EA Tremolo. Distortion problem.

[ZeusMalt]

New to the forum, new to DIY, hi y'all

So the problem is distorted sound in EA trem. I've seen couple of threads concearning the matter, but i'd like to get some conversation on the matter and how to solve it. I used the ROG version that can be seen here: http://runoffgroove.com/eatremolo.html

Here's what i found (or think i did):
The distortion comes from the Q2. In some thread the problem was solved by changing the 100k vol pot to 100k resistor, replacing the 1M resistor (trim-Q1) with 1M pot:
lug1-trim
lug2-0.1u cap from input
lug3-Q1
--> decreasing the signal at Q1?.

So here's what I'm thinking. As i already have the 100K pot assembled, can I keep that as the volume and change the 1M resistor to 1M pot as mentioned and just adjust that enough to loose the dist? That way i wouldn't have to resolder the 100k pot and if I wanted to i could use the pedal as a booster by removing the Q2 and adjusting the trimmer to match the original layout and get the "full potential" out of the Q1 as a booster 'cause that doesn't get distorted? ... ... Someone understand what i'm trying to say?

[ZeusMalt]

Hello me. You just tried out putting that 1M trimmer and here's what you found out: it really does help. To a point, where the resistance starts cutting ur precious tone.
Conclusion. The ea tremolo - rog version (at least for me) is very "hot" and breaks up when hitting the strings harder. After adding the trimmer the little breaks were actually quite cool.
You can get rid of them for good by using a volumepedal or a compressor before the tremolo and lower the signal a bit before it and then using the tremolos vol to boost it up again.

[kaycee]

What you can do is put the volume pot on the front end instead of the back, it has more or less the same function there and can stop the clipping IIRC

[ZeusMalt]

how? Compared to what i did and what i told... What pot and where?

[GGBB]

I think what kaycee is referring to is to replace the 1M pulldown resistor from the input with a 1M pot and take the input to the 0.1uF cap off of the wiper.  This is a known EA Tremolo mod.  You can then replace the 100K volume pot with a 100K resistor, and take your output from the junction of that resistor with the 0.47uF cap.  Or leave the 100K pot so you have both pre and post volume controls.

There are probably more appropriate ways to get rid of the distortion by lowering the "gains" of the FETs.  I'm not sure how to do that but I suspect it involves changing the 4.7k, 180, and/or 1.2k resistors.

[kaycee]

Nope, although all of what GGBB says may well be applicable as well

Simply take off the volume pot and add it to the front end as a pre-gain type control. Then you can limit the signal going in and hopefully stop the clipping. You'll still need the 1M pulldown on the front, otherwise the 'open' cap will most likely give you a pop on switching. You may also need to add a resistor to ground after the 47n on the output - maybe 150k

Its quite a while since I played around with one of these, last one I wired up with a momentary as well as bypass, but was getting bleed through, its approaching the top of the debug pile again

[ZeusMalt]

Hmm... You could also match a resistor after 1M and input cap? Volume to the max, and then add a R to get rid of the distortion. I'll try that out! I'll socket the R and test...

[GGBB]

Simply take off the volume pot and add it to the front end as a pre-gain type control. Then you can limit the signal going in and hopefully stop the clipping. You'll still need the 1M pulldown on the front, otherwise the 'open' cap will most likely give you a pop on switching.

If you mean connect the pot in parallel with the 1M pulldown, with signal coming out of the wiper, then this is functionally no different than what I described and Zeus has already tried.  The additional 1M pulldown would be unnecessary in this situation since when in parallel with the pot it would only serve to reduce the total "pulldown" resistance.  The pot itself alone would actually function as a pulldown resistor.

If you meant something else then please elaborate.

You may also need to add a resistor to ground after the 47n on the output - maybe 150k.

Actually if you remove the output volume pot, you should definitely replace it with a resistor otherwise there my be some tone loss/change.  The 470n cap and 100K pot function as a high pass filter at about 3Hz.  Without the pot or a substitute resistor, that filter changes which could result in a tonal shift of some sort.  The resistance of the pot will also have some interaction with whatever is connected after the pedal, so again changing that resistance will have some tonal effect.  If you don't want to risk changing the tone, stick with 100K.

My ROG EA Tremolo works and sounds great - the only thing I did differently with mine was the depth control and a few changes to add a second rate control.

I will also mention that on mine, unity gain is just above half way on the volume control (which is audio taper so that's really only about 10-15% up in resistance), so if I were to ditch the volume control and take output at full level, that would be quite a bit of boost. which would require a fair bit of input trimming to compensate.  I'm just guessing here, but I think that moving the volume control to the front is not a good idea in this circuit since the reduced input signal would end up at the output carrying more noise proportionally than when you apply gain to the full input signal and reduce the overall level at the output including the noise.  This circuit is not exactly whisper quiet, so I would expect to find a fair bit more noise with the volume control on the input.  But if it works for you, go for it.  Another option would be to replace the output volume with two resistors in series taking the signal from the middle so that you reduce the signal a little at the output and not have to turn it down so much at the input.  Something like 68K+33K.

Hmm... You could also match a resistor after 1M and input cap? Volume to the max, and then add a R to get rid of the distortion. I'll try that out! I'll socket the R and test...

Yes - if you're finding that you need to turn it down a lot to get to unity gain, but when you turn it down just a little it sounds okay just too much volume, your idea might work.  Another thing to try is a small amount of series resistance on the input - maybe 1-10K.

All that being said, I still think the better approach would be to lower the gain of the FETs.  Try changing 4k7 to 10k.

[ZeusMalt]

.... Duh. I'll definetly do that. That's exactly what's needed.
I study electricity, but not electronics so i lack knowledge on electronics. Should have thought about that but i found a thread where someone had messed with the input and.....
Thanks!

[GGBB]

I was curious about this too so I did a little digging and cam across the AMZ Mosfet Booster.  Judging from that, I think if you raise the 180 (and lower 1k2) you will get less gain.  I think a small resistor in series with the 22uF cap might also work.

[ZeusMalt]

I put sockets on the 4k7 and found out that changing it to a SMALLER value got rid of the distortion. Equals to same thing that you came up? Reduces the potential difference over fet or something?
To the next question... U said something about depth-mod? What did you do and do you know how to change the depths starting point so that it's more transparent in the beginning? So less depth when the knob is all the way down?

[GGBB]

I put sockets on the 4k7 and found out that changing it to a SMALLER value got rid of the distortion. Equals to same thing that you came up? Reduces the potential difference over fet or something?

Excellent - what value did you end up with?  Yes - something like that  I was pretty sure one of those would be the ticket.

To the next question... U said something about depth-mod? What did you do and do you know how to change the depths starting point so that it's more transparent in the beginning? So less depth when the knob is all the way down?

Depth mods are easy. Shallower depth at the bottom is achieved by lowering the 68k. if you want the whole range to be shallower, lower 68k and 120k.  If you want it to be deeper, raise 68k and 120k.  If you want a wider range overall, lower 68K and raise 120K.  You get the picture - the 68k sets the min depth, and the 120k sets the max depth, both values relative to the overall total resistance of the 68K + the 120K + the 250K depth pot.  In my pedal the depth pot is implemented as a variable resistance instead of as a divider. so as I increase depth, I am adding resistance at the bottom therefore the overall resistance of the network goes up.  This seems to have no impact other than a slight reverse-log taper to the depth control (which is good) and a required change to the min and max resistor values so that the min and max depths stay the same.  Also, the max resistor is implemented as a 100k pot which functions as the chop depth extender.  Although I implemented the depth control a little differently, in normal mode it covers the same range as the original ROG pedal.

What I did in my pedal was I added a second footswitch to toggle between two modes: normal and turbo, as well as a second second speed knob "turbo" and a "chop" knob which is really a depth extender.  When in normal mode, it is a standard EA Tremolo with a slightly lowered max rate (implemented as an "overlap" trim resistor).  When in turbo mode, the turbo and chop controls become active, and the original "normal" speed control is bypassed.  The original depth control is active in both modes, but when in turbo mode, the chop control becomes active and what it does is allows you to boost the original depth setting up to max depth.  The turbo mode speed control is a faster range of speeds than the normal mode speed control, but there's no reason why they couldn't be the same.

Here's the schematic:

http://i1190.photobucket.com/albums/z460/GGBB1/Pedals/Tremolo/TurboTwinTremolo.gif

[ZeusMalt]

Aight... back to the square one :/  Damn. First of all, changing the 4k7 doesn't help. Why? Because the point of the trimmer is to set the voltage from Q1 to ground to 4.5-5. Changing the 4k7 doesn't help.

Aaand off we go to the new problems-> changing the 180 definetly helps. But i guess that resistor is on the "signalpath"? And rising that value would turn in to a toneloss? The better option would be lowering the 1k2?

Aaand as i managed to socket the 180 and rise the value of it (quuite alot) and get rid of the distortion, at the same time i lost the depth of the tremolo. I don't know if that happens when you lower the 1k2...? At that point i also tried socketing the 68k and 120k as you mentioned, but i seemed to get no or really small changes when changing the values of those resistors. Although i can't be sure if everything is connected anymore... I'm starting to get rly bad headache from this Now i have to get everything "unboxed", socket EVERYTHING and try everything.... and i wasn't even doing this for myself... so once i get everything set up i propably have to do everything again since everything looks like sh** at this point.

[GGBB]

Aight... back to the square one :/  Damn. First of all, changing the 4k7 doesn't help. Why? Because the point of the trimmer is to set the voltage from Q1 to ground to 4.5-5. Changing the 4k7 doesn't help.

Aaand off we go to the new problems-> changing the 180 definetly helps. But i guess that resistor is on the "signalpath"? And rising that value would turn in to a toneloss? The better option would be lowering the 1k2?

Aaand as i managed to socket the 180 and rise the value of it (quuite alot) and get rid of the distortion, at the same time i lost the depth of the tremolo. I don't know if that happens when you lower the 1k2...? At that point i also tried socketing the 68k and 120k as you mentioned, but i seemed to get no or really small changes when changing the values of those resistors. Although i can't be sure if everything is connected anymore... I'm starting to get rly bad headache from this Now i have to get everything "unboxed", socket EVERYTHING and try everything.... and i wasn't even doing this for myself... so once i get everything set up i propably have to do everything again since everything looks like sh** at this point.

I honestly don't know - I don't understand the circuit enough to suggest anything more than try this or try that based on the limited knowledge I do have.  The way I believe it works is that Q3 is the rate, and Q2 is the depth, and together they vary the gain of Q1 to produce the tremolo effect.  Beyond that I am largely guessing.  I wish some of the experts here would weigh in but they are not.

I would guess that if for some reason Q1 is oversaturating/clipping, you essentially have compression which limits the depth of the effect - I think this would sound like more on than off tremolo.  On the other hand, if Q1 is under-driven, you are not going to be able hear the full depth either - I think this would be heard as mostly off sound.  Both cases would end up sounding like not much depth, but the depth is not really the problem.  In other words, until you sort out the clipping, don't worry about the depth. 

Thinking this through, what "drives" (sets its gain) Q1 is Q2, a JFET, which is acting as a voltage controlled resistor.  So the gain of Q1 is controlled by the 4k7, the 180, and Q2, as well as Vc via the trimmer.  The 22uF cap between the 180 and Q2 drain blocks DC so that only AC (audio) gain is affected by Q2.  I think the 1k2 resistor would act (as far as audio is concerned) as a parallel resistor to the drain-source resistance of Q2, so it has an effect too, but I'm not sure it's as important as Q2 itself.

Here's a big guess - if all that is right, then you should be able to lower the gain of Q1 by adding resistance between Q2 source and ground.  A very small value is probably enough - maybe 100ohm - I think.  This would avoid having to mess with the ratios of the 4k7, 180, and 1k2.

And now that I have written all that, I am thinking, since the trimmer also controls the gain of Q1, what happens if you try to adjust the trimmer to get rid of the distortion (don't worry about the 4.5-5V rule)?

[Davelectro]

What kind of jfet are you using for Q2? I think you should try one with higher Vp (2n5458, mpf102, bf245B...)

[ZeusMalt]

Aight! Here's what has happened.

Only thing i had to do was to change 180r to 470r. That got rid of the distortion with j201 on q2. Didn't test thoroughly enough to say if it's ok with sk170, but im guessing it should be a bit higher with that.

4,5 to 5volts on q1 is required for the thing to work. More or less than that leads to no sound.

I'm not able to heck right now, but would j201 happen to have higher vp than k170?

[Davelectro]

I think 2sk170 is very similar to j201 in terms of Vp.

BTW, my EA Trem uses a bf245A with Vp= -1.8V and it's perfectly clean.

[GGBB]

I used an MPF102 - also very clean.

[kaycee]

Simply take off the volume pot and add it to the front end as a pre-gain type control. Then you can limit the signal going in and hopefully stop the clipping. You'll still need the 1M pulldown on the front, otherwise the 'open' cap will most likely give you a pop on switching.

If you mean connect the pot in parallel with the 1M pulldown, with signal coming out of the wiper, then this is functionally no different than what I described and Zeus has already tried.  The additional 1M pulldown would be unnecessary in this situation since when in parallel with the pot it would only serve to reduce the total "pulldown" resistance.  The pot itself alone would actually function as a pulldown resistor.

If you meant something else then please elaborate.

No, I meant use the pot as a variable series resistor. Take the input in through one leg, tie the other two together and then into the circuit, no signal going to ground, doesn't make any difference to the pulldown resistor.