Ibanez PH99 Classic Phase (6-stage optical phaser) - biasing & LEDs

[aion]

I've got one of these on the bench, trying to see if I can find an alternative opto (home-rolled or vactrol) that will perform similarly to the HTV P873-G35-911 optos used in the original.

LFO excerpt with trimmer:


Full schematic:
https://dirk-hendrik.com/temp/ibanez_ph99.pdf

1. I had assumed that the bias trimmer would adjust the LED brightness and shift the resistance range of the LDRs. But with the depth on full, and a multimeter connected to the LDR leads, I didn't notice any shift whatsoever in the resistance sweep range as I adjusted the trimmer. What's this trimmer biasing, and by what spec might they have set this in the factory?

2. There aren't a lot of options for 3-leg (dual element) optocouplers. I'm going to try the VTL5C3-2 and see how it does, but that's really the only option for 3-legged ones, so if that doesn't work then I'd have to either double the number of LEDs or just pair each LED with two discrete LDRs (which is clumsy to heatshrink).

I tested this out and connected another LED in parallel with one of the existing ones, but it seems to have reduced the brightness of the existing LEDs and raised the sweep resistance range by about 20-30%. How would I compensate for extra LEDs added to the LFO output? Would it be better to do another chain in parallel, i.e. two sets of two? The Fv of the HTV LEDs is around 1.8V.

3. What's the impact of different resistance ranges? The two HTV optos in this unit are pretty closely matched (sweep range 7k-59k and 5.84k-56k) but I swapped them out with a few others I had on hand of the same type, and those were a lot lower - one was 4.5k to 32k, and the other was 3.6k to 21.5k. This means their filters are (very roughly) around 2x the frequency of the others, from 730 Hz to 5.20kHz, while the originals were more like 400-3.3kHz... which makes me wonder whether Ibanez/Maxon may have had some selection process, since the two in the original were pretty close to each other, but not at all similar to mine. Seems like if there was a biasing setup with individual control over each LED's range of brightness, it might be easier to compensate for differences.

[PRR]

LEDs are not Edison Bulbs. They do not parallel well. Put them in series. Watch the voltage needs: four red LEDs will approach 8V which is tight under 9V power and other stuff in the string. You may end up with a more elaborate driver.

I'm stunned there is no NFB around TR1. Fancy-pants linear LED drivers will have a series resistor, often in emitter, and loop back around an op-amp for exact (1%) linearity.

As I see it, your trim fades from O_BIAS to O_BIAS, no change. OK, there is an AC component also, but after a slow opto on a slow meter you may not see it.

[aion]

LEDs are not Edison Bulbs. They do not parallel well. Put them in series. Watch the voltage needs: four red LEDs will approach 8V which is tight under 9V power and other stuff in the string. You may end up with a more elaborate driver.

Could that be solved by adding another transistor in parallel with TR1 to drive the second pair of LEDs, set up exactly the same? (+V→LED→LED→collector)

I'm stunned there is no NFB around TR1. Fancy-pants linear LED drivers will have a series resistor, often in emitter, and loop back around an op-amp for exact (1%) linearity.

As I see it, your trim fades from O_BIAS to O_BIAS, no change. OK, there is an AC component also, but after a slow opto on a slow meter you may not see it.

It looks like it's setting a bias voltage (1M base resistor, trim to +V and ground). Would that be at all similar to BBD biasing where you set the DC point of the waveform for maximum headroom? The trimmer was factory-set to about 60% on the rotation. I remember BJTs often perform better at slightly above half-supply. There are a lot of asymmetrical reference voltages on the input buffers of '80s Maxon designs (e.g. the MT10's 9k1/22k) and I assumed this was why.

Also, one correction from the first post... there are actually two trimmers and I was adjusting the wrong one (SR2, the minimum feedback trimmer) so scratch what I said about it having no impact on the LFO. SR1 is a bit harder to get to so I haven't tried messing with it yet, but wanted to mention that in case it's relevant.

[chance.lima]

Good luck with this, I look forward to seeing this work. I designed a vero project for this about 4 years ago, I have some VTL5C3-2 vactrols waiting for me to get going on this. I'm not sure I'll get around to completing it before you figure it out.

[jubal81]

If you're going to try 'naked' LDRs, I don't see why you couldn't use one LED with the 4 LDRs around it, like a lot of other optical phasers.

I'm guessing this is for a DIY project and that's a configuration and part sourcing effort easy for DIYers to understand.

If I ever bite the bullet and buy a 3D printer, I'd really like to design a good optocoupler housing for multiple LDRs.