Somebody who knows op amps.

Uncle Krunkus

Could anyone, who knows a lot about circuit design, please have a look at this and tell me if anything there rings alarm bells? If there's nothing wrong with it, could it be improved? Could it draw less current? Be less noisy etc.
I know it's not strictly a synth circuit (though it could easily be adapted)

Some crazy dude I know wants to put a pre-amp/dual parametric EQ inside his bass guitar. Well I said, "You've got to be kidding man! Do you know how much work I've got to do on the Sorcerer?" He said, "I don't care! I need to be able to play my bass when I'm not wiring the Sorcerer!" He started pushing, and I started shoving and, well, there were a couple of black eyes, you get the picture. Rolling Eyes

Do not assume anything about it is right. This is the first op-amp based circuit I've tried hitching together (from other peoples ideas)
It's also the first real drawing I've done with Ray's Schem Publisher. I'm getting better at it, but could still improve.
Thanks heaps in advance to anyone who can be bothered to check it out. Smile

ParaSchem.gif

Description:

Filesize:

14.09 KB

Viewed:

291 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Scott Stites · Posted: Sat Sep 02, 2006 7:55 am Post subject:

Hey Uncle K,

Just looking at it superficially, mind you, but if you are planning on powering the op amps from +/-9V, you don't need to create a virtual ground at 1/2 Vcc - that would only be necessary if you are powering the op amps from +V and ground.

I suppose you could go one of two ways - use virtual ground and a single supply, which means your buddy only has to use one battery, or use two batteries and use ground as your reference. For me, I have an easier time doing things with a split supply, but that's just me. The single supply certainly would have the advantage of using less batteries.

Either way, once you get it going, you might want to try Ray's trick of using LF44X op amps - lower current requirement, longer battery life, more money your friend has for buying you beer.......

Take care,
Scott

Uncle Krunkus · Posted: Sat Sep 02, 2006 9:14 am Post subject:

Nah,
no dual supply for this one, (although that is generally my preference too) It's all going to be mounted inside my, um,... his bass guitar, so he just wants one 9V for power.
Thanks for the tip about the LF44Xs. I think I've still got a couple left over from the SoundLab. They are hard to find over here.
Does it look okay apart from that?
Thanks Scott

zipzap · Posted: Sat Sep 02, 2006 12:31 pm Post subject:

I leave the shematic to the experts, but i just thought of one thing:
I saw a one fet preamp built inside a 1/4" jack in a book.
The pramp cable was using a symetric mic-cable, powering the fet the way it´s done with 48v phantom power (or some different way?) - From the other end of the cable.
Anyway, if space is a problem inside the bass this is possible.

piedwagtail · Posted: Sat Sep 02, 2006 7:57 pm Post subject:

Uncle K,

I take it he is paying for development time costs....let alone construction work per hour and X4 component costs....
and costs for when he returns it for tweaking....(which will be in the multiples)

I now see the Buchla 200e series as actually remarkable value,even if i can't afford it.
In a roundabout way,i'm saying it's hard enough perfecting what you actually want to do,let alone perfecting something you'd rather not being doing. Smile

Robert

Uncle Krunkus · Posted: Mon Sep 04, 2006 2:23 am Post subject:

What's happened to Uncle Jan?
I was sure that he would dive right in and tell me a few ways to improve this idea. Maybe he's away for a few days.

Uncle Krunkus · Posted: Tue Sep 05, 2006 6:05 am Post subject:

Well I've breadboarded the front end of this circuit, and there's problems with it. I took the output off the wiper of the balance pot through a 100nF cap.
It seems to work okay, the balance pot even sounds good, but when you pluck one of the bass strings beyond a certain strength, there is sudden and nasty distortion.
The VU meters on my mixer show that the signal is not very big until the distortion occurs, when they suddenly go off the scale. I tried changing the 56Ks to 20K to reduce the gain just to be sure, but it makes it even more distinct ie; hardly any gain, low levels, until the incoming signal gets past a certain point, and then, wham!
My gut feeling is that this means there's a problem with the impedance matching on the input? yeah? Or on the output? Maybe it just needs the added buffer after the balance/mixer?
Maybe the 100Ks on the input need to go up to 500K or 1M?
Any ideas?

Scott Stites · Posted: Tue Sep 05, 2006 10:48 am Post subject:

Uncle K,

I've refrained from offering any advice, because I've just never built a stompbox intended for anything other than synth level, so I'd be a really poor advice giver kind of guy. Having said that, I think the input impedance should be a lot higher than what you have there? Secondly, I've always been told that the feedback of an opamp doesn't like anything much under 10K, but I could have dreamed that as well.

Anyhoo, at the point you're at, I'd try a divide and conquer - IE, take a look at what's happening on the input - disconnect the output of the balance pot from the rest and see if the voltage surge is happening there. If that's all hunky-dory, I'd move on to the second section and try to figure out what's going on there.

That's my questionable advice for the day Very Happy

Cheerio,
Scott

Uncle Krunkus · Posted: Tue Sep 05, 2006 2:15 pm Post subject:

I went for 100K on the input because that's what Ray used for his guitar trigger circuit. I'll play around with it today, (and add the second half) and see what happens. Thanks Scott.

Uncle Krunkus · Posted: Tue Sep 05, 2006 11:58 pm Post subject:

I worked out that on a single supply version like this the 100Ks (R1&R7) need to go to 1/2V, not ground. I added two 1M resistors to ground before the input caps. I saw this on another circuit but I'm unsure if it's necessary.
Now the horrible distortion is gone, and has been replaced by a fairly soft overdrive when you play the bass hard. I think this is due to impedance matching problems still in the front end. The main thing is that the obvious fault is gone.
I'll re-work and re-post the schem soon.

Uncle Krunkus · Posted: Wed Sep 06, 2006 6:04 pm Post subject:

Here's the updated schem.
I fixed the distortion by changing the feedback resistors to 56K & 20K. I'm not sure why this worked. I'm kinda flying blind, but I do feel I'm learning alot as I go.
So at this point there is an undistorted signal at the wiper of the BAL pot. Slightly raised gain. Very nice.
I'll start looking at the EQ section today.

ParaSchem.gif

Description:

Filesize:

14.23 KB

Viewed:

286 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Uncle Krunkus · Posted: Thu Sep 07, 2006 7:20 am Post subject:

I breadboarded the first EQ section today, and it seems to work okay, except for a problem. With the gain turned up full (gain) and the resonance turned up full, and the frequency dropped down to the bottom, a pluck of a deep string will bring on this uncontrollable feedback/self oscillation. I can only stop it by turning the gain back and disconnecting the power. I assume it's some kind of self oscillation, but I don't know why it's happening.

kkissinger · Posted: Thu Sep 07, 2006 11:50 am Post subject:

Hi. I know just enough to be "dangerous".

I'm just going to hazard a guess just for the fun of it and I will be curious to find out what the solution is.

I would suspect that IC2 is being driven into oscillation because the un-inverted side's output is a little hotter than the inverted side. With the resonance and gain all the way up, a portion of the positive output could backfeed via the two 10K pots (effectively 5K since they are in parallel). If a portion of the positive signal is being backfed to the positive side of the IC, then you risk a latchup or otherwise uncontrolled oscillations.

Perhaps the solution would be as simple as larger values for the 10K pots...

Ok... that is my wild guess. Let me know how it turns out.

v-un-v · Posted: Thu Sep 07, 2006 1:12 pm Post subject:

Andrew, this all sounds perfectly normal and I applaud you for making a circuit that manufacturers go to great lengths to avoid!! Cool

Trimmers? Some internal trimmers to tame the gain I reckon? I think all that's hapenning is oscillation is being induced because of a certain resonant frequency that the circuit is particully prone to everytime that note is plucked.

k's stuff about the uninverted side being too hot sounds like a likely canditate for a couple of trimmers too.

Trimmers! my 2 pence Very Happy

FWIW, Bob Moog once said in an interview that one thing he regretted leaving out of his minimoog design was a fixed filter bank. Three of these parametrics could be a nice little module too Smile

ian-s · Posted: Thu Sep 07, 2006 1:58 pm Post subject:

Someone once told me that the 2nd hardest thing in electronics was getting a circuit to oscillate. The hardest was stopping a circuit from oscillating.

Uncle Krunkus · Posted: Thu Sep 07, 2006 2:20 pm Post subject:

I've only done one channel at this stage, so only one of those 10K gain/cut pots is there.
I think it has something to do with the frequency being turned right down, ie closer to DC. And it does need the initial pluck of a string to kick it along, so maybe it's a byproduct of the EQ section being overdriven or something like that.
I'll play around with it a bit more tonight.

mosc · Posted: Fri Sep 08, 2006 5:52 pm Post subject:

Jan is on his annual vacation. Should be back in a few days more or less.

blue hell · Posted: Sat Sep 16, 2006 5:53 am Post subject:

The changes around R2 and R7 would be the cause for the distortion to disappear. You seem to have mixed up ground and negative supply voltage, when the circuit is fed from a single supply the 1/2 V rail should be considerd to be "the ground" reference, not the "minus" connection of the battery. But anyway, that's fixed now.

Even with seemingly simple circuits it's quite easy to make a fool out of oneself by making any remarks, but anyway the following reflects my thoughts on it - for as far as not already mentioned by others.

The schematics do not show how the 1/2V virtual ground reference was made, so I'll just ask. Did you put in bypass capacitors over the resistive divider ? And what value resistors did you use, are they maybe on the large side ? Mistakes here could well drive circuit into oscillation in a hard to understand way - as AC will go everwhere when the 1/2V signal is not solid or not decoupled for AC properly. Resistors being too large would probably lead to (very) low freuency oscillation (assuming decoupling is present).

I understand the reasons for going with a single battery circuit, but the point above make a good reason to opt for the two battery solution - its easier. Even then there would have to be some electrolytics over the power rails - a batteries AC impedance is not particulary low.

Otherwise I have no direct clues for the circuit being bad, when everything fails maybe an opamp buffer put after the P3 (BAL) pot could make the 2nd stage behave better.

From a quick glance the circuits around IC1-C and D seem to emulate LC filters. This might oscillate for certain control settings (circuits like this seem to be a bit tricky always), but that could/should be easily fixed following Tom's (v-un-v) trimmer suggestions. It should always be possible to go out of oscillation by just turning the "causing" pot back - when otherwise all is OK. I'll assume you copied this from a "reliable" source Very Happy

blue hell · Posted: Sat Sep 16, 2006 6:09 am Post subject:

Uncle Krunkus · Posted: Sat Sep 16, 2006 6:09 am Post subject:

The 1/2V line comes from a divider made of two 10K resistors and a 47uF from -ve(ground) to the 1/2V line. I also put a 100uF across the -ve and +ve (not sure why, just seemed like a good idea at the time Laughing

)

The really weird thing is that turning the "causing" pot back doesn't stop the oscillation. It's like it gets stuck, and I have to disconnect the power (for a second or two)

I'm thinking I should check what's going on with the spectrum analyser. There's a good chance that the theoretical centre frequency goes down to DC when the oscillations start. Which I obviously don't need.

blue hell · Posted: Sat Sep 16, 2006 7:14 am Post subject:

Maybe the 47 uF is a bit on the low side, I 'd try 470 uF maybe with some 100 nF or so decoupling capacitor in parallel for better HF decoupling.

Or maybe even better, not as a permanent fix but as a nexperiment, try to operate the circuit from a good +/gnd/- power source.

All of this just to make really sure the 1/2V line is not causing the trouble. Later on you could always try to make the 1/2V line harder by throwing in an extra opamp to buffer it.

Uncle Krunkus · Posted: Sat Sep 16, 2006 7:33 am Post subject:

Just to correct what I said before.
I checked it out just then, and turning the gain back does stop the oscillations. But even with it turned back, playing hard intermittently starts it up again. It actually sounds like an octaver which has got dodgy tracking problems. Like a square wave layered over the signal which comes and goes.
The ground connection does go to the -ve battery connection yeah? Not the 1/2V line? Cos in a +/gnd/- setup the ground for the input and output go to the middle which would be the same as the 1/2V line. Does that make sense?

I'm starting to think this whole pre-amp/EQ on the bass is one extra project I don't need to get sidetracked with at the moment.

blue hell · Posted: Sat Sep 16, 2006 8:41 am Post subject:

Just to be sure ... from a +/gnd/- supply the + should go to +ve, the gnd to 1/2V and the - to -ve. Where +ve corrsponds to battery + and -ve to battery -.

For a test there would be no need to leave in the resistive divider - that one just draws some extra extra current then for no purpuse nor harm.

And the other thing that came to mind ... I do hope you connected the inputs between the input terminal and 1/2V ? As well as the output terminal being between the output terminal and 1/2V ?

Again, the 1/2V line should be considered to be the ground in the circuit, do not connect ground lines to -ve, everything is referenced to the 1/2V line.

Uncle Krunkus · Posted: Sat Sep 16, 2006 3:31 pm Post subject:

That's exactly the point I was making!
I've had the input and output referenced back to 0V!
Oh, thankyou Jan! You da man!
It all seems so obvious now. I was worrying that I hadn't properly gotten my head around the derived dual supply stuff, and I was right! Laughing

It's this fundamental op-amp stuff which I knew you would clear up! Smile

blue hell · Posted: Sat Sep 16, 2006 3:51 pm Post subject:

Ok, but does it work now ?