Slightly OT: DOD FX90 Analog delay troubleshoot

[Ben N]

So I pulls out my old FX90 after it has been in storage for something like four years. And it sounds ok, but the controls are a bit odd. First, the delay knob goes the wrong way. Oh, yeah, now I remember: it was always like that. The delay seems a it shorter at max than I recall, but who knows? But what is really annoying is that the repeat pot is off--it very subtly adds repeats up to about noon, then abruptly goes dry: one echo only, for the next quarter turn, and for the last 20 degrees or so it jumps to near-self-oscillation. Is this just a bad pot, or is there something else that I should be looking for? This thing has a mess of trimmers, but I can't tell if any of them may have any bearing on this. TW, I found the schematic here. I think it is the same version, although  know there were several.

TIA

[oldschoolanalog]

First clean the innards of the pot & see what happens. If the problem persists, I have a couple of working FX90's. I could compare V's, etc. for you if necessary.
Dave

[Ben N]

Thanks, Dave, I appreciate it. I'll get that pot out and try to get inside it asap. I really can't imagine that this is the kind of troubleshooting where voltages would help, but maybe it can be fixed at the pot. BTW, should I need one, do you know where to get a replacement for this pot? Also, as an FX90 afficionado, can you confirm for me that some of them have the delay time go up as you turn to the right and others to the left? Do you know of any mods to get the most out of these puppies, e.g. maybe changing out the TL022 for a TL072 or other lower noise opamp?

Thanks again, and I'll report back once I have checked the pot.

[Stompin Tom]

Hmm... I have one, too. I like it for a cheap analogue delay. The biggest drawback of the unit is the way it cuts highs (I stopped using it because of this)... I guess the buffer ain't very good. My delay pot works the normal way, turn right, get more delay time. I liked it best with the delay time trim pot turned just to the point where the delays start to freak out a bit... you know, get a little modulation going on. Sounds very pleasing and interesting to my ears. I honestly don't remember what the other two trimmers in mine do, but I suppose it's possible that one of them determines the range of the repeat pot? I remember having to tweak one of them to get infinite feedback... I'd try fiddling with the trimmers, then spraying the pot with deoxit or the like... you could also measure the pot with a multimeter to see if it works throughout it's entire sweep. If it's broke, just replace it with any old pot... You'll have to add some off board wires.

[Ben N]

You can see the delay time trim in the schematic easily, but I don't see a trimmer that (to my ignorant eye, anyway) relates to repeats. Actually, it would be pretty cool to know what they all do.

[rustypinto]

I guess the buffer ain't very good.

The schematic posted above doesn't have a buffer. I wonder if any other versions included one? It was common for the Boss/ Ibanez echos to have one at the input (which everyone copied).

I see three trims on the schematic, are there more in someone else's unit? Clock wise from the top; BBD bias (5k), Delay time trim (1M), expander input level (5k). I don't see a repeats trim, but if you turn the 22k resistor from pin 12 of the NE570 to that 0.56uF cap into say 100k trim, you can have a much finer control on the feedback loop gain.

[Ben N]

No buffer? What's the TL022 doing? Kind of a necessity for the switching, no? Or am I miusunderstanding your meaning?

[rustypinto]

At the input of the schematic you posted, 1/2 of the TL022 is an interesting inverting amplifier which is almost unity gain. It has a filter on it at the junction of the 330k and 150k resistors in the feedback. This is not really a buffer in the most traditional sense of an emitter follower or op-amp buffer (eg. inverting input looped back to output with input at the non-inverting input). The other half of the TL022 is a type of dry/wet mixer circuit, also inverting so there isn't any phasing at the output jack.

The switching is even more interesting: the momentary switch drives a funny latch circuit which opens and closes the gate on TR7. When the effect is on, TR7 connects the virtual ground for the TL022 to the wiper of the "Mix" pot, and when off it bypasses the mix control and you simply get the input signal. But the topology of the TL022 is irrelevant to the switching.

[oldschoolanalog]

The max delay time of this pedal is comparable to any other single 4096 stage BBD (MN3005, MN3205, SAD4096) analog delay in my collection that I compared it to. The shorter delay at max than you remember could be a function of having used delays with more max delay time since you last used the FX90 (think AD80, AD3208, Rebotes, Memory Man, etc.). The delay might be the same as it always was; you may have just developed a different perspective if what max delay is/should be. It has been 4 years.
If you must have a board mounted pot this should fill the bill. However, legs on this pot are a bit long (I compared with the actual part. This is not guesswork.); but with a bit of careful bending or filing it can be made to fit nicely.
http://www.smallbearelec.com/Detail.bok?no=710
Otherwise...

You'll have to add some off board wires.

No problem

I see three trims on the schematic, are there more in someone else's unit? Clock wise from the top; BBD bias (5k), Delay time trim (1M), expander input level (5k). I don't see a repeats trim, but if you turn the 22k resistor from pin 12 of the NE570 to that 0.56uF cap into say 100k trim, you can have a much finer control on the feedback loop gain.

There are 3 trimpots in mine also. I agree w/Stompin Tom's assessment of their functions. Tomorrow I'll open 'er up & check out the functions of the trimpots in more detail (using an f counter, 'scope, etc.) If you want. Let me know...
I say experiment w/the f'back for sure; first get the pot thing sorted out.
All the Best!

[Stompin Tom]

I guess the buffer ain't very good.

The schematic posted above doesn't have a buffer.

Huh. Never really looked at the schemo or compared it to my unit. I'm surprised at this, but kind of makes sense considering my experience with the loss of highs. Maybe I'll break that bad boy out and just stick a buffer on the front end! Better yet, a small booster if there's room.

[rustypinto]

What kills me about that circuit is all the treble being thrown away on that input amplifier. The 22k and 0.005 (uF?) add a pole and start rolling off frequencies after ~1.4kHz (if it is 0.005uF). Also, there's another filter built around TR1 which rolls off around 3kHz. Why the TR1 filter is before the compander is way beyond me - normally its between compander and BBD to serve as an anti-alias filter.

If you weren't too attached to this delay, i would remove the 0.005uF cap on the input half of the TL022 so you don't lose so much treble. I suspect this may have been added to reduce clock hum. If its really a ~300ms delay, the lowest clock frequency is ~7kHz and you should get sufficient clock hum attenuation from the TR1 and TR4/5 filters (sufficient would be ~3.5kHz roll off, by Nyquist, which is already achieved).

[Ben N]

OK, just to let you know, I replaced the pot with a 100k I had in my parts bin, and it works fine. the original pot was a dead short through something like 20% of its rotation, so I figured that is more than a cleaning will fix anyway. Plenty of room for the extra wire, and I think in the future I will try to squeeze a small buffer in there (maybe on the in jack?). The only problem is that the pots I had were all solid shaft, and the native ones are knurled and smaller, so the original knob would not go on. I had a Radio Shack knob to fit, but the Repeat and Mix knobs rub against each other a little, and one turns the other if you don't hold them. That's a new twist on interactive controls, I guess.  I didn't look for the 0.005u cap, but maybe next time. Thanks for your help, all.