I've always wanted a Ross compressor, and finally I've made myself one, using the GeoFex layout (as hosted on GGG: http://www.generalguitargadgets.com/diagrams/geo_d&r_comp.pdf).

I've used the GGG recommended modifications:

    *  C2 should be value of 220pF, the parts list lists it as 0.0022uF (2200pF).
    * C11 should be value 0.001uF instead of 0.01uF
    * C12 should be 0.01uF instead of 0.1uf

As for further modifications, I've used the BC550C transistors (the closest approximation available here). I've soldered everything as in the GeoFex project file:



The problem is, I get little or no sustain, only volume boost. As far as I can see in the voltages, pin 6 seems to be the one most "off":

Sustain at "0", jack in input. Power to PCB - 8,98v (using a battery)
IC - 1 - 0v
      2 - 4,65v
      3 - 4,64v
      4 - 0v
      5 - 0.58v
      6 - 1,66v
      7 - 8,97v
      8 - 0v
Transistors (transistor marked as on this picture):



Q1 - C - 8,52v
       B - 0,94v
       E - 0,44v
Q2 - C - 7.88v
       B - 1,66v
       E - 1,13v
Q3 - C - 8,52v
       B - 0,23v
       E - 0v
Q4 - C - 8,51v
       B - 0,23v
       E - 0v
Q5 - C - 8,95v
       B - 8,51v
       E - 8.10v

Comparing this to, say, FuzzCentral voltages is a bit over my head... but I'd really love to figure out what's wrong, as I love the sound of Ross compressor. Could I ask for your help?

Your description and the voltages seem to indicate that the current that should be causing the gain of the 3080 to rise is not getting to pin 5. But it's not as simple as an open circuit to pin 5, or pin 5 would not be at 0.58V. Can you tell us the voltage on each end of R12 and the value of R12?

Sure, thanks a lot for your help, and thanks for a great work with Ross comp layout and explanations!



(In case that the picture won't display: R12 is 26,9 kOhm, voltage going to IC pin 5 from R12 is 0,58v, voltage coming from "Sustain" pot to R12 is 0,97v.)

You're welcome - now I just hope we can get yours going.

The voltages for R12 let me do two things: verify that there is some current coming through, and compute how much.

The current is (0.97V-0.58V)/26,900 = 14.9uA. That's 'way small for it not having any signal in it. But there is some current there, so that means that the sustain pot - or SOMETHING - is connected to R12, so there's not a broken wire to it. Does the voltage on the high end of R12 vary when you rotate the sustain pot?

What I suspect is that there may be something wrong with the sustain pot. You show the emitter of Q5 at 8.1V, which is where it should be with no signal. The 14.5uA value works with 8.1V and 0.97V to give 491K for the actual sustain pot value, which is about right for the 500K sustain pot, but that value should change as you rotate the pot.

Your mention of Pin 6 as being 'off' is a clue - though the output of the OTA isn't at fault (especially as it doesn't have a voltage, just a current, you bias it to whatever voltage you want).

What IS wrong is the bias voltage.

You show the V+ supply as being 9v (or close enough to call it that). The bias supply is a 56k in series with a 27k and the bias is taken from their junction. This works out to:

27k/(27k)+(56k) = 27k/83k = .325 or 32.5% of the voltage across the two resistors appears at the junction: There should be in the neighborhood of 3 volts as the bias supply. What this means is the bases of Q1 and Q2 should be biased to the 3V bias supply, leaving the emitter one diode drop below that, or around 2.4V. And with the input buffer and the output buffer/phase splitter, each having equal collector and emitter resistors, if 2.4V is dropped across the emitter resistor, and remembering that any current out of the emitter is nearly the same current into the collector, with the same resistor as well as the same current at the collector as at the emitter, the voltage drop will therefore be the same though in opposite directions, leaving the bias at the collector to be 6.6 (giving a phase splitter stage nearly the most headroom that you can get - Q1 just uses the same bias supply to save cost but it is set up the same way).

BUT - in your situation the bias supply is about half of what it is supposed to be - which means the upper of the two bias resistors is most likely too high a value (something COULD be sucking too much current, but as your supply is nearly 9V, and as R.G. pointed out, just a trickle of current is going into the Iabc port of the OTA, I don't think that is the case). I did some quick math and if you put in a 100k resistor instead of a 56k resistor you would get:

27k/27k+100k=27k/127k= 0.2125 or 21.25% of the bias supply which is around 1.9V - which is close to your bias voltage.

BUT if you happened to have used a 20% tolerance carbon comp resistor for the 100k and it was at the top of the tolerance range, its actual value was 120k:

27k/27k+120k=27k/147k= .1836 or 18.36% of the bias supply which is: 1.653 when the battery is 9V. OR - damn near what you are getting.

Which all makes sense in your case - if the bias supply for the OTA's inputs is too low a current value, the gain of the OTA is lowered dramatically and the sustain is effectively removed because even the lowest setting of the sustain pot, the current mirror that makes up the 'tail' of the input's differential amplifier/long tailed pair will always sink more current than the diff. amp can let through - meaning changing the current through the tail doesn't change the gain - which is key to getting the circuit to work.

Another clue is that Q1 and Q2 are biased just BARELY on, there is only about a 0.5V drop across the base to emitter resistors which says that there is enough voltage to turn the diode on, but not enough current to turn it on very hard. And because the entire input network of the Ross/Dyna circuit is there to REDUCE the signal level so as not to overdrive the 3080, by the time it hits the OTA inputs, what's left of your signal in combination with the lowered gain of the input circuit of the OTA, isn't even enough to clip the OTA's output. Thus, you have gain, no sustain, and most importantly - no, or barely any noticeble, clipping. And you would clip the HELL out of the signal if the OTA stage was running full bore:

The gain of an OTA stage is:

(gm) * Rload,

Where 'gm' means transconductance and is equal to: (19.2) * (Iabc).

Where Iabc is the 'Amplifier Bias Current' and is equal to the current in to pin 5 of the OTA (or the current through the 27k resistor you measured for R.G.).

So given the values you mentioned and R.G. worked out -

If there is 0.0000149 of current into the Iabc port:

The gm = 19.2 * (0.0000149) = 0.00028608 (which I'll call 0.00029 for the sake of our brains), this means that the transconductance of the stage as you have it is: 0.29 milliSiemens*

If we plug that number into the gain equation for an OTA:

Gain = (gm) * (Rload) = (0.00029) * (150,000) = 43.5

Which would seem to be a lot of gain and would clip, but considering that the input network divides your signal down considerably, and combine that with the reduction in gain from two emitter followers running at a low current (which raises their r_e, lowering the gain through voltage division with the emitter resistor), and one can easily see how a gain of 43.5 could end up not having all that much effect.

So, R.G. and I have offered two easy things to check:

1 - The values of the bias supply resistors
2 - The Sustain pot

I suggest doing my idea first if only because it is a lot easier to translate the color codes on the resistors than it is to test the pot itself, PLUS - if I'm wrong, R.G. is almost certainly right. But I'm pretty sure it's the bias supply.

Good luck and let us know if you get it working!

Regards,

Jay Doyle


* They changed the unit of measure from umhos ('ohm' backwards) to Siemens, which, consistent for the semiconductor world, makes abosolutely no sense; what exactly IS a Siemen??? The inverse of resistance, noted by spelling 'ohm' backwards, which is EXACTLY what transconductance is, makes so much more sense - but why would we want anything to make sense? We want to FORCE people to spend the money on school just so they can learn all of the nomenclature...

Excellent post, JDoyle! Here are my measurements at R13/R15 junction and after that junction (I hope that the picture is visible):



As for values, R15 is 27k (red-pink-black-red-brown), R13 is 56k (green-blue-black-red-brown), R4 and R3 are 470k (yellow-pink-black-orange-brown).

How do those voltages look? Do they suggest any fault - a bad element, perhaps?

As for the "Sustain" pot - it looks OK (tests at 490-something ohms), but it does not perform as I have expected, voltage-wise: at minimal sustain (pot turned fully counter-clockwise), the voltage on pin 6 is 1,66v. At maximal sustain (pot turned fully clockwise), the voltage on pin 6 is 1,07v. I thought that (in Ross compressor build) maximal sustain would produce a higher voltage on pin 6? I think I will replace it tomorrow and see what happens.

I think your sustain pot is wired backwards, but that's not the whole problem.

There is something wrong with R3/Q1. R3 is a bias resistor that should have quite low voltage across it. It's dropping almost a volt. What's the value of R3? Are there any shorts, solder problems, etc, around R3 or Q1?

Quote from: R.G. on June 02, 2008, 06:10:17 PM
I think your sustain pot is wired backwards, but that's not the whole problem.

Ah, it is quite a possible scenario with me... It's wired like this one:

Quote from: R.G. on June 02, 2008, 06:10:17 PM
There is something wrong with R3/Q1. R3 is a bias resistor that should have quite low voltage across it. It's dropping almost a volt. What's the value of R3? Are there any shorts, solder problems, etc, around R3 or Q1?

R3 is 470k (yellow-pink-black-orange-brown). It also reads on the multimeter as 470k, as well as R4. I've reheated the connections, as well as Q1's connections (as well as several connections around Q1, R3 and R4), and - still no change in readings.

Since both R3 and R4 are dropping a volt, are they both suspicious, or is R4 supposed to drop a volt because of its connection with C4?

QuoteSince both R3 and R4 are dropping a volt, are they both suspicious, or is R4 supposed to drop a volt because of its connection with C4?

Good work. You're going to be telling other people how to debug their pedals some day.

We're almost there. R3 and R4 are supposed to have a drop of maybe 0.2V total across them. They're there to provide the base current for Q1 from the bias voltage. The voltage on R5 should be about 2V; that takes 200uA (2V/10k) to do. Q1 should have a current gain of 400 or so, meaning that it needs only 200uA/300 = 0.5uA of base current (all this is very rough estimation mind you.) So with 0.6uA though them, R3+R4 should only have a voltage drop of V = (470K +470K)*0.5uA = 0.47V or 0.235V each. Instead, there is (2.74-1.80)/470k = 2uA going through them, and the voltage at the base of Q1 is sagging badly.

There are only a few possibilities. If the circuit is as shown in the schematic (i.e. no shorts or opens, and the resistors within tolerance) then Q1 has to be too low on gain by a bunch. It's gain would be 100 or so. Looking at NXP's datasheet shows a typical hfe at 200uA of over 400. So one possibility is a bad - or heat damaged - transistor. The datasheet also shows that the pinout of the BC550C is the exact reverse of the US-style 2N numbers, so it is possible that you may have put it in backwards; that would match the printed outline on the parts placement diagram.

You have measured R3 and R4. A possible failure is that R5 is really 1K, not 10K. That's worth checking. Finallly, if there is not the same voltage on both sides of R2, then current is leaking out the side of R2 that's supposedly blocked by a capacitor. Voltage across R2 is worth checking too.

R.G. has the list of what you want to do, though I have a feeling that Q1 is your problem - if it was damaged or in backwards you would have a forward biased diode that your signal is going through, which has quite a low impedence.

You wouldn't get the typical distortion one gets from a diode because the inputs of the OTA are also, essentially, diodes - the 13600, the next gen. dual OTA that came after the 3080, has 'linearizing' diodes that have a function similar to what your Q1 'diode' could be doing and actually extending the linear range of the OTA! Basically by 'predistorting' the signal, when the signal hits the OTA input and IT distorts, it distorts in the same, or very similar way, counteracting the diode distortion created earlier. However, you don't want this for the Ross Comp., it severely reduces the gain by reducing the input signal, and you need all the gain you can get to have the Ross work right... [It also increases noise and really harshens the sound of overdriving an OTA, I really don't like using the diodes]

After you check the resistors that R.G. mentioned, double check the pin out and replace Q1 with a spanking new transistor and see what happens...

Good luck!

On the USA pinout, base in the middle, reversing the transistor will make the transistor still work but in inverted mode. The "hfe" in this mode is pitifully small. That would account for the apparent low gain.

Chances are, if Q1 is backwards, they all are. At least that's how I would do it... 

Quote from: R.G. on June 03, 2008, 12:23:57 PM
On the USA pinout, base in the middle, reversing the transistor will make the transistor still work but in inverted mode. The "hfe" in this mode is pitifully small. That would account for the apparent low gain.

Chances are, if Q1 is backwards, they all are. At least that's how I would do it... 

OK, got it! And I'm right there with you on screwing the whole lot if I screwed one!

But isn't that still a lot of current for an inverted mode setup? Not that I've ever used a Q like that and have any idea what the readings would be...

But you are right, that would definitely do it!

Quote from: R.G. on June 03, 2008, 12:23:57 PM
Chances are, if Q1 is backwards, they all are. At least that's how I would do it... 

They all were backwards.  I feel like an idiot - when choosing the transistors, I was looking for low noise ones, and completely forgot to check the pinouts! I just followed the layout - "flat side of the transistor towards the right part of the PCB". That should teach me a lesson for future builds...

I didn't fry anything, did I? I've replaced all of them, should I check any other element?

So, it finally does that legendary squish! I loved my Boss CS-3 (unmodded), but always found it a bit "shallow". For me, Ross goes in that "bright and deep" category, and most certainly has "that" sound.

And now to check if everything works as it should. The new voltages are:

Sustain at "0", jack in input. Power to PCB - 8,98v (using a battery)
IC - 1 - 0v
      2 - 4,65v
      3 - 4,65v
      4 - 0v
      5 - 0.58v
      6 - 2,78v ("Sustain" on max gives 2.42v)
      7 - 8,97v
      8 - 0v
Transistors (transistor marked as on previously posted picture):

Q1 - C - 6,96v
       B - 2,38v
       E - 1,99v
Q2 - C - 6,72v
       B - 2,77v
       E - 2,23v
Q3 - C - 8,82v
       B - 0v
       E - 0v
Q4 - C - 8,82v
       B - 0v
       E - 0v
Q5 - C - 8,94v
       B - 8,81v
       E - 8.42v

How do the voltages look? And, countless thanks, gentlemen!

Here's how my Ross looks like. It's a WIMA & metal film build, I wanted it to be as good and as quiet as it could be. OTA is LM3080N (I've got an CA3080E as a backup, I just hoped LM3080N is a bit quieter). The box is Hammond 1590S - it's a bit taller & smaller than a 1590BB. As for the name, I didn't want to pass it as an AnalogMan's creation, I just wanted it to be easily readable and recognizable

(Sorry, doublepost!)

(Sorry, doublepost!)

Congrats man! R.G. certainly can call it.

Your voltages look fine - and if it sounds good, it is fine.

Swap out the LM with the CA and see if you get a reduction in noise, you may, you may not. And if you can find a CA3080AE, that is your best bet...

(Unless you want to do it discrete... )

Again man, Congrats! The Ross Comp is what got me into this hobby in the first place...

Regards,

Jay Doyle

Quote from: Rocket Roll on June 03, 2008, 02:00:15 PM
They all were backwards.  I feel like an idiot - when choosing the transistors, I was looking for low noise ones, and completely forgot to check the pinouts! I just followed the layout - "flat side of the transistor towards the right part of the PCB". That should teach me a lesson for future builds...

Oh yeah....NONE of us have EVER done anything like THAT.  Oh no.  Not me....never....really.......seriously......um, can I go now?

Congrats.  You've been baptized my friend.

Quote from: Rocket Roll on June 03, 2008, 02:00:15 PM
They all were backwards.  I feel like an idiot - when choosing the transistors, I was looking for low noise ones, and completely forgot to check the pinouts! I just followed the layout - "flat side of the transistor towards the right part of the PCB". That should teach me a lesson for future builds...

Not to worry. I meant it when I said that if I got one backwards, I'd get them all backwards. 

Here's a better one: back when I was using sockets for ICs, I spent a couple of pretty full days debugging a hack-wired board from the bottom (where I could get at all the connections) only to find that I had not stuffed the ICs into the sockets on the top side before I fired it up. My face was red for a long time after that one. 

I meant what I said a few posts back. You have the right attitudes.

Learning is always expensive, no matter how you pay for it.

Quote from: R.G. on June 03, 2008, 04:18:31 PM
I meant what I said a few posts back. You have the right attitudes.

Learning is always expensive, no matter how you pay for it.

Thank you, that means a lot to me, especially when it comes from you! I'm trying to be as careful as possible, and I care a lot about (and really enjoy in) proper DIYing. Plus, I think it's much better than being a "common consumer". So, I'm learning...

If I had the chance to play a Ross, I'd be all over this project much sooner! It's the sound, literally. Doublestops and pedal bends sound just like they should.

My personal embarassment - two days of debugging, with time off to calm down in between. I had no idea why I wasn't getting ANYTHING from the circuit...

Until I realized I never hooked up a battery to power it...