ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

Re: Anybody built this compressor?

5/5/2000 12:36 AM

Reid Kneeland

Re: Anybody built this compressor?

quote:
"That's not really a cathode follower in this case. "

Well since it's a diode, that goes without saying. I just meant that when you first look at the circuit, it seem bizarrre.

Reid

5/5/2000 12:11 AM
Chris Barrow	The second half of the second tube has the rectified DC coming in on it's plate. At a certain voltage level, the tube will begin to conduct to groung through the cathode. This makes a voltage divider with the 470K series resistor and lowers the signal level. Instant compression! Looks like the 1 Meg pot is an attack/decay control, varying the resistor varies the time constant to charge/discharge the cap in parallel with it. The other pot is compression amount. I hope I got it right. Looks like a cool project. Chris Barrow

5/5/2000 12:11 AM

Chris Barrow

The second half of the second tube has the rectified DC coming in on it's plate. At a certain voltage level, the tube will begin to conduct to groung through the cathode. This makes a voltage divider with the 470K series resistor and lowers the signal level. Instant compression! Looks like the 1 Meg pot is an attack/decay control, varying the resistor varies the time constant to charge/discharge the cap in parallel with it. The other pot is compression amount.
I hope I got it right. Looks like a cool project.

Chris Barrow

5/5/2000 4:55 AM
Bryan James	very good explanation. was wondering how that would work. i do know that is another ldr less compressor that's in the pittman book, it's on the gibson bass amp. the question then becomes, which was ask earlier what voltage?? and also what mods to make it a stand alone effect?? -Bryan (boy like i don't have enough projects)

5/5/2000 2:40 AM

Ken Gilbert

quote:
"I don't completely understand it either, but the second section of the first tube is being used as a diode (the grid is floating)."

i always thought that was a no-no--leaving elements disconnected. i never figured out WHY, but that's what i picked up. i suppose you could always tie the grid to the plate if it's merely a diode. it would be a BETTER diode, in fact.

kg

5/5/2000 3:44 PM
Scott Swartz	It looks like to me that Chris Barrow’s description is correct. If we assume this is the case, there are a couple of problems with this circuit, both relating to control voltage feedthrough: 1. The time constant of the side chain will have to be slow enough so that the control voltage going is DC, ie no “ripple”. If there is ripple, it will modulate the output signal at the ripple frequency, which I guess will be half the fundamental frequency of the signal. I’m not sure what that would sound like, but it its probably undesirable. 2. Even if the control voltage is DC, the DC charge on the output capacitor will be forever changing which may couple through the output capacitor and upset circuits downstream. This is similar to when you probe with an AC coupled oscilloscope and the trace deflects off the screen until the DC potential on both sides of the input cap stabilize. There was a voltage divider type design from the 40s that is in Audio Anthology No. 2 that worked around these problems by using differential amplifliers and a tranformer to isolate the DC shifts, but it was very complex Now for the shameless plug section of this post. I have written a tube compressor article that will be published in the next Glass Audio for a LDR based tube compressor. I submitted my final revisions to the article a couple months ago and they have told me it will be the next issue, but sometimes they shift things around. Anyway, since they are paying me for the article, I can’t post the schematic (buy the mag!), but I can give you a hint. Take a normal gain stage and direct couple a cathode follower (Fender Bassman EQ driver). Purchase a Vactec VTL5C6 from Newark Electronics and hook up the resistor part of it as a feedback resistor, connecting between the output capacitor and the grid of the gain stage. Now you just need a side chain to drive the LED in the VTL5C6. The Stompbox Cookbook by Nick Boscorelli has a lot of good side chain ideas. This design has none of the feedthrough problems discussed above.

5/5/2000 3:44 PM

Scott Swartz

It looks like to me that Chris Barrow’s description is correct. If we assume this is the case, there are a couple of problems with this circuit, both relating to control voltage feedthrough:

1. The time constant of the side chain will have to be slow enough so that the control voltage going is DC, ie no “ripple”. If there is ripple, it will modulate the output signal at the ripple frequency, which I guess will be half the fundamental frequency of the signal. I’m not sure what that would sound like, but it its probably undesirable.

2. Even if the control voltage is DC, the DC charge on the output capacitor will be forever changing which may couple through the output capacitor and upset circuits downstream. This is similar to when you probe with an AC coupled oscilloscope and the trace deflects off the screen until the DC potential on both sides of the input cap stabilize.

There was a voltage divider type design from the 40s that is in Audio Anthology No. 2 that worked around these problems by using differential amplifliers and a tranformer to isolate the DC shifts, but it was very complex

Now for the shameless plug section of this post. I have written a tube compressor article that will be published in the next Glass Audio for a LDR based tube compressor. I submitted my final revisions to the article a couple months ago and they have told me it will be the next issue, but sometimes they shift things around. Anyway, since they are paying me for the article, I can’t post the schematic (buy the mag!), but I can give you a hint. Take a normal gain stage and direct couple a cathode follower (Fender Bassman EQ driver). Purchase a Vactec VTL5C6 from Newark Electronics and hook up the resistor part of it as a feedback resistor, connecting between the output capacitor and the grid of the gain stage. Now you just need a side chain to drive the LED in the VTL5C6. The Stompbox Cookbook by Nick Boscorelli has a lot of good side chain ideas. This design has none of the feedthrough problems discussed above.

5/5/2000 6:20 PM
Reid Kneeland	Thanks for your observations, Scott. I'll be watching for that article. I'm now determined to build this thing and see what it does (and more importantly, hear what it does). As for the ripple modulating the output signal, who knows? It might sound cool! That's the fun thing about instrument amps - what would be a miserable failure in general audio can often be put to great use as an effect! In this case, it's simple enough to throw one together and check it out. Reid

5/5/2000 6:20 PM

Reid Kneeland

Thanks for your observations, Scott. I'll be watching for that article. I'm now determined to build this thing and see what it does (and more importantly, hear what it does).

As for the ripple modulating the output signal, who knows? It might sound cool! That's the fun thing about instrument amps - what would be a miserable failure in general audio can often be put to great use as an effect! In this case, it's simple enough to throw one together and check it out.

Reid

5/5/2000 6:39 PM
GFR	If it's a half wave rectifier it's exactly the same frequency as the input signal (FWR would give twice the frequency). If it modulates the signal, it will give mostly second order harmonics, it can sound very good. Most compressors have this effect to some degree, with LDR compressors it's just not too much noticeable because LDR's are slow (so you don't have control leakage but you can't have fast attacks or fast decays). The Dynacomp has LOTs of this kind of distortion - many people like it for guitar (myself included). Also, some people like if some of the control voltage leaks, specially at transients, as this can give the impression that your signal still has "attack" even if it's very compressed.

5/5/2000 6:39 PM

GFR

If it's a half wave rectifier it's exactly the same frequency as the input signal (FWR would give twice the frequency). If it modulates the signal, it will give mostly second order harmonics, it can sound very good. Most compressors have this effect to some degree, with LDR compressors it's just not too much noticeable because LDR's are slow (so you don't have control leakage but you can't have fast attacks or fast decays). The Dynacomp has LOTs of this kind of distortion - many people like it for guitar (myself included).

Also, some people like if some of the control voltage leaks, specially at transients, as this can give the impression that your signal still has "attack" even if it's very compressed.

Page 2 of 3