Delegate (Warden) Upgrades

[Chuck D. Bones]

Good question. I don't know. I just checked my pix and they are 22nF in my build. 220nF may work fine, but they seem to be overkill.

OK, here's what happened. In breadboard testing the Boneyard Edition, I found that increasing C7 & C8 was beneficial so I told Mr. PedalPCB that 220nF was the preferred value. That's why the Build Docs say 220nF. That conversation took place 17 months ago and had slipped from my memory banks. Increasing those caps speeds up the peak detector and allows the user to set a faster attack response, if that's what you want. You can still dial back the attack speed with the ATTACK knob.

In summary, the Build Docs are correct. 22nF will work; 220nF will work better. I should update my Delegate.

 

[jmwreck]

In summary, the Build Docs are correct. 22nF will work; 220nF will work better. I should update my Delegate.

Ok, thanks for the info, I'm building it soon.. Not to derail the thread, but how does the Delegate/Boneyard edition differ to Keeley (bass/guitar) Comp if you've tried one?

 

[Chuck D. Bones]

They are very different beasts, both in circuit design & performance.

 

[cooder]

@Chuck D. Bones ^ I think you were a bit polite there... always the gentleman.

 

[Chuck D. Bones]

The Keeley is a feed-forward, VCA-based design. It will be faster and more precise than the Delegate or any other optical compressor.

 

[lrgaraujo]

I know I'm probably wrong here, but I like to avoid charge pumps as much as i can, and run everything on 9V. Any chance this will work ok on 9V?

 

[Chuck D. Bones]

I hear ya, charge pumps are not my favorite either. You can delete the charge pump and you will give up 4 or 5dB of headroom. Maybe that's ok. If you want to keep the same performance in the peak detector, you'll need to cut a few resistor values in half.

 

[xefned]

We all need power supplies with a few dedicated 18v outputs.

I haven't had any charge pump issues yet, but I'd hate to go onstage and suddenly discover a combination of pedals/PSU/house lights that set the charge pumps a'whirring with noise.

 

[lrgaraujo]

I hear ya, charge pumps are not my favorite either. You can delete the charge pump and you will give up 4 or 5dB of headroom. Maybe that's ok. If you want to keep the same performance in the peak detector, you'll need to cut a few resistor values in half.

Just finished my latest breadboard project, so timing feels right. I'll probably pester you with questions in the near future

 

[lrgaraujo]

@Chuck D. Bones, I've started breadboarding something similar to this one and I plan to use some NSL-32 I've got. Since I'm changing the photocell/led pair, I'm expecting that I might have to change some caps and resistors to make it work. Is C10 in the store schematic the main suspect?

 

[Chuck D. Bones]

I recommend that you use all of the stock component values. The NSL-32 should work fine as long as it's fast enough. If it's not fast enough, you'll get a slow release and breathing. The is nothing you can do to fix that except replace the opto with a faster one. Have you measured the light and dark resistance? To measure the light resistance, power the LED side with 9V and a 4.7K resistor in series. Measure the resistance of the LDR side. Should be <10K. Now remove the power from the LED side and watch the resistance reading go up. The faster the better, Should get above 1M in a fraction of a second. We want >2M dark resistance.

 

[lrgaraujo]

Just did what you said and it passes the test, thanks a lot!!
I'll finish laying it on the breadboard and listen to how it sounds

 

[Chuck D. Bones]

The controls interact, particularly ATTACK, RELEASE & RATIO. Not all combinations of settings will produce musically useful results. I recommend starting with ATTACK & RELEASE at 10:00, RATIO at 3:00 and SUSTAIN at noon. Increasing ATTACK & RELEASE makes them slower. RATIO controls how much the circuit can reduce the gain, while SUSTAIN sets how much the circuit can increase the gain. RATIO kind of works like a Blend control. Backing off on RATIO makes the sound less compressed. Adjust TONE & LEVEL to taste. The opamp does not have to be OPA2134. TL072 will work almost as well. Q1, Q2 & Q3 can be any high gain Si NPN. BC549C is a good option.

Have fun!

 

[lrgaraujo]

I'm more interested in the envelope detector part of the circuit, actually. I'm trying a discrete jfet approach for the gain stage. Let's see how it works out

 

[Chuck D. Bones]

Cool. I'd like to see that. Headroom is important in compressors, so keep that in mind when you design your gain stage. If I had this to do over again, I'd configure the charge pump to make -9V instead of +18V. Or maybe -9V, +18V and a +4.5V reference, à la Klon.

 

[Erik S]

I recommend that you use all of the stock component values. The NSL-32 should work fine as long as it's fast enough. If it's not fast enough, you'll get a slow release and breathing. The is nothing you can do to fix that except replace the opto with a faster one. Have you measured the light and dark resistance? To measure the light resistance, power the LED side with 9V and a 4.7K resistor in series. Measure the resistance of the LDR side. Should be <10K. Now remove the power from the LED side and watch the resistance reading go up. The faster the better, Should get above 1M in a fraction of a second. We want >2M dark resistance.

My Delegate is getting close to the finish line, and I need to select an LDR.

I came here for looking for some tips, and I'm guessing that "The faster the better" comment is probably the key?

I have GL5528s, KE-10720s, and KE-10715s on hand.

I tested some samples of each - all took my meter out of range at 4M when I hooked them up and shut them in a drawer, although they took different amounts of time to get there.

GL5528s passed 4M fastest at around 1.5 sec
KE-10720 in 4 sec
KE-10715 in 13 sec

Under flourescent light at workbench height:

GL5528s were between 1.5k and 3.8k with most close to 2.5k
KE-10720 right at 1k/ 1.1k
KE-10715 right at 1.1k/ 1.2k

GL5528 won the race by a mile. The tolerance for light-on resistance isn't great, although I just need one, and I can cherry pick the lowest one if that's the goal.

I did see at least one build report with positive results from a KE-10720, and those were giving me the lowest resistance under light. Is it possible that would be a better pick? Or stick with the speedy ones?

 

[Chuck D. Bones]

As long as the LDR resistance gets below 10K, you're good. The GL5528 should work well.

 

[Erik S]

As long as the LDR resistance gets below 10K, you're good. The GL5528 should work well

Gl5528 it is! Thank you Mr Bones!

 

[apc42069]

OK, here's what happened. In breadboard testing the Boneyard Edition, I found that increasing C6 & C8 was beneficial so I told Mr. PedalPCB that 220nF was the preferred value. That's why the Build Docs say 220nF. That conversation took place 17 months ago and had slipped from my memory banks. Increasing those caps speeds up the peak detector and allows the user to set a faster attack response, if that's what you want. You can still dial back the attack speed with the ATTACK knob.

In summary, the Build Docs are correct. 22nF will work; 220nF will work better. I should update my Delegate.

Rez erection time (sorry)

Do you mean C7 & C8?

C6 looks tone stack related to my naive eyes.

Thanks

 

[Chuck D. Bones]

Do you mean C7 & C8?

Yes, I have just corrected the typo in my post.