Jürgen Haible LFO redesign

[armdnrdy]

I've been working on a modified version of the Shulte Compact Phasing A. I've been following Jürgen Haible's lead with his Krautrock Phaser build but...with a few modifications. (More on that later)

While working on this thing...I realized that Jürgen implemented a modification in his redesign that I could use in another project that I was working on. The Tycobrahe Pedalflanger.

My plan was to create a build for the Pedalflanger that would fit inside a standard wah enclosure. I ended up "shelving" the project because the original uses a dual 1M/2M pot for the foot pedal control. (1M for sweep and 2M for rate) I tried to source or make such a pot but after much research...struck out.

The original Shulte Compact Phasing A uses a 2.2MA pot for the Oscillation Period (Rate) control. Jürgen redesigned the LFO to: "get precisely the same behavior," as he put it. He redesigned it to use a more easily sourced 50KA pot instead of the 2.2MA.

Now this got me thinking....what if I modified the LFO for the Pedalflanger to accept a 1M pot?   I could probably find a dual 1M pot that would work in a pedal.

I found Jürgen's explanation of the LFO/Rate pot redesign on another site. I have the information in my grasp but...I don't grasp the information!    I could use some help on exactly what he did here.

This is what Jürgen wrote:
Rate pot redesign:
I'd go for a 2.2M pot, too, but it's hard to get them in the PCB mount 11mm version. But the way I designed it, you get precisely the same behavior with a smaller pot value (50k) _and_ a smaller capacitor value.
The "secret" is driving the potentiometer not directly from the opamp output, but from a voltage divider. the output impedance of this voltage divider (R2 * R1 / (R2 + R1) ) must have the same value as the resistor that was in series with the original pot, _scaled_ with the ratio of the new pot resistance divided by the old pot resistance.


Here is my redraw of the original LFO:


Here is my redraw of Jürgen's redesign.



Disregard the 100KC rate pot. Jürgen's redesign included a 50KA pot to behave like the original. The original had the fastest speed at the CCW position. The 100KC that I added "fixes" that and doubles the sweep time at the slowest setting, without decreasing fastest sweep rate.

[slacker]

I don't know if this is any help but I can get from his quote to the values he used but there's a bit I'm not sure of, I think there's a bit of info missing in his quote.

2M2 / 50k = 44 "ratio of the new pot resistance divided by the old pot resistance."

4.7k /44 = 0.106k "resistor that was in series with the original pot, _scaled_ with the ratio of the new pot resistance divided by the old pot resistance."

Then the ((5k1+4k7)* 110R) / ((5k1+4k7) + 110R) = 108R or 0.108K which is near enough 0.106k

You could get the same answer using totally different resistor values though, which is why I think there's a piece of the puzzle missing.
I cheated a simmed the divider versus the original setup and for any pot setting you get about half the current with the divider that you get with the original setup. This makes sense because he halves the value of the integrator cap from 2u2 to 1u to keep the range the same.
So I think what he basically did is halve the cap and double the series resistor from 4k7 to (5k1 + 4k7) to keep the fastest speed the same, then adds the 110R to scale the divider to fit the new pot value.

[armdnrdy]

Thanks for taking the time Ian. I really appreciate the effort!

I think that I might breadboard two PedalFlanger LFOs side by side to compare the original to the modified version.

I should be able to get a rough idea if they behave in a similar fashion on a scope.

The PedalFlanger LFO doesn't include C7 so..I would just double the resistance in series with the new rate pot with a single resistor, add the calculated divider resistor, and half the integrator cap?

[Scruffie]

For the 3007/3207 EHX EchoFlanger i'm doing I just used the A/DA flanger style voltage divider to remove the C2M pot if that's any use, resistor value seemed inconsequential.

[slacker]

I would just double the resistance in series with the new rate pot with a single resistor, add the calculated divider resistor, and half the integrator cap?

Yeah that's it, actually I don't think the doubling the resistor halving the cap step is necessary, just doing the resistor divider part should work. The halving the cap might have been just so he could use a smaller cap for some reason.

[armdnrdy]

Thanks for that Scruffie...I'll take a look.

Ian,
Even though a 2.2µf non polar cap isn't extremely common....it isn't exactly non obtainable. Jurgen's BOM for the Shulte Phaser included parts from Mouser.

Remember that he described the redesign:
But the way I designed it, you get precisely the same behavior with a smaller pot value (50k) _and_ a smaller capacitor value.



So...I'm not real sure if it's needed or not.
I've been extremely busy with work but, I have this whole weekend off. Hopefully I can bread board the two LFOs side by side to compare.

[armdnrdy]

After trying to source a dual 1M/2M potentiometer, (found one online but the 2M was way out of spec.)
and working out the details to many other projects, I finally returned back to this one...
the Tycobrahe Pedalflanger.

I worked up the calculations as per Jürgen Haible's Shulte Compact Phaser LFO redesign, and Ian's (slacker) help.

To incorporate a 1M rate pot, I reduced the integrator caps by half (2.2µf), added a 5.1K resistor to VR (from calculations) between the rate pot and the series resistor (R26), doubled R26, and then adjusted R26 resistance to match the original LFO's lower rate.

On the scope the amplitude and frequency look very close. I believe that due to the tolerance of the original 2M rate pot...no two Pedalflangers would have the exact same rate.

I'm very satisfied with the result, and find this conversion a useful tool to work around some of the older designs that use unobtainable controls.

I also tried the ADA flanger divider that Scruffie suggested.
I found that it did not produce favorable results when compared side by side with the original LFO.
It would work but...requires some component "tweeking"

Original LFO:


Modified LFO:

[Scruffie]

After trying to source a dual 1M/2M potentiometer, (found one online but the 2M was way out of spec.)
and working out the details to many other projects, I finally returned back to this one...
the Tycobrahe Pedalflanger.

I worked up the calculations as per Jürgen Haible's Shulte Compact Phaser LFO redesign, and Ian's (slacker) help.

To incorporate a 1M rate pot, I reduced the integrator caps by half (2.2µf), added a 5.1K resistor to VR (from calculations) between the rate pot and the series resistor (R26), doubled R26, and then adjusted R26 resistance to match the original LFO's lower rate.

On the scope the amplitude and frequency look very close. I believe that due to the tolerance of the original 2M rate pot...no two Pedalflangers would have the exact same rate.

I'm very satisfied with the result, and find this conversion a useful tool to work around some of the older designs that use unobtainable controls.

I also tried the ADA flanger divider that Scruffie suggested.
I found that it did not produce favorable results when compared side by side with the original LFO.
It would work but...requires some component "tweeking"

Original LFO:


Modified LFO:

Hmm, perhaps before the second layout is prototyped I should look at trying your method then, I didn't have too much variation in sweep width to the original units i'd measured though, although I did have to go to a 500k pot from 2M which in your case, not so useful.

Certainly a circuit snippet to keep in the bank though, thanks for your work on it.

[armdnrdy]

Hey Scruffie,

I tried different combinations with the ADA voltage divider. I used two 5.1K resistors for the divider.
I tried the original integrator cap value (2-4.7µf) and halved (2-2.2µf), original value for the series resistor (10K) and the 18K.

None of this got me in the ballpark like Jürgen's redesign did.

It was a great help having the original LFO and the modified LFO side by side on the breadboard for analysis and adjustment. I was able to achieve very close behavior, frequency, and amplitude.