electronaut
Well-known member
OK folks... I hope I'm not pissing off the moderators by playing hack-moderator a little, but in an effort to keep the discussion of analag's tube opto-compressor on topic, I started a new thread.
Here is the schematic of analag's compressor that he, like so many other generous contributors, offered up for discussion:
Some notes by analag:
- D1 is part of a Vactrol Opto-isolator. R3 is actually a 100k pot with the LDR strapped across it, much like the La2a. R1 and R2 is also a pot.
- I use a tube connected exactly as you see in my schematic, you may use a resistor in series with the LED say 120-470 ohms, this will tune the slope of the unit. In this particular compressor the same tube tube has been functioning since 1998 with no problems, in my studio. This is a rather simple but excellent device in terms of sound quality and compression give it a try.
- Cathode follower and grid feedback aka plate follower equals lower output imedance and very low noise.
[quote author="gyraf"]Analag - what kind of opto is used?
Some quite slow opto type would probably be needed here in order not to distort at low frequencies..?
I guess that the opto's LDR part goes in parallel with the output level pot (marked "gain") - forming a voltage divider with the 100K resistor that goes to the output pot?
Is there any particular reason why you don't use a simple transistor or opamp stage (maybe even running off heater voltage) for the LED-driving - to save a nice ECC82 tube - and to reduce heat and current?
Jakob E.[/quote]
analag replied: VTL5C4/2 optos are used. I use the fastest ones I could find, it does not distort at low frequency. If you notice, the signal that drives the gain reduction amp comes from the output of the first voltage stage or U1 and at the same time the LDR is connected to U1's input. This seems to eliminated the frequency distortion you suggest might be taking place.
[quote author="chrissugar"]The VTL5C4 is not the fastest vactrol, it has a very slow release, 1.5 seconds. That long release probably simulates in a way the memory effect of the T4B optocoupler from the LA2A and that is probably the reason why you don't have distortion at low freq because it responds slowly.
The VTL5C1 is the fastest vactrol available.[/quote]
[quote author="PRR"]> There's also no current limiting resistance in series with the LED...
That 12AU7 is a very effective resistor. Without looking at the curve, I'd guess you can't get 30mA out of it, and that only for peaks: well within sight of the ~20mA rating on most LEDs.
> Current can never flow out of it.
There are such things as "AC" optos, with two back-connected LEDs. That will work. No matter that it isn't shown that way on the schematic: these drawings also don't show which resistor is the opto-R, they are hastily drawn.
Also the single diode WILL reverse-breakdown. At around 7V. So all that changes is that the threshold is higher than a hasty glance suggests. Reverse-breakdown is not harmful (not at the current the 12AU7 will pass).
I think it has a lot more resistors than I would use (even allowing for the several resistors that seem to be merely simulation place-holders). I'm baffled by the use of both cathode feedback and plate-grid feedback. Sure it works, but there are other ways to do it. Is it more euphonic? I don't know. U5 seems to be working awful hard considering its load. The B+ bypass for U1 seems pointless. U6 will suck great spikes of current and should probably be further up the supply chain.[/quote]
Here is the schematic of analag's compressor that he, like so many other generous contributors, offered up for discussion:
Some notes by analag:
- D1 is part of a Vactrol Opto-isolator. R3 is actually a 100k pot with the LDR strapped across it, much like the La2a. R1 and R2 is also a pot.
- I use a tube connected exactly as you see in my schematic, you may use a resistor in series with the LED say 120-470 ohms, this will tune the slope of the unit. In this particular compressor the same tube tube has been functioning since 1998 with no problems, in my studio. This is a rather simple but excellent device in terms of sound quality and compression give it a try.
- Cathode follower and grid feedback aka plate follower equals lower output imedance and very low noise.
[quote author="gyraf"]Analag - what kind of opto is used?
Some quite slow opto type would probably be needed here in order not to distort at low frequencies..?
I guess that the opto's LDR part goes in parallel with the output level pot (marked "gain") - forming a voltage divider with the 100K resistor that goes to the output pot?
Is there any particular reason why you don't use a simple transistor or opamp stage (maybe even running off heater voltage) for the LED-driving - to save a nice ECC82 tube - and to reduce heat and current?
Jakob E.[/quote]
analag replied: VTL5C4/2 optos are used. I use the fastest ones I could find, it does not distort at low frequency. If you notice, the signal that drives the gain reduction amp comes from the output of the first voltage stage or U1 and at the same time the LDR is connected to U1's input. This seems to eliminated the frequency distortion you suggest might be taking place.
[quote author="chrissugar"]The VTL5C4 is not the fastest vactrol, it has a very slow release, 1.5 seconds. That long release probably simulates in a way the memory effect of the T4B optocoupler from the LA2A and that is probably the reason why you don't have distortion at low freq because it responds slowly.
The VTL5C1 is the fastest vactrol available.[/quote]
[quote author="PRR"]> There's also no current limiting resistance in series with the LED...
That 12AU7 is a very effective resistor. Without looking at the curve, I'd guess you can't get 30mA out of it, and that only for peaks: well within sight of the ~20mA rating on most LEDs.
> Current can never flow out of it.
There are such things as "AC" optos, with two back-connected LEDs. That will work. No matter that it isn't shown that way on the schematic: these drawings also don't show which resistor is the opto-R, they are hastily drawn.
Also the single diode WILL reverse-breakdown. At around 7V. So all that changes is that the threshold is higher than a hasty glance suggests. Reverse-breakdown is not harmful (not at the current the 12AU7 will pass).
I think it has a lot more resistors than I would use (even allowing for the several resistors that seem to be merely simulation place-holders). I'm baffled by the use of both cathode feedback and plate-grid feedback. Sure it works, but there are other ways to do it. Is it more euphonic? I don't know. U5 seems to be working awful hard considering its load. The B+ bypass for U1 seems pointless. U6 will suck great spikes of current and should probably be further up the supply chain.[/quote]
