The Idea

I read a post somewhere on here where PRR stated that the Fairchild 670 is probably not specified for todays applications in the studio. I hope I didn't misinterpret what he said and if I did, well then you have my apology in advance. Here is what I'm thinking...JJ Electronic is reissuing the 6386 sometime in the near future.http://www.jj-electronic.com/pdf/6386LGP.pdf. What if we tried to change up the old mythical design a little bit by using a more powerful sidechain amp for a much wider range of TC (Time Constant) settings. Different transformers for more transparency. Just a thought.

only if we do the sidechain in sandstate.  It also has to be SMD..  ;D

I have yet to see anybody just take a big solid-state power amp and drive a bridge as a sidechain amp. That would easily be a proof of concept for a more finished design. I bet it would sound fine. It's what Narma would have done today. Low-z, lotsa volts.

Not that's anything alike but the peavey tube comp did use a el34

Larrchild said:

I have yet to see anybody just take a big solid-state power amp and drive a bridge as a sidechain amp. That would easily be a proof of concept for a more finished design. I bet it would sound fine. It's what Narma would have done today. Low-z, lotsa volts.

Click to expand...

That's exactly what I'm going to do. I just bought a PCB for a stereo LM3886 amp for that. But I'll be testing the thing with PCC189's.

The lazy (and frugal) builder would find a cheap 50w/ch power amp and transfer it's guts to a bigger case with the audio stages.

> I read a post somewhere on here

No idea what you are remembering.

> using a more powerful sidechain amp for a much wider range of TC (Time Constant) settings.

A bigger amp buys faster attack. Who here is having trouble with 50 microseconds?

Effects limiting often uses attacks 1,000 times slower.

http://www.groupdiy.com/index.php?topic=26847.0

Probably misunderstood from here. And it is a very fast compressor when driven with a 50W amp.

Haven't asked a dumb question for a while so here goes...

Even for the Fairchild what we're talking about is charging/discharging a capacitor.
The charge has to be fast so we drive using a high current source.

But...

We could just lower the cap value to reduce the current requirement, couldn't we?

So what is the constraint that requires the Fairchild to use a 2uF cap with 220k||68k discharge?
Why not a 200nF cap with 2M2||680k discharge?

What am I missing here?

cheers
Nick

Your way.



Fairchild way.

One of the things observed is, the Fairchild way presents a less distorted waveform when compared to the source. This is just an approximation.

Ok call me thick...

Granted I've only talked about discharge time constant however my point was that the requirement for high current drive to charge the cap would be reduced by lowering the cap value.

What I believe the "your way" diagram shows is what would happen if you reduced the cap value to 200nF and increased the discharge resistors proportionally but left the charge cicuit the same - sure, this would charge too fast - but it's not what I'm talking about. Correct?

Again - what am I missing that makes using a 2uF cap driven by a big amp a requirement?

Thanks.

Edit: diagram added

nickt said:

We could just lower the cap value to reduce the current requirement, couldn't we?

Click to expand...

http://www.groupdiy.com/index.php?topic=11910.0

JDB.

What about using a small cap/big resistors and buffering the control voltage going to the grids? A TL072 might not like the voltages present, but this ought to be possible with a few cents worth of transistors. Or is there some kind of mojo in the maybe nonlinear control voltage/grid current relation that would be lost?

> We could just lower the cap value to reduce the current requirement, couldn't we?

Yes, but maximum release time would be shorter.

We could increase the size of the release resistor, but grid-current is only "small", not "zero", and a very-large release resistor times +/- stray grid current = uncertain grid voltage. As an extreme, a transformer-loaded stage as in Fairchild "could" go run-away and melt.

And Fairchild has multiple grids in parallel, more grid current.

Instead of grid current, tubes are often rated for a maximum grid resistor. OTOO 1Meg. Fairchild has four grids, 1Meg/4= 220K standard value.

So in the simple case, the maximum safe grid resistor, divided by the ratio of max release-time to min attack-time, sets your attack "resistor", including amplifier and rectifier resistances.

And the maximum control voltage, divided by attack resistor, is your peak charge current.

My tubes are rated 1Meg max grid resistor, I have two, my max grid resistor is 500K; this is also typically my release resistor.

Say I want 0.5mS attack and 5 Sec release, 10,000:1 ratio.

The attack "resistor" is 500K/10,000 or 50 ohms.

Say I need 25V control voltage swing. I need 25V/50R= 0.5 Amperes, at 25V, is 12.5 Watts peak. I need at-least 6.26 Watts Sine RMS amplifier.

The Fairchild has a faster attack, even more grids to drive, and low-Mu tubes working at good high voltage, needs like 40V of control voltage. Hence the 15W Sine RMS sidechain hammer.

Yes, buffering a hi-R time-constant network to the tubes would allow less heroic hammer. 13-cent TL072 would allow 1/10th of the sidechain amplifier power, IF (as you say) it can swing the control voltage (the Fairchild's CV is beyond the reach of cheap chips). Or you could derive a 0-4V voltage on the cap, then amplify it up to 0-40V. And you will have more/less trouble with leakage around hi-Z networks; the 99-cent cheap phenolic switches may show faster release on very humid days.

Does it make the music better? I don't see much need for attacks faster than a millisecond except as the final stage before an "Oh, CRAP!" recorder where the slightest overload will ruin a take. Tape don't mind overload. Digi-systems should clip without distress. The primary need for a fast limiter may be phonodisc cutting, because if grooves touch the master is useless.

Thanks PRR - your response is educational beyond the call of duty (as usual) - great stuff.

Also jdbakker - wish I'd seen that thread before.

PRR breakin' it down like it's 1999......

nickt said:

wish I'd seen that thread before.

Click to expand...

Me too !!!

analag said:

...What if we tried to change up the old mythical design a little bit by using a more powerful sidechain amp for a much wider range of TC (Time Constant) settings. Different transformers for more transparency. Just a thought.

Click to expand...

Bro, you design it, they will come.  You offer it, we will build it. 

Yes yes yes.

Hi Guys,


  sorry to troll over ancient ground, but . . .


    I think i must be missing something here. Can you use any old 20W amp for this? do you just drive the bridge rectifier, or do you need to drive a transformer first? (If it that easy, why does anyone BOTHER with a Tube amplifier?)


  just out of interest, what is the load that the side-chain amp is driving in the fairchild?


    I am very inspired by SuperMagoo's PoorChild, and want to lash-up something similar, and would rather a cheaper amp! Go SuperMagoo!!



      Kindest regards,


        ANdyP

> Can you use any old 20W amp for this?

Almost.

> do you just drive the bridge rectifier, or do you need to drive a transformer first?

The DC side of the rectifier must be grounded. Look how rectifiers work. If the DC side must be grounded, either the AC side must float or you need two push-pull AC sources.

Alternatively, you "could" rectify at lower voltage and current, then use a SS gain-buffer to jack-up the CV to drive all those grids.

> (If it that easy, why does anyone BOTHER with a Tube amplifier?)

At the time it was designed, there was no choice.

Nobody since then has had the gumption to RE-design the mojo.