Sustainer Ideas

[Hank McSpank]

I don't have a problem with making 12 individual single unison drivers. The main thing I'm trying to avoid is have each one powered by its own amplifier. Is this reasonable?

It's reasonable....but what's going to be key here is the coil characteristics & how you connect them all together. For example an inductance figure of 1mH (circa 150 windings) & a DC resistance of 8 ohms seems to be the ballpark for a general purpose driver ...the inductance figure is derived (in the main) from the number of windings your driver coil has.

Therefore if you want to connect say 12 drivers in series, you'd want to shoot for 1/12th of those figures per coil (therefore each individual coil should be about 0.66 ohms & 83uH!). You'd need to experiment with differing coil gauge as you'd be beating a trailblazer's path there! My gut feeling is that a coil of 83uh, that the wire would need to be very thin to be able to get 150 turns - so thin, I reckon it'd not handle the current needed to get a higher frequency type string 'excited' sufficiently.

I'd also have to assume that trying to span/excite so much string real estate with one 'virtual coil' (12 mini coils in series) would need a fairly chunky driving amplifier.

if you don't need individual control per string (as per the EMPP), then I'd go with your 12 drivers connected in parallel or a combination of series/parallel (You'd still need to be careful with your coil characteristics, but they'd be more do-able vs pure 'series' connection)

Well, I tried a couple more designs. One was a thin steel core, 4" long, wired with 30 AWG to 8 Ohms, solidified with PVA glue, with a ceramic block magnet attached to the blade.

Another was like you suggested, wire wound around a plastic sewing bobbin, again potted with PVA glue, with a bolt in the middle of the bobbin and a neodymium magnet attached to the bolt.

Both worked okay, not great, although using the glue eliminated the direct sound I had been getting with my earlier attempts.

However, neither was nearly as good as the Radio Shack Miniature Audio Output Transformer, I cores removed and E cores aligned, with a neodymium magnet attached. This driver is by far the best of those that I've tried thus far.

So...thinking aloud, would it at all be reasonable to wire up 12 of these transformers in a combination of series and parallel like:

7 in series

3 in series

2 in series

and those three groups connected in parallel?

With 8 ohms per transformer, that combination seems to yield a total impedance of 8.1951 ohms. Of course, this doesn't consider inductance.

EDIT: I had incorrectly assumed that the transformers had an impedance of 8 ohms. In fact, after measuring, they have an impedance of 1.5 ohms each! So, how does that effect what I should do?

IMHO, you're focusing on the least important aspect of making a driver (the DC resistance!).

Re the sewing machine bobbin...what wire diameter are you using?

The problem you're going to have is, unless you know the inductance & the total winding count of your own particular coils, then you really are going to be in the dark - & walking a (very) solitary & uncertain path. You can put the coils in any combination you wish, so long as the DC resistance matches whatever your final power amp wants to see *but* it may not sustain - what's absolutely key with sustainers, is not so much the DC resistance (that bit merely placates the needs of your power amp chip), but the delicate balance of number of driver windings (a bit like the three bears...this one has too many, this one not enough this one is just right!), strength of magnet (again three bears!), type of core etc.

If you read through this thread you'll keep coming up with posts that wire gauge X is the most suitable - however such recommendations apply only to coils of a similar design. For example, many on here have established through trial & error, that circa 0.20mm diameter copper wire works best with a single driver breaching all six strings (I've seEn it said often on here that 0.2mm diameter wire is the only size that works ....this IMHO is quite incorrect *unless* we are all making drivers of exactly the same style/dimesions). theredfore 0.2mm wire gauge is the one to go for *if* (& only if) you are making a driver of similar proportions as the one most talked about on here.

That said, what's been generally established on here, is that (for the main) a driver with about 150 windings & an inductance of about 1mH (maybe a little more) is about right. Therefore for my part, where I've been making single string drivers, to keep the winding count circa 150 turns & the inductance about 1mH, I have to use thinner wire than most (about 0.12mm-0.15mm)...your mileage will certainly vary.

Therefore in summary...if you're making your own coils towards a different end goal, then you're going to need to find the wire gauge that suits your own situation (ie that satisfies the aforementioned characteristics). You'll be on forever & a day with the trial & error route...so I strongly I suggest you invest in a cheap inductance meter from China (mine cost a mere £14 delivered to my door from the far east! http://www.tinyurl.com/ompaon )

Mock duck - I can't comment about bass guitars (I don't use 'em) my gut feeling is that you won't need any extra oomph...because the strings that sustain best on a normal six string guitar are the lower frequency ones anyway! There's certainly a lot more ferrous material in the bass string for the sustainer coil to act upon Also the strings aren't as taught, so again I reckon this (in theory) ought to mean that no extra power will be required.

re where to place the sustainer 'intensity' pot - that depends on the particular solution you decide to deploy, but it's probably fair to say that in the main, the intensity pot would best fit between the preamp output & the power amp input. Re its 'resistance' ...that depends on the optimum load that your preamp wants to 'see' but pots in the region of 5k-10k upwards are normally in the general ballpark

Edited May 11, 2009 by Hank McSpank

[mock duck]

Hey I got an idea: would it be possible to put an extra volume knob between pickup and preamp, additionally to the one between pickup and output jack? This way I wouldnt need a switch, I can just turn the sustainer off by turning the pot to zero, and change it's intensity by using the pot?

And just to make sure: I would use a 5k or 10k linear pot for that?

[Dogue]

IMHO, you're focusing on the least important aspect of making a driver (the DC resistance!).

Re the sewing machine bobbin...what wire diameter are you using?

The problem you're going to have is, unless you know the inductance & the total winding count of your own particular coils, then you really are going to be in the dark - & walking a (very) solitary & uncertain path. You can put the coils in any combination you wish, so long as the DC resistance matches whatever your final power amp wants to see *but* it may not sustain - what's absolutely key with sustainers, is not so much the DC resistance (that bit merely placates the needs of your power amp chip), but the delicate balance of number of driver windings (a bit like the three bears...this one has too many, this one not enough this one is just right!), strength of magnet (again three bears!), type of core etc.

If you read through this thread you'll keep coming up with posts that wire gauge X is the most suitable - however such recommendations apply only to coils of a similar design. For example, many on here have established through trial & error, that circa 0.20mm diameter copper wire works best with a single driver breaching all six strings (I've seEn it said often on here that 0.2mm diameter wire is the only size that works ....this IMHO is quite incorrect *unless* we are all making drivers of exactly the same style/dimesions). theredfore 0.2mm wire gauge is the one to go for *if* (& only if) you are making a driver of similar proportions as the one most talked about on here.

That said, what's been generally established on here, is that (for the main) a driver with about 150 windings & an inductance of about 1mH (maybe a little more) is about right. Therefore for my part, where I've been making single string drivers, to keep the winding count circa 150 turns & the inductance about 1mH, I have to use thinner wire than most (about 0.12mm-0.15mm)...your mileage will certainly vary.

Therefore in summary...if you're making your own coils towards a different end goal, then you're going to need to find the wire gauge that suits your own situation (ie that satisfies the aforementioned characteristics). You'll be on forever & a day with the trial & error route...so I strongly I suggest you invest in a cheap inductance meter from China (mine cost a mere £14 delivered to my door from the far east! http://www.tinyurl.com/ompaon )

Thanks again.

Wow, yes, somewhere along the way I got so focused on the impedance that I neglected the other characteristics...which would probably explain why I haven't had as much success as I'd like. I've previously made some single string drivers that have worked, but I guess it was probably just luck due to core size, etc. At some point I forgot about the 150 turns and instead started using a pickup-winding calculator to determine how many turns to make the impedance 8 ohms.

Some of the single string drivers (which were the most successful ones I've built) ended up being 120-160 turns, 30 AWG. The sewing bobbin I've tried I think I ended up with about 400 (!) turns. I did however make two other sewing bobbin drivers of around 150 turns, 3.5 ohms, which I haven't yet tested. The large drivers I used about 60 turns.

I've been using 30 AWG for everything.

My ideal driver would be about 7" long. Is there a good way of calculating wire size to satisfy the goals of turns, inductance, and impedance?

[psw]

Hi there...

Long time readers of the thread will perhaps have wondered what has happened of late...so for people who subscribe to this thread I thought I'd share some personal news.

Obviously things have gotten on top of me in recent times...to top it off the house I rent is going to be sold so a while back I was given a 60 day notice...in the middle of a housing crisis and rapidly increasing rents due to the economic crisis.

It has been something of a panic of course and having difficulty remaining where I am I have made a far left field life changing decision that I hope will improve my health and attitude.

I will be moving in a couple of weeks to an island! Largely rural and small it is a little cold (is famous for it's penguins!) but a beautiful spot...I will be close to a beach and can even see the sea from my bedroom window. Although cold and isolated in winter it is a well known destination for tourists and holiday makers and the little 8,000 population swells to 40,000 when the weather picks up. 3.5 million visit the island each year to see the wildlife, surf and such.

As a city boy all my life, this will be a dramatic and sudden change but one I hope for the better. The house I have got is beautiful and hopefully the time away (at least a year) in the back of beyond will do me some good and help me reassess my priorities and get back to a more open mind.

I may even resurrect my playing, get a handle on recording...mainly though I will be trying to eat well and exercise, take in the fresh air and let go of some of the major problems that have associated from the loss of my children and in more recent times my work.

Anyway...not sure how the internet access will be, most likely my normal email and other connections will require termination due to my ISP not being able to operate in this remote area in a couple of weeks..but I will be looking in with interest when I can for sure.

Good luck to all with their projects and thanks to all those who have been supportive or simply put up with my long winded posts over the years. There are links in my signature to a few other related threads and a bit of a presence related to this thread at various sites like aron's stompbox forum and guitar nuts 2.

pete

[mrjstudios]

I wish you the best Pete, and I hope you can continue to contribute to the thread in the future!

And enjoy your new location!

Edited May 12, 2009 by mrjstudios

[Hank McSpank]

Some of the single string drivers (which were the most successful ones I've built) ended up being 120-160 turns, 30 AWG. The sewing bobbin I've tried I think I ended up with about 400 (!) turns. I did however make two other sewing bobbin drivers of around 150 turns, 3.5 ohms, which I haven't yet tested. The large drivers I used about 60 turns.

I've been using 30 AWG for everything.

My ideal driver would be about 7" long. Is there a good way of calculating wire size to satisfy the goals of turns, inductance, and impedance?

Well, there's your problem. Four hundred turns of wire is going to give you a fairly big inductance. I've just checked & established that 30AWG is about 0.25mm...so if you halved the wire diameter size (0.125mm), you'd need approximately 200 turns to get the same DC resistance - but that amount of turns is still perhaps a little too much.

From recollection, with my sewing machine bobbin, I used about 140 turns of 0.15mm (approx 34AWG), which yielded a DC resistance of about 4.2 ohms (which is fine for my particular output chip) & an inductance of 1mH.

You really only have two options...

1. Stick with your 30AWG but you must make big coils (eg rectangluar & with the longest bit at least 60mm), that way you get to nail your amp's required DC resistance fine yet witout too many turns.

2. Buy some thinner wire!

[pedant alert] You keep speaking of impedance but it's actually DC resistance you appear to be referring to. Impedance does matter (a lot), but the impedance is not the DC resistance of your coil...I'll not bore you (there's plenty of info about) & there are some good calculators where you can extrapolate the interaction of each (but it won't tell you the wire length). For example http://pr.erau.edu/~newmana/imped.html [/pedant alert]

Hey I got an idea: would it be possible to put an extra volume knob between pickup and preamp, additionally to the one between pickup and output jack? This way I wouldnt need a switch, I can just turn the sustainer off by turning the pot to zero, and change it's intensity by using the pot?

And just to make sure: I would use a 5k or 10k linear pot for that?

No, don't go there...you've already got a volume pot in your guitar. If you start putting other pots in, you're going to get some interaction going on. The 5-10K pots I spoke of are for use within active circuits - you can't just drop one into a passive guitar wiring circuit with some *major* impact.

Edited May 12, 2009 by Hank McSpank

[Fresh Fizz]

Hi Pete,

I tried to google up that island you mentioned. No luck. There seem to be penguins everywhere!

Anyway good luck to you, Pete.

I hope that you will feel at home on your island very soon

fresh fizz

[mock duck]

So I put the pot for changing the intensity of the Sustainer into the amplification circuit? Is the preamp in the Fetzer/Ruby the JFET and the LM386 the main amp?

[psw]

Phillip Island

This island is one of Australia's big attractions...the penguins though are the smallest in the world...very cute! It forms a breakwater to Bass Strait so has rugged coasts and surf beaches on the south side and calm tidal waters on the north in western port bay, so there is a lot of variety in environments in a very small space. It has a lot of Koalas and small roos and wallabies about the place and a lot of large sea birds. There are wildlife parks where the animals are very tame and will let you feed and touch them so popular with the tourists and children. It gets a bit cold though in winter (like now) and mild in summer.

It is a big step for a city boy like me to leave but everyone on the island through my short visits have been very friendly. As far as music goes, I'm not sure...however the country towns do seem to still have some live scenes that have died out with the poker machines that came to town a while back and perhaps I will meet some local musicians round and about.

...

As for my various projects, I am not sure...I have sold off a lot of things...my main guitar is the tele still so the sustainer will be used occasionally and perhaps I will be enthused to develop things further. I have another guitar that I was working on that was to feature a piezo thing that I have been working on. My strat is still needing it's wiring completed which featured the ultra thin coil driver and a piezo thing...maybe I was reaching too far to add so much to a guitar.

However, I really need to scale down my activities and prioritize other parts of my life for a while...I may not be able to do much work in this line of things, we will have to see.

...

Anyway...thanks for the well wishes.

...

The sustainer will work in a simplified form quite well with attention to detail and within the limitations of performance one might expect. With care it will give good sustain on all strings but generally lower strings will win out in a chord...this is much the same as with the sustainiac or fernandes systems but can be built smaller and simpler. To get "improved performance" it perhaps depends on what you are aiming for. The moog thing is interesting but details are almost nonexistent at this point. It would appear to be a refined hex system but it is very difficult to tell exactly what is going on and it's limitations (can it only work with the bridge pickup for instance?) ... but then perhaps it does point a way in experimenting further in that direction.

...

I did a little work with bass guitars...generally you need less power but having a circuit better able to deliver lower frequencies helps a bit (I put a 470uF output cap into my circuit for instance). Those big loose strings are easy to drive at low frequencies and most basses have enough room to surface mount the driver far enough away from any pickups and so avoid the need to route or interfere with the controls.

You would be best to have an on off switch but without the need to bypass a push pull pot can handle it, a harmonic switch would also be important and again could be a push pull. If not wanting to add to the controls an intensity control could be used with a trim pot and preset. Turning it down would not be enough to preserve battery life and wouldn't be suggested...these things have a big drain on battery life and loss of performance once the battery gets too depleted so battery access is a must.

From what I have tried however, I have not found a particular musical use for a feedback bass. It would be nice if I could have gotten a nice bowed effect perhaps...maybe I need to experiment still further with the thing or maybe it would take someone with more commitment to lead the way for bass sustainers. Sustainiac did at one stage make a bass sustainer btw.

...

Otherwise not a lot to add at this stage...the move will be in a couple of weeks...thanks again for the well wishes it is a major change of life and something of an adventure for sure. Anyone who may be venturing down this way be sure to drop me a line...I'm only a couple of ks from the penguins...I might go a little more crazy waiting for the summer to come for some company. In the meantime the plan is to eat well (I found a farm produce place nearby and there are no fast food places on the island) and exercise every day in the fresh air, run along the nearby beach or swim in the ocean in summer...maybe play a little guitar...and get some peace of mind...with any luck it will work out, if not at least I will get a tan...

pete

[col]

So...thinking aloud, would it at all be reasonable to wire up 12 of these transformers in a combination of series and parallel like:

7 in series

3 in series

2 in series

and those three groups connected in parallel?

With 8 ohms per transformer, that combination seems to yield a total impedance of 8.1951 ohms. Of course, this doesn't consider inductance.

EDIT: I had incorrectly assumed that the transformers had an impedance of 8 ohms. In fact, after measuring, they have an impedance of 1.5 ohms each! So, how does that effect what I should do?

Ignoring for now the 'EDIT' info, lets say each coil did have resistance of 8ohm and inductance of 1mH. Your series/parallel wiring will give an overall DC resistance of roughly 8ohm and an overall inductance of roughly 1mH => 1 / (1/7 + 1/3 + 1/2) = 1.02439. So not only will the DC resistance be the same, the impedance will also be roughly the same at different frequencies.

Unfortunately, it still probably won't do what you are expecting. The problem that is still there is that the strength of the field that the coils will generate depends on the current through the wire.

The three parallel branches each have a different dc resistance (7 in series is 56 0hm, 3 in series is 24ohm and 2 is 16 ohm) so each branch will get a different amount of current.

Eg. each of the seven coils will get less than a third (0.286) as much current as goes through each of the set of two series coils. For this to work out in a practical system, you would need some very clever layout to take this disparity into account.

This doesn't mean you shouldn't use combinations of series and parallel - just that it's going to be much easier if you ensure that the parallel branches are all equal.

cheers

Col

[mock duck]

So in general, there isn't really a necesity of an intensity pot for the sustainer?

[Dogue]

Well, there's your problem. Four hundred turns of wire is going to give you a fairly big inductance. I've just checked & established that 30AWG is about 0.25mm...so if you halved the wire diameter size (0.125mm), you'd need approximately 200 turns to get the same DC resistance - but that amount of turns is still perhaps a little too much.

From recollection, with my sewing machine bobbin, I used about 140 turns of 0.15mm (approx 34AWG), which yielded a DC resistance of about 4.2 ohms (which is fine for my particular output chip) & an inductance of 1mH.

You really only have two options...

1. Stick with your 30AWG but you must make big coils (eg rectangluar & with the longest bit at least 60mm), that way you get to nail your amp's required DC resistance fine yet witout too many turns.

2. Buy some thinner wire!

Hank: Thanks again.

Bigger coils are not a problem at all -- in fact, if I can get one to work well, it would be preferred. My initial large octave drivers were 7" (178 mm) long. And obviously one driver would be logistically easier to manage than 12 individual small drivers.

I'm sure I overrated the importance of 8 ohms for my amp -- I'm pretty sure it could handle quite a bit lower DC resistance.

Now, I've been using about 60 turns of 30 AWG for the large 7" (178 mm) drivers to reach 8 ohms. So, if I wanted to aim for 150 turns, should I use a larger diameter wire? If I'm still aiming for 8 ohms with 150 turns, it seems 27 AWG (0.36 mm) works out. Although, again, I'm not sure how the inductance of this coil is calculated.

(edit:) Upon your suggestion, I looked up a bit re: relationship between DC resistance and impedance. If I'm aiming for an 8 ohm impedance, then should I aim for roughly 6 ohm DC resistance? In which case, even slightly thicker wire would be appropriate?

col:

You're right; I presumed there was something that I wasn't considering in prioritizing DC resistance. And since I think my amp will handle lower DC resistance, it should be possible to arrange the series/parallel groups more equally.

I did just a brief test yesterday, sending signal into one of those modified audio transformers. As I mentioned, this was the best result I've gotten yet with any of my drivers. Although, as I discovered, it's very low DC resistance, measured at 1.5 ohms on my multimeter (although the package says something like 0.7 +/- 20%, and given the following series measured resistance I'm confused to why a single transformer read 1.5 ohms -- I checked several, and I got the same reading for each). I then connected 5 in series, and the measured DC resistance was the expected 3.5 ohms. However, the output was less satisfactory than the single driver. I suppose this makes perfect sense somehow, which seems to be the current which you explain: I presume the 5 in series are received 1/5 as much current as the single driver?

Edited May 12, 2009 by Dogue

[col]

I did just a brief test yesterday, sending signal into one of those modified audio transformers. As I mentioned, this was the best result I've gotten yet with any of my drivers. Although, as I discovered, it's very low DC resistance, measured at 1.5 ohms on my multimeter (although the package says something like 0.7 +/- 20%, and given the following series measured resistance I'm confused to why a single transformer read 1.5 ohms -- I checked several, and I got the same reading for each). I then connected 5 in series, and the measured DC resistance was the expected 3.5 ohms.

This is probably just down to your multimeter. Unless its a very high spec unit, its accuracy down near 1ohm is unlikely to be good. You've done the right thing by connecting a few together and testing that.

However, the output was less satisfactory than the single driver. I suppose this makes perfect sense somehow, which seems to be the current which you explain: I presume the 5 in series are received 1/5 as much current as the single driver?

5 coils in series will have 5 times the inductance of 1 coil. This meas that as the frequencies get into the hundreds of Hz, the impedance rises very quickly to the point where you won't get any useful drive.

Unless you have all the specifications for the material you are using for the driver core, and have a degree in physics and maths, the only way you are going to get a remotely accurate figure for the inductance of your coil is to measure it with an inductance meter!

Main things to remember if you are trying to understand how it all fits together:

It's impedance that matters rather than resistance

DC resistance makes up part of the impedance.

The rest is a combination of capacitive reactance (mostly from the output coupling cap) and inductive reactance which comes from the driver coil.

To calculate these, you will need to know your complex number arithmetic. Or else, use a meter and an

online calculator

.

Inductance is very important:

As impedance rises, the magnet gets stronger, but the frequency response gets worse quickly

the ideal driver coil will have an inductance as high as it can be while still presenting an impedance that the amp can handle at the highest frequency your design intends to sustain.

Current through the wire is equally important:

The more current through the coil the better.

In general, the higher the impedance at the output of the amp, the less current it will give you. Ideally your amp will be able to drive very low impedance loads very efficiently, allowing you to maximise the current.

An LM386 doesn't seem to handle variation in output impedance very well at all hence all the talk about it being a horrible chip.

If you can keep the output impedance at or very near to 8ohm, it will be fine. much higher and you get lots of distortion much lower and the efficiency of the amp decreases a lot.

If you are not going to just make a driver based on Petes design, you will have to either bite the bullet and buy an inductance meter, then get to grips with the basic math and physics of coils, or be willing to spend hundreds of hours making many many iterations of coils and circuits as you try to home in on a suitable spec. OR just get really lucky.

Good luck

Col

[Dogue]

Col, thanks a lot.

I just ordered an inductance meter as well as some different gauge magnet wire. Also currently working on several 1.75" long drivers experimentally...

[StormLeader]

Man, this thread just keeps on sustianing!

I'm sorry, I just had to pop in here and add that.

-Stormy

[Fresh Fizz]

I did some tests with the LM386. This is the simplest limiter I could come up with.

Boy, that IC gets hot! It gets easily as hot as my circuitry with the TDA7231A which produces more output. Heatsink required... Or class D.

Also a pic of the wave shape on computer. You can see the even order distortion, but it's not that bad for a simple device like this. Input signal was a 1000 Hz sine wave. 100 Hz or 5 kHz gives similar results.

Cheers

Fresh Fizz

[Hank McSpank]

I did some tests with the LM386. This is the simplest limiter I could come up with.

Boy, that IC gets hot! It gets easily as hot as my circuitry with the TDA7231A which produces more output. Heatsink required... Or class D.

Also a pic of the wave shape on computer. You can see the even order distortion, but it's not that bad for a simple device like this. Input signal was a 1000 Hz sine wave. 100 Hz or 5 kHz gives similar results.

Cheers

Fresh Fizz

Well that scope trace signal is looking a lot cleaner than the crud I saw on my scope (unfortunately, I didn't take a screenshot of my LM386 results). I can see evidence of JFET induced distortion on that trace ...but what's with the vertical lines on the final two positive cycles (the falling slope)....is that some kind of oscillation? (or just a bad screen shot!)

RE the heat...I'm not surprised you're getting a hot chip - it's nothing to do with that specific chip per se .... but more the fact that you're pumping 5Vpp across an 8 ohm coil (or is yours 4 ohms?). If my maths serves me right 5Vpp, is 3.536V RMS, therefore @8 ohms that means about 440mA = 1.56W!!! To put this into context, I'm getting good sustain on the lower 4 strings with an 8 Ohm driver @ about 0.7Vpp! (5V is insane!). I reckon you'll be able to move your driver a few metres away from the guitar & still see your strings sustain! (I hope none of your relatives visit with a pacemaker fitted when you're testing that circuit

Edited May 20, 2009 by Hank McSpank

[Fresh Fizz]

Well that scope trace signal is looking a lot cleaner than the crud I saw on my scope (unfortunately, I didn't take a screenshot of my LM386 results). I can see evidence of JFET induced distortion on that trace ...but what's with the vertical lines on the final two positive cycles (the falling slope)....is that some kind of oscillation? (or just a bad screen shot!)

A bad screen shot....

You're looking at a digital fotograph. I'm still using my good old Handyprobe (connected to the parallel port) running on MS-Dos (floppy). It worked well in Windows 95 on an AT286 but not so in XP. The Y-axis (voltage) is accurate but the X-axis is in need of a new calibration. The vertical lines is another peculiar phenomenon, when wave shapes get weirder it occurs even more. It's not oscillation.

RE the heat...I'm not surprised you're getting a hot chip - it's nothing to do with that specific chip per se .... but more the fact that you're pumping 5Vpp across an 8 ohm coil (or is yours 4 ohms?). If my maths serves me right 5Vpp, is 3.536V RMS, therefore @8 ohms that means about 440mA = 1.56W!!! To put this into context, I'm getting good sustain on the lower 4 strings with an 8 Ohm driver @ about 0.7Vpp! (5V is insane!). I reckon you'll be able to move your driver a few metres away from the guitar & still see your strings sustain! (I hope none of your relatives visit with a pacemaker fitted when you're testing that circuit

No it's less, 4Vpp, 2.8V rms, 1W @ 8 ohms. But you know, I'm not so sure about that figure. After a while the driver runs hot(ter). Copper wire's resistance goes up? My goal was to develop a simple design that at least is capable of producing enough output to produce sustain. Now I know that 1W is the max with the LM386. By increasing the feedback (trimpot Rv) the output level can be lowered. I only hope there is enough output signal to go low enough, otherwise some extra DC voltage has to be added somehow. But not the Aussimart way, that makes long release times impossible.

Cheers

Fresh Fizz

[Hank McSpank]

No it's less, 4Vpp, 2.8V rms, 1W @ 8 ohms.

As an aside, your scope trace shows the peak to peak voltage nearer 7.5V, (5.3V RMS)...which works out at 662mA into an 8 Ohm coil = 3.5W - positively nuclear!

Edited May 21, 2009 by Hank McSpank

[Fresh Fizz]

No it's less, 4Vpp, 2.8V rms, 1W @ 8 ohms.

As an aside, your scope trace shows the peak to peak voltage nearer 7.5V, (5.3V RMS)...which works out at 662mA into an 8 Ohm coil = 3.5W - positively nuclear!

The 4V peak-peak was wrongly stated, ofcourse it's 8V pp. But to calculate RMS you need to divide by 2.82 , not 1.41 .

With a symmetric signal you take one half of the sine wave (peak to X-axis) and divide by 1.41 .

It's 1W, really!

FF

[Hank McSpank]

The 4V peak-peak was wrongly stated, ofcourse it's 8V pp. But to calculate RMS you need to divide by 2.82 , not 1.41 .

With a symmetric signal you take one half of the sine wave (peak to X-axis) and divide by 1.41 .

It's 1W, really!

FF

Oops...my bad - you are of course completely right!

[Hank McSpank]

5 coils in series will have 5 times the inductance of 1 coil. This meas that as the frequencies get into the hundreds of Hz, the impedance rises very quickly to the point where you won't get any useful drive.

But if the that higher impedance is because of more coil turns, then ok, there'll be less 'drive' (i'm figuring you're meaning current), but all other things being equal the ampere turns will remain the same - therefore same amount of flux force at the strings? (the question mark is because I'm theorizing here!)

It's impedance that matters rather than resistance

DC resistance makes up part of the impedance.

The rest is a combination of capacitive reactance (mostly from the output coupling cap) and inductive reactance which comes from the driver coil.

To calculate these, you will need to know your complex number arithmetic. Or else, use a meter and an online calculator.

Inductance is very important:

As impedance rises, the magnet gets stronger, but the frequency response gets worse quickly

the ideal driver coil will have an inductance as high as it can be while still presenting an impedance that the amp can handle at the highest frequency your design intends to sustain.

Current through the wire is equally important:

The more current through the coil the better.

In general, the higher the impedance at the output of the amp, the less current it will give you

. Ideally your amp will be able to drive very low impedance loads very efficiently, allowing you to maximise the current.

There's some good stuff in there that'd be a good starting point for discussion! (eg more current through the coil might be 'the more the bbetter', but not if you're running off batteries, where you just want the bare minimum current possible to get a good 'grip' on the strings)

Incidentally, I've been in relative radio silence...not because I've lost interest - far from it - but more, life is getting in the way! (also, I've spent an inordinate amount of time getting my PIC based coil winder running with 'bells & whistles' not to mention that darned CNC machine - TOP TIP....DON'T EVER BUILD ONE!). Anyway, I'm getting to the point where I can start getting back to sustainers tagain.

I'd like to try & stimulate some discussion around ampere turns/ magnetomotive force (OR MMF which, which as it goes, I don't know a whole lot about).

I'm slowly wading my way through this thread hoping that the odd pearl will surface (I jumped in at about page 30 & I'm up to about page 75 - woaah....what a lot of repetitive woolly 'noise' there is in there too ...in just about every other post!). I find it surprising - .bar one poster back then - that MMF hasn't been discussed much at all (certainly not up to page 75 anyway!).

Ultimately it's the MMF that's going to produce magnetic force which will make our drivers physically move the guitar strings ....a common theme seems to be

"My sustainer works on the lowest 4 strings fine, but not the top two

" ....then surely we need to establish roughly how much MMF these strings need to get them sustaining well? (removing some of the 'variables' where possible - eg distance between driver & string etc).

From my understanding, those drivers that span six strings generally have the following 'known' characteristics (I say 'known'...not 'best')...

About 150-200 turns

A DC resistance of 8Ω (this is the figure that everyone seems to focus on & I'm puzzled why - it's the least important!)

An inductance of about 1mH (this one is certainly more significant)

The variables we can't remove are ...

type of magnet

core material

..because a lot of folks will simply use what they can get or have to hand!

Variables that could be removed are (for the purposes of collecting data)...

Distance between sting & driver (ie if when taking measurements, we all ensure we use the same distance - eg 3mm)

pickup feeding the circuit - use a Sig generator to feed into the circuit (as opposed to a pickup)

There hasn't been a huge amount of technical data put forward wrt the measured signal levels across the driver which results in good sustain (I realise this term is subjective, so perhaps it's better to talk about the 'edge of sustain' ....this being where the driver has just got enough grip of the string to start sustaining it).

In my opinion, unless we can start getting some of this data together, then we're all just shooting in the dark & making clones of a supposed 'efficient' driver (& that term actually gets my goat, as without some meaningful technical data to back it up, it's just people bigging their own driver up).

So this is a call to all those of a more technical ilk, that have access to a scope, to start noting down & posting up some meaningful data!!

Once we know the current running through the driver *and* the number of turns of the driver...then we can start talking 'ampere turns' & MMF.

For example, let's take a

single string

8 ohm driver, 1mH of about 200 turns as an example with 3mm distance between string & driver (you can slot a drill bit in between the gap to confirm the actual distance) let's say that to get the 'edge of sustain' for the top E string, that 400mV peak to peak sine wave is needed.

Bear with me on this bit as I may have my figures wrong.....

Firstly 'ampere turns' is meant to be DC current related (and measured in a vacuum!)...& what we're dealing with here is AC current (& no vacuum)...but let's park that pesky fact for now!

If @200 coil turns it takes a 400mV peak to peak 'drive signal' to move a top E string at the 'edge of sustain', then assuming a coil of 8 ohms we can summise...

Impedance is 8.264 Ω (top E = 330Hz, coil inductance = 1mH, DC resistance was 8Ω)

RMS voltage is 282mV (ie our 400mV peak to peak)

Therefore driver current is 34mA (282mV/8.264Ω)

Therefore our 'rough & dirty' ampere turns figure for the 'E' string is 6.8aT (200 turns x 34mA)

Once we have figures for two highest strings, then we can start honing drivers that will cater for these troublesome strings (one idea ...if you didn't all have solid cores running through the entire core length of your driver, is simply to wind a few more local turns at the higher string end of the driver - that'd put more ampere turns at the end causing the problem)

Thoughts / Comments?

Edited May 29, 2009 by Hank McSpank

[col]

5 coils in series will have 5 times the inductance of 1 coil. This meas that as the frequencies get into the hundreds of Hz, the impedance rises very quickly to the point where you won't get any useful drive.

But if the that higher impedance is because of more coil turns, then ok, there'll be less 'drive' (i'm figuring you're meaning current), but all other things being equal the ampere turns will remain the same - therefore same amount of flux force at the strings? (the question mark is because I'm theorizing here!)

I've been through this discussion with myself more than once.

Unfortunately, It doesn't work that way. And equally unfortunately, its not simple

I'll try to remember some of the details.

# more coil turns means not only more inductive reactance, but more DC resistance - to counteract this effect, you can use chunkier wire, but there are practical limits to how chunky the wire can get.

# As the impedance rises with coil turns, the efficiency and distortion characteristics of our basic optimized for 8ohm power amps falls away. (can we design a power stage that is better suited for driving a higher impedance load?)

# the inductance and therefor inductive reactance of a coil increases with the square of the number of turns whereas MMF rises linearly with the number of turns. So for a single coil, there is a point as you add turns where the loss through increased reactance is more than the gain through increased flux. Where this point is depends on your desired highest frequency.

# When there are multiple coils, it gets too difficult for me . If you could disregard magnetic coupling between the coils, you could say 5 coils has 5 times the inductance of 1 coil, but that's not going to be the reality... its probably a case of trial and measurement... as you add more coils, and pack them tighter, the coupling is going to increase, as are the limitations on the wire guage imposed by physical space restrictions...

Glad to hear you're still working on this. I've been taking a break. Too many other things on just now, and I've not been playing my electric at all, so the motivation to keep working on the sustainer is pretty low. I've not given up though!

cheers

Col

[Fresh Fizz]

About 150-200 turns

A DC resistance of 8Ω (this is the figure that everyone seems to focus on & I'm puzzled why - it's the least important!)

An inductance of about 1mH (this one is certainly more significant)

For example, let's take a single string 8 ohm driver, 1mH of about 200 turns as an example with 3mm distance between string & driver (you can slot a drill bit in between the gap to confirm the actual distance) let's say that to get the 'edge of sustain' for the top E string, that 400mV peak to peak sine wave is needed.

Bear with me on this bit as I may have my figures wrong.....

If @200 coil turns it takes a 400mV peak to peak 'drive signal' to move a top E string at the 'edge of sustain', then assuming a coil of 8 ohms we can summise...

Impedance is 8.264 Ω (top E = 330Hz, coil inductance = 1mH, DC resistance was 8Ω)

RMS voltage is 282mV (ie our 400mV peak to peak)

Therefore driver current is 34mA (282mV/8.264Ω)

Therefore our 'rough & dirty' ampere turns figure for the 'E' string is 6.8aT (200 turns x 34mA)

Once we have figures for two highest strings, then we can start honing drivers that will cater for these troublesome strings (one idea ...if you didn't all have solid cores running through the entire core length of your driver, is simply to wind a few more local turns at the higher string end of the driver - that'd put more ampere turns at the end causing the problem)

Thoughts / Comments?

400mV pp seems pretty o.k. for me! 400mV / 2√2 ≈ 141mV. Driver current should be 17mA. That's 0.141*0.017=2.4mW ! Does that mean that 6 single string drivers are more efficient than one large driver? 6 * 2.4 mW = 14.4 mW compared to the nearly 1 W I use.

I. How fast does your sustain set in? I can imagine that when you use that little power it can take some time to achieve a sustaining tone. I hate to wait for sustain to happen.

II. How focussed is the magnetic field? If you bend the string do you still have sustain? And I'm curious how you solve that issue with bending strings. Is that a matter of PIC, like when you play the high E-string your E and B driver work. (I don't know if that makes sense, I'm not familiar with PIC). Or do you build drivers with some sort of overlap (when a string is bent it is still above the driver). Or are you fully focussed on sustaining chords?

III. There is a chance that when your sustainer is operational the drivers could have a negative effect on one another. You stated that you needed 400 mV pp for your E-string, but what if your B-string is sustaining at the same time? It could be that because of interference of the 2 signals it's a lot harder to get good sustain.

So maybe I should operate my sustainer (Steady Steed) with less power. (There is a trimpot inside my strat.) Could be that with less power I still have a good sustain but without the funny fizz effects on the low strings when playing chords.

I will do some test with the LM Squeezer to get the signal limited at a lower level (like 400 mV pp) and will post it.

Greetings

Fresh Fizz

[Hank McSpank]

Hi Fresh,

I'm definitely having some finger trouble with those RMS conversions! (I keep tapping 400mV peak to peak into online calculators, whereas it should be 200mV peak!)

Ok, first the bad news - that 400mV peak to peak signal I mentioned was just an example. That said, 400mV peak to peak figure I gave is probably from an actual reading I took (I seem to recall that's the type of signal level I was seeing across my driver ... during my first 'stint' of sustainer experimentation, I did a lot of initial experiments with my associated notes scribbled on the back of scrap paper ...I really must get more disciplined with my notation this time!)

Alas, at the moment I can't double check - my sustainer 'REDS' facility (Research, Experimentation , Development Stop. ) is out of action at the minute ...due to a hulking hobby CNC machine I'm building taking over *all* my bench space.

Now, to your specific questions...

(bear in mind, I'm still on single string drivers until my CNC is complete, then I can start cutting some bobbin top/bottoms that'll hold six of my individual coils wired in series)

"I. How fast does your sustain set in? I can imagine that when you use that little power it can take some time to achieve a sustaining tone. I hate to wait for sustain to happen."

You may not recall, but I'm using a PIC to digitally sample a rectified incoming pickup signal. Once sampled, my little PIC program can determine how much gain is needed & how fast....therefore it's possible to have the sustainer kick in as fast as required , but as it goes I prefer to fade the sustain in, as the string's natural sustain ebbs (with the 'melding point' level set to my preference)...it's all pretty seamless. BTW, I have no fizz whatsoever in my sustained notes....clear as a bell.

I've got to go back & revisit the supporting analogue circuit again once my CNC is ceremoniously launched (dumped!) from my bench...I have a whole heap of new analogue ICs to play with. There's still a lot of work to be done, but I'm striving towards the bare minimum of potentiometers or presets at all (hopefully none - aiming to be all done digitally with just a couple of small push buttons. I'm still someway off from that...my main focus is the 'patent applied for' McSpanKStainer complete in the shortest time!

II. How focussed is the magnetic field? If you bend the string do you still have sustain? And I'm curious how you solve that issue with bending strings. Is that a matter of PIC, like when you play the high E-string your E and B driver work. (I don't know if that makes sense, I'm not familiar with PIC). Or do you build drivers with some sort of overlap (when a string is bent it is still above the driver). Or are you fully focussed on sustaining chords?

Ok, this is an interesting one. You must bear in mind that I set out on this journey with a Hex driver in mind. My tests showed that with a single string driver scenario, the string can be bent a good way from the small driver & the sustain would still 'hold'. Too allow this, it needs a fast AGC circuit & a sufficiently powerful enough output stage (that said, when you're only dealing with one string...the driver's power needs drops significantly). So, as the string moves away from the area of driver flux 'focus', my circuit really is designed to crank up the gain fast & spam out a bit more flux & keep that string sustaining.

III. There is a chance that when your sustainer is operational the drivers could have a negative effect on one another. You stated that you needed 400 mV pp for your E-string, but what if your B-string is sustaining at the same time? It could be that because of interference of the 2 signals it's a lot harder to get good sustain.

There's every chance...but when you think about it...most of the time , guitarists have a 'mono' sustainer requirement (occasionally 2-3 strings for a fleeting moment)..ie this being when a guitar player is soloing. Therefore at any one time, there will normally be just one dominant note played by the guitarist...in a hex driver scenario, that'll map to one dominant driver 'ruling the roost' for that moment in time. For chords, I'm not so sure that driver interaction will be so benign. I'm hoping that string/driver resonance plays a large part. eg, the driver will only impact on the string it's getting a feed from (assuming a hex input) ...the other strings because they aren't resonant...will hopefully disregard. It may well be that there won't be as much focused MMF (not a porn term in this instance :-) ) presented at the string due to inter-driver flux fights...but at the end of the day it'll be very much a case of suck & see. I'm presently creating some simple CAD designs for holding multiple single coils to be held at exactly the string's mid point...I now only have the Dremel tool holder for my CNC machine to complete & I can then finally start cutting these driver holder CAD designs out of acrylic....

The screen shotABOVE just a simple CAD/CAM 2D representation of the top/bottom six coil holder design, (the dotted red box represents the acrylic it'll be cut from....BTW I'm being tongue-in-cheek with the wording!!!). It's full single coil in size...why? For no other reason than my benchtop guitar is a strat has a big hole where the neck pickup used to be...therefore I'll fill it with this shape! I'll be scaling the driver down to suit once I have definitive coil dimensions.

I'm now up to page 110 of this (somewhat highly repetitive thread! Each page is like Groundhog Day), but as of yet, I haven't read of anyone back then using hex drivers with a hex pickup feeding six channels into the driver circuitry.

The big stumbling block I have with the 'standard issue' driver that has been adopted by many on here, (ie a one coil driver breaching all strings), is that for most 'sustainer' situations (certainly when soloing) - at any one time, you really only need a dominant string to be sustained ...but the driver adpoted by many here, presents the required one note across all six strings - highly inefficient & which I reckon is wasting a lot of scarce battery power as the driver attempts to sustain strings which aren't meant to be sustained ....far better to find a way of 'routing' the dominant required note for sustaining to the correct string(s), therefore two options I can think of....

1. Take a bog standard merged/mono guitar pickup signal & bandpass filter it to get the right frequencies 'routed' to the right strings (obviously there'll be some overlap, as for example a high 'B' could come from the B string itself or the two strings below fretted high up) & have six drivers (and preamps/power amps)

or (far better)...

2. Take a hex input feed from the likes of a midi/hex pickup etc.

I may pursue a non Hex input solution (too many 'decision point forks' with this project ....& I blow hot & cold as to which path to beat first!) ...but if I do go with a 'driver across all six strings' approach, one avenue I'll certainly explore is to customize the 'turns' needed per string across the strings by using six coils below in series ...less turns on the coils for the lower strings....more turns for the higher strings. This won't save power, but at least it'll help get the string balance right.

Re the power need for your driver...I can't say what's best in your case - but that's the whole point of my 'call to arms' for those of us ith scopes.

If you can scope the signal across your driver for each open string in turn, at a known distance (eg use a drill bit to slide between driver top & sting bottom to establish the distance) where you just get past the edge of sustain (a subjective term, but can be described where the sustain is just enough to keep the string sustaining)...with sufficient data we should be able to glean how much power is needed per string etc. This will at least allow us all to address the "My top two strings don't sustain at all" type posts we see.

It's a little puzzling to me how a thread can be this old & yet still have no meaningful technical data!

I'll not pretend to be a clued up electronics wizard with lots of Formulae waiting in the wings to solve all sustainer problems (& after all, I've calculated some RMS voltage levels erroneously twice!), but I'm sure that discussing/applying the basic stuff (power levels, impedance, inductance, 'ampere turns' etc) will go along way to finally solving this sustainer conundrum that (unless I've missed it), still hasn't been satisfactorily nailed in all that time?! Sure there are DIY sustainers a plenty out there, but everyone I've read of seems to have a problem-ette of some sort (distortion/fizz, weak/no sustain on some strings/fret positions, no harmonics, bad EMI, pop problems etc). There's only so many times I can read "0.2mm wire is the magic ingredient" without any explanation why, or "it's the thin efficient coil you see" - without anyone even trying to establish its efficiency. There must be about 20% of all posts in the first one hundred pages alone harping on about pva vs epoxy vs superglue vs wax during the winding process (sure you don't want the coil vibrating, but purlease!!)- but few are chiming in & questioning the technical aspect of the driver ...ie power levels, turns required per string (note: not the the driver turns in total across all strings), inductance blah blah...nope, everyone appears to follow the you need to wind the magical 0.2mm gauge wire thinly until you get 8 ohms! line! Yet everyone seems to have a hiccup of some sort with their sustainer?!!

Edited June 1, 2009 by Hank McSpank