SOLVED Delegate (Boneyard Edition): LED not changing with signal

jeffwhitfield

jeffwhitfield

Well-known member
Hey folks! So...finished putting together a Delegate compressor last night. But after testing it a bit, seems something is askew. One thing I noticed is that the LED isn't changing with the signal. It just stays on no matter what...not even a flicker. Any suggestions on how to test things to reveal the problem? I mean...I could trace the board...but this could very well be a very simple thing. Figured I'd ask in case anyone happens to have had a similar problem. :P

IMG_0332.jpeg
 
Sort by date Sort by votes
Ok, so I think it's working...sort of.

I switched out one of the 2N5089's. Seems one of them was mismatched. It's a 2N5089...but not the same brand. Probably doesn't make a difference...but you never know.

Did some tests and, yes, the LED is indeed changing based on the value of the Release. It's rather subtle though. Go over a certain point, like around 11 o'clock, and it doesn't appear to have much of a difference with anything above that.

Compression is rather subtle too. Did some tests on my bass...basically doing some slapping and popping to test the compression of transients. Seems like I have to turn the ratio and sustain up quite a bit before the transients get knocked down.
 
Upvote 0 Downvote
I used a white LED in mine. It was a board I etched, but same principle. Also, did you test it while the back was on? I'm a bass player as well, I've made that mistake.:ROFLMAO:
 
Upvote 0 Downvote
Like @jimilee said you have to test it in the dark or with the enclosure fully closed to keep the LDR in full darkness.

Couple of things that may help:

- Replace that LED for a white one and bend it to have the tip of the LED pointing at you LDR.
- Test you LDR to see if it is produce 500K in the dark

I also understood that the speed of the LDR to react to light is a major factor in this circuit. You may want to test a few different ones and see if you get better result.
 
Upvote 0 Downvote
I don't think the problem is with the LDR. I just got a real Warden pedal to compare it to and, after cracking it open, I noticed right away the difference. With the EQD Warden, the LED is quite consistent. It's barely on when no signal is present. Mine, not so much.

To test it, I set the Ratio to 12 o'clock, Sustain all the way on, and Attack and Release all the way off. The Warden reacted well to transients and the LED came on as expected. Mine sort of came on...but not very bright. I switched out the LED with a green 5mm one and the result is pretty much the same. I think the issue is in the LED itself. You may be right...a clear one might work better, which is exactly what EQD uses in theirs. Got some LED's on order so should know more soon. :D
 
Upvote 0 Downvote
Alright, so...tried a few different LED's...a white one, a 'warm' white one, and a chartreuse one. Strangely enough, the chartreuse sort of worked better. I'm thinking this has something to do with the resistance on the LED. Without any signal, the LED is on...like on enough that when a signal is present, the change in the LED is quite minimal. By comparison with the Warden, the change is quite drastic.

So then...something tells me there's a resistance issue somewhere. Question: Are there any difference between resistor values between the silkscreen and the documentation? I did see a few difference with the capacitors (which I went with the documentation over the silkscreen). But what about the resistor? Are they all the same?

Also, would changing R17 make any difference with the default intensity of the LED? I'm wondering if I need to increase R17 a bit to compensate for the extra brightness of the LED.
 
Upvote 0 Downvote
Alright, a bit more testing. Green diffused LED's seem to work best for sure...but...still not working in a way that makes sense.

In this test, everything was at noon but with Attack and Release at 9 o'clock. Warden showed LED differences with signal activity. My pedal...not so much. With Attack and Release off, both seem to be pretty much the same. So...hmm...
 
Upvote 0 Downvote
Changed the 220n's to 22n's...no real change to speak of. Green LED isn't as bright but doesn't seem to reflect much change with everything at noon and Attack/Release all the way off. A white LED is alway on to some degree...which it shouldn't. I would think that regardless of the LED, it shouldn't be on at all. So...time to audit the resistors and do some tracing. :P
 
Upvote 0 Downvote
Did some voltage testing and came across an oddity. The LED has a voltage reading of about 16v. That can't be right. Should be lower, right? Seems to be coming from D101 and D102...or leg 8 of IC1. Zgrav might be right...too much voltage?
 
Upvote 0 Downvote
benny_profane said:
IC1 is a charge pump. Vcc is 2x the input (minus the Vf of the series diodes). With an input of ~9V, Vcc should be ~18V.
Click to expand...
Makes sense...but I'm wondering if different diodes make a difference. The Warden uses 1N4001's. The Delegate calls for 1N5817's, which have half the forward voltage of a 1N4001. The Warden reads 15.4v of the leg of the charge pump going to the LED's, whereas mine is at around 16.5v. So, it looks like there's 1v difference...which could explain things. That extra volt could be what's causing the LED to not work properly. Could be that using 1N4001's or even 1N4002's give it the forward voltage needed to knock it down a little bit.

Then again, I'm not an engineer so, like EQD's Jamie, I'm flying by the seat of my pants. :P
 
Upvote 0 Downvote
So, yeah, 1N4001 does indeed drop the voltage a bit. However, still having a problem. I think it might be the Release pot that's the issue. It's weird. All the way off it seems to work properly. But the minute I turn it up even just a little bit, the LED goes dark and doesn't wanna light up much with the signal. I think that's probably the issue. Will see.

Fun times! :D
 
  • Like
Reactions: fig
Upvote 0 Downvote
The diodes in the charge pump block aren't an issue. This circuit is constructed so that Vcc is 18V, Vref is 9V, and GND is 0V. (It should have been made as a split rail supply such that Vref bias could be 0V—but that's not what we're discussing here.) If you try to drop Vcc, you'll end up impacting the bias of everything (in particular, the voltage swing of the op-amp).

If the LED isn't responding to dynamics, there's an issue with the signal reaching the LED. Looking at the top right of the schematic, verify all of the parts and connections there.

Also, keep in mind that this circuit has some significant changes compared to the EQD version. The principle is the same, but the differences are important.
 
Upvote 0 Downvote
benny_profane said:
The diodes in the charge pump block aren't an issue. This circuit is constructed so that Vcc is 18V, Vref is 9V, and GND is 0V. (It should have been made as a split rail supply such that Vref bias could be 0V—but that's not what we're discussing here.) If you try to drop Vcc, you'll end up impacting the bias of everything (in particular, the voltage swing of the op-amp).

If the LED isn't responding to dynamics, there's an issue with the signal reaching the LED. Looking at the top right of the schematic, verify all of the parts and connections there.

Also, keep in mind that this circuit has some significant changes compared to the EQD version. The principle is the same, but the differences are important.
Click to expand...
Gotcha. I'll do some traces off that. The behavior now is that, with everything timed but Attack and Release, if I dime the Release then I get barely even a blip off the LED. That can't be right. The brightness should be the same whether it's full on or off, right? The only thing that should change there is the length of time the LED stays on.
 
Upvote 0 Downvote
jeffwhitfield said:
The brightness should be the same whether it's full on or off, right? The only thing that should change there is the length of time the LED stays on.
Click to expand...
No. Check out an Attack, Decay, Sustain, Release (ADSR) graph. ADSR graphs show the behavior of a note (e.g., piano keystrike, plucked guitar string, etc.) as a function of amplitude in the time domain. With the ATTACK and RELEASE controls, you're altering the behavior of those parts of the note's envelope. Unless very extreme, they are not binary on/off states, but rather ramping (visualized with varying levels of brightness). The LED's brightness affects the value of the LDR in the feedback loop of the op-amp. This is how the dynamics are modulated.

Also keep in mind that an LDR is going to be more sensitive than your eye. Especially if you're looking at an LED in a lit room, you might not see the subtle changes.
 
Upvote 0 Downvote
Could it be working as intended? If you assemble the enclosure and don’t look at the light, what do you think?

Edit: I want to build a delegate but not sure which version. Maybe I’ll breadboard both and see how they compare.
 
Upvote 0 Downvote
You must log in or register to reply here.
Back
Top