Has anyone successfully used the TAPLFO to vary a FET?

[midwayfair]

Just wondering if anyone has done this before I toss it on the breadboard. I'm just wondering if I can set it up so that it drives the FET in an EA Tremolo-type situation without getting a bunch of noise. I think I tried it at some point with Boss-style FETs and there was a high-pitched whine.

[R.G.]

No, I haven't.

However, remember that an LFO doesn't exist apart from the circuit it modulates - not usefully, anyway. You have to match the LFO peak to peak level and the DC offset of the LFO to be what the thing being driven needs.

The whine is almost certainly leftovers from the power and ground to the processor chip or the PWM on the output that makes the LFO signal. Whine means that you need better decoupling, power/ground wiring, shielding, or some of all three.

[midwayfair]

No, I haven't.

However, remember that an LFO doesn't exist apart from the circuit it modulates - not usefully, anyway. You have to match the LFO peak to peak level and the DC offset of the LFO to be what the thing being driven needs.

The whine is almost certainly leftovers from the power and ground to the processor chip or the PWM on the output that makes the LFO signal. Whine means that you need better decoupling, power/ground wiring, shielding, or some of all three.

Thanks for the response.

For reference, I'm using the tremolo circuit in the datasheet here:
http://www.electricdruid.net/datasheets/TAPLFO2Datasheet.pdf

up to the collector of the 3904, which is connected to +5V through a 10K, and the collector is attached to the junction of R26 and Q9 in your circuit here:
http://www.geofex.com/FX_images/FakeFenderVib2.pdf

(I'm having some trouble getting it to oscillate out of phase at the moment, but that's beside the point as I'm sure I've just overlooked something doing it from memory.)

I have a perfectly fine ~5V swing, and it creates tremolo. It works with a MOSFET, FET, and BJT about equally well. It just makes a little "beep" every time the LFO cycles.

The TAPLFO datasheet has a 100nF pin as filtering between +5V and ground; adding it increases the whine substantially and also makes a wooshing/static noise. So I've taken it out for now.

I'm on a breadboard, so there's only so much I can do as far as wiring and shielding. The PWM makes a tiny bit of noise in most circuits I've made with this, but it's far louder in this circuit and I can't seem to dial it out without killing the effect entirely.

What's confusing me is: how is that noise even getting into the drain output of the FETs? I've never encountered another situation where feeding a control voltage into a circuit's negative feedback introduces noise heard in the audio path.

[drolo]

Are you feeding the PWM signal directly into the FET ?
The PWM frequency is around 19KHz if i remember right. That will surely be audible. Try to filter out the PWM either with a passive lowpass or with an active one similar to what is shown in the TAPLFO datasheet. That one is setup for a bipolar supply but i'm sure it can be converted to work with a single supply.

[midwayfair]

Are you feeding the PWM signal directly into the FET ?
The PWM frequency is around 19KHz if i remember right. That will surely be audible. Try to filter out the PWM either with a passive lowpass or with an active one similar to what is shown in the TAPLFO datasheet. That one is setup for a bipolar supply but i'm sure it can be converted to work with a single supply.

Hm. I did manage to kill the whine with a 100uF from the base of that TR1 3904 to ground. Pretty much anything smaller and it made a loud thump. I could go as low as about 2.2K on the 10K preceding it before I started getting a whine. That's pretty nuts -- it's a couple DECADES below 20Hz before the noise disappears.

I think I've concluded that it's not worth the trouble and should just stick with LDRs. Oh well.

[drolo]

I think I've concluded that it's not worth the trouble and should just stick with LDRs. Oh well.

I did some tinkering with fets as well and came to the same conclusion 
I guess the only proper way to go about it would be the active lowpass ...

[ElectricDruid]

Are you feeding the PWM signal directly into the FET ?
The PWM frequency is around 19KHz if i remember right. That will surely be audible. Try to filter out the PWM either with a passive lowpass or with an active one similar to what is shown in the TAPLFO datasheet. That one is setup for a bipolar supply but i'm sure it can be converted to work with a single supply.

Hm. I did manage to kill the whine with a 100uF from the base of that TR1 3904 to ground. Pretty much anything smaller and it made a loud thump. I could go as low as about 2.2K on the 10K preceding it before I started getting a whine. That's pretty nuts -- it's a couple DECADES below 20Hz before the noise disappears.

I think I've concluded that it's not worth the trouble and should just stick with LDRs. Oh well.

As someone else said, if you're not filtering the PWM output before you feed it to the FET, you might well have problems. The LDR solution only works as well as it does because the slow response of the LDR gives the filtering. There's a minimal-parts passive filter in the LoopEnv datasheet you could use (the LoopEnv chip uses the same 19.5KHz PWM output):

http://www.electricdruid.net/datasheets/LOOPENV1Datasheet.pdf

Page 6 - just three resistors and three caps.

As to the PWM being audible...you must be younger than me! My hearing tops out around 13KHz these days, which means I've lost approximately one octave of treble. I blame too much Jimi Hendrix.

[midwayfair]

As to the PWM being audible...you must be younger than me! My hearing tops out around 13KHz these days, which means I've lost approximately one octave of treble. I blame too much Jimi Hendrix.

Must not be the PWM, then. It's a single squeak at the top of the waveform each time the thing cycles, and it's WAY lower than 19KHz. I'm guessing it's closer to ... 3-5K? Sort of sounds like letting a tiny bit of air out of a balloon at a time. It becomes less noticeable at higher speeds, where it starts to sound more like a "flp flp flp" from a ticking LED. If I were hearing the PWM, it would be a nearly constant tone, right?

[mth5044]

As someone else said, if you're not filtering the PWM output before you feed it to the FET, you might well have problems. The LDR solution only works as well as it does because the slow response of the LDR gives the filtering. There's a minimal-parts passive filter in the LoopEnv datasheet you could use (the LoopEnv chip uses the same 19.5KHz PWM output):

http://www.electricdruid.net/datasheets/LOOPENV1Datasheet.pdf

Page 6 - just three resistors and three caps.

As to the PWM being audible...you must be younger than me! My hearing tops out around 13KHz these days, which means I've lost approximately one octave of treble. I blame too much Jimi Hendrix.

Thank you for posting that. We had been in contact a bit ago with trying to get the bessel filters in the tapLFO datasheet to play nice on a single rail supply, but I could never get it to work right. This seems much more reasonable. Thanks again.

[jez79]

I have had success today using the taplfo to drive phaser fets.
I used the simple passive PWM filtering mentioned above, and the biasing scheme from the phaser schematic.
The depth goes a little too far, with the bottom swing going beyond fet range, so the sweep squares at the bottom of it's swing. Will limit depth knob voltages so it doesn't go that far...

[ElectricDruid]

I have had success today using the taplfo to drive phaser fets.
I used the simple passive PWM filtering mentioned above, and the biasing scheme from the phaser schematic.
The depth goes a little too far, with the bottom swing going beyond fet range, so the sweep squares at the bottom of it's swing. Will limit depth knob voltages so it doesn't go that far...

An alternative solution would be to muck with the passive filter resistor values to increase the losses in the filter. Then you'd finish up with a LFO that didn't produce such a loud output.
The 1x/10x/100x pattern of the filter is chosen to *avoid* loading one stage down with the next one, but you don't have to...

The advantage of reducing the level with the filter is that you still get the maximum resolution out of the chip. Although it won't make much odds for small reductions.

Incidentally, how much of the range *can* you use before the FET squares it up? Which FET are you using and how much level can it handle? I should do a post on my website about these passive filters, and this type of thing would be good to include.

HTH,
Tom