SAD1024 clock noise help

[puretube]

Did we already mention "the old anti-tick trick"?:
100n ceramic cap, one side directly soldered to pin8 of IC11, the other side (via short wire) to ground or directly to pin4).
(Cap >35V).
IC12 could be candidate for the same procedure, too.

[Rob Strand]

As far as getting the clock range correct.   That circuit is a little tricky to adjust because both trimpots (VR3 and VR3) affect both ends of the scale.

Suppose you have checked all the voltages and they look OK.  ie.  +/- 9.6V or so from the LFO and from the Q22's emitter when adjusting the Manual control.

In your case, where you can't the low end of the clock frequency down, it might help to set VR4 to a low resistance then use VR3 to set the upper frequency.  After that check the lower frequency.    If the upper frequency can't reach 1MHx increase the resistance of VR4 a but then try again.

If ultimately you simply cannot get the low frequency end to work I recommend changing R89 from 2.0k to 2.2k.   This will definitely help get a low clock frequency and hopefully there should be enough adjustment in VR4 and VR3  to fine tune the low frequency end an adjust the high frequency end.   Playing with R89 is very sensitive so only make small changes.

Overall the Korg circuit and trimpot configuration is much less reliable than the method used on the Boss BF2.   The circuits are similar but the way the BF2 does it lets you set the low clock frequency end of the scale very reliably.

[four_corners]

As far as getting the clock range correct.   That circuit is a little tricky to adjust because both trimpots (VR3 and VR3) affect both ends of the scale.

Suppose you have checked all the voltages and they look OK.  ie.  +/- 9.6V or so from the LFO and from the Q22's emitter when adjusting the Manual control.

In your case, where you can't the low end of the clock frequency down, it might help to set VR4 to a low resistance then use VR3 to set the upper frequency.  After that check the lower frequency.    If the upper frequency can't reach 1MHx increase the resistance of VR4 a but then try again.

If ultimately you simply cannot get the low frequency end to work I recommend changing R89 from 2.0k to 2.2k.   This will definitely help get a low clock frequency and hopefully there should be enough adjustment in VR4 and VR3  to fine tune the low frequency end an adjust the high frequency end.   Playing with R89 is very sensitive so only make small changes.

Overall the Korg circuit and trimpot configuration is much less reliable than the method used on the Boss BF2.   The circuits are similar but the way the BF2 does it lets you set the low clock frequency end of the scale very reliably.

My voltages don't seem to match up with what you've mentioned.

Q22 emitter range (basically the opposite lug on the Intensity pot from IC12's pin 1) is +3.5v to +11v when moving the Manual pot from 0-10.

I'm having trouble pinning down the voltage from IC12 pin 1, as that is the sine wave that changes with the Speed pot. I'm assuming because of filtering the amplitude/voltage changes with the frequency as the Speed pot is moved, so I'm not really sure what I should be setting the Speed pot at the correctly check IC12 pin1 voltage.

[four_corners]

Im becoming the king of breaking things, haha.

I was lazy and using a metal screwdriver instead of just looking for my plastic one, and it slipped off VR4 and shorted something, my guess is the dual transistor directly next to it. Heard a pop, and then no flanger.

I'm going to just take it out, put a 5 pin socket there, and see if I can make one of these dual transistors with a pair or some other transistors.

[anotherjim]

I'm confused. I'm not sure you can set the modulation range to any set plan any other way from full manual control with the LFO Intensity pot full off (no LFO - does that work?). As you have noted, the LFO amplitude decreases with increasing speed due to the lowpass filter into IC12a rounding down the peaks of the original triangle. Manual control has its own limits, which are different but consistent.

To tweak with the LFO running, you have to judge by ear. with the Intensity control on full. Most would consider the LFO preferable to the manual control anyway. Also, slow speed shows up the range of sweep better. If the synth has a noise generator, use that as a test source. The more high-frequency content, the better to hear flanger action around the fastest BBD clock. To me, the manual pot only makes sense with a changeover switch between manual and LFO and not a mixture which is too interactive.

...and where is the "I can't look" meme when you need one?


I think you did add some extra capacitors, but there isn't a lot of room. Yes, every power pin should have its own local ceramic bypass cap to 0v and if the designers were building it now with SMD they probably would have. This suggests an idea - SMD chip capacitors - if there's a ground trace near a power pin, you can scrape a bit of solder resist off (if there is any) and tack solder an SMD chip capacitor on.

If the LFO ticking is still a bother, you might build the IC11 circuit on a daughterboard with fully filtered and decoupled power rails.

[Rob Strand]

My voltages don't seem to match up with what you've mentioned.

Q22 emitter range (basically the opposite lug on the Intensity pot from IC12's pin 1) is +3.5v to +11v when moving the Manual pot from 0-10.

My guess is bad soldering on the collector of Q21 (PNP).  That would create voltages very close to what you are seeing.

I'm having trouble pinning down the voltage from IC12 pin 1, as that is the sine wave that changes with the Speed pot. I'm assuming because of filtering the amplitude/voltage changes with the frequency as the Speed pot is moved, so I'm not really sure what I should be setting the Speed pot at the correctly check IC12 pin1 voltage.

It's best done using an oscilloscope (with input coupling set to DC).
You can try to measure it with a multimeter when the speeds is set the slow but if the speeds can go below 0.5Hz the numbers aren't very useful.

I'm going to just take it out, put a 5 pin socket there, and see if I can make one of these dual transistors with a pair or some other transistors.

If you can remove the dual transistor you can do some checks with your multimeter to see if it is damaged.   It is possible to make one using two transistors.

[four_corners]

I'm confused. I'm not sure you can set the modulation range to any set plan any other way from full manual control with the LFO Intensity pot full off (no LFO - does that work?). As you have noted, the LFO amplitude decreases with increasing speed due to the lowpass filter into IC12a rounding down the peaks of the original triangle. Manual control has its own limits, which are different but consistent.

To tweak with the LFO running, you have to judge by ear. with the Intensity control on full. Most would consider the LFO preferable to the manual control anyway. Also, slow speed shows up the range of sweep better. If the synth has a noise generator, use that as a test source. The more high-frequency content, the better to hear flanger action around the fastest BBD clock. To me, the manual pot only makes sense with a changeover switch between manual and LFO and not a mixture which is too interactive.

...and where is the "I can't look" meme when you need one?


I think you did add some extra capacitors, but there isn't a lot of room. Yes, every power pin should have its own local ceramic bypass cap to 0v and if the designers were building it now with SMD they probably would have. This suggests an idea - SMD chip capacitors - if there's a ground trace near a power pin, you can scrape a bit of solder resist off (if there is any) and tack solder an SMD chip capacitor on.

If the LFO ticking is still a bother, you might build the IC11 circuit on a daughterboard with fully filtered and decoupled power rails.

Believe me, if I can get this working again, I'm just going to forget about this tick, and like you said the LFO is more preferable to the Manual anyway, I just thought I'd be able to eventually get it, but of course I now have a non functioning flanger.

I think I might have found the issue though, I'm getting continuity to ground all through this red path...



Also pin14 on the 4013 is showing up as ground.

This all should be positive voltage correct? I either have a shot part that is shorting to ground, or I scratched something and shorted voltage to ground.

[anotherjim]

If 4013 pin 14 was short to ground, you might read other pins to ground with continuity thru the CMOS protection diodes which would normally be reversed with power on 14. Is the continuity reading a dead zero or a diode voltage drop? Also possible +15v was absent since I'm not sure how pin14 gets grounded without also being disconnected from the supply!

[four_corners]

If 4013 pin 14 was short to ground, you might read other pins to ground with continuity thru the CMOS protection diodes which would normally be reversed with power on 14. Is the continuity reading a dead zero or a diode voltage drop? Also possible +15v was absent since I'm not sure how pin14 gets grounded without also being disconnected from the supply!

When I have my meter on the continuity setting, i'm getting 3ohms between pin14 and ground, but it is beeping.

When I pull the 4013 (it is socketed), i'm no longer getting continuity. I didn't want to run it on power too long in case I'm actually frying something, but I could check with power on.

EDIT: looks like it is that specific motorola 4013, the original one. I havent powered it up with the CD4013, but I'll try here in a second.

[four_corners]

I have a feeling I killed the BBD.

If pin 12 and 13 on the 4013 (TP 6 and 7) are good, that means everything before it should be good right (for example, Q18 and Q19)? I'm getting the full correct waveform that changes with Manual, and speed makes it bounce/modulate.

IC12, pin1 gives a sine wave, but the issues seem to show up out of the pin 7 area. 'Sometimes' it gives a correct sine wave, but strangely it will sometimes be quite noisy/glitchy or break up very easily.

Pin15 on the BBD also does this weird glitchy sine wave. The strange thing too is that if I hook up the scope and turn on the synth, the wave will be a bit more clear, but if i take off the probe and put it back on, it is a mess. This happens on pin 2 as well, where it looks like a mix of the modulating LFO and the synth signal, with a ghost sort of waveform trailing behind it but in time. If I unhook the probe, then rehook it, it is a mess.

The only other thing I can think is that CD4013 truly dont work in this circuit, and it needs to be the Motorola one to actually function.

___________

At the same time, I feel like it might not be the BBD though.

I am getting SOMETHING at TP8, which is the output of the BBD, it just isn't pretty. I feel like I should be hearing something and not just only the clean synth signal. I might need to follow the path between the incoming signal, into the BBD, out of the BBD, and towards the effected signal output/feedback signal loop. Maybe I should put together a little audio probe, so I don't have to just use the visual of the scope? I guess anything I am hearing I'd see on a scope though.


I feel like I owe you all a bunch of money or something for all the time you've spent trying to help me! Sorry to keep saying 'thank you' over and over, but thank you!

[puretube]

Things are getting nasty.
Do yourself a favour and cut the northern connection of R103 to IC12 (located to the right side of VR10) for a while.
This disconnects the direct connection of the LFO to the BBD audio-path.
This also disables the influence of the clock-frequency-generating voltages from their control-sectio to the BBD-bias.
Doesn`t matter audibly for the moment.
Almost no "normal" flanger has this feature.
(I would only implement this feature in an "expensive" higher-fidelity flanger,
to allow a continuously high audio-level throughout the BBD, for S/N reasons).

Now you can concentrate on the clock-generation and its control-section pots and trimmers,
without their influence on ticking or otherwise upsetting the audiopath itself.

The "Bias-Adj" VR6 meanwhile should be set by ear to a "sweet spot", where it distorts a little as possible.

[four_corners]

Things are getting nasty.
Do yourself a favour and cut the northern connection of R103 to IC12 (located to the right side of VR10) for a while.
This disconnects the direct connection of the LFO to the BBD audio-path.
This also disables the influence of the clock-frequency-generating voltages from their control-sectio to the BBD-bias.
Doesn`t matter audibly for the moment.
Almost no "normal" flanger has this feature.
(I would only implement this feature in an "expensive" higher-fidelity flanger,
to allow a continuously high audio-level throughout the BBD, for S/N reasons).

Now you can concentrate on the clock-generation and its control-section pots and trimmers,
without their influence on ticking or otherwise upsetting the audiopath itself.

The "Bias-Adj" VR6 meanwhile should be set by ear to a "sweet spot", where it distorts a little as possible.

I disconnected R103 to IC12.

The thing I'm confused about is that if I set Intensity to zero and play a note, and probe the output side of R62 (red arrow), this is the waveform I get (basically the same as the input waveform before R61)...




But, if I turn up the Intensity to 10, and probe the same spot, I get this...



This obviously looks like a modulated signal. If I change the speed, I can see the waveform modulation change (the image above is with speed set pretty high).

Shouldn't I be hearing this in synths output if I have the Flanger turned on and I'm getting this reading at R62? I think I even probed at R81 and got roughly the same thing, or at least some sort of modulated waveform.

There are a ton of diodes and transistors dealing with the switching of the effect, but I have no clue how I would have messed any of that up. It also isn't the literal switch or anything, as I can hear a bit of filtering happen when I turn the effect on, but just no flanging at all (and no clock ticking at all, which I usually could hear even when not playing notes before).

[Rob Strand]

The modulation isn't really modulation but a low frequency signal added to the sawtooth.

Can you probe the output of IC9 pin 1.  If the BBD is passing audio you should see your sawtooth (kbd note) there.

What I think might be going on is the LFO is modulating the output of IC12 at pin 7, you should probe that point, then that low frequency signal is applied to *both halves* of the BBD pin 2 and pin 15.

If the BBD is only passing the low frequency modulated bias then you will only see that at BBD pin 2 and BBD pin 15, no sawtooth   However what I'm afraid is there's something wrong with the first half of the BBD, so  the audio signal is not passing through the BBD from pin 2 to pin 5,6 and then pin 15.   Instead the bias modulation is applied to pin 15 and that passes through the second working half of the BBD.   The LF bias modulation then appears at pins 11 and pin 12, and then through to IC9 pin 1.

In short there might be something wrong with the first part of the BBD and that's blocking the audio making its way all the way through.   The sawtooth you are seeing at pin 3 IC8 comes directly through resistor R63 (100k), not through the BBD.

You will need to do some measurements to confirm what signals you see at each point.  It's only a theory.

[Rob Strand]

FWIW, IC8 (pins 5,6, 7) and Q17 might be stopping the sawtooth getting through so you should check the signal at those outputs as well.


FYI:  If the low frequencies are getting through via IC12 at pin 7 and R107, then that would explain the LFO tick!   (I mentioned it back in reply #71.)

[puretube]

Wait.
That low-frequency "modulation" added to the Saw-sound could be the DC-shift variation of the internal BBD-bias mentioned in reply #32. If this is the case, you don`t see it on the input-side of the BBB.
It origins accumulating through the buckets and shows up only towards the output of the BBD.
Since you cut R103, it is no longer being fought by the output-sine of IC12. This sine now only modulates the clock-generator-frequency.

[puretube]

Not all is lost!
I`d try this with the 2-channel scope:
look at testpoint3 at R62. Sync the scope channel A to this and set the scope-freq. to show only a few sawteeth of this DRY keyboard signal.
Same settings for channel B, looking at pin1 of IC9. This is your (delayed) WET output.
It should show the same sawteeth as channel A, but this time slowly wandering back and forth at the rate of the LFO-speed. (you might have to play with the scope`s sync-options of this channel, to see it shifting compared to channel B.
Now:
if channel B doesn`t show a keyboard-saw at all, but only the (slow and small) sinewave of the BBD-bias that is being "modulated" by the clock-frequency,
there are 2 possibilities:
1. The BBD doesn`t work anymore as mentioned by Rob,
or:
2. The BBD-bias at the input-side (where R70 & R72 connect) is way off, and the BIAS-ADJ. VR6 can`t catch the "sweet spot" anymore, cause we cut the R103 connection, and now some additional DC-potential is missing at pin6 of IC12, which came from the wiper of intensity-pot VR10 (and also had some LFO-sine riding on it - which we wanted to eliminate for a while, to exclude ticks creeping in at this point).
So in the latter case (2.) we need to add some constant clean DC-level to pin6 of IC12:
through a 510k resistor, the other end hooked to the wiper of a 100k (trim-)pot, whose outer ends connect to ground and to + 15V.
Playing with this pot should get the range of the BIAS ADJ. VR6 back into its well-working region.
(There`s a chance, this 100k help-pot needs to be set all the way to 0V - can`t say for sure).
Now: if the BBD is still okay, the delayed, back and forth wandering keyboard-saw should be seen on channel B,
and the flanging should be back audibly - without ticking.

[four_corners]

Not all is lost!
I`d try this with the 2-channel scope:
look at testpoint3 at R62. Sync the scope channel A to this and set the scope-freq. to show only a few sawteeth of this DRY keyboard signal.
Same settings for channel B, looking at pin1 of IC9. This is your (delayed) WET output.
It should show the same sawteeth as channel A, but this time slowly wandering back and forth at the rate of the LFO-speed. (you might have to play with the scope`s sync-options of this channel, to see it shifting compared to channel B.
Now:
if channel B doesn`t show a keyboard-saw at all, but only the (slow and small) sinewave of the BBD-bias that is being "modulated" by the clock-frequency,
there are 2 possibilities:
1. The BBD doesn`t work anymore as mentioned by Rob,
or:
2. The BBD-bias at the input-side (where R70 & R72 connect) is way off, and the BIAS-ADJ. VR6 can`t catch the "sweet spot" anymore, cause we cut the R103 connection, and now some additional DC-potential is missing at pin6 of IC12, which came from the wiper of intensity-pot VR10 (and also had some LFO-sine riding on it - which we wanted to eliminate for a while, to exclude ticks creeping in at this point).
So in the latter case (2.) we need to add some constant clean DC-level to pin6 of IC12:
through a 510k resistor, the other end hooked to the wiper of a 100k (trim-)pot, whose outer ends connect to ground and to + 15V.
Playing with this pot should get the range of the BIAS ADJ. VR6 back into its well-working region.
(There`s a chance, this 100k help-pot needs to be set all the way to 0V - can`t say for sure).
Now: if the BBD is still okay, the delayed, back and forth wandering keyboard-saw should be seen on channel B,
and the flanging should be back audibly - without ticking.

Okay, got some more info! haha. I tried the 2-channel scope idea, but I'm going to step back a bit as the info below I think sheds more light.

First off, pin1 of IC9 doesn't seem to be correct, as it has no version of the incoming signal in it, just a wacky version of the LFO (Intensity and Speed change it) If I play a note, it doesn't change the waveform at all, it is just a constant wave. Here is the scope of Pin1 of IC9.



So Rob was correct about the previous image I posted. It wasn't actually a flanged modulation, just essentially the incoming signal attenuated through R62 and mixed with the LFO (but not the BBD flanged signal).

So here is something interesting. This is a video of my scope directly out of the BBD pin 11/12, at TP 8.

https://www.veed.io/view/44d3f7bd-8cb4-4e5d-b2ae-c009006c40b8

Does this looks like the correct flanged BBD signal? Or possibly this is just the second half of the BBD, and not the first half with the audio.

The weird thing though is if I barely touch the Speed or Intensity pot, all hell breaks loose (maybe my fault? Not a good probe connection/ground? Or touching the pot is doing some sort of grounding thing). Here is a video of that...

https://www.veed.io/view/72ae4139-083c-4f27-b568-99f3e2ce342a

This must mean something is going on between VR2 and IC9, right? This makes sense as well because the Feedback pot and VR1 don't seem to change the BBD output I showed at TP8 (the testpoint in the videos above). Basically it seems like before a wet signal can get to the Feedback circuit to loop back around, it is dead (as IC9 pin1 is connected to the Feedback loop).

Maybe desolder IC9 and socket a new TL022/4558?

I'm a bit confused how the LFO is showing up at IC9 pin1, especially since I disconnected R103.

[four_corners]

FWIW, IC8 (pins 5,6, 7) and Q17 might be stopping the sawtooth getting through so you should check the signal at those outputs as well.


FYI:  If the low frequencies are getting through via IC12 at pin 7 and R107, then that would explain the LFO tick!   (I mentioned it back in reply #71.)

On the base and emitter of Q17 I get the keyboard signal sawtooth wave. The collector is +15V. IC8 pin 7 I get the same waveform as Q17, pin 5 and 6 I'm getting nothing but noise.

IC12 pin7 is just messy waveform of some kind right now so I'm not sure if the LFO is getting through (I dont think it is, but it might), this is correct though as I have R103 disconnected, right?

[puretube]

So Rob was correct about the previous image I posted. It wasn't actually a flanged modulation, just essentially the incoming signal attenuated through R62 and mixed with the LFO

.
No doubt the sawteeth are the Dry input sig. via R62.
But IMHO that low added sine can`t be the LFO sig. itself, since R103 is cut.
But like I said it can be the BBD-bias-DC "dancing" at the speed of the LFO, because the LFO still modulates the clock-frequency, which in turn causes to move the bias-DC up and down.

[puretube]

Does this looks like the correct flanged BBD signal? Or possibly this is just the second half of the BBD, and not the first half with the audio.

Not be the "flanged" signal, because the real flanging sound happens not before R62 & R79 join.
But it does look like a time-modulated keyboard saw, i.e. the variably delayed (not: "flanged") copy of your input signal varying in pitch at the rate of the LFO.
This signal was what I expected to see at pin1 of IC9 (as well as on pin7 of IC9, and as a matter of fact on TP8 too, just with lots of clock-frequeny-artefacts spoiling a sharp scopetrace).
Now while you get a keyboard-signal at this TP8, look at it with one scope-channel, and "follow" the signal from there with the other channel (same settings), to see where it suddenly seems to vanish. (VR2 wiper, both sides of C34, R73, R75, R76, R77, R78, pin1. - You won`t see much on the pin6-side of R74).

Leave Feedback-pot at zero, because it might show a parallel ghost-trace.

Set the scope timebase so only a few sawteeth show up, so that the low-frequency "wobble" doesn`t disturb the view.