Help diagnosing Digitech Multiplay 20/20 Clock/Timing issue?

[any]

Well, it's that time again... finally got round to trying to revive my all time favorite 'commercially made' pedal, the Digitech PSD 20/20 multiplay.

Ran an audio probe through the signal path and I think I narrowed it down to the NE570N compounder.
I get sound on pin 2 and 3 (back from the digital part of the circuit), but pin 7 (output) is completely silent.
I've ordered a couple of new NE570N from uncle Ali X, but then I kind of wondered if it actually has to do with
one or more caps being faulty. (after reading a post on the forum about a dead Boss DD-2 with a similar
issue, being half of the compounder working fine, the other half being faulty:
https://www.diystompboxes.com/smfforum/index.php?topic=119454.msg1114673#msg1114673 )

So took the voltages for the NE570N as follows:

P1= 0v
P2= 1.8v
P3= 1.8v
P4= 0v
P5= 1.8v
P6= 1.8v
P7= 4.3v
P8= 1.8v
P9= 1.4v (fluctuating quite a bit up and down?)
P10= 1.8v
P11= 1.8v
P12= 7v
P13= 1.8v
P14= 1.8v
P15= 3.7v
P16= 1.8v

Pinout as per image


Now I assume pin 1 is suspect as pin 9 has the same function (rect cap) on the other 'working' half and is fluctuating at 1.4v)
Is anyone able to shed some light? thing is, pin 9 goes 120pf cap to ground whereas pin 1 goes to a 2n5088 transistor with a .47uF(?) cap
and a 2m resistor to ground. I've attached the schematic showing the NE570N underlined in green for the 'working' side (going to the adc0820)
and underlined in red for the 'dead' side.



I might take a chance and replace the .47 cap on pin 1 to see what happens, but rather learn more first.
Also, the 2n5088 is actually involved in the bypass switching through a 4007 ic so I wonder whether the issue actually might be there,
as in not actually setting pin 1 correctly. The LED lights up when I switch it on though...

As a nice aside, I read on a post over at http://falseelectronics.blogspot.com/2017/11/digitech-pds-1000-2000-repairs.html
that the digitech pds delays don't use any proprietary ic's or micro controllers and as such can be completely built from scratch if needed.
Since I haven't seen any pds 20/20 for under $300US this might actually become interesting.
The full schematic is available at experimentalists anonymous btw:
http://www.experimentalistsanonymous.com/diy/Schematics/Delay%20Echo%20and%20Samplers/Digitech%20PDS2020.pdf

Help greatly appreciated!

[any]

Oh the gremlins... just been re-checking voltages on the transistor around pin 1 and all a sudden got 1v at pin 1 and hey presto, sound comes out. 
BUT, now the switch for delay time doesn't work... only longest setting 2sec. the shorter ones give nothing, just dry signal.

Anyone smart enough to enlighten me how the delay time is set in this pedal? Why does the longest setting work but not shorter ones?
The weird thing is in the schematic the only lines running to the timing/clock part seem to be the shorter settings where as the longest setting doesn't
connect directly to the clock?

Seems like everything works as intended bar the chorus and flanger times, but makes me wonder if It would be a good idea to re-cap just in case as
I really don't like this "intermittent issues" stuff. Anyone experience with re-capping old effects?

Will try and prod some more and see if I can educate myself...

[ElectricDruid]

The two halves of the NE570 are mirror-images of each other, so the equivalent of pin 1 on the other half is pin 16, not pin 9.

Pin 9 is the THD adjustment, and pin 8 is its equivalent on the other side. Those pins wouldn't be the first place I'd look for the problem, certainly.

The "Rectifier Cap" that hangs off pin 1/16 is definitely a big possibility. C9 on one half, C22 on the other half, both 0.47uF.

I would say if you've got intermittent issues, absolutely don't re-cap. Taking loads of stuff out and then putting it back in is hardly going to increase reliability, and it will introduce a ton of new unknowns. Get the pedal working first. The desire to re-cap is just a vague hope that a scattergun approach might solve the problem. It might, but far *more* likely is that it breaks something else. And you'll never know what or be able to fix it, because you didn't have a known-good pedal when you started. It's a nightmare scenario, so don't go there.

The delay time switch thing doesn't sound so insoluble. Have you got the relevant page of the schematic we could have a look at?

HTH

[any]

Thanks for the wisdom ElectricDruid, makes sense now... ;-) the full schematic is the last link in my post. I'm going to get some 4007's at the local bug dealer (Jaycar here in NZ) as there only $1.40 and swap them out. Read over at another blog (falseelectronics, link also in my first post) that there was a similar problem with a pds 1002 that traced back to a faulty ic that did the switching. Think in the pds 1002 it was a hex inverter where as in the 20/20 it's a dual input nand but anyway, worth a shot.

Sounds like a fun project to actually do a DIY board for this delay anyway, especially with the stupid prices people want for one nowadays.

I'll report back after some more trouble shooting.

[any]

Ok, so learning more and more but also finding new questions...

How does the clock actually work on this thing? 

Changing the 74HC00 Nand gates didn't do anything, so went back to the schematic (link at the bottom of original post),
and found that 'line D' (the common line for the time range selector) goes to the output enable pin on a 74HC374 (Octal D flip/flop).
So measuring the voltage there (pin 1) shows 0v at the 2 second delay setting, but 5.73v at both the 250msec and 16msec settings.
(soo effectively disabling the output?) Does that make sense? is the clock for the 2 second delay seperate from the other settings,
and how does it actually work in this pedal? I can see the timing section (the best thing about this pedal imho) feeding into
two 74HC195 4-bit parallel access shift registers but I can't get my head around how the delay time range is set.

If the 2 second setting is working am I right to assume that the 74HC195's are working fine? And if it was the 74AC04 Hex inverter
would it make sense that the 2 second setting still works?

Who's got some insight? 

[ElectricDruid]

the full schematic is the last link in my post.

Ah, thanks - I missed that and only saw the page you posted as an image. I'll have a read and see what I can see.

[any]

the full schematic is the last link in my post.

Ah, thanks - I missed that and only saw the page you posted as an image. I'll have a read and see what I can see.

Beware, the link on Collins site is a pdf with two different versions of the schematic though (one september, the other December)
The one's saying "manufacturing original" seem to be the right ones. (although I've found some variation in part values with my unit compared to the schematic)

Cheers,

[any]

Well just checked the Hex inverter and replaced it but no change... Interesting detail, the schematic call for a 74AC04 (high speed) hex inverter,
but there was a 74HC04 in there (which apparently is 3 times as slow) so I guess part substitutions are pretty common.
Does mean i need to figure out if it's the 74HC195 4 bit registers..

[ElectricDruid]

I'll have a guess about the clock and the Delay time switching:

It looks to me like the analog LFO and clock part (based on 74AC04, top right) is the same in every mode. E.g. it doesn't actually change speed.

What changes is the addressing to the memory. The clock is fed to two counters, a 4040 and a 4024, which gives a maximum 16-bit address. The whole lot is used for the long delay. For the shorter modes, the two lines that are fed back (A9 and A13) are used to reset the counters (so looping the RAM back to the beginning before we've addressed it all). The 74HC00 NAND gates and the voltages coming from the Delay range switch are used to gate these 'reset' signals and turn them on or off.

How does that sound?

If you've already replaced the 74HC00 NANDs and the shorter delays still aren't working, I'd check the voltages on the E and F lines to make sure that the Delay Range switch is creating proper logic level voltages. Perhaps some of the diodes or the 4.7uF caps around the Delay switch are dead? Checking the voltages there would give you a clue.

[any]

Well, shoot me... it's working (for now?) but no clue why.
I've measured voltages to the 74HC00 Nand and they all kept checking out, apart from the output to the reset/clear staying low all the time. Decided to put the old IC back in just in case and got it going again! 

So I'm leaning towards re-seating the IC's being the likely solution?
Still leaves me a bit uneasy though... I'd rather find a fault so i know the issue is resolved then this kind of 'and now it works again' scenario's where i keep worrying for the 'and now it doesn't' to happen...

Such an amazing pedal though! I'm thinking about cloning a circuitboard as all parts are non proprietary. That way i could build a new one for way less of what a vintage unit would cost me nowadays. A friend of mine was raving about an Electro Harmonix Freeze pedal, which is pretty much setting the multiplay to chorus (250msec) and hitting the repeat footswitch at the right time. So many great sounds in there.

Anyway, thanks ElectricDruid for your input! Much appreciated!

[ElectricDruid]

Glad to hear it's working, that's great.

Tarnished sockets/chip legs does sound quite likely, I agree. If you're concerned about it, you could try pulling the ICs out again, cleaning the legs up a bit with a hard eraser, and then sticking them back in. A bit of a service like you've done will often keep something going for several more years. In my experience, it just goes with the territory with vintage electronics - you will need to keep opening it up periodically and having a fiddle about to keep things running nicely.

[any]

Well, at least it'll soon be part of my setup again! (green plastic card for debugging purposes)
got a 10 minute loop sampler tacked on in the same box too, already had a blast in the couple of minutes testing the setup.

[any]

Round 2... not really an issue, but since i got this working I've been wondering about the pds clock.
Is this a fairly standard clock setup? and In case @ElectricDruid is checking in, can the taplfo (or ideally something arduino based, see my other thread in digital)
be spliced in (or replace) the pds clock in order to have tap tempo and the same modulated delay control the Multiplay has, but with the benefit of digital controls?

I've been digging and found some stuff about clocking the 74HC195 with arduino but I'm unsure of the clock frequency that is used in the pds20/20.
Also found an old thread from Earthtone that kinda relates? ( https://www.diystompboxes.com/smfforum/index.php?topic=94949.0 )
Anyway, if someone has anything to chip in that would be cool ;-)

Also, could a signal be injected where the 'width' pot currently sits? Looks like another potential point to intercept a signal?

Clock part of the schematic attached:

[ElectricDruid]

TapLFO uses the tap to set the *speed of the LFO*, not any delay time, so it's not going to do what you want. At least, not directly.

That said, Arduino or something similar should be able to do a tap-tempo delay clock. Adding modulation as well is where it gets tricky, because for that you need the changes in clock output to be small and smooth enough that you don't hear any artefacts in the delay.

But the basic tap-tempo isn't too difficult. You measure the time between taps with the Arduino, and then use the resulting period to calculate the required clock frequency. Of course, you do need to know what clock frequency you want, so some measuring or experimenting to get a "Clock Freq X = Delay time Y" type equation is required.

The "Width" pot is a flanger-esque blend between the LFO output and the static voltage provided by the Delay Time pot and buffer. That voltage is then buffered again and fed to the clock circuit. I'd say yes, you could switch out either end of that Width pot for a different source of voltage. Just watch the levels. I notice the Delay Time pot is only connected to 5V, and heavily limited top and bottom, so it only covers 2V around the 2.5V midpoint (e.g. 1.5V to 3.5V). Similarly, it looks like the LFO is biased around 3V from the 5V supply (R81 and R82 far left), so I doubt it has a wide range either.