gurus req to check out AGC Sustainer Circuit

[psw]

Pete from Project Guitar's Monster Sustainer Thread here...

http://projectguitar.ibforums.com/index.php?showtopic=7512&st=1635

Looking for electronics gurus to check out this circuit suggested (and currently breadboarded and working) by contributor "col".



here it is in testing / development mode...



and here is another pic showing the "thin driver" modeled after my sustainer principles that is used to drive the strings to infinite sustain...


in his description, col says...

In the top down pic, there are three ICs, the left most one is the LM386, the furthest right is an LF358N which is just a standard dual op-amp, I'm using that for buffering the guitar input - one half buffers the input straight to output, the other half sets up the signal for the sustainer circuit with about 3x gain (although that is not final, its one of the variables that gets tweaks a plenty )
The big old 16pin IC in the centre is an LM13700 which is a "Dual Operational Transconductance Amplifier with Linearizing diodes and Buffers" - sounds grand eh ?  I'm using it as an automatic gain control (AGC). The gain in the input buffer sets the low level threshold for the AGC which basically gives a big output from a small input and a small output from a big input. Thats the "conditioning" I was talking about and it's all kinda hacked together with a rudimentary understanding, guesswork and happy accidents, and I'm certain someone with a better theoretical background could whip it into much better shape.

So there's a challenge...not really a "stompbox", but interesting just the same. We are getting projected battery drain very low for such a device and it has proved to be able to produce the effect required.

The Sustainer Thread has been going for a number of years and is still very active. I have a working guitar with a very simple LM386 amp and high gain preamp to drive a "thin coil" driver that sits on top of the neck pickup. Col's circuit refinements promise a bit more control and refinement as well as lower power consumption and is an interesting development.

A circuit featuring a chip (LM13700) described as a "Dual Operational Transconductance Amplifier with Linearizing diodes and Buffers", has got to draw attention...maybe it has other applications!

Anyway, the schematic will be undergoing some refinements over at the PG forum and more input would be appreciated...if you've a mind too... pete

[col_012]

Hi, Col here - seems Pete posted my circuit here

So I guess I should explain what it is supposed to do and how it is supposed to do it.

The basic idea is massage the signal from the guitar so that the system as a whole works as efficiently as possible and has as even a response as possible over all strings and all frets.

The circuit consists of a buffer for the pickup output to be sent to the guitar jack. A gain stage to bring the signal up to a useful level. Next is the Auto gain control which is where all the interesting stuff happens, then Last, a power amp that feeds the 8ohm electromagnetic driver.

What I wanted to do initially was make the strong signal weaker, and the weak signal stronger - so the obvious choice was some sort of compressor. After much research (I'm a bit of an electronics novice),trial and tribulations, I found a simple looking design for an AGC amp in the LM13700 datasheet. I had one of these chips in my box, so I got this up and running, and it helped, but didn't give the awesome result I was after 
After some tweaking, headscratching and lots of mulling, I had an idea... Instead of just compressing the dynamic range - giving an equal level out for high and low level inputs - what if it was inverted, so that a low level input gives a high level output and vice versa. This seemed to me like an ideal way of driving a guitar string... a strongly vibrating string needs and gets no drive, while weaker and weaker levels of vibration get progressively more drive.
The AGC amp uses just one half of the LM13700, it works by feeding some of the output back back to its linearising diodes via one of its darlington buffers, this works well, but because it's all in a feedback loop, the best it can do is heavy compression. No inversion.
I kind of understand how this works when I read the blurb... So I though what if I was to split this linearising signal and feed some forward into the diodes of the other half of the chip.
So the first half of the LM13700 works as a compressor. Then its output is then fed into the input of the second half, which has the linearising signal from the compressor injected into its diodes. That means that the second OTA amplifies or attenuates the signal depending on the control signal from the compressors feedback loop....
Anyhow this seems to work, and provides the dynamic range inversion I was hoping for. many of the components have been chosen by guesswork and trial and error... I'm starting to understand some more of the subtlety now, even since I posted that schematic...
The 47n cap that connects to pins 2 and 15 of the OTA controls the attack, and 47n is way to small, however, this encourages chords and lower notes to bloom into nice harmonics, so I may use a more 'correct' value and have the smaller one switchable as a 'harmonic boost' effect... I have now go a better idea of the useful range of gain from the input stage - so I have set up a trimpot that acts as a threshold for the AGC - it allows the peak output of the response curve to be varied between roughly inputs of 18mV up to about 60mV.
One other thing I have done is balance the DC level of the LM13700 better - I have since removed the 2u2 cap attached to pin7.

What I would really like is if anyone can take a look over the circuit and check for any obvious newbie errors - are the voltage divider resistors at sensible levels... filter caps etc. Is there anything inherently wrong with this design that will give it a short life span or make it unstable in certain circumstances.
(remember that this doesn't get into the audio signal chain, so some noise is not an issue (i think?). Any way to make simplifications, or useful improvements without dramatically increasing the parts count... One thing thats bugging me is that there are those chunky resistors on the inputs of the OTAs, I've got the first down to 100k, but it seems like all that amplifying and attenuating is inefficient and shouldn't be necessary...
It would be nice to have a way of tailing off the signal completely at a certain threshold - say around 150mV?, and squelching everything below a certain level...

Also can anyone thing of a simpler, better, more efficient way of achieving the same result - we want the circuit to be as small as possible, and  the LM13700 is a chunky chip.

Well, thats enough from me, thanks,

Col

[DDD]

Hi psw,
Try your sustainer without any AGC, and you'll have a lot of fun for sure.

[puretube]

lots have been written about 13700 & 13600,
as well as counterpart 3080.
lots in the datasheet/appnotes, too.

"search" is friendly...

[col_012]

lots have been written about 13700 & 13600,
as well as counterpart 3080.
lots in the datasheet/appnotes, too.

"search" is friendly...

Yep, the LM13700 is a very old much used piece of technology.
As I noted, this circuit started as a modification of one of the expamples from the datasheet.

As far as searching, I've done many hours of searching, but not found anything that does what I'm trying to achieve with this circuit (LM13700 based or otherwise). It's likely that other circuits can do what I want as undocumented side effects, or with minimal modifications. Thats why we're asking for help on this forum.

fwiw the circuit works reasonably well as is, but I would really like an expert opinion and some criticism - maybe someone else who is interested in the DIY sustainer project can provide that

cheers

Col

[Bernardduur]

Very nice!

What does the three pots do? I only count one on the schem

[col_012]

Very nice!

What does the three pots do? I only count one on the schem

Thanks.
About the pots:
one is to tweak the input gain, one was for tweaking the resistor from pins 1 & 16 to 9v (was trying to find a reasonable value for that resistor), the third is the attenuator between the AGC and the LM386.

Col

[psw]

Hi psw,
Try your sustainer without any AGC, and you'll have a lot of fun for sure.

Well yes...

The Sustainer Thread has been going for a number of years and is still very active. I have a working guitar with a very simple LM386 amp and high gain preamp to drive a "thin coil" driver that sits on top of the neck pickup. Col's circuit refinements promise a bit more control and refinement as well as lower power consumption and is an interesting development.

Yes...it is a lot of fun and a simple amp works well. I have sustain on all strings, a small circuit (a bit of a "pop" on shutdown (and no the answer won't be found by searching here as I have done so, this is not a stompbox/signal processor and has it's own problems)) and reasonable battery consumption for what it does.

What col is trying to do with this circuit is to refine it a bit more. I would encourage those who have the know to take a little look at the thing, consider what it is trying to achieve, suggest ways that the same effect could be achieved in otherways, or troubleshoot or suggest ways in which the circuit could be improved to get the effect required.

I too have used various "compressors" including a purpose built compressor/limiter and various stompboxes as preamps. It works ok, but is not necessarily better than my simple setup. Col has come up with something a little different and I think it is worthy of a closer look. It is working and I am sure will be able to be refined a little more...then we can do a bit of comparison...

Meanwhile, there have been further developments by other members on a thin driver, twin coil, EMIbucker(TM), rail style device.

The combination of these refined drivers and a circuit like this suggests some very interesting developments in sustainer technology... pete

[DDD]

Taking AGC away = less power consumption by the by ;-)

[col_012]

Taking AGC away = less power consumption by the by ;-)

In this application, taking AGC away = significant _increase_ in power consumption in theory and practice.

Col

[psw]

Yes.. people need to have a think about what is trying to be achieved here. We are trying to get the string to sustain infinitely. This can be done with simple amplification (a preamped LM386 for instace works well) but this is a refinement. The idea here is that power is only applied to vibrate the string as required...when the string has enough momentum to vibrate sufficiently by it's own momentum, this circuit reduces the power applied to drive the string. With my purely amplified method, the signal just builds and builds to the extent the string can physically vibrate...this idea is to produce a more controled drive signal. The power consumption savings are a by-product of this action of applying amplification only as required. I'm don't know if it would be better t some how attenuate the poweramp section, but it is a sound idea in principle.

What this type of circuit may offer is the ability to utilize even more powerful amplification for a more immediate response...amoungst other posibilities. It at least offers a more elegant drive signal responsive to what the string is physically doing and thus a more controlled sustainer circuit. This at least makes it worthy of consideration...the fact that lower battery consumption may result is simply a bonus, not the purpose of the circuit as such.

pete

[DDD]

psw,
I'm very impressed with your sustainer as well as your detailed instructions how to build it. So we can say that sustainers are quite simple things for DIYers today.
At the same time I venture to recommend you to direct your efforts not to the electronic circuitry but to further improvement of the driver. The driver is a key thing and surely there is a lot of possibilities to make your sustainer better (including consumption etc).

[H.Manback]

Looks cool, but it's a bit over my head at the moment hehe.

I do have a suggestion for the power consumption. If it proves to be very difficult to get the consumption to an acceptable level, why not make a phantom power system? It would require a stereo jack and microphone cable, and a box that you would put first after your guitar which has a stereo in (with the phantom power on it) and a regular mono jack out. It might not be the neatest solution, but it's a possibility.

[col_012]

psw,
I'm very impressed with your sustainer as well as your detailed instructions how to build it. So we can say that sustainers are quite simple things for DIYers today.
At the same time I venture to recommend you to direct your efforts not to the electronic circuitry but to further improvement of the driver. The driver is a key thing and surely there is a lot of possibilities to make your sustainer better (including consumption etc).

There's driver development going on at the moment in an attempt to reduce EMI related issues (there is coupleing between the driver and the pickup - together they act at a parasitic transformer causing low level fuzz which is annoying when using a clean sound.

As far as the circuit, my initial reason for the design of the AGC was to attempt to balance the response accross the strings and up the neck.
It doesn't matter how good the driver is, the strings  with less mass will not respond as readily to the drive, so without AGC, the G and A on my axe go mental if I use enough juice to get the high B and E going.
Even with a hex design and 6 pre-amps there would still be significant balance problems. The system is highly sensitive to the size of the gap between the driver and the strings. Even with a relatively low action, and a well cut nut, the difference in that gap between playing on the low frets, and high up the neck is enough to have a big impact on the balance and tonal response of the system.

Btw, when using the system without AGC, my setup was sucking about 150 - 160 mA, enough to kill an 9v alkaline VERY quickly. With the latest incarnation of the AGC circuit set up just right, I can get a good effect with better response with a maximum drain of around 22mA ! (I still have to double verify this after replacing the battery in my meter)

btw, does anyone have any suggestions about how to reduce the parasitic transformer effect? also ideas about how to tweak the driver design to contain and focus its magnetic field better.
Oh, and is anyone here good with class D amp designs ?

cheers
Col

[psw]

...I venture to recommend you to direct your efforts not to the electronic circuitry but to further improvement of the driver. The driver is a key thing and surely there is a lot of possibilities to make your sustainer better (including consumption etc).

Quite right...circuitry is not my thing...but it needs circuitry and a more refined circuit means a better sustainer. I am still content to use conventional amplification and work on the driver as the key element...it is! But these things should be considered if someone has the will and ability to explore the areas.

Don't get to hung up on the power consumption thing. Mine will work well for weeks with intemitent play... This circuit is about different and improved performance...

As for the drivers...just posted an idea of a rail driver with bi-lateral driver magnetics...so plenty to explore in that area...

cheers pete

[DDD]

psw,
as far as I know Mr. Alan Hoover of Maniac Music has solved the problems. He uses D-class amps as well as bilateral drivers in his sustainers. To reduce crosstalk he also uses a little piece of steel to adjust inductance of one of the halves of the driver - very simple and effective way.

[psw]

  As has Kenji Tumura...and Floyd Rose...all of which not only use AGC and Limiting but complicated phase compensation circuitry and quite sizable drivers.

My "Thin Driver Design" seeks to address the "problems" within the driver design itself so as only to use simple amplification strategies...my circuit is smaller without resorting to SMD and the battery consumption respectable with a common LM386 at it's heart!

In this regard it has been successful. On my guitar I have a unique standard passive pickup with a 3mm thick driver on top that works very well. I do think we can do even better, and I think there are further areas to be explored (as you so kindly and rightly pointed out in your previous post) in driver design...as there is in circuitry considerations...as is the subject of this thread!!!

I do take your point though...I am attracted to working on the physical driver to create the effect rather than the circuitry.

Col though has proposed something that does offer a degree of control that would be/is quite different from the effect that I have created (disregard power consumption for now...and D class is really a SMD technology, there are 1,000 interesting chips I'd love to try...if they weren't so small). I think it is worthy of consideration for the dimension it adds to the device and I am keen to try it out at least and compare it to my clipped compression preamp circuitry in my guitar. Maybe, quite possibly, I will prefer my guitar's present circuit option and I still feel like developing it further.

When I am able too...I plan to build this and a few other options in modular form and plug them into my guitar which has been purpose modified for just such a purpose...for now, I'll let col do the development, and you guys give the advice!  pete

[DDD]

psw,
By the by, I'm not sure that the Fernandes sustainers as well as Floyd Rose and Maniac Music ones sustain all of the notes on all of the strings.
And what's your opinion?

[psw]

psw,
By the by, I'm not sure that the Fernandes sustainers as well as Floyd Rose and Maniac Music ones sustain all of the notes on all of the strings.
And what's your opinion?

At the risk of hijacking this thread...I have no opinion...cause I have never located one to compare down here !!!

It is a little ironic though because Manic Music used to supply Fernandes their early sustainers (so I believe, before their commercial falling out!)

Anyway...though Fernandes has done much to promote sustainers, especially in standard guitar models...for aftermarket models I think, and suggest Sustainiac (especially if you live in the USA) has the edge...and a little more versitility.

I get conflicting reports. Do you mean you can sustain notes on any string on any fret?...well even mine does that. If you mean that it sustains all notes equally at once (not just play them but actively feed them back) I am a little sceptical. Without a polyphonic driver (as I attempted to create with my hex devices) and preferably a hex pickup and amp system, I am not sure that this is possible. Bass notes in the monophonic signal do tend to predominate in a chord...the rest of the chord sounds, but is not sustained infinitely. Reports from people who own various sustainer models seem to bear that out...

Having not played one, I stand to be corrected...

I think there are some ideas in driver design that may help, and I posted one just today and various people are building things that could lend further towards developing this further...including col.

I also agree with you that the driver is the key. I posted something on this to bring discussion back to what has been acheived in the basic single coil "thin driver" concept and to build on that. The battery consuption in my guitar is acceptable (at least to me) for what it does, and the effect is pretty good. I think what col is proposing does hold some promise, but I'd like also to think that a small and basic circuit design could still be used to drive improved basic driver concepts...

To get and fit a fernandes system into a guitar would set you back about Aus$500, or so I am told. My DIY option cost me roughly $40 all up and I get improved functionality, totally passive standard guitar technology and access to something beyond my reach otherwise. I also get to help develop the concept further and to promote it in general. I should think Manic Music and Fernandes should be encouraged by my work for I know for a fact that many people have gone out and got their systems on the back of the discussions we have over at PG and the interest the thread generates.

In short, I am an admirer of what they, the commercial creators, have done and would like to play a role in promoting and taking the concept even further. In the future, I'd like there to be more accent on what you can do with such systems...the music and technique...as I am sure this would help fuel the technology into still further development.

I also love the ebow concept for it's sheer simplicity and functionality by the way...but there again, you are looking at nearly A$400!!!

So, onward and upward. I am on an extended hiatus from building these devices at present (I leave that to others) but hope to get back into it and have some more "simple", and I hope effective, ideas brewing that I intend to make freely available for others to develop even further.

Meanwhile...any comment on col's circuit would be appreciated...or indeed, anyone who might have some suggestions for sustainer circuits, or ideas along those lines are more than welcome... The sustainer thread has been going for a few years, and I don't see it stopping anytime soon...sustain on... pete

[col_012]

Can anyone explain how to calculate (and change?) the input impedence of the first LM13700 stage of this circuit ?
Would it be possible to give it an impedence high enough so a normal guitar pickup could be connected to it without significant tone sucking - bypassing the input ?

Col