How to "Trigger" an LFO?

[tommy.genes]

I want to add a switch - possibly external - to my Ibanez DML-20 Modulation Delay to be able to turn the modulation on and off (aside from just turning WIDTH all the way down). Now I'm thinking it would be cool if the same switch could also "reset" the LFO to either the top or bottom of the cycle so as to get a predictable response when the modulation is turned on. This would be similar to the Eventide Instant Flanger, although I am aware that device uses an envelope follower and not an LFO.

The LFO section of the DML is shown below. Since I'm not familiar with the function of an LFO circuit, is there some point I can tie to either ground or Vcc to accomplish what I'm describing? Is what I'm describing even clear? Would I be better off grafting in an envelope follower as is done in the Instant Flanger?

Thanks,
-- T. G. --

[R.G.]

Split R80 (47K) into two 22K resistors. Tie the junction of the two to ground to force the LFO to stop. Open it and it will start. An NPN transistor with emitter to ground and collector to the two 22K resistors will do this just fine.

[tommy.genes]

Thank you! 

-- T. G. --

[R.G.]

My pleasure. Let me know if that doesn't work and we'll try again.

[markusw]

May I ask how the transistor arrangement of the Wobulator's LFO trigger works?

http://www.aronnelson.com/DIYFiles/up/Wobulator_v3.pdf

As I understand, the JFET shorts the integrating resistor. But what are the two BC549's doing? 

Regards,

Markus

[slacker]

they look like they're some sort of pulse generator or switch debouncer to me, could be completely wrong though.

[R.G.]

They're simple transistor switches, two inverters in series. When the signal at Trigger In is low, output at the JFET gate is low and JFET is off. When the signal at the Trigger in is high, output at JFET gate is high and the JFET is on, shorting the integrating capacitor. The caps to ground just slow things down a bit.

I didn't want to mess with the extra stuff to do a JFET cap shorter on the integrating cap. That's the standard way to reset an integrator from synth practice. That's why I went for the single transistor to ground on the split integrating resistor.

[markusw]

Thanks for your explanation, R.G.! 

Good to know that there is a one-transistor-solution.

[tommy.genes]

Let me know if that doesn't work and we'll try again.

I got to play with this last night. First off, the schem I have is for the original DML but I have the DML-20, so the component numbering was different. For the LFO at least, the circuit is the same and I was still able to test RG's suggestion. I also didn't use a transistor for these initial tests. I just left a bare wire sticking up from the new "node," added another wire to a convenient ground point on the board and touched the two together to shut down the LFO.

When the node is first grounded, wherever the LFO was in its sweep, the delay time would ramp up (i.e. the pitch would ramp down) to the longest time available in that delay range - even beyond what would have been set by the WIDTH knob. As long as the node was grounded, the pedal would function as a fixed delay, and releasing the node would cause the delay time to shorten (i.e. the pitch would ramp up) until it was modulating normally as set by the WIDTH knob.

One small problem: at certain settings, a substantial "bacon sizzling" sound was introduced to the output, sometimes developing into whining feedback. At first I thought I had just made a bad choice of ground and was introducing clock noise to the output, but there's really no clock frequency in the LFO section. I'm now thinking this might be an inherent problem to the delay chip (MC4101F) since I recall reading about similar problems with PT239X-based delays when used at their max delay settings. Perhaps I will have to add a small circuit to prevent the delay time from going all the way to this outer limit.

Overall the effect is pretty cool. I would like to add a footswitch so I can try it with both hands free for playing the bass, to get a better feel for how it actually works. Hooking it up to a scope (I don't have one, unfortunately) would also confirm that the LFO is operating in the manner I think it is. However, and you knew there had to be a caveat, I think it would be more musical if the switch started the LFO at the shortest delay time / highest pitch and swept toards longer delay / lower pitch. I'm not sure if this can be accomplished using the onboard LFO circuit (will have to do some reading), but I wouldn't mind building a new LFO or ramp generator circuit (will have to do some research).

Thanks again to R.G. for the tip.

-- T. G. --

[R.G.]

I think it would be more musical if the switch started the LFO at the shortest delay time / highest pitch and swept toards longer delay / lower pitch. I'm not sure if this can be accomplished using the onboard LFO circuit (will have to do some reading), but I wouldn't mind building a new LFO or ramp generator circuit (will have to do some research).

No problem. Instead of ground, tie that point to +9V. That forces the output of the LFO to be the other direction. It should rest at shortest delay then.

[Dirk_Hendrik]

When the signal at Trigger In is low, output at the JFET gate is low and JFET is off. When the signal at the Trigger in is high, output at JFET gate is high and the JFET is on, shorting the integrating capacitor. The caps to ground just slow things down a bit.

Yep.
That was as far as I got when reversing a Wob. That section in the schematic is the same as what I found for which reason I tend to trust is is traced correct by Markus as well as by me.

Trouble lies in the fact that I do not understand how the thing comes to a reliable "high" as the input, when not grounded, just floats.

[tommy.genes]

No problem. Instead of ground, tie that point to +9V. That forces the output of the LFO to be the other direction. It should rest at shortest delay then.

Thank you once again!

I has suspected it might work this way, but it was late and I erred on the side of caution, wanting to learn a bit more before dropping voltages in any old place. The LFO is actually running at +5V, but I can obtain that easily enough off the board (there is an onboard regulator providing the 5V to power the delay IC). I was thinking I might have to add another reg or charge pump to obtain -5V, but it appears the LFO only runs between 0V and +5V.

To use a transistor as a switch, I would connect the collector to +5V and the emitter to the node?

It now occurs to me that, by keeping the delay "resting" at its shortest setting as opposed to its longest, this might indirectly solve the problem of noise at the longest setting.

Hopefully I'll get a chance to try it tonight and I will report when I do.

-- T. G. --

[R.G.]

Use a PNP, emitter to +5, collector to the point you're pulling up. Pull the PNP base down to ground through a resistor to pull the control point up.

[tommy.genes]

Well I still didn't use a transistor switch, but using a sub-mini toggle, I was able to get THIS.

It's quirky with how the switch interacts with the D-TIME and WIDTH controls (WIDTH is actually a blend pot between the LFO and the static D-TIME setting), and the final version may include a DPDT switch to completely remove the LFO from the circuit when it's off, as well as "setting" it for when it is turned on. Still, I consider this a very successful "proof-of-concept" test. It's going to take some time, and I'll be doing a few more experiments on this thing before I finally box it all up again, but I will post a new thread once it is finished.

Thanks one last time for the help.

-- T. G. --