Sea Urchin based delay - Is this a PT2399 problem?

[subvader12]

Greetings everyone!
I've been working for a couple of days on a delay pedal based on the Sea Urchin, the thing is I re-draw the schematic to make a custom PCB, etched one of it and just finished without a positive result.
Once I finished soldering all the components and wires, I connected it but it only gave one repetition at the output. I had pins 3 and 4 of the PT2399 disconnected between, and tried jumping them to GND as I saw on older posts, but it didn't changed the sound and still there was a single repeat, so I took the output jack cable and started testing the audio from pins 9-16 of the IC (I saw 4 outputs there), the first time I tried it some of them didn't have sound (only pin 14 was passing the signal). I started looking for bad soldering points and good continuity between trails as I was testing for audio and it now started passing signal at the outputs of the PT2399 and all voltages are 2.5v, but they all sound different specially the OP outs, 9 and 12 as tracing them between components gave low volume at some points and it didn't sound a feedback effect (only one repeat at all outputs), is this normal?
I then continued looking for bad connections and misplaced components but they all were just fine, I was giving up when I tried the sound at the normal output of the circuit and boom! Delay! (Really no idea what happened there) I was so happy I played for a while and then disconnected it, after few hours I was trying it again and came back to the starting problem, only one repeat.... I saw a post saying the Sea Urchin tends to lock the PTs, but could it be problem here? I'll attach the schematic I'm working with, maybe I'm missing something here. Any help will be very appreciated as I'm starting to lose my mind on this one.

[ElectricDruid]

No, I don't think it's a PT2399 problem. The feedback path looks wrong to me.

The output signal should come out of pin 14 (yours does, via a couple of passive lowpass RC filters which will hit the level) but then it's taken *into* the wiper of the feedback pot. Should be to the top. The wiper should then go to a resistor that feeds into the filter on pins 15 and 16. So I think the top and the wiper of the feedback pot are switched.

The filter on pins 15 and 16 of the PT2399 has been broken up. You've got one of its capacitors by the chip, and then the rest of it floating about over by the feedback control. That makes it very difficult to work out whether that filter is correctly assembled. I think it's right, but it's hard to be sure. It's a multiple feedback filter, so as far as possible it should *look like* a multiple feedback filter.


I also spotted a couple of minor things:

1) The 9V supply is taken from *before* the noise filtering, so that's not doing anything useful.
2) Are R11 and R12 the right way around? I've got a similar schematic that has the 20K where the 10K is and vice versa.


In general, I find the schematic very hard to read and it doesn't surprise me that you're having trouble. There's a lot of connections just marked with labels, and the labels don't tell you anything. It says "PT2399-14" instead of something useful like "Wet signal" or "Repeats signal" or something. The signal path should be as obvious as you can make it.

Also the PT2399 is upside-down, and the TL072 symbol is good for a board layout, but difficult for a schematic.

Sorry if this comes across as super-critical, but I think your schematic is making life harder for you, and it should be making it easier!

[subvader12]

No, I don't think it's a PT2399 problem. The feedback path looks wrong to me.

The output signal should come out of pin 14 (yours does, via a couple of passive lowpass RC filters which will hit the level) but then it's taken *into* the wiper of the feedback pot. Should be to the top. The wiper should then go to a resistor that feeds into the filter on pins 15 and 16. So I think the top and the wiper of the feedback pot are switched.

The filter on pins 15 and 16 of the PT2399 has been broken up. You've got one of its capacitors by the chip, and then the rest of it floating about over by the feedback control. That makes it very difficult to work out whether that filter is correctly assembled. I think it's right, but it's hard to be sure. It's a multiple feedback filter, so as far as possible it should *look like* a multiple feedback filter.


I also spotted a couple of minor things:

1) The 9V supply is taken from *before* the noise filtering, so that's not doing anything useful.
2) Are R11 and R12 the right way around? I've got a similar schematic that has the 20K where the 10K is and vice versa.


In general, I find the schematic very hard to read and it doesn't surprise me that you're having trouble. There's a lot of connections just marked with labels, and the labels don't tell you anything. It says "PT2399-14" instead of something useful like "Wet signal" or "Repeats signal" or something. The signal path should be as obvious as you can make it.

Also the PT2399 is upside-down, and the TL072 symbol is good for a board layout, but difficult for a schematic.

Sorry if this comes across as super-critical, but I think your schematic is making life harder for you, and it should be making it easier!

Thanks for taking the time to actually see through my schematic,

For the first part, I understand your point of the wiper as I'm seeing at the Rebote 2.5 and it goes from pin 14 to top of it, but the Sea Urchin schem I based this on have it on the wiper and then top as out, mechanically it won't matter as the resistance between those two points should change equally, no? I'll look into it nevertheless. Also, I'll try and change the values of the passive RC filter there to see if signal gets stronger without sound decadence.

For the second part about the filter at pins 15 and 16, comparing to the OP-AMP image you attached, it seems fine, sorry if I broke up the schematic with labels but in my head it makes sense and doesn't get so messy and will try and change the labels for more descriptive ones (ironically, I was trying to get things cleaner), so I think the problem isn't laying on that part of the circuit.

I forgot to mention I'm using a 12v power supply, so the TL072 its getting (14v somehow at input) +7v and -7v, is it fine for it?
Can you post that schematic where R11 and R12 are inversed? So I can check for possible faults in the original drawing I have too.

You get me motivated to try and solve this mystery! Thanks again for the response!

[duck_arse]

welcome to the forum, subvader12.

first thing, make sure you always connect pins 3 and 4 ogether and to ground. and then don't think about them.

as ED says, it's difficult for us to follow your connections on the circuit diagram, as your diagram conventions are not the same as our diagramming conventions, or at least the conventions we all expect. and so, when you say your opamp supply .... the PT2399 ins and outs bias at about 2V5, whereas your opamp ins and outs are biased at some other voltage, 6V? 7? so, you need blocking caps between the opamp and the pt2399 to isolate the different levels.

it looks as tho you have pins 15 and 16 shorting to pin 1 of the opamp. no consistency between the drawings.

[subvader12]

welcome to the forum, subvader12.

first thing, make sure you always connect pins 3 and 4 ogether and to ground. and then don't think about them.

as ED says, it's difficult for us to follow your connections on the circuit diagram, as your diagram conventions are not the same as our diagramming conventions, or at least the conventions we all expect. and so, when you say your opamp supply .... the PT2399 ins and outs bias at about 2V5, whereas your opamp ins and outs are biased at some other voltage, 6V? 7? so, you need blocking caps between the opamp and the pt2399 to isolate the different levels.

it looks as tho you have pins 15 and 16 shorting to pin 1 of the opamp. no consistency between the drawings.

Thanks for the welcome! Glad to be part of this amazing community.

The drawing I made is the same as the Sea Urchin you find online, I re-did it just for making a custom PCB in the software I use.
All the voltages from pin 9 to 16 of my PT2399 are 2.5V. The blocking cap I think you're mentioning is C3 (electrolytic) that goes from pin 1 of TL072 (positive) to the 15 and 16.

I was messing with the circuit this afternoon and came to something:

Somehow, R12 and R13 got disconnected from the C11 joint that goes to GND, and it started the delays again...



But... Now there's an annoying noise maybe coming from the DC supply, or a bad GND joint. Or are the pins 9 to 13 of the PT2399 filtering all that noise out? I don't really know.

I'll attach a sound sample so you can maybe help me identify this issue: https://drive.google.com/file/d/0B1HnZDpxapJ-cUVhZTZpNHFaZGpPakV0NTVNUFhkMXNmR1V3/view?usp=sharing

[ElectricDruid]

Sounds like a bad connection to me. Could be ground, but it could be elsewhere. Check all your soldering and external wiring.

[subvader12]

Sounds like a bad connection to me. Could be ground, but it could be elsewhere. Check all your soldering and external wiring.

I can totally make another PCB as I'm practicing my etching skills anyway if that's a bad connection (it is messy now to work with this one), but can I leave R11 and R12 disconnected? Will it affect the overall circuit in the future? It sounds amazing minus the noise generated, good sustain on the feedback and clean sound on highs and lows. What do you think?

[bean]

My old drawing of the "Sea Urchin" is frankly not that great. Accurate but that PT2399 library part is confusing for some people. Here it is re-drawn in a more conventional way with a few new mods added. Hope this helps.

fixed

[duck_arse]

apologies, subvader12, I didn't even look at that top corner section.

[bean]

I went ahead and did up a single sided layout for the 2019 version since I was planning on doing one anyway (at some point). For 125B.

Here's the direct link to the project with etching artwork: http://www.madbeanpedals.com/EP/downloads/delay/SeaUrchin2019.zip
http://www.madbeanpedals.com/EP/index.html

[subvader12]

My old drawing of the "Sea Urchin" is frankly not that great. Accurate but that PT2399 library part is confusing for some people. Here it is re-drawn in a more conventional way with a few new mods added. Hope this helps.

fixed

Hi Bean! I'm a great fan of your work and it makes me so happy to know you're here willing to help people in the same spot as me. I was comparing your drawing with mine searching for possible errors but everything seems fine in the one I made, I totally loved the mod section and already added D3 to limit my oscillating clipping, seems to work fine!
I have a few questions though, I see you use ceramic caps at C4, C13 and C14, but original is electrolytic, any reason for this?

Also, I see an addition of C23 and C18 100nF caps, are they for filtering voltage noises coming from the supply?

As my latest question have not been answered, do you have an idea of why disconnecting (I'll use your resistors name) R9 and R10 gives me the actual delay in the circuit, otherwise it'll only be one repeat at output? Are the values attenuating the repeat signal, making it almost silent?

[bean]

Sure thing! Happy to help.

I see you use ceramic caps at C4, C13 and C14, but original is electrolytic, any reason for this?

It's just personal preference. Whenever possible I'll sub a 1uF film cap in place of an electrolytic if it's in the path of the signal. There are a few instances where this could be a bad idea but here it's fine. The 1uF make for a more expensive build so electrolytic are a better choice when production costs are being considered but as DIY we can do things how we like

Also, I see an addition of C23 and C18 100nF caps, are they for filtering voltage noises coming from the supply?

I put 100n bypass caps in parallel with all power supply parts on my builds. Electrolytics are not so hot for filtering high frequency noise and putting a small cap (film or MLCC) in parallel with them is just a bit of extra insurance. With the well regulated power supplies we have now it's not really a problem but it only adds a few cents cost so why not do it.

Do you have an idea of why disconnecting (I'll use your resistors name) R9 and R10 gives me the actual delay in the circuit, otherwise it'll only be one repeat at output? Are the values attenuating the repeat signal, making it almost silent?

If I had to guess it's because of a short on your board or maybe in your IC? The two delay output pins are 9 and 12 and by disconnecting R9 you should be disconnecting the actual delay output from the filter used in LPF2 (pins 13 and 14). Unless there is something I am forgetting I'd say that is not normal but can't say exactly why it's happening to you.

Actually, one trick to make a small improvement in fidelity on PT2399 delays is to skip LPF2 altogether and instead use an external op-amp for that filter stage. There are a couple good reasons to do this and one bad: 1) and external op-amp will run at a higher voltage (9v vs 5v on the PT2399) and 2) it gives you more options on exactly what type of filtering you want to do on the delay output. I've spent a lot of time fooling around with the PT and using LPF2 can be a big source of noise, IMO. Anyway, the one bad reason not to do anything I just said is added parts and cost.

In this version of the Urchin I added a 1n914 diode on the regulator to bump up the PT supply voltage a little closer to 6v. Not a big difference, but any little bit helps and it seems to run fine at 6v on the other delays I've tried it on. I can't remember where I saw this idea originally. Probably here.

[ElectricDruid]

Hi Bean! I'm a great fan of your work and it makes me so happy to know you're here willing to help people in the same spot as me. I was comparing your drawing with mine searching for possible errors but everything seems fine in the one I made, I totally loved the mod section and already added D3 to limit my oscillating clipping, seems to work fine!
I have a few questions though, I see you use ceramic caps at C4, C13 and C14, but original is electrolytic, any reason for this?

Also, I see an addition of C23 and C18 100nF caps, are they for filtering voltage noises coming from the supply?

As my latest question have not been answered, do you have an idea of why disconnecting (I'll use your resistors name) R9 and R10 gives me the actual delay in the circuit, otherwise it'll only be one repeat at output? Are the values attenuating the repeat signal, making it almost silent?

Notice that in Bean's drawing the feedback pot is the way around I suggested - signal comes out of pin 14 to the top of the pot, and the wiper takes signal out to be mixed into the input at pin 15/16. The other way just didn't look right.

If removing R9 and R10 makes things better, you've probably got a mistake in that filter. It should be fine with them in, so if it isn't, there's a problem somewhere. A short? A bad joint? A wrong cap value?

[ElectricDruid]

Also I double-checked what I said about the 10K and 20K resistors being the wrong way around. The filter will actually work either way around, but it changes the gain from unity to x2 (+6dB). The way you have it is set up for +6dB, which makes sense. Sorry for the false alarm.

T.