA082 based preamp in one of my designs, but with it running into the good old 386 circuit you can see in the latest layout I posted. I ultimately didn't ever finish the design, because the more I read about the TLA082, the more it sounded like it would distort just about as much as a J201 or similar JFET. But there was a few things that did seem to be promising about the TLA082. Some of the driving problems you have sound like overpowering the driver (take it from me, I have tried that and as much as you want more power to = more sustain, it doesnt.) Any reason why you can't use your preamp running into a 386? I beleive that with a good driver, the 386 power stage has more than enough power and it is a familiar and easily manipulated power amp. Also, they are very easy to get about anywhere around the world. I think maybe some of the odd problems you are experiencing could be caused by the fact that your new power chip has extra features in it... but I can't be sure. It might be worth trying to run your new preamp into a 386 power section and see if that solves some of the problems. Thanks for the encouragement. No reason I could not run the preamp into the LM386. In fact I did, but in a much different configuration. I believe will come back to it at some point. Link to comment Share on other sites More sharing options...
psw Posted January 21, 2009 Author Report Share Posted January 21, 2009 Well...i don't think anyone would be put out with a new or improved circuit...lol...and, you are actually doing what has been proposed for a long time by myself and others...so keep at it. The moog thing is a different beast again. Some have suggested that they just reverse the flow...but of course we know that is how we produce harmonics by dampening the fundamental. It is possible that a lot of flow is pushed into individual coils but that would take a bit of power...maybe a DC signal...hahaha. Either way, you would need to have individual sensors and separate circuits and driving etc in order to work out the logic to selectively dampen strings. it's amazing how it has captured peoples imagination even if no one has owned up to actually playing one...hehehe Anyway...before you get to carried away...do try and get a basic circuit and driver working and installed as the more complex, the more problems and identifying things can be very hard. Getting something reasonably basic working and developing that further can really help. Not being about to alter the gain structure on the chip could be a problem...there needs to be some eye on power consumption, even if planning on remote power...don't necessarily bet on it. Still, the circuit gives people an idea of awhat an IC alternative to the fetzer could look like. My present circuit still uses a transistor, but not in the fetzer design at all. I had planned an IC version as a mark II perhaps...but sometimes simple is good. Hopefully I will build up a few soon...but ironically the heat is now stopping me! (still at over $1,000 to fix my car today, i beter generate some cash some how!) pete Link to comment Share on other sites More sharing options...
Hank McSpank Posted January 21, 2009 Report Share Posted January 21, 2009 (edited) Having mulled this a while, I'm sticking with the LM386 format as it can be physically condensed easier (as the chip is available in SMT). Ideally, I'd like a stereo 9V 1w power amp chip in SMT format...but I haven't one as of yet. Why stereo? Well, I still intend pursuing the Hex pickup route & then I'd only need three ICs (and space is of a premium) Just one comment about the AGC - I'm pretty sure this is *not* the correct circuit needed for a sustainer's preamp circuitry. I reckon what you really need is a limiter. An AGC will try & boost the incoming signal when it's low...but with a Sustainer, you're going to have incoming signal in spades from the off! What you really need is a way of limiting the output to a fixed limit. Here's how I'll be tackling it... 1. Make up the LM386 power amp stage first (breadboard) 2. Make initially one single string driver (though ultimately six needed) 3. Using a Sig Gen & scope, drive the amp stage direct to see what signal level the LM386 needs for best performance at each string's limits (open sting & highest fret) 4. Design *individual* string preamp circuit to fulfill the signal level info derived from step 3. (ie six 'per string' dedicated preamp circuits) My line of figuring here is that each string requires different amount of physical 'drive'. Also each string has a limited band of (fundamental) frequencies. Therefore it follows that each string will need a different amount of gain &/or driver for the unique frequency range it's capable of producing. I'm only upto about page 70 of this long thread.... so can anyone tell me if it's discussed exactly how much power is required to move a string? The datasheet shows that an LM386 typically yields about 500mw at 9V @8 ohms, but that 'power' will be dispensed across all strings. I'm intrigued to know how much power say a top E string needs to get it moving sufficiently. Ok, so there's a lot of variables (efficiency of driver, type of string, distance from the string), but how about for a typical situation? (ie a reasonably efficient driver, fairly close to the string etc)....what type of electrical power are we talking? I'm thinking that's it's actually likely to be quite low (say 200mw-300mw?). Such info would help greatly in working out some hex cicuit & coil prototype variants! Edited January 21, 2009 by Hank McSpank Link to comment Share on other sites More sharing options...
psw Posted January 21, 2009 Author Report Share Posted January 21, 2009 Hi Hank... Well...SMT has so many fantastic options these days...it is the main reason a lot of things can't be explored by typical DIY experimentation...eventually you will find discussions putting forward chip alternatives particularly envying SMD devices with built in equalization and limiting...afterall these things and more a typical of devices such as mobile phones and ipods...along with class D amplification that provides a more efficient use of the available power. As for AGC...this is perhaps being misused, or not understood. Automatic Gain Control (AGC) is an unbrella term for things like compression and limiting and variations like feedback and forward feed kinds. This last and rare form was very well argued for by col with his circuit design. My design has a relatively crude hard limiting kind of thing...it literally will turn off the signal when it exceeds thresholds...this is still AGC, in fact, in some ways it is approaching a noise gate in reverse...but it operates at some speed of course as no sooner is the drive stopped but the drive reduces and so the circuit is engaged again. There are any number of stereo chips about...I think there may even be a stereo LM386 about...or at least equivalents...but in SMD most devices are stereo as these are the typical applications. ... Anyway...you are going into relatively uncharted waters with the hex thing. I wish you and donovan could explain to me why you feel it necessary to do this as you are making things far more than six times as difficult and throwing away a lot of the established research and so having to deal with it again. I believe you are also being a bit naive about things still...why?...because I know I was too. In part because I like you fell in love with the idea and pursued it for over a year...and partly, although somewhat successful and developing quite radical new designs and greatly shrinking things down in size, I had not at that stage made a simpler design and had a very cohesive reason to pursue this course. It is similar to the idea of making a hexaphonic pickup but mixing it all back down to mono...it begs the question as to why would you do that. For one thing, each coil is going to be less than that of a larger coil, you are moving away from known parameters (thus having to reinvent the formulas) and you would not get enough power unless the coils were chained together in series to work passively. A similar thing is in place with the driver...at present you are not quite getting the importance of the driver and it's specifications...sure mine is not the only workable design...but once you go to smaller tiny coils your whole recipe needs to change...no point taking the 0.2mm wire gauge I suggest on a big coil and thinking this is best for six individual coils. Or, that any 8 ohm coil will do! Also...you guys have not seemed to grasp the concepts of magnetic and electromagnetic coupling crucial to this device...any magnet in close proximity to another will be effectively coupled, often to the extent that they are wired together or worse...have a think about how the pair of coils in a transformer works...this is a very important concept. The proposal to put six drivers all close together and magnetically coupled and thinking they will simply work completely independently is really wishful thinking and you need to do a lot more contemplating on this aspect (it is also the mechanisim that requires extensive ground lifting bypass switching). The "fizz" effect is also a direct symptom of magnetic coupling and not yet come to grasps yet... Of course, the easiest ways to appreciate this kind of thing is to build a simple working known design regardless of where you seek to go...words are not enough to describe the effects really. There are in general some huge assumptions and as someone who got carried away with the hex thing, it is worth really contemplating and generally experimenting with these concepts.. pete Link to comment Share on other sites More sharing options...
Recommended Posts