DOD 670 Flanger - Repair

[spartadrive_in]

YES! I have another broken pedal and could use some help.

After replacing the switch the unit does power on and clean signal passes through. When engaged the effect does not go through. This unit has an LED that flashes according to the the SPEED. This LED does flash accordingly.

My guess is the ICs? One is bad. Am I right to think this?

Thank you.

1981

[Enzo]

I'd verify my power supply was good. Appears to run off +15VDC, so is it indeed more or less 15 volts at the regulator, and is it clean? Then are the reference levels OK, meaning are all the signal pins on the various op amps sitting at about 7-8volts?

And the REticon chip is the heart of this thing, so look at pin 6 for DC and see if it is about 7-8 volts there too. The input bias control sets the voltage ther. It should be appoximately half the supply voltage.

Note the NE571 IC has its own power coming through that 100 ohm resistor center left on drawing. Is there indeed 15v on pin 13 of the 571? Hoping the 100 ohm is not open nor the cap there shorted.

I'd verify there was a clock signal at pin 1 of the Reticon IC.

I'd trace signal through the circuit. There must be signal at pin 7 of the input chip or you would not have clean signal. SO is there signal coming out of the NE571 at pin 5? How about coming out the Reticon IC at pins 3,4? Then how about at pin 1 of that op amp in the lower right corner of the schematic? That feeds the other half of the NE571, so is there signal coming out of it at pins 10,11?

SO, we checked for power, made sure there was a clock pulse to step the signal through the circuit, and then followed the signal through to look for where it gets interrupted.

[Mark Hammer]

Enzo's hunch seems spot on. The input stage built around that first half of the RC4558 is always in circuit and directly connected to the output via a 4k7 resistor and 4u7 cap, so you should always hear a clean signal and the level of that clean signal should not really change in either bypass or effect mode.

The switching of the flanger is highly unusual, and I mean highly. It is common for commercial flangers to and choruses to simply interrupt the delay signal and prevent it from reaching the output mixing stage. Quite often, this is done with a FET, and DOD used exactly such a circuit since the very early years of the product line. Here, however, they do something truly odd. They ground out one of the control pins (pin 16) from the NE571 compander chip such that the on board gain cell simply never provides any audible output. When the switch is stepped on again, that ground connection is lifted from the compander, and now it is the LED that gets connected to ground. When that happens, the compander commences working as per usual, and the LED can now conduct current, which it does at the rate of the LFO sweep. Cute.

Now, since the 4558 opamp will still permit signal to pass passively, whether it is in bypass mode or not, you should check to verify that pin 8 of that chip measures 15vdc (unfortunately, power connections are not shown in the schem). You should also check to see if the 3-pin regulator (LM78L15A) reads 15v at its output (that should be the rightmost pin, looking at it from the "front"). The reference levels Enzo refers to are best exemplified by the voltage showing up on pin 5 of that first IC, but most particularly at the junction of the pair of 22k resistors and 10uf cap where that hand-drawn arrow points to.

Other weak points in the circuit could be the 4001 chip, which is generally not very comfortable with anything above 15v. If one of those power supply caps (470uf or 10uf) is off-spec from aging, it is possible that the 4001 might have been blown by excessive voltage. And while the LM78L15 should be pumping out 15v, there is nothing that says it wouldn't momentarily put out a little more under unusual wall power conditions (remember that the power transformer simply scales down the wall voltage it feeds the regulator and does not feed +/-18v come rain come shine).

Pray that the weak link is not the SAD512D, because those are quite rare and costly. You'd be better off spending the money for a chip on an entire functioning flanger pedal.

[spartadrive_in]

WOW. That's a lot of great information. Before I made this thread I checked that I was getting 15V after the regulator already and I did get exactly that.

I will go over all those pins that you guys stated and check for those voltages.

"What are you doing this Friday night?"
-"Oh, me? I'll be wrist deep in my pedal."

[spartadrive_in]

OK. Voltages Checked

I will answer these in order of who said and what said.

(Everything tested with effect engaged)

ENZO:
Power Supply: 15V
Reticon PIN 6: 6.5V (The bias would not go any farther than about 6.9V (counter-clockwise) so I decided to set it at 1V less than we should expect?)
NE571 PIN 13: 14.3V
Reticon PIN 1: Clock Signal= 0.25V (according to data sheet should be minimum at 5V?)
NE571 PIN 5: 1.77V
Reticon PIN 3,4: 0.07V
OpAmp PIN 1: 1.5V
--NE571 PIN 10,11: 6.7V


Mark Hammer:
4558 OpAmp PIN 8: 14.35V
LM78L15A: See Pic Below

Here are the voltages on all pins of the 4001 Chip
PIN Voltage
1 14.8V
2 14.8V
3 0.3V
4 0.3V
5 14.8V
6 14.8V
7 0V (GND)

8 0.3V
9 0.3V
10 14.8V
11 14.8V
12 10.5V
13 10.5V
14 15V (VCC)

I've included the data sheets. I have the 4558 also but was too large to upload here.
Also, a pic of the underside.

Hope this helps!!

[Mark Hammer]

Thanks for the pix. Though I have all the datasheets already, it's useful to see you have them too. Though some of the voltages provided are not particularly diagnostic or helpful (to me, in any event), what is clear is that the power supply is working as it should. It is not immediately apparent to me, however, why the supply voltage should be different for the analog chips than for the 4001 CMOS chip (14.8 vs 14.3v). Unfortunately, the schematic does not provide power connection info, so I have no idea if there are any additional supply-related components not shown that migtht account for that difference.

The presence of a flashing LED suggests all is right with the world as far as the LFO goes, but lets verify that just to be sure. Set the speed for a medium rate; something that will allow you to see a rise and fall in voltage on your meter, but not too slow. Stick the probe from the meter on the wiper (middle lug) of the width control. As you rotate the width pot towards the LFO side (clockwise, I guess), the resulting voltage swing should be wider, and the Depth control will have little impact. Rotate it fully in the other direction, and what you should see is that speed has little or no impact and changes to the depth control should produce a different voltage at the wiper of the width pot.

Or at least that's how it should work. If it does, then we know it's not the sweep generation circuitry.

If I might suggest one thing, it would be to get yourself some methyl hydrate (paint thinner), some cotton-tipped swabs, and get rid of that flux surrounding all the relevant joints so you can see stuff more clearly. The methyl hydrate will dissolve the flux quickly, and you can clean up the board easily.

Finally, the 120pf cap in the corner of the copper side of the board is part of the HF clock circuit. Is it possible that one of the leads of what is clearly a bent cap is touching something it shouldn't be?

[Enzo]

Your clock signal is not something to measure with a meter. it is a string of pulses. Scope it. Even with a volt meter, it seems like the clock pulses would register as the meter reading bumping at the speed of the LED blink

[spartadrive_in]

Your clock signal is not something to measure with a meter. it is a string of pulses. Scope it. Even with a volt meter, it seems like the clock pulses would register as the meter reading bumping at the speed of the LED blink

OK. Well I don't have an O-Scope so how would I do it with a VOM? What would I need to look for?

[Enzo]

Does the reading wiggle? Does changing the SPEED setting change the wiggle speed? Look at the circuit, after the LED there is another op amp, which drives a transistor, which feeds the logic chip. The slowest setting will be the easiest to see. If you can see the wiggle at the LED, then follow the circuit through the parts I mentioned and see if the wiggle is present at each step.

[Mark Hammer]

Your clock signal is not something to measure with a meter. it is a string of pulses. Scope it. Even with a volt meter, it seems like the clock pulses would register as the meter reading bumping at the speed of the LED blink

Clock, no. LFO...sort of. This is one of those instances where analog meters rule in the absence of a scope.

My money is still on a fried 4001 stemming from off-spec caps.

[Enzo]

The LFO signal clocks the bucket brigade chip. TO the audio it is the LFO, to the chip it is the clock. Whatever we call it it should be pulsing. I would think a signal tracer would allow audio "scoping" of this signal.

And suspecting a bad 4001 is why I was looking for clock pulses, so we are on the same page really.

[JohnGreene]

The LFO signal clocks the bucket brigade chip. TO the audio it is the LFO, to the chip it is the clock. Whatever we call it it should be pulsing. I would think a signal tracer would allow audio "scoping" of this signal.

And suspecting a bad 4001 is why I was looking for clock pulses, so we are on the same page really.

Not quite. The LFO modulates the clock. The 2N4124 changes the frequency of the clock being generated (or not in this case) of the oscillator formed by the CD4001.

Something strikes me odd about your measurements. It's possible that by probing the CD4001 you are changing the state of the gates but I would have expect pins 8, 9 and 11 to all be the same since they are connected together.

All that aside, I would expect to read the same DC on pins 3 & 4 of the reticon part (outputs) as pin 1 (input) if the clock was running. Since they are not, I'm going to also throw suspect at the CD4001. Either that or the Reticon part has died.

--john

[Enzo]

You're right, I wasn't looking close enough.

[tangelolemon]

Hey everyone.

Sorry to revive an old thread, but since the documentation was here already I figured it would be a good idea.

I have one of these flangers and it worked fine for a long time until recently, it stopped flanging. It now just gives a high vibrato squeal when the "width" control is maxed. The squeal's pitch change is modulated by the LFO and the speed changes as the speed control is turned.

I only have 4.8v on pin 6 of the Reticon chip.

Clean signal passes through both in bypass and engaged mode, and the LED pulses to the speed of the LFO.

Any ideas of where to look?

[beatnik]

power supply

i'm bumping this again.

i have this flanger bought from the states and it has 110v primary transformer.

But i am in italy so i need to convert it to 220v.

what voltage for the secondary?

15v is enough to make the regulator work properly?