Ross Compressor Question?

[modsquad]

Is it normal for the Ross Compressor to begin to distort as you pass about 3 o'clock positin on the level.   I made one for somone that uses it for a clean boost also.  He complains that the level is too low and has to increase the level control to make it do what he wants.

[Plinky]

I just got my Ross clone working this week and it's pretty quiet no matter where the level or sustain is set. Have you checked voltages and compared them with the reference voltages on the Fuzz Central site? http://fuzzcentral.ssguitar.com/ross.php
That would probably be a good starting point.

[petemoore]

Is it normal for the Ross Compressor to begin to distort as you pass about 3 o'clock positin on the level.
  IIRC the level control is a variable voltage divider on the output..the input and circuit see the same input levels no matter where the output level is set. The compression should be relatively clean, with the level control simply attenuating output.
  It is 'normal' [not necessarily desirable though] for anything that boosts signal output to cause distortion to occur in anything it is input to, such as an amp or other circuit.
  Testing the voltages on compressor is generally a good practice IMO.
 
 

[Pushtone]

there are a ton of posts about the Ross Comp distorting.

Here are a couple. There are many more.

http://www.diystompboxes.com/smfforum/index.php?topic=4920.msg28297#msg28297

http://www.diystompboxes.com/smfforum/index.php?topic=20405.msg125742#msg125742

http://www.diystompboxes.com/smfforum/index.php?topic=49112.0

http://www.diystompboxes.com/smfforum/index.php?topic=61830.0

The easy solution is to turn down the guitar volume control and make it up with the Ross level control.
First time I plugged mine in it threw me to.

But it does sound good.
Good and dirty.

[Mark Hammer]

The CA3080 is known to have a low tolerance for high input signals.  So known, in fact, that any circuits using it generally attenuate the input signal before it reaches the 3080.  Having said that, any attenuation provided is merely a proportional reduction (e.g., cut by XX%) of the input signal, not an absolute one (i.e., reduce to this average level no matter what), so it is theoretically possible that an input signal CAN push the CA3080 hard enough to distort.

What people often hear as "distortion", however, can often be envelope ripple in the rectifier circuit.     That is, the averaging out of the input signal amplitude to provide a smooth control signal proportional to the input is not as smooth as it might be, and small "wiggles" in that control signal adjust the gain in the 3080 up and down at rates which are audible.  It sounds like a buzz or Scottish burr, and is generally most pronounced or noticeable during the final decay of a held/sustained chord/note.

Smooth compressor sound is akin to getting good octave fuzz in a number of octave-up units.  If there are noticable discrepancies in the component values used for each complementary half of the rectifier circuit, there is greater risk of ripple and less likelihood of good octaving.

[JDoyle]

The Level pot should only adjust the actual output level of the unit, as Pete said. If it is the cause of the distortion, then I would look at what is following the comp because it sounds to me like the Ross is overdriving the following input. That is the only reason I can think of for the level pot to induce distortion. If it is properly hooked up...

Also, I am sceptical of the posted reasoning behind previous posts relating to why it is distorting. I guess not sceptical, more thinking that there are two types of distortion that can occur here, overloading the input and having a problem with the circuit. Yes, on signals above 0.6V the 3080 starts to distort. This is because the input signal is starting to bring the transistors in the differential input into the range that the transistors in the current mirror operate. The current mirror is directly attached to the Iabc port (pin 5) and is what adjusts the gm of the chip by adjusting the current through the differential pair.

But you REALLY have to nail a 3080 with a large signal, I would say 2Vp-p, before you start to hear what I would consider 'distortion', it resembles that of diode clipping - it is extremely obvious. Up until that point there is definite squashing of the signal, but it's resultant distortion is (in my opinion) no worse than the natural distortion that occurs when a compressor does what it does. Remember, the compressor is essentially distorting the input anyway. Also, even if your guitar were to put out a 2Vp-p signal, it wouldn't be that large for very long because of natural decay AND the compressor wouldn't be boosting that signal; it's boost wouldn't kick in until the signal started to die away and was considerably smaller, smaller enough to not come close to overloading the input.

SO: if the distortion occurs at the start of a picked note (directly from the guitar, not boosted in any way) and then disappears as the compressor kicks in and boosts the decaying signal, then yes, you are overloading the 3080s input. BUT if the distortion occurs throughout the entire note, including when the boost starts to kick in, then something else is wrong because after the initial attack, the signal is going to be significantly smaller when the comp starts to boost it and wouldn't be enough to overdrive the inputs.

I hope this makes sense.

I have made quite a few circuits based on the Ross with 3080s and while I have gotten distortion when I was trying to, I have never had a problem with it when using a straight Ross comp. But I play a Strat.

Check your voltages and make sure that the trim is set properly (with guitar vol on 0 and the comp on, adjust the trim until you hear the MOST noise, counterintuitive but exactly how it is supposed to work).

Regards,

Jay Doyle

[Mark Hammer]

Your arguments are precisely the reason why I said "theoretically" possible.   Like you, I suspect anything heard as distortion in this instance is NOT from overload.

And while envelope ripple does get "heard" as distortion, this is less likely to occur with full-wave rectification than with half-wave rectification, and since this unit is a FWR circuit, the odds are not high that's what it is.  More than likely, the problem, if indeed there is one, exists somewhere else in the circuit.  The fact that it appears to occur with adjustments to the very last component before the output jack also suggest something else.

I might point out, as an aside, that the distiortion one gets from a 3080 is not unpleasant.

[markm]

Check your voltages and make sure that the trim is set properly (with guitar vol on 0 and the comp on, adjust the trim until you hear the MOST noise, counterintuitive but exactly how it is supposed to work).

The trim pot adjustment throws alot of people off as to how it should be adjusted.

[JDoyle]

What people often hear as "distortion", however, can often be envelope ripple in the rectifier circuit.     That is, the averaging out of the input signal amplitude to provide a smooth control signal proportional to the input is not as smooth as it might be, and small "wiggles" in that control signal adjust the gain in the 3080 up and down at rates which are audible.  It sounds like a buzz or Scottish burr, and is generally most pronounced or noticeable during the final decay of a held/sustained chord/note.

Smooth compressor sound is akin to getting good octave fuzz in a number of octave-up units.  If there are noticable discrepancies in the component values used for each complementary half of the rectifier circuit, there is greater risk of ripple and less likelihood of good octaving.

Mark, these are actually quite different and I don't think this applies to the Ross Comp. In an octave up box the goal is to double the frequency of the input signal, to do that you rectify the signal. In that case the rectifying cap is relatively quite small because you just want to smooth the ripple without attenuating the signal itself. In the comp, the rectifying cap is very large in comparison because it's goal isn't to do anything to the signal itself but to adjust the gain of the 3080 inversely proportional to the strength of the input signal.

Additionally, the FWR of the Ross is really quick to respond to large signals (by dumping the charge on the rect. cap through the parallel b-e paths of the two rectifier transistors, which is about 75 ohms) but A LOT slower when charging back up as the signal strength decreases (through the shared 150k resistor). So - the cap really doesn't have all that much time to charge through the 150k resistor (when the input signal is swinging in the negative direction) before it is dumped out through the b-e junctions (when the input signal is swinging positive).

In other words, what would normally be heard as 'pumping' or 'breathing' in a octave up box can not happen in the Ross Comp FWR because the charge/discharge paths of the rectifying cap have different impedences and therefore charge/discharge the cap at different rates. It would sound like someone taking an extremely quick, sharp breath followed by an even more extreme long exhale.

Again, I hope this makes sense.

Regards,

Jay Doyle   

PS - I am a big believer that the 3080 is the most underutilized distortion block. It's distortion is smooth, input voltage dependant, changes over to diode-like clipping at high levels, and can be made assymetric by adjusting the bias away from 1/2V+. There is A LOT of experimentation that can be done with it. Too bad the 3080 is out of production and it's price has increased by a power of 10. Luckily it is pretty easy to make one discretely with the same response, if not the same freq. response (which is excessive for our use anyway).

[markm]

Too bad the 3080 is out of production and it's price has increased by a power of 10. Luckily it is pretty easy to make one discretely with the same response, if not the same freq. response (which is excessive for our use anyway).

You don't happen to have a discreet 3080 design that you could share do you J? 

[Mark Hammer]

Makes perfect sense, Jay.  More sense than what I said.

What I was trying to get across was that when the two complementary halves of a phase-splitter-type octave-fuzz are not well-matched the octave is not nearly as prominent.  When the two complementary halves of a FWR circuit like that used in this application are not roughly equivalent, its ripple behaviour starts to approximate that of a SWR.  The common thread being that balancing the two complementary halves of the rectification is important to yielding the optimum sound for each application, even though a number of parameters and sonic goals are obviously very different.

Too bad the 3080 is out of production and it's price has increased by a power of 10. Luckily it is pretty easy to make one discretely with the same response, if not the same freq. response (which is excessive for our use anyway).

You don't happen to have a discreet 3080 design that you could share do you J? 

Check for discrete VCA circuits on any of the modular analog synth sites.  can't give you a link right this second, but there are plenty around. Which ones most closely approximate the 3080's response I will leaveup to Mr. Doyle to identify.

[markm]

That would be great if he would! 

[JDoyle]

The common thread being that balancing the two complementary halves of the rectification is important to yielding the optimum sound for each application, even though a number of parameters and sonic goals are obviously very different.

OK. Mark, I promise, I'm really not trying to be argumentative here... 

In the Ross/Dyna circuit, I don't believe that there is a need to worry about the balance of the FWR. The only way that it could become a factor is if one of the transistors in the FWR had abnormally small gain or was just faulty, and I don't even know if that would matter because the decay is so slow and the attack so fast. The way those are set up, with EXTREMELY high gain so that they slam off with the presense of any signal, along with the extremely slow decay time in the integrator cap, in itself is going to balance the rectified signal, or at the very least obviate the need for any balance.

While both octave up boxes and comps work with FWR circuits, they are used in different ways. In the octave up box, it is used to directly alter the input signal, and any integrator cap used is there to smoth out the resultant signal equally for both the upward and downward swings of the signal, so balance between the two halves is EXTREMELY important as you said, in the compressor, the rectified signal is used as a control voltage, not the output voltage itself, and therefore the upward and downward swings of the rectified signal have different functions and aren't part of the output signal. Upward corresponds to the attack of the control signal while downward corresponds to the decay.

In an octave box the attack and decay are equal to ensure that the positive and negative swings of the input signal are reproduced as close as possible (in fact there really ISN'T an attack or decay, it's just the output signal being cleaned of the excessive highs that occur at the sharp 'flip' point in the rectified signal), whereas in the comp we want to CONTROL those factors (the upward and downward direction, and the speed it goes in those directions) to give us adjustable attack and decay or at least different attack and decay times.

[I haven't tried it, but I bet that even if you took one of the transistors out of the FWR in the Ross circuit, you would have nearly an identical response with just a tad longer attack time (which is so short to begin with that I doubt you would notice) or the OTA will be driven to a higher minimum gain with the input of a mid-sized signal because the cap doesn't have  time to completely discharge before the control voltage is back at 9V for the missing half of the rectified signal. But I really don't think that it would affect the decay at all.]

Again, hope this makes sense.

Regards,

Jay

[JDoyle]

Too bad the 3080 is out of production and it's price has increased by a power of 10. Luckily it is pretty easy to make one discretely with the same response, if not the same freq. response (which is excessive for our use anyway).

You don't happen to have a discreet 3080 design that you could share do you J? 

I have a discrete OTA design. I mentioned that as part of my comments on distortion with the 3080. It is pretty easy to build a discrete OTA with the same overload/distortion response as a 3080, getting the rest of the properties to approach those of the 3080 is a lot harder, requires matched transistors (both NPN and PNP), and some really well matched resistors (better than 1%).

For distortion purposes: take a simple discrete op amp set up with a gain of 1, and add another differential amplifier in front of it, the collectors of each transistor in the new diff amp are connected directly to the bases of the transistors in the discrete opamp (the + and - inputs just flipped by doing this, so that the base of the transistor in the new diff amp whose collector is connected to the + input of the discrete opamp is now the - input for the OTA). THEN replace the resistor connected to both emitters in the diff amp with a transistor, the collector connects to the emitters of the two transistors and it's emitter is directly connected to ground. The base of this new transistor is the Iabc port for our new OTA.  Enjoy! (the control voltage into the Iabc port has to be at least 0.6V above ground or the OTA shuts off).

Jay Doyle