Vari Mu Compressor attack & release times ???

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

gary o

Well-known member
Joined
Dec 28, 2004
Messages
1,531
Location
uk
Gates 26U collins STA level 356E am864 BA6A La1D Chiswick Disa altec ands so I notice these have different attack & release times & daul modes Ive read in places some have faster attack times & I see the resistors & caps in the side chain circuits & I know they effect the attack & release

does the output power that drives the rectifier effect attack speed ?

And does the rect it self effect it ...say 6AL5 verses 6U6

And the gain reduction tube...do they have different attack times ..say the famous 6386 would that have a different attack than say the older 6SK7

Im just trying to make myself a vari mu with as much flexability ..choice as I can

I suppose Im looking to make the fastest attact time so I can then slow it down via a pot & the shortest to longest release time via pots or whatever
maybe some swichable daul constanants robbed from good sounding old proven limiters....I mainly record vocals by the way

Im also thinking of having daul input stages .....a6386 stage & a 6SK7 stage & switch between the two

Thanks for reading
 
More questions Im afraid..

Does the less resistance between rectifier and inpu tran centre tap mean faster attack time ??

Does anyone know the attack release times of any vari mu limiters ?

I think the BA6A is .6ms att rel 333ms ?? & Collins 26U 0.5 to 3.0 m secs Im not sure about my millis & micros tho gets confusing

The Coliins 26U would have no resistance in path between rectifier and CT of tran if its attack pot was turned down

The BA6A has no pot and would have same no resistances in the rectifier to CT path in its Single SW position......So 0.5 & .6ms is that about the same ?


Thanks
 
> the gain reduction tube...do they have different attack times

Vacuum tubes are "infinitely fast" for all audio purposes.

> does the output power that drives the rectifier effect attack speed? And does the rect it self effect it

Basic R-C theory.

Attack time depends on the sum of all the resistances through rectifier to time-constant capacitor. That includes your adjustment pot, the rectifier resistance, and the impedance of the audio source driving the rectifier.

This is why the Fairchild had a rectifier amplifier bigger than the output driving amp.

> with as much flexability ..choice as I can ....

Most Good Audio Design is about finding out what you REALLY need, and building just that much into the gear.

> I suppose Im looking to make the fastest attact time

The Fairchild's 15-WATT rectifier driving amp, and the caps that Narma used, are "fast enough" for disk-cutting. In these days of 96KHz sample rates etc, you could argue that you need something bigger.

> I mainly record vocals

Do your recording systems CLANG when they overload? Or just clip? The old RCA film-sound ribbon modulators would do terrible things if they hit the stops.... but most recording systems just clip/squash. And for vocals, that's not a big problem. 5mS is usually plenty fast for vocal (indeed, most musical sounds) and 100mS allows an "edge" of clipping at the start of the vowel (speech overload/limiting is all about the vowels) to accentuate what you are about to limit-down.

And a too-fast attack will duck-down on unaudible bits of spit, giving low output for no apparent reason.

Go get Banjo Center's $89 limiter, one with numbers on the knobs, and see what range of times you really need, before you go bonkers building 50-Watt rectifier drivers.

BTW: Narma would have used a chip to drive the Fairchild's limiter; he was just working 25 years too soon. Using 6L6 and 6AL5 for sidechain is just eroticizing the vacuum.
 
Man..... PRR...... you really know your stuff. All of that is off the top of your head, isn't it? Hats off to you my friend. I am but a grasshoppa but I endeavor to be a Master like you one day. Much respect to you. DW.
 
Yep he really is the Daddy!.... Yeah thanks PRR really usful info for me .....when I said I want flexibility I wanted that so i could find the magic numbers that sound best for vocal,s i didnt know good attack release times was hopeing to fiddle & I hear what I like...As I understand these old vari mu amps werent designed for vocals so i wondered if they would sound better if I fiddled...answer as PRR knew already is Nope!..Since the first post I bread boarded about 8 different RC networks & tried them all on my homemade BA6A connected between the rectifier & the input tran CT, I recorded same vocal vs chorus thru each but nothin beat the BA6s original daul poistion haha I mmade up the RC network from Gates STA level AM864 collins 356 a Disa.....Im still gonna copy the input part of a limiter that uses a 6386 & graft it onto the BA6...was thinking of copying the collins 26U input 1st
 
How do you put an attack pot onto the fairchild side chain. I assume you put it in series with the rectifier. Would it have to be a high wattage because this is a 10-15 watt rectifier driver amp???? What value do you guys recommend??
 
> How do you put an attack pot onto the fairchild side chain.

You don't. There were less than a hundred made, they work fine the way they were made, it is irresponsible to muck-up a Fairchild.
 
Hi Tardishead I would think it would e the same as other vari mu boxes pot between control voltage and input tran centre tap....have a look at other vari mu circuits here that have attack pots.
 
I am not clear. when the "driving Rectifier amp is bigger (power) would be the fastest time of attack?
For that reason, the Fairchild compressor would be faster than any other?
 
Hiya that is how i understand it.....PRR has talk about this...very intersting...& slow or slower isnt alwys a bad thing ....vocals & bass can be good.

Hers a link that may help explain, Im not too good with the maths myself but very interesting

http://www.groupdiy.com/index.php?topic=27226.0
 
PRR said:
BTW: Narma would have used a chip to drive the Fairchild's limiter; he was just working 25 years too soon. Using 6L6 and 6AL5 for sidechain is just eroticizing the vacuum.

I'm not so sure about that. Remember that vacuum tube diodes such as the 6AL5 and 6H6 have a 3/2 power transfer function whereas solid state rectifiers have an exponential one.  That's got to have a significant effect on the "knee" of the compression curve.

---Joe
 
Uncharged time constant caps are low z and need a lot of initial current to charge up quickly. Delivering a lot of current quickly at low-z is not the providence of the 6AL5.
 
Larrchild said:
Uncharged time constant caps are low z and need a lot of initial current to charge up quickly. Delivering a lot of current quickly at low-z is not the providence of the 6AL5.

I agree. I wasn't commenting so much on the design of the Fairchild as I was about the "eroticizing the vacuum tube" part of that quote. What I was trying to say is, part of the sound of a lot of the vari-mu limiters we all love so much is the soft knee provided by the use of vacuum tube diodes. If you replace the tube with a solid state diode, you're going to get a significantly different sidechain response, and maybe not for the better.  While a 3/2 power device might be able to be simulated with op amps, the easiest way to get it is to use a device that has that transfer curve intriniscally.  Just an observation.

---Joe
 
> vacuum tube diodes... have a 3/2 power transfer function whereas solid state rectifiers have an exponential one.

Fairchild is solid state diodes. That must be why it is unpopular.

> That's got to have a significant effect on the "knee" of the compression curve.

"Attack time depends on the sum of all the resistances through rectifier to time-constant capacitor."

If the rectifier resistance is a large part of the sum, you have a very lame "peak detector".

Yes, since a real-good peak detector is costly, many "lame" peak detectors have been used, particularly for utility work where sloppy limiting was preferable to gross over-modulation. Their "flaws" may be "features" in the hands of careful musical engineers.

See last chart in GE 6AL5 data. As Larrchild says, 6AL5 is a weak diode, and puts out less than you put in, but at any given load is fairly linear with any reasonable signal level. As it has to be to work well as an AM detector.
 
It also depends how "remote" your cutoff is.
If your tube is asleep at -20VDC, a 6AL5 is fine.
If you need 90-100v of negative bias to shut er down, the 6AL5 might run off the curve.
 
Back
Top