Sustainer Transducer Used As Driver As Well As Sensor

[Fresh Fizz]

Hi folks,

Since Hideki entered the PG forum with the Moog Guitar patent: here I have been thinking if it's possible to build a sustainer that uses the sustainer coil in a duo function: as driver as well as sensor.

I took the liberty to copy col's last post from McSeem's hexaphonic project.

And I wonder if it's possible to build a basic mono Moog. I don't think it's pulse-width modulation. Couldn't it be time division multiplexing (and demultiplexing)?

Using an oscillator and some CMOS switches, could that work? Switch between pickup and driver function?

I only don't understand why the Moog doesn't sound fizzy while the sustainer signal is that much truncated.

hmm... thinking out loud time...

Yes, like back in the good old days.

does the signal in the moog go straight from pickup to amp, or does it go through some dsp that 'reconstructs' it ?

I have no idea whatsoever. The first thing I would want to know is what a chopped signal does to a string. Could be as simple as a ring modulator driven by a 196 Hz sine wave and a 2 kHz block wave as carrier signal. The ring mod signal goes to the sustainer driver in neck position. Then I use my bridge pickup to listen what my open G-string sounds like. (Clean ?)

if the pickup signal is only sampled while there is no sustainer drive present, then the switching frequency can be removed as it is a known quantity. The 'holes' in the output could then be removed either with filtering or some sort of dsp using interpolation or some such...

I suppose the crux is how high the switching frequency can be. I guess this depends on how quickly the field of the driver/pickup can be collapsed and reconstructed. For the system to work with an analog output - like a traditional guitar - the frequency would have to be high so that any switching noise can be filtered out without impacting the guitar tone....

The way I see it 2 things need to happen.

1. Switching from sensor mode to drive mode

2. Delaying the signal picked up in sensor mode and use that signal in drive mode.

Yes, the switching frequency is a problem. Our driver coils are not able to handle high frequencies too well.

Cheers,

FF

[McSeem]

First, there's no way to use time multiplexing with a mono pickup/sensor; this part has to be polyphonic. Second, there's just straight physics: I tried to use pickup coils as the driver, it didn't work. The coils I use now for hex mag pickup have about 300 ohms DC resistance and produce up to 10 millivolts RMS, which is pretty low. They can theoretically work as drivers, but you need a lot of voltage. The driver coils have about 12 ohms and they are fine with about 0.5-1 volt of the signal. However, if you try to use these coils as sensors, they will barely provide you 10s of microvolts. Much lower than tape recorder pickups, which I remember was always a problem - to amplify it with low noise. So, my opinion is everything must be designed in a reasonable manner. The easiest way is to use a true hexaphonic loopback. If not possible -- at least a hexaphonic sensor with time multiplexing for the mono driver. And, of course, a separate signal pickup.

[RM2488]

I don't see why it cannot be done. At least, theoretically, it seems possible. As for the fizz, if the sampling frequency is high enough, there won't be any fizz. Definitely higher than the Nyquist frequency, probably more than double of the Nyquist frequency for the highest harmonic (within reason) of the highest note on the high-e string, as processing and subsequent string driving take time.

That being said, McSeem is right- the current DIY drivers are not suited for this function. However, the Moog patent does state that the sensing and driving is both done by the same transducer.

[psw]

A lot of this kind of idea and what I seem to be getting from the Moog patent from a brief reading is that it is similar to some stuff I was doing during the hex era part of the sustainer thread. A lot was not gone into in that area as I was largely working alone and a lot of things were hypothetical till the basics were possible such as making effective hex drivers small and independent enough. I did discuss a few of these ideas with LK who was most active at the time...

Most importantly and similar I think to the way the Moog this works it would seem, was my idea of a "drive engine" circuit. The would appear to have both driver and pickup coils. I tested my hex things as low impedance pickups and they did 'work' to that extent though that wasn't the aim.

The idea of the "drive engine" was to address a number of 'issues' by operating each drive element separately and sequentially so as to avoid interaction. So, a sequencer chip would send a signal for a small time frame, then to another driver further away, then another and another, perhaps in pairs...with the momentum of the string carrying through the non-drive states. This is much as my present "limiting' things work too, saving power by shutting off when drive is sufficient.

So, taking this further, you could potentially sequence through sampling the string and then holding that to replay back in driver mode. There would be some complexity involved for the phase characteristics I would imagine as the sample and play would be in separate time frames, but it is possible that it would work even in a basic form.

Potentially it might work in a mono version, we know that the basic mono sustainer will work with an effective coil design without any phase compensation and very simple amplification. We know that it can be turned off and on to limit the runaway feedback option in relation to amplitude.

More complex would be to say have a circuit that could ascertain the frequency of a note being played (tuner like circuitry) then a sample and hold (sample that sound with delay type circuitry) then replay that sound briefly in line with the frequency to keep in phase...then repeat the process at reasonably high speed.

There are plenty of pitfalls, you need to set up a clock for the sequencer that wont introduce clock noise and fast enough to keep up with things so it will stay in phase, plus enough power to drive the string with enough momentum (though I don't think this is a problem as these things would ideally sample and play very fast)...and then there are the characteristics of the coils as well, there will likely be some backlash from the capacitance of the coils having stored energy and having coils capable of responding at a suitably high speed.

The "mono" thing is trickier still as the signal will be quite mixed and probably prohibit this kind of complexity, though it still may have merit. It would take a bit of experimenting, but already the DIY project has shown for instance the effectiveness of the 'piggyback pickup driver' for instance and so in that you already have a pickup and driver elements on the same core. From diagrams I saw, it would seem that there are both driver and pickup coils but sharing the same core, so very much like the "piggyback' thing in miniture.

My feeling on a lot of these things were to put it to one side, there seems little reason to pursue a mono source hex driver particularly when the effect can already be achieved by quite simple and conventional means. You would want some tangible improvement. However, if you could do it with a hex thing, perhaps even using the same coils to both sense and drive (though it would appear that the Mood don't do that) sequentially, then there could be some tangible polyphonic improvements from such a scheme worth pursuing.

It's one of those conundrums, you are taking a sample from the string in the same place as you are going to drive it, but then you are shifting that drive in time considerable. So, the advantages of 'phase' being sampled in the location are undermined by such a radical shift and going to require a fair bit of circuitry to deal with, this in itself may bring in delays of processing that will require compensating for...so a tricky enterprise. Plus, there are some fairly well know physical effects that will be occurring, the drivers will be inducing currents in nearby coils and affecting magnet and core materials, currents in coils will build up a capacitance charge and a polarity that will need to be "relaxed" or they themselves will be creating delays and inefficiencies.

The "driver engine" idea (like firing the drivers as if the sparks in a combustion engine was the analogy) was mainly aimed at these kinds of aspects, giving each driver some time to 'relax' before being fired again. The other side of the hex driver from a mono source was to perhaps condition the mono signal to be more biased for each string frequency, through filtering or driver construction itself, as well as for a more even response with different drive power attuned to the string it was driving for a better string to string balance.

I think if you are going to go that far and with such complex circuitry and seeking some kind of tangible benefit over the existing working simple technology, then you'd be clear as to what is being aimed for. To this extent you are likely to need to go "hex" completely.

In my own work, fairly early on we did set out some criteria to aim for, much was to not alter a guitar excessively and to retain the pickups and technology already in the instrument. I think that such criteria is vital so you can gauge direction and progress in a project. So, thing like low mod, battery operation, low mod (so small circuit boards), tangible benefits to adding more complexity, all such things need to be taken into account or a new set of criteria or aims set out to keep things in check and not running wild.

For a lot of my aims, it is not just an anathema to require an enormous amount of processing power to get a similar result as a more direct analog system as much as it is for me to require supplying remote power or a circuit so big that it wouldn't fit in the guitar. For my criteria, something like my telecaster fulfilled most aims entirely, compact surface mounted driver, no mods to speak of except for rewiring and a switch, compact circuitry, good power consumption, fitting the battery into a tight guitars control system. Plus, within the capabilities of DIY and no detrimental affect on the functioning of the instrument in its conventional role.

Others may have other criteria, I believe this was discussed around the 30's pages of the thread, but mine were fairly extensive. Lower on the list were things like improved polyphonic sustain and some more exotic functions and control. I had hoped in my hex work to make such significant improvements with an elaborate driver that it would justify the work and expense but I don't think that is necessarily so without far more put into it, rapidly outstripping my know how and being prohibitively expensive and undermining many of the core criteria. It was possible to make quite compact surface mounted hex things but to run them from a mono source in a simple form they were not proving any more effective than a single coil driver and exhibiting a lot of 'quirks' that were detrimental to performance. Mine were for instance, very susceptible to alignment, so they needed to be built for a particular guitar, but then what of string bending?

The more complex the system and the aims, exponentially complexities arise. It is even possible that some can not be overcome (a magnetic driver will put out EMI for instance), but regardless such things are extremely likely to break the core criteria I set for myself. And, in the end, many of these things will result in "one offs" for the maker, hardly heard and probably not repeated or verified by other makers...so not fulfilling the aim for a DIY project as such.

But, if you have a mind to do "it" why not. So, you might want to consider the idea of switching in an out of driver and pickup functions and investigate as a first port of call, simple sequencing chips.

One other advantage of the "driver engine" that I remember now is that you'd most likely only require one amplifier and preamp stage as with the basic sustainers as this too could be switched through and avoid a lot of duplication and power use...

...

Anyway, it's a thought from way back in the early days of the thread before most people involved now were involved with then or perhaps heard of or considered. It would appear to be something of the jis of how the Moog people are doing it and that the square drive thing is perhaps (I don't have time to sift through things completely) is more about generating a signal to trigger and time the sequencing of pickup and driver functions and signals.

[col]

The way I see it 2 things need to happen.

1. Switching from sensor mode to drive mode

2. Delaying the signal picked up in sensor mode and use that signal in drive mode.

Yes, the switching frequency is a problem. Our driver coils are not able to handle high frequencies too well.

Cheers,

FF

One idea I've mentioned in the past that could help here is to create a sandwich stack. (yep, more brainstorms I'm afraid)

You'd have 3 coils:

one main 'pickup' coil - a low powered pickup coil, but still many more turns than a driver coil

two small driver coils, each identical, one on top of the pickup coil, and one on the bottom.

The idea is that you feed an opposing drive current to the lower coil.

The symetry of the two driver coils with respect to the pickup coil should cancel much of the drive signal current that would otherwise be induced in the pickup.

At the same time, the asymetry with respect to the strings should allow the top driver coil to still drive them and provide sustain.

(there was something tricky going on with the core structure as well, but I don't remember the details exactly)

The main problem with this idea was that the manufacturing tolerances required make it impractical. I had wondered if a microcontroller could be used to tweak the balance to the driver coils in real time in order to prevent oscillation, but I doubt it would be possible (I have pages of notes somewhere).

However, if the drive signal was being alternated with the pickup signal, this stack format might help - it may allow the system to function at higher frequencies due to the cancellation of the drive signal in the pickup.

-----------------------------

For whole idea to work, you'd need to sample at 2.5KHz or more.

As much of the clock cycle as possible would be spent driving. As small a 'frame' as possible would be used to take a pickup 'sample' from each cycle.

As the sampling frequency rises and the sample frame size is reduced, the inductance of the driver coils must be reduced in order to allow the driver coils to react fast enough to prevent bleed-through - the cancellation of the stack helps here.

The lower the inductance, the more current you need to provide the same magnetic field strength.

The bit I'm least sure about is: Does the cancellation property of this sandwich stack idea mean that the relatively high inductance of the pickup coil doesn't slow the response of the magnetic circuit down?

cheers

Col

[Fresh Fizz]

Thanks for the replies guys. I myself have a lot of doubts if this could work. But as RM2488 mentioned it's already been done by Moog.

So it's theoretically as possible as me building a guitar and winning the GOTM competition.

Some of the issues mentioned I even haven't thought of.

I want to share some other idea I have after reading your comments.

One idea I've mentioned in the past that could help here is to create a sandwich stack. (yep, more brainstorms I'm afraid)

You'd have 3 coils:

one main 'pickup' coil - a low powered pickup coil, but still many more turns than a driver coil

two small driver coils, each identical, one on top of the pickup coil, and one on the bottom.

The idea is that you feed an opposing drive current to the lower coil.

The symetry of the two driver coils with respect to the pickup coil should cancel much of the drive signal current that would otherwise be induced in the pickup.

At the same time, the asymetry with respect to the strings should allow the top driver coil to still drive them and provide sustain.

(there was something tricky going on with the core structure as well, but I don't remember the details exactly)

"..."

The bit I'm least sure about is: Does the cancellation property of this sandwich stack idea mean that the relatively high inductance of the pickup coil doesn't slow the response of the magnetic circuit down?

What I am thinking of is using 3 coils. Two driver coils (bifilar wound) and one sensor coil. Both driver coils have their own driver amp. One of the amps has the signal inverted (inverting opamp, coil leads swapped).

The switching should work like this:

1. driver mode: 100% non-inverting amp, 0% inverting amp

2. sensor mode: non-inverting amps switches to 50%, inverting amp switches to 50%.

PRO's(?): no single driver coil that runs idle (100% to 0%)

side effects of switching up & down should offset one another.

If the level of the inverting amp exceeds that of the non-inverting amp there should be string damping.

Cheers

FF

[psw]

I think the moog patent needs to be studied a lot closer to find out what is actually being achieved and how and more of the finer details, the extent of what I read was washing over a lot of these details, the drawings perhaps giving a few clues...but there still seems to be a lot of assumptions with the "moog" thing...at least now there is something to study to work out what is going on there.

I'm not following exactly what the proposal is there FF and col. To me it would seem likely that the drive sandwich might largely cancel each other out and running the pickup and driver simultaneously would seem to be hugely problematic. If there is any EMF to move the strings coming out of the driver however it is arranged or the number of coils, surely this will be "heard" by the 'pickup' it is sharing the coil with and likely overwhelm the signal from the vibrating strings surely?

Also not clear as to what the goals here are. For the work and complexity, what is the pay off of it. I can understand the desire say to have a more polyphonic sustain across all strings, for that you are probably going to go towards a hex system, at least in the drivers but likely in the signal as well with a hex pickup system too. I could understand a desire to have a driver/pickup kind of thing that would allow a selection of pickups or placement of the driver in an unorthodox position such as the middle or perhaps the bridge. I certainly followed that kind of thing for a long time. But, if the end result is a mono device that operates and performs much like what can already be achieved through simple means, what is the 'aim' of the adventure?

Just trying to apply a little Ockham's razor to things you understand.

There is a perception that the signal must be taken at the same place along the string and there is some sense in that and of course the patents are strewn with such concerns for compensating for that and the qualities of the coils. But what has been shown again and again in the DIY versions is that it is possible to use simple means to produce the "effect".

There is room for "improvement" but I'm not grasping what the improvement being aimed for in these schemes are. It can be enthralling to be 'building a better mousetrap' but one needs to be clear to define what "better' is...I mean, if the end result is just that you catch a mouse, then in what way is the trap "better"?

[Fresh Fizz]

@psw

#1: My road to success is trial and error, develop some basic idea, build an early prototype and then figure out why it doesn't work. I'd rather not study patents, too boring.

#2:'Our old school' sustainers make use of 2 separate transducers (driver and sensor) which operate continuously.

The Moog kind of sustainer uses only one transducer. But the transducer cannot do both functions at the same time therefore this kind of sustainer needs a technique to switch between the two functions.

Col's sandwich stack and my version deal with another issue, the big difference that exists between signal levels in driver and sensor mode. The combined sandwich drivers have a hum cancelling effect on the sensor coil. This should greatly reduce the interference of the decaying signal across the driver coils (right after the switching) with the signal across the sensor coil.

#3: Well, the goal is mentioned in the topic title. Figure out how it could be done and then ,if possible, build it.

Hexaphonic is always an option but in an experimental stage you want to keep it as simple as possible.

#4: ?

#5 and 6: No need for improvement! Going fishing all day isn't better than going to a shop and buy a fish. Same goes for carving a bear out of wood instead of watching a real one in the wild.

Cheers

FF

[psw]

@psw

Col's sandwich stack and my version deal with another issue, the big difference that exists between signal levels in driver and sensor mode. The combined sandwich drivers have a hum cancelling effect on the sensor coil. This should greatly reduce the interference of the decaying signal across the driver coils (right after the switching) with the signal across the sensor coil.

Well...could you expand on this aspect then because I'm not understanding I suspect.

To me there seems to be a logic flaw. If a driver is putting out enough EMF to move the string, any magnetic sensor is going to sense this fluctuating EMF resulting in oscillation or squeal, no? If the cancellation of the two driver coils is complete, then there will be not enough EMF to move a string surely? I can kind of see the idea of having the two drivers separated but my experiments in sandwich designs and similar things never could overcome this kind of intrinsic quality problem...

I'm not that keen on studying patents these days either, though the referencing of the Moog thing and not studying it for at least the 10 minutes it took to get the general gist of it and see that there have even with this information some false assumptions seems neglectful. The basic thing is much as I described and proposed years ago, sample and hold and activating things exactly as described in the "driver engine" idea so that interactions that they describe between close drivers and sensors are eliminated through the sequencing of activation.

However, they use "time" separation instead of the physical separation required in more traditional sustainers with sensors and drivers separated by space. This would eliminate interactions between sensors and drivers as they would not be on at the same time...except that any other sensor will just as easily 'sense' these driver signals just as in any other system.

What seems to be proposed is an attempt to run both sensor and driver on the same core simultaneously and that's where I just don't understand the logic that will overcome such seemingly exclusive interactive forces. I can see what col is thinking, perhaps having one of the drivers on top of the sensor it will still have enough effect to drive the string, this would remain to be seen, that from trial and error methods (for which I have been resoundingly criticized on many occasions) I have conducted it would not seem hopeful.

but you know, good luck

[Fresh Fizz]

@psw

Col's sandwich stack and my version deal with another issue, the big difference that exists between signal levels in driver and sensor mode. The combined sandwich drivers have a hum cancelling effect on the sensor coil. This should greatly reduce the interference of the decaying signal across the driver coils (right after the switching) with the signal across the sensor coil.

Well...could you expand on this aspect then because I'm not understanding I suspect.

To me there seems to be a logic flaw. If a driver is putting out enough EMF to move the string, any magnetic sensor is going to sense this fluctuating EMF resulting in oscillation or squeal, no? If the cancellation of the two driver coils is complete, then there will be not enough EMF to move a string surely? I can kind of see the idea of having the two drivers separated but my experiments in sandwich designs and similar things never could overcome this kind of intrinsic quality problem...

I'm not that keen on studying patents these days either, though the referencing of the Moog thing and not studying it for at least the 10 minutes it took to get the general gist of it and see that there have even with this information some false assumptions seems neglectful. The basic thing is much as I described and proposed years ago, sample and hold and activating things exactly as described in the "driver engine" idea so that interactions that they describe between close drivers and sensors are eliminated through the sequencing of activation.

I mentioned the name Moog like: Moog did it so we could do it. The false assumptions are yours. I spent more than 10 minutes on the Moog patent and watched several YouTube clips to get an idea of its performance. I'm not familiar with your "driver engine" idea, but if you say that it works on the same principles as the Moog then why you don't understand?

FF