Dark Rift Delay - Distortion/Fuzz in Wet Signal

absolvism

Member
Hello!

I just finished building a Dark Rift, and everything seems to be working... except the repeated/wet/delay signal has a distortion/fuzz in it..

So the first pluck always sounds normal, but then all the succeeding delay signal has a very pronounced fuzzy/scratchy undertone to the whole thing. I've tried different amps and cables and it appears that it's definitely in the pedal itself. If I set the mix to 100% dry, there is no fuzz, it's just in the wet.

Besides pulling the whole thing apart and starting over, does anyone have any ideas or directions to troubleshoot that they might be able to help point me in? I can record example audio if that would help.

Thanks so much in advance,
Ellis
 
I listened to the sound clip a few more times. Around 11 seconds in, I hear the dry signal clean, the first repeat is clean, and the repeats after that are distorted. This points to the path thru the REPEATS control. Pretty sure I know the cause of the distortion. I have a question into The Man himself. If I'm right, this is easily fixed and there's nothing wrong with your PT2399.

As a check on my theory, turn REPEATS all the way down and see if you get any distortion.
 
As a check on my theory, turn REPEATS all the way down and see if you get any distortion.

At least with my pedal, there actually is dirt in the first delay (and the sole repeat w REPEATS at 0). If I play very softly, I might not get as much dirt, so that's what you may have been hearing in the other clip.

Here's a clip with REPEATS at 0 with some stronger plucks. https://ellisedwards.com/content/1_repeat.mp3

Thanks for the help, Chuck!
 
OK, that shoots down that theory. The PT2399 is still the prime suspect.

Since I have PT2399s from 2 or 3 different sources, the next time I build a pedal with a PT2399 in it, I'm installing sockets!
 
OK, that shoots down that theory. The PT2399 is still the prime suspect.

Since I have PT2399s from 2 or 3 different sources, the next time I build a pedal with a PT2399 in it, I'm installing sockets!
Ok, so I tried 2 other PT2399’s and am still having the issue.

The board was a bit fluxy (I didn’t have flux remover on hand yet when I built this one and IPA was leaving a mess) so I was going to try to clean the board and replace the wiring, BUT I managed to pull off the GND pad from the foot pedal wiring area (see pic). The pad on the other side is intact. Is there a workaround or is the board shot? Can I attach that GND wiring to a different connection? Sorry to add more problems to this troubleshooting ?
 

Attachments

  • A60CABFE-667C-45D9-9146-85CF675DCB70.jpeg
    A60CABFE-667C-45D9-9146-85CF675DCB70.jpeg
    290.7 KB · Views: 25
Here’s full pics of both sides of the board minus ICs and ext wiring btw
 

Attachments

  • 43AE8EC3-E204-491B-9FCC-6B454201AC19.jpeg
    43AE8EC3-E204-491B-9FCC-6B454201AC19.jpeg
    402.1 KB · Views: 44
  • 4CE93608-ED63-4B5A-B3EB-2FC416307919.jpeg
    4CE93608-ED63-4B5A-B3EB-2FC416307919.jpeg
    412.3 KB · Views: 44
  • F7168C27-8C68-470D-B8C6-1126BDF2C60C.jpeg
    F7168C27-8C68-470D-B8C6-1126BDF2C60C.jpeg
    343 KB · Views: 42
Ok, so I tried 2 other PT2399’s and am still having the issue.

The board was a bit fluxy (I didn’t have flux remover on hand yet when I built this one and IPA was leaving a mess) so I was going to try to clean the board and replace the wiring, BUT I managed to pull off the GND pad from the foot pedal wiring area (see pic). The pad on the other side is intact. Is there a workaround or is the board shot? Can I attach that GND wiring to a different connection? Sorry to add more problems to this troubleshooting ?
Is the thru plating still in the hole, or did that come out with the wire? You can either pick up ground from somewhere else, or resolder to the remaining pad. I'd strain-relieve the wire after soldering with a drop of epoxy so the wire doesn't pull off the remaining pad if stressed.

If IPA is leaving a mess, it's because you're not using enough. I flood my boards to wash the flux away.

Those stamped steel sockets are not the best. Poor retaining force, especially after inserting and removing the ICs a few times. I checked all the parts on the top half of the board and they look ok. Pad 3 of the MIX pot looks like the solder may not have wetted to the pad. Hard to tell due to glare.
 
Is the thru plating still in the hole, or did that come out with the wire? You can either pick up ground from somewhere else, or resolder to the remaining pad. I'd strain-relieve the wire after soldering with a drop of epoxy so the wire doesn't pull off the remaining pad if stressed.

If IPA is leaving a mess, it's because you're not using enough. I flood my boards to wash the flux away.

Those stamped steel sockets are not the best. Poor retaining force, especially after inserting and removing the ICs a few times. I checked all the parts on the top half of the board and they look ok. Pad 3 of the MIX pot looks like the solder may not have wetted to the pad. Hard to tell due to glare.
I appreciate you taking a peek at it.

I feel like it’s just the silver coating on the pad of the one side that came off. I think the inner plating is still there. I’ll try resoldering to it first and epoxy gluing as you recommend.

2 quick followup questions if that’s ok—

At what stage do you flood wash your PCBs? What I’m wondering is: other than pots and ICs, what other components would be too sensitive to be installed prior to IPA flooding the board?

And what you recommend in place of the stamped steel sockets?
 
I wash my board before installing the pots to avoid washing flux into the pots. IPA won't hurt pots, but dirty IPA might. If I have to do cleaning with pots in place, I direct the flow of dirty IPA away from the pots. I use a toothbrush to scrub any stubborn spots. Not aware of any other parts that are sensitive to IPA.

Small Bear, Mouser, DigiKey and others sell machined pin sockets. They cost a little more, but are much more reliable. You want to keep dirty IPA out of sockets too.
http://smallbear-electronics.mybigcommerce.com/ic-socket-machine-pin-mill-max-8-pin/
 
Ok so I’ve torn this thing apart and banged my head against every wall—
I’ve:
• Reflowed bunk or questionable joints
• Swapped in/out 2 other PT2399 (from Smallbear)
• Replaced every bit of wiring
• Replaced hardware
• Removed fluxy film from rear
• Added footswitch daughter board
• Prayed to every god
.. and this damn thing is still distorting very audibly in the repeat signal

I’ve verified that all the joints look good and appear to make solid connections. I can’t even imagine where to go next but I don’t want to throw this in the garbage :(

Any thoughts? Is it possible/realistic the 2 additional PT2399 chips from SB are still bad? That seems like the obvious culprit but I’m 0/3, and 2 bad ones from SB seems hard to believe.

Could there be a quality issue with parts? Like is anything I’ve got in there not recommended? Could the clip-type sockets be a problem? Are the blue 1u caps crappy? Could either cause dist in the signal?

Other pedals have worked out totally fine but this has me baffled.
@bschobs: Did you ever get yours fixed??

5A2A8954-54B5-4961-9BF9-4CAAD38E96B0.jpeg
AD10C58F-65C3-4B05-9D60-E4639F9EBC7C.jpeg
 
A few shots in the dark:
Verified that the 5V supply on pin 1 of the PT2399 is in spec.
Try subbing a different dual opamp for IC3, something other than an LM833.
If you have an audio probe, try setting REPEATS to zero and listen to pin 15 (input) and pin 12 (output) of the PT2399. This will confirm whether or not the PT2399 is the source of the distortion.

I wonder where leaving pin 13 on the PT2399 floating could be causing problems. Princeton is pretty economical with applications information in their datasheet. They don't say one way or another whether it is ok to float any of the pins. Generally, it's a bad idea to leave the inputs of an opamp not connected to anything.
 
A few shots in the dark:
Verified that the 5V supply on pin 1 of the PT2399 is in spec.
Try subbing a different dual opamp for IC3, something other than an LM833.
If you have an audio probe, try setting REPEATS to zero and listen to pin 15 (input) and pin 12 (output) of the PT2399. This will confirm whether or not the PT2399 is the source of the distortion.

I wonder where leaving pin 13 on the PT2399 floating could be causing problems. Princeton is pretty economical with applications information in their datasheet. They don't say one way or another whether it is ok to float any of the pins. Generally, it's a bad idea to leave the inputs of an opamp not connected to anything.
Thanks for the reply and the ideas to investigate. I’ll probe into these as best I can and report back. I truly appreciate it
 
Install a 5mm green led between pin 7 and ground. Long leg to pin 7, short to gnd (pin 3 on the 2399).

This distortion in the repeats is well known across many diy forums. Modern thinking also suggests using 1 5v regulator per chip as oppose to both coming off the same 5v line.

You can test the led thing with a crocodile clip. Make a small hook at the end of the long leg, hook it around pin 7, the clip the other leg to ground. No soldering, no mess, but it worked for me on a different pedal using a similar topology.
 
Hey @Chuck D. Bones - I checked the V to pin 1 on the socket for the PT2399 and it showed 4.81. Is this an ok value? Datasheet implies it can go down to 4.5. Also --- I did one more test before I was going to try @BurntFingers idea with the green LED (which I will still try) and it was interesting.

So I pulled out the 2399 entirely and ran signal through with it engaged, and I found that with the mix on and variable levels of repeat I can get the distortion without the PT2399 in at all, on the initial note hit. It's a little hard to coerce out, but especially played a bit harder on 6th and 5th strings it's definitely distorting in a way it should not.

I have built a probe, but not yet probed. I found it interesting that I could get it to fizz without any PT2399 in. It seems to suggest to me that there's something else in the circuit or build that causing the fizz and maybe the PT2399 is amplifying that signal error? Does this make sense to anyone or point to another suspect component?
 
I don't think that's a fair test. With the PT2399 out, IC3 will not bias correctly and the signal can flow backwards thru the Repeat path.

Not surprisingly, there are some shortcomings to EQD's design. IC3 is biased by the PT2399. Not good because the PT2399 runs on +5V and it's output is biased at 1/2 of that. You measured +4.8V on IC2-1, which means that IC2-12 is probably around +2.4V. IC3 is DC-coupled to IC2, so it's also biased at +2.4V. That leaves very little headroom. The Sea Machine, by comparison, biases all of the opamps at +4.5V. I built a Sea Machine and tweaked the TIME control so I works like the Space Spiral, only without the headroom problem.

The LED will help the PT2399's headroom problem, but if you're driving this pedal with hot pickups, then there's still the opamp headroom issue to deal with.

This is no fault of PedalPCB. They faithfully replicate the original pedal, warts and all. With some of those pedals, it's up to us to improve on the original's design.
 
I don't think that's a fair test. With the PT2399 out, IC3 will not bias correctly and the signal can flow backwards thru the Repeat path.

Not surprisingly, there are some shortcomings to EQD's design. IC3 is biased by the PT2399. Not good because the PT2399 runs on +5V and it's output is biased at 1/2 of that. You measured +4.8V on IC2-1, which means that IC2-12 is probably around +2.4V. IC3 is DC-coupled to IC2, so it's also biased at +2.4V. That leaves very little headroom. The Sea Machine, by comparison, biases all of the opamps at +4.5V. I built a Sea Machine and tweaked the TIME control so I works like the Space Spiral, only without the headroom problem.

The LED will help the PT2399's headroom problem, but if you're driving this pedal with hot pickups, then there's still the opamp headroom issue to deal with.

This is no fault of PedalPCB. They faithfully replicate the original pedal, warts and all. With some of those pedals, it's up to us to improve on the original's design.
Oh for sure I don't think PedalPCB is at any "fault" of any kind. Just the reverse - I think he is a literal hero. And I appreciate everyone's help- I hope to be at a point where I can pay it forward someday. I just keep holding out hope that there is something that can be done that will tame this distortion/improve on this pedal & it can end up useable. (Also, the other PedalPCB boards I've built out have gone together perfectly FWIW - such an awesome company)

I didn't realize that a test without the PT2399 in is not a real or fair test. Disregard entirely.

I measured IC2-1 to IC2-12 and it reads at +2.45V. Am I understanding correctly that the IC3 headroom bias issue is a suspect for the distortion? Is that an unmodifiable scenario? The guitar I was testing w/ was a LP with pretty tame pickups.

I have some various opamps coming in the mail on a slow boat, so I can can swap IC3 for something else in some time and report back, as well as re: audio probe.

I had also read about someone in a similar topology changing one of the pots from 50k -> 25k, but since it's fizzing even on low times/low repeats/low mix, that doesn't seem like it could help or am I not seeing things right?
 
It would be good if someone who owns a Space Spiral or Dark Rift could chime in and confirm or refute whether this is normal behavior for that pedal.

My modified Sea Machine (Sea Horse) makes a little noise and distortion at longer delays, but that's the consequence of slow sampling. The PT2399 was designed for reverb applications. Longer delays will not sound as clean. We could tune the anti-aliasing filters down lower, but then the tone would be darker. It's a compromise.

If you want a clean, crystal clear delay, then FV-1 is the way to go.
 
There's no way this is normal behavior- I listened to SS demos and it has nothing to do with the longer delay, which would is understandable.

Here's a video showing the problem. I am playing fairly hard in order to accentuate the problem, but there is no setting where this is not sound-ruining. You'll be able to see clearly what I mean:


I also still think the highest probability is that I've done something stupid or rookie (or 3 bad PT2399s), and it doesn't have anything to do with the circuit or the nature or standard noise of PT2399. I've read heaps of posts about PT2399's in the last week and it seems like this is a complaint that does arise, so I'm hoping working through this might save other PedalPCB customers some annoyance down the line, and saves me (and maybe some others) from having to take a total L on this. If that happens, I'll ring up the FV-1. Sorry this is dragging on.
 
Back
Top