EA tremolo debugging and mods

[swever]

I've just boxed up my EA Tremolo (http://www.generalguitargadgets.com/effects-projects/tremolo/ea-tremolo/) and discovered it has way too much gain to the point that it starts to overdrive loud notes.

I added the Trem/Boost switch and with it in Boost mode I get unity volume with the volume control set just below 9 o'clock. When I engage the Trem mode, it seems to modulate around that level, so it gets even higher at positive peaks of the modulation.

Could this be caused by adding the Flashing LED mod? I am not completely sure if I remember it overdriving before I started boxing and adding the flashing led and the trem/boost switch.

Also, I had a problem with my depth control - it would go from heavy vibrato to almost square wave stutter. I ended up using a logarithmic 250 pot for depth and changing the R19 to 10k and R7 to 820k. I am not 100% sure this could not have caused the q1 gain to go too high.

I believe the stage around q4 is a linear gain buffer, so I can leave it alone, right? There probably must be a simple way to adjust the gain of q1 stage to make everything nice and clean.

[PRR]

Turn down the guitar.

Increase the value of R5 180 to like 500r.

However this thing is subject to way different performance depending on parts tolerances. The Depth control network probably has to be trimmed to the specific JFET you have.

[swever]

Increase the value of R5 180 to like 500r.

That worked perfectly!


I've been suggested to substitute R5 and R6 for a pot. I tried a b2k and it does a very good job. It interacts with both volume and depth, but it's still very usable and I can get anything from crystal clear to light heat to prominent overdrive - in both tremolo and boost modes.

[swever]

After some experiments I settled at this. It gives a very light tremolo effect even at the lowest setting of this pot, and a nice overdrive at it's maximum. I used a ГТ404Б npn germanium for q1 and it made the overdriven settings sound majestic. It's a subtle difference, but it's definitely more round and smooth now and at the same time more dynamic and lively, with a nice slightly mid-focused punchy attack.



The only thing I've discovered I yet would like to improve is the behavior of the speed control around it's maximum. For some reason, whenever I adjust the speed from any position to the last couple of degrees of it's rotation the depth of modulation starts to fade away quickly and it almost disappears after a few seconds. Once I dial it down a bee's dick the modulation returns to normal. It's also observable on the lfo indicator led.

Is there any way to address this?

[GGBB]

That's kind of normal and has to do with the charge-discharge cycle time of three 1uF caps in the LFO. To raise the LFO speed, you need to increase EDIT: speed up the cycle time. The cycle time is a product of the cap value and the drain resistors which are the 15k R10 and the speed pot plus 1k R11. If it's merely the fact that you lose LFO at higher speeds that bothers you, set the pot at the point just before it drops out, measure the resistance across the pot, and then add that value to the 1k (round up). If you want more speed, keep in mind that as-is this LFO design won't ever be able to give you fast choppy tremolo sounds, but you can either try lowering the 15k a bit, or using smaller caps. Halving the cap values gives you twice the speed. Note that doing either will also make your slowest speeds faster - so it's a trade off. You might be able to get back the low speeds by increasing the speed pot to 250k or higher, but there's only so much you can get in terms of range out of this circuit.

[Kipper4]

Reply three.

http://www.diystompboxes.com/smfforum/index.php?topic=99318.msg870594#msg870594

[swever]

What I'm trying to acheive is to prevent the LFO "lock-up" (modulation depth goes to almost null within a couple of seconds). I've found it happens at both extremes of the speed pot's rotation. Once it's gone there (especially at the minimum), it takes a significant amout of time to recover back to normal operation. Also, it does not recover by just by rotating the pot it a bee's dick as I stated before – it needs to go somewhere around mid-rotation, maybe even closer to the opposite extreme.

I've done some experiments late at night yesterday. I found that I can decrese the maximum speed by increasing the value of R11, and thus prevent the lock up at maximum. I did not find a way to "limit' the minimum speed and prevent the lock up there. My guess it's R10? It was late and I was tipsy and tired so I am not sure any of that makes sense really.

[swever]

Reply three.

http://www.diystompboxes.com/smfforum/index.php?topic=99318.msg870594#msg870594

Unfortunately I am really bad at decyphering that harcore science . I think R.G. did mention something along the lines that there is a way to increase the lfo rate control range by using a dual gang potentiomenter somewhere, right?  I guess that would be a variable resistor in place of (or in parallel with) R10, correct? Sould it increase together with R14 or should it be wired the opposite direction?

[swever]

Bump 

This is how it looks at the moment:


Had no time to work on it today and test my theories. Hopefully tomorrow.

One guess that poped up in my mind is maybe it's the MLC capacitors that I used in the lfo section (this is called PSO, right?) that cause the lock up? Is there any chance it's going to get better if I swap them for films?

I checked this demo on youtube and did not really notice this kind of behavior.

[GGBB]

There shouldn't be any "lock up" at slow speeds, but very slow may be indiscernible from off. Depth plays a role as well - shallow depth with very slow or very fast speeds is difficult to make out. Cap type should not matter - the time constant is dependent on value not type - but in practice perhaps differences in something like ESR could have an impact, though it would likely be a small one.

When you adjust one end of the speed range, you also adjust the other end, so it is a balancing act. It is further complicated by the anti-log response - a specific change in resistance - say 1k - has a tiny impact on the min speed relative to a huge impact on the max speed. For that reason, fix the min speed problem first, then tweak the max speed.

If you can't solve both satisfactorily, try a 50k speed pot (or parallel the 100k with a 100k or larger resistor) - this will lose roughly 10% of the low speed range. (A drawback of the parallel resistor approach is an increase in the sensitivity of the speed control toward the higher speeds.)

[PRR]

> increase the lfo rate control range by using a dual gang potentiomenter



You "should" diddle three resistors, but who has 3-gang pots?

Diddling one of three resistors 100:1 gives 100^(1/3) or 4.6:1 range.

Diddling two of three resistors 100:1 gives 100^(2/3) or 21:1 range.

Diddling three of three resistors 100:1 gives 100^(3/3) or 100:1 range.

(The 'third resistor' is actually input of Q3 and other bits-- not a rigorous design but adequate for a gizmo with a knob on it. So we can't use a 3-gang pot without major complication.)

This may not give the rates you need. Change all *three* 1uFd caps to transpose the whole range.

[Transmogrifox]

Cap type should not matter - the time constant is dependent on value not type - but in practice perhaps differences in something like ESR could have an impact, though it would likely be a small one.

MLCC capacitors have another potentially substantial impact that wasn't mentioned.  Typical general purpose capacitors with high dielectric constants ( X5R or X7R and others), the capacitance changes with applied voltage.  If you're using 50V or 100V MLCC caps, probably the effect is negligible.  If you're using 10V or 16V MLCC capacitors it could be significant enough to make a noticeable difference in how the oscillator behaves.

http://www.murata.com/en-us/support/faqs/products/capacitor/mlcc/char/0005
http://ds.murata.co.jp/software/simsurfing/en-us/#app=71e3&abf8-selectedIndex=2 <-Search a part and you can get a graph.

If you like the overdrive effect, then maybe the following isn't useful to you, but here is my take on the EA tremolo design that causes it to behave in the way I like it.  I personally thought the excessive gain and resultant distortion of the EA tremolo doesn't belong in a tremolo effect.  However, you can get back the gain with a 1k pot instead of (my schem) R8 of 470 ohms.  The main thing I addressed was to be able to get a depth of -60 dB (rev B design).
http://cackleberrypines.net/transmogrifox/rEAgeneratedTremolo/pages/index.html

I didn't do anything with the oscillator, so it won't give you any insight in that respect, but maybe worth looking at it anyway.  Read the write-up if you're interested in more detail:
http://cackleberrypines.net/transmogrifox/rEAgeneratedTremolo/pages/projfiles/rEAgenerated_Tremolo_Design_Rationale_RevB.pdf