Variable Mu

[analag]

To be built, when I get around to it.

analag

 

[gyraf]

Looks good!!

:thumb:

Jakob E.

 

[Kit]

Cool , whats the reason for the mosfet?

 

[analag]

Because I want a stabilized B+ to the 6BC8. If I were to use a resistor for the voltage dropping purpose, with the wildly varying current draw of the 6BC8 the B+ would be all over the place. At idle, the current draw to that stage is around 7ma and at full compression, less than 1ma. This MOSFET will ensure longer tube life and a tighter compression characteristic.

 

[Viitalahde]

[quote author="gyraf"]Looks good!! [/quote]

:green:

Yes, it actually does. I don't see myself building another vari-mu anywhere near, but someone might get a killer unit from this one. :thumb:

 

[skipwave]

[quote author="analag"]Because I want a stabilized B+[/quote]

That is something I've been wondering about with these vari-gain limiters. None of the old GE/RCA/Collins/Altecs had regulated B+, so did that contribute to their "character" or was it just a limitation of the technology of the day? Probably up for debate, so... opinions?

This got me thinking about it:
Larrchild wrote,"Funny you mentioned the plate resistance effect. One day I accidently had a few k of resistor from the reg'd dc voltage to the 6386's. What should have been a stiff supply was slightly choking and starving the tube on peaks, and giving it a totally different limiting sound. Later,I realized this wasn't coming from grid control and made a note."

Link to that complete thread.

 

[analag]

How are the PCC189 vs 6BC8 anyone...Jakob?

analag

 

[Svart]

Analag never ceases to amaze me with the sheer amount of design work he does. :green:

:thumb: :thumb:

 

[Viitalahde]

One thing I wonder about DIY designed vari-mu's: Why not design the front end for single triodes like 6BA6? (not sure of the model number). Would be easier to find a match.. I remember a ratio of 2 out of 14 until I found an acceptable match from PCC189's.

 

[dale116dot7]

Just a couple things that I noticed. The detector diodes may leak too much - the 6AL5 tube is really nice there - it doesn't leak appreciably. Check the leakage specs of the diodes and also the reverse recovery time may do something funny at high frequencies. You might want to consider a zener diode in the MOSFET gate circuit to actually regulate the output voltage there like the Fairchild 670 did, also.... I dunno. The metering goes a bit funny if the 275 volt supply drifts with load. Looks nice though, a lot like a 436. The release time constant looks interesting.

 

[ioaudio]

thank you very much for sharing analag!
is this the same circuit which was discussed earlier in another thread?

regards, max

 

[Bauman]

I have an eye on it :thumb:

 

[analag]

It is a 436 derivative without the 6AL5, A zener to regulate the voltage at the MOSFet gate circuit, I should have put it in this schematic, but because I only use regulated B+ I wouldn't need it. So...my bad.
[quote author="ioaudio"]is this the same circuit which was discussed earlier in another thread?[/quote]
No...but that's the one I'm building now.

analag

 

[NewYorkDave]

There's a glaring omission in the sidechain circuit. I'm surprised none of you spotted it. Hint: without it, the comp will not be useful for much except perhaps as a special-effect drum room-mic smasher.

Another thing I noticed right away is that the plate-grid feedback on the 6SN7s is not very useful, since the feedback factor is almost the same as the open-loop gain of the tube itself. If your simulator is telling you otherwise, you should be very suspicious of the model.

 

[chrissugar]

Wouldn't be a better idea to use an attenuator before the input transformer, and a threshold control pot, that apply variable voltage before the two rectifier diodes. Anyway, dual pots don't track too good so for better balance I think it is good to avoid them.

chrissugar

 

[NewYorkDave]

the input audio level, and also the control signal.

Actually, no. The tube grids do not draw appreciable current from the sidechain, so the series resistance introduced by the pots has no significant effect on the grid bias. The Altec used a similar arrangement. The tracking of dual pots is definitely an issue, though; a dual stepped attenuator would be better for maintaining balance.

But you did spot what I alluded to earlier: no threshold adjustment, no reverse bias on the rectifier diodes. Also, 1N4007 wouldn't be my first choice. There are other diodes intended as signal rectifers that would probably work better.

 

[chrissugar]

[quote author="NewYorkDave"]

the input audio level, and also the control signal.

Actually, no. The tube grids do not draw appreciable current from the sidechain, so the series resistance introduced by the pots has no significant effect on the grid bias.[/quote]

Yes Dave, I realised that, and quickly edited my post but you was faster than me. :grin:

chrissugar

 

[analag]

I used dual pots for simplicity in the posted schemo, rotory switch is the proper way to go in this application.
Dave, all pointers and examples are welcomed

How is that.

analag

 

[Larrchild]

Ok. What about this. A 6AL5 to replace the diodes. Low Loading.

Then, to get gain and transform impedance, one of those FET's you dig.

A low Z source to the time constants. Yay. Fast risetime.

 

[skipwave]

Excuse the newb question, but I'm just learning about vari-mu designs. Is regulated B+ a necessity?

Thanks for the great discussion. I will be tempted to try analag's design with 6BA6s instead, since I have about 30 of those. :grin: