RG's ersatz fender vibe v2 vero/troubleshooting

[ulysses]

hey guys/RG

i've just done a vero of this circuit
http://www.geofex.com/FX_images/FakeFenderVib3.pdf

i built it and i'm having a problem. i'm wondering if others have had the same problem or if there could be a problem with my vero/build

when i turn the speed pot above 90% it goes to infinity, and then halts the lfo. if i turn back down the lfo won't come back. i have to power the pedal off for a minute and then back on for the lfo to function.

ive used 2n5458's and bd549c's throughout

anyone else seen this problem?

when i've got it all sorted, i'll share the vero

cheers, ulysses

[R.G.]

Phase shift oscillators (this LFO is one) get funny about starting and running if the phase shift resistors get out of range.

First, it is possible that there is a circuit mistake. Check your breadboarding. The not starting again is odd, but not impossible.

A cure for your circuit might be to increase the value of those two 1K resistors, R32 and R33. Stick in 10K and see if the problem vanishes.

These resistors limit how little the resistance to ground can get from the junctions of the phase shift caps. Obviously, if the resistances where the speed pots and R32/R33 are get to zero, the phase shift signal is grounded and it can't work. So there is some minimum value. A lot of math involving the voltage and current gains of the transistor would tell you an extimated minimum, but resistors are cheaper than thinking, right? 

Drop in a couple of 10Ks and see if you have enough speed range.

[PRR]

Agree about expecting a builder-error.

Agree that the no-restart is odd. It has like twice the gain it should need to barely-start, so it should start with vigor.

Sure agree that resistors are cheaper than thinking.

Agree that 100:1 range on Speed pot may be a bit much. 10K for R32 R33 stoppers seems reasonable and should give over 4:1 range of Speed. Or turn it just until it quits, then measure the pot resistance at that point. Add 1K, round-up and try that value stopper.

What kinda caps for C17 C18 C19? Polarized Electrolytic "may probably" work, or may get upset about the wild swings of polarity. Film caps would be a safer bet. Or Ceramics although the wide variations with voltage and temperature "could" give drifty rate or ugly waveform.

[ulysses]

thanks very much for the replies guys

i swapped out the 1k resistors for 10k, and that did solve the problem- but now the range is pretty poor - i should just measure the pot at max as PRR said, but i would like to know if its my build as i'm keen to share it with others, but not if there's a problem with it. ive been over it twice on the vero layout, once on the build, so i'll go over it again on the build when i get some time. i would like it to work as listed on the schem before i share the vero - don't want other people running into trouble.

i tried swapping out the bc549's for 2n5088's for Q7,8,9. didnt make any difference.

i did question whether using tant's for C16,17,18,19 would be causing the problem - they are not polarity marked - do you guys think that could be the problem?

cheers, ulysses

[R.G.]

i swapped out the 1k resistors for 10k, and that did solve the problem- but now the range is pretty poor - i should just measure the pot at max as PRR said, but i would like to know if its my build as i'm keen to share it with others, but not if there's a problem with it. ive been over it twice on the vero layout, once on the build, so i'll go over it again on the build when i get some time. i would like it to work as listed on the schem before i share the vero - don't want other people running into trouble.

I did a simulation of that, and it seems to work well, even with 1K's. Something is different.

i did question whether using tant's for C16,17,18,19 would be causing the problem - they are not polarity marked - do you guys think that could be the problem?

I used film caps for those. It is possible that could be an issue. Electros do leak when reversed, even a little. Tantalums are better than aluminums, but still it's not good. Electros WANT to be held polarized. Do you have any 1uF films?

[midwayfair]

What you need to do is use your multimeter to measure the exact point where the circuit stops oscillating at the top of the range and the bottom of the range. Use parallel resistors to get as close as possible to the high end and change the 1K series resistors for the low end.

When I built this, the most range I was able to get out of the speed control was about 63K on the high end across both gangs of the dual gang pot, and used 3.3K resistors as the limit (though I seem to recall I could have gotten away with 2k2). This is obviously VASTLY different than RG's expected range. However, I also don't recall seeing any other posts about this before yours, and I really don't know how many people have built this.

Note that RG's schematic already assumes a greater range of speed than the Univibe had, and IIRC the univibe uses 2u2 caps in the LFO instead of 1uF.

Just increasing the 1K resistors really isn't the best idea anyway - the range of the speed pot isn't linear. As the resistances get smaller on both gangs of the pot, the rate increases at an increasingly fast rate. At slower speeds, it takes more resistance to effect a greater change in rate. Meaning that the difference between 0K and 1K on the pot at the fast end is much greater than the difference between 99K and 100K or 49K and 50K.

Also, if it IS a layout issue, it's hard for anyone to help determine that without seeing the layout. (Not that I'm offering to backtrace a vero layout.)

[duck_arse]

if the R32 leg is taken to +9V, it can (will) then satisfy all the caps polarities. is this a valid fixx?

tants being polarised will be polarity marked. I've found 3 caps same value, diff brands will tend to stopping more-so than same brand/value does/will.

[ulysses]

i just checked the caps - and they look like tants, but the receipt i have says they are monolithic caps. i will grab some film caps next time i go to the city.

im trying to figure out how the lfo circuit works, but i dont see it - is it easy enough to explain? why do such large film caps need to be used?

thx again for your help

cheers

[duck_arse]

searching followed by reading:

http://www.diystompboxes.com/smfforum/index.php?topic=106566.0

[ulysses]

searching followed by reading:

http://www.diystompboxes.com/smfforum/index.php?topic=106566.0

thx for that - i'll have a read

[ulysses]

if anyone feels like verifying the vero you can download the unverified copy of the layout here:
https://www.dropbox.com/s/wxx1wdaj61vzfu6/ErsatzFenderProVibe2LayoutUlyssesNOTVERIFIED.jpg?dl=0

cheers

[ulysses]

When I built this, the most range I was able to get out of the speed control was about 63K on the high end across both gangs of the dual gang pot, and used 3.3K resistors as the limit (though I seem to recall I could have gotten away with 2k2). This is obviously VASTLY different than RG's expected range. However, I also don't recall seeing any other posts about this before yours, and I really don't know how many people have built this.

interesting -thx for this

[R.G.]

In a phase shift oscillator, the limits on oscillation speed are usually set by the amplifying device's ability to provide enough gain to overcome the voltage attenuation in the feedback network (i.e. the speed caps and resistors) and to provide enough current to drive them.

If you think about it, with fixed capacitors, if the resistors go to nearly zero, the RC phase shift goes to very high frequencies, but the attenuation gets huge, as the resistors are nearly a "short" for the LFO signal, so the amplifier needs a BIG current ability to drive them, and a larger voltage gain to make up for the losses. On the other end, when the resistors get really big, the feedback network starts looking more and more like a series of three caps in series.

Making the amplifier have more gain seems like an obvious answer, but that quickly leads to distorted outputs. Making the amplifier have higher input impedance and more current drive helps, but you can easily make the whole thing very complicated, losing the advantage of simplicity.

It's a three-way contest between waveform distortion, speed range, and simplicity. Like so many things, there isn't one single answer that's an optimum for all cases, only solutions that fit one set of criteria better than another.

Remember: fast, good, cheap; pick any two.

[ulysses]

changed the R32 & R33 to 2.2k - now working fine.

cheers

[PRR]

> now working fine.

With enough range-of-rate for musical purposes?

Are you in a playful mood? Are you on a power-supply (not battery)?

Tack 1K resistor from Q8 Emitter to ground. (Across R27 and Depth.)

This will add 5mA to the power demand. A significant hurt to a battery's capacity but nothing to a wall-supply.

Now go back to R32 R32 = 1K (or tack a pair of 2.2K across the 2.2Ks you just added for an easy 1.1K here).

Will it now go all the way? (Fast.)

Theory: Q8 must drive Q9 to drive the JFET gates, easy. But it must also drive the phase-shift network. Simple analysis: When Speed-pot is minimum it has to drive R33, which was 1K. But Q8's load resistor is 9.7K as-drawn. The external load is 1/10th of the amplifier's internal resistance. This is Heavy Loading. If signal levels are small, it can work. But an oscillator like this builds-up until it smacks the power rails, or is lamed otherwise. Large Signal condition.

Less-simple: the C-R network means Q8 does not see just R33. Load is higher due to C18, lower due to C19 R32 C17 R30 in parallel with R33. One way and another, my guess is that the phase-shift network runs close to 1K, and that 9.7K of DC resistance under Q8 is cutting things close. Another 1K of DC resistance under Q8 makes Q8 honkin' powerful, so 1K in the phase-shift network doesn't bother it.

> now working fine.

OTOH, if you are happy, that's groovy also.

[ulysses]

thx for the replies guys

the speed wheel seems perfect to me now - goes faster than what i would use with 2k2's

with 1k0's it would pretty much go to infinity at 3/4 turn, so fast you couldnt tell it was even oscillating - beyond 3/4 turn and it would kill the lfo

even if we could stop the lfo dying, you wouldn't want to go beyond 3/4 turn anyway when using 1k0s

im not sure why there are differences between the computer modelled / mathematically calculated example and a real build??

anyway it's all good now

super busy getting ready to play a festival so i dont have time to do any tinkering atm

cheers