Altec 436 and EMI RS124 - release times?

[lion]

First post. Hoping for help to understand a couple of technical details which has me confused.

Curious about vari-mu compressors - mainly the Altec 436 and EMI RS124 - I've done some research, including going through a number of interesting and enlightening threads here.

Comparing the Altec and the RS124 - the release switching and times on the RS124 has me puzzled.

I understand the formula for calculating the time constant (to 63% - usual level) is: uF x M ohms = time in seconds.

In the 436 circuit the TC related component values are C4/1uF, R9/270K and P2/1M, and using the formula the available release times calculates out to a range of 0.27 - 1.3 secs (1uF x 0.270M = 0.27 Sec and 1uF x 1.270M = 1.27 Sec). All making perfectly sense and happily in agreement with Altec specs.

In the RS124 timing cap C4 is halved to 0.5uF, and the release pot setup is replaced with a rotary switch with resistors for 6 fixed time settings. The timing resistor values are clearly notes as 1M (R9), 0.680M, 1M, 2M, 3.3M, 5.6M (R20-24). All according to EMI schem posted by John Hinson.

AFAICS with the switch in the fastest release setting only R9/1M resistor is involved. In the max release switch position all 6 resistors forms an accumulating series string =12.68M in total.

Putting the values in the formula. Minimum release turns out to 0.5 sec (0.5uF x 1M = 0.5 sec), which seems plausible. But the calculated max release time lands on a wooping 6+ secs (0.5 x 12.78M = 6.34 Sec), which surely can't be right!

Someone please tell me what I'm I missing here?

Erik

 

[Michael Tibes]

Hi Erik,

maybe you could attach the schematics, it might help to answer your question.

Michael

 

[emrr]

6 seconds is entirely correct for classical music. 

 

[DaveP]

Your figures are correct.

At that time EMI had to record a whole range of music, not just pop groups.

Nowadays, compressors are used in a different more creative way (with shorter timings) to obtain effects, back in the day, these techniques were only just being worked out.  In the early days they were designed to replace an engineer riding the gain control.

If you look at the Altec sales literature, you can see that it was a low cost way of regulating sounds systems, rather than a high end recording compressor like a Fairchild.  From the book Recording the Beatles, the RS124 was mainly used for the guitar tracks.

DaveP

 

[lion]

Thanks gentlemen
I thought 6+mS release times was too excessive - and of no real use, not even for classical music. Nice to learn something new.

Can I pick your brain on another 436/RS124 issue as well.

Re attack time constants - here's a couple of statements I've found:

"The attack and release times in model 60070B are approximately three times faster than those of the others, so provides a more audible compression effect. You may find that models 60050A and 61010B are more suitable for bus or mix compression and model 60070B more suited to channel compression".

"The three RS 124 plugins from Abbey Road (now discontinued) that were supposedly modelled on the actual units at Abbey Road are all slightly different and seem to have attack times of somewhere around 15-35 mS, 25-60 mS and 50-90 mS".

"The Chandler Limited RS124 includes three favorite historic RS124 compressors still in use at Abbey Road Studios today, in a single unit for the first time ever. Included amongst the nine attack options are three positions marked in red. These red colored settings reflect the serial numbers (60070B, 60050A, and 61010B) of historic RS124s".

During my research I have noted that there are a number of people who have successfully increase the attack time on the Altec 436 - simply by reducing C4 from the original 1uF to 0.5uF - giving a resultant attack time of 25mS.

However, I have also noticed that attempts to go lower - say down to 15mS attack - in practise has proved difficult, as far as I understand it, due to some instability problems causing an audiable (thumping) noise on recovery, sound chopping out etc.

So, what I'm wondering is, does anyone know how the EMI techs managed to have a useable 15mS attack setting which presumably was the case with the 60070B models?

Erik

 

[DaveP]

Having made a couple of these, I can explain the problem like this:-

Imagine a power supply where you reduce the value of the reservoir capacitors.  At some point you will start to get ripple on the HT line.  As you reduce the value of the control voltage capacitor, the same thing happens and the gain control tubes start to amplify the ripple spikes.  The whole compressor is a feedback loop, so any a.c. on the  CV cap  can start oscillation.

The intentional feedback that was incorporated into the RS124 between the anodes/plates does help to lower the output resistance of the compressor and speed up the response time, but obviously a limit will exist somewhere (internal wiring may be better on some versions).  The 1uF cap gives a safety margin to allow for production tolerances.

A useful modification to the compressor would be to incorporate some regulation on the HT/B+ supply, along the lines of the UA 175 and others.  This will tighten up the compression slope and smooth out some spikes, but then it won't be an RS124, so you just have to accept its limitations if you want the genuine article.

DaveP

 

[lion]

Dave, thanks for sharing your knowledge - I'm learning a lot today.

The intentional feedback that was incorporated into the RS124 between the anodes/plates does help to lower the output resistance of the compressor and speed up the response time, but obviously a limit will exist somewhere (internal wiring may be better on some versions)...

I take it you are refering to the 100K resistors (R26-27 on the EMI schem) between the 6BC8 and 6CG7 plates, right?

Would you say they make any real difference with respect to how much you can successfully cut the attack times by reducing the timing cap - and/or the R12/33K serie resistor?

Erik

 

[DaveP]

Yes, we are referring to the 100k resistors.

If you read page 300/301 of Morgan Jones book, in the chapter on rectification, it has some useful equations there.

Where the charge on a capacitor is Q ,  Capacitance in Farads is C and Voltage is V.
                                                                             
Q=CV

Q also = Current in Amps I x t in seconds,  Q=It

Therefore CV= It and so CV/I=t.  or CR=t

Putting some figures in, we get 0.000001F x  33000 = 33mS.

The formula for calculating the ripple on a capacitor is :-

I/fxC=Vripple, where I=current in Amps, f=frequency in Hz and C= Capacitance in Farads.

Using the Altec as an example, and hypothetically saying current is 100 uA

0.0001/100Hz x 0.000001= 1V of ripple voltage.

If you reduce the frequency or the capacitor then the ripple rises  along with the chance of oscillation.

I have no idea what the actual current is, but the formula shows the principle.

Regarding the feedback.  The output resistance is reduced by the same factor as the gain reduction.  So if the feedback is 6dB then the output resistance will be halved.

There are actually three components to the resistance acting on the capacitor.  The attack resistor, the output resistance of the amp and the resistance of the diode at whatever current is passing at the time.

From the above you can see why feed capacitors are carefully chosen (they don't pass too much bass, see formula)  The gain reduction will act on the bass, but the DC control voltage might be generated by the higher frequencies.

I hope that makes it clearer, you should be able to work out what you need to know from this, thanks to Morgan Jones.

DaveP

 

[lion]

I can't say I fully understand everything you said, but I'll work on it. More study and learning ahead.

Thanks a lot Dave. Much appreciated.

Erik