Another vari-mu: Chiswik Reach Compressor

[mitsos]

please post here.

 

[mulletchuck]

Djn, upload the files to the group Gmail account for all to share 

 

[skal1]

where i can not see it


skal1

 

[djn111]

Hi,

Sorry can't do I am not the original designer, so pm and you get them

DJN

 

[maxwall]

DJN

PM box is full ?

 

[jackies]

:
Oh cmon already people!
Anyone here used a computer before?
All you have to do is to go to Nadege's photobucket album and view pictures on the photobucket site.
How difficult is that?

 

[maxwall]

jackies

same files, Nadege and DJN ?

not sure since I have not seen DJN's

 

[jackies]

I'm pretty sure DJN was referring to Nadege's pcb design.
Let him correct me if I'm wrong.

 

[KickBack]

Hi

I am like a real oldster, and I have worked on lots of tube gear.

I have a pile of info stuck up there about tubes.

I also have Modded a lot of tube gear and discovered thier dar and dirty secrets. That altec Compressor has a serious flaw. The original tube, the 6bs8, is one of a series of tubes, 6bs8, 6bz8, 6bk7, 6bq7, 6aq8, and a couple others that have the ability to be controlled by varing the grid bias.

Vary-Mu is kind of misleading because any tube varyies Mu or gain some with bias change. Most tubes are fairly stable and gain is a function of grid and plate resistors, plate voltage, and screen voltage.

Thes 6 Volt dual triodes were designed for tuners. RCA comes out with a new TV set and designs the 6bc8. Then GE comes out with thier latest and designs the 6Bz8 etc. and a whole series was done up. Then years later the why and other types that relate is lost.

The problem with these compressors is this. I have run into this when working on Vintage gear and just more secrets lost in time. This compressor was sent out with tested, special versions of  the first tube. The balance in the two triodes as to gain is critical because if the gain is off, then one side of the output stage drives too hard, and the compressor "Thumps", and sometimes oscillates.

Even old guitar amps came with special versions of tubes, so they do goofy stuff sometimes when using new versions.

The solution to making this a stable, and very decent compressor, is to use a pentode-triode type, like a 6GH8, 6Bl8, 6AN8, 7199 etc. the pentode as first gain stage, the triode as phase splitter. Then the gain control voltage goes back to the pentode, and the signal is split very balanced, so no funny Bizz Mit the Thumpys or other fun.

You can use the same socket, just change out the dual triode.

Again, 6bc8,  6bz8, 6bk7, 6bq7, 6aq8, 12ay7, 12au7, 6bs8, 6bc8, and 12at7 are all fairly close in specs, and used original for some tuner and some audio circuits. 12au7, 6bk7, and 6bq7 are the warmest sounding.

I had a studio send two of these to me about six years ago freaking out, and I found that, since they were used mostly for PA systems for schools, they never had to deal with complex signals, so the imbalance does not cause much of a problem, and in schools or other PA use, the thumps get lost as no low end in these systems.

 

[KickBack]

I meant 6bc8.

The changes made to the 6bc8 stage are ways to try and correct the imbalances when the gain of the two triodes changes, but the tube itself, if not the right version of this tube, will not have the same plate curves for each triode.

Also, back then, companies made special versions of tubes that solved their design problems.

Fender this with amps they made before they had adjustable bias. I have to tweak older Fender amps to bias them as they were original sold with 6v6, 6l6 etc. tubes that biased up with thier circuit.

This also happens because the cathodes of the 6bc8 are tied together. About 100 to 330 ohms in series with each kathode would make the two triodes more stable. The gain fall off, slight, can be taken up with 56K plate resistors.

Also plate to grid negative feed back, crosscoupled, would improve it's compression action.

Best of ruck

 

[djn111]

Hello,

Yup you are right

DJN

 

[gary o]

Hi there Kickback you real oldster hehe who seems to no what he is talking about unlike my self which is why Im here to learn, got some questions for you or anybody else.....

Kickback I dont no if you have seen the chiswik schematic that is here

http://purusha.smokinggunrecording.com/XVXDY/1schems.pdf

Youl have to forgive me my knowledge is patchy

Iv made a point to point version of this with a few little changes Im just trying to learn and end up with a nice sounding compressor for my vocals and mines is workin great Im happy but would like to learn more and maybe improve.......

Im using Ecc85 daul triode seems to work well.....

Kickback your saying a pentode wired as triode would be better ?? I dont dispute , want to hear more, I no theres plenty vari mu circuits that do that...

Doesnt the 1K pot between the first tubes plates tune out imbalances ??

Ive seen altec schematics with an added pot between the first tubes cathodes, Is tha a further improvement ? or do we only need the plates balanced?

The tubes used for the Altec/chiswik including my ECC85 are not true vari mu tubes I have read ......wouldnt it be better to use a tube that is, maybe octal 12SK7?

Kickback you said about the pentode as phase splitter ? does that mean we dont need centre tap on input transformer ?

Also plate to grid feedback cross coupled ? I think I understand plate to grid feedback.....as in Chiswiks 12AT7 output tube......

What is cross coupled ??

were you talking about feedback on the gain reduction tube ???

Questins questions questions sorry Im just very hungry to learn I believe in this little altec circuit and Abbey road did not that Im a fan.....I DIYed a RCA BA6A had Sowter make transformers specially for me the thing is great and so it should be its much more complicated I love it its sounds great its sounds different to thicker bass heavy to the Altec chiswik circuit but on vocals  its doesnt sound worse just different thats why i think this thing is great, punches above its weight.

 

[maxwall]

Kickback

Thanks for that explanation , very informative.

So 6BC8 or the listed dual triodes should be sub'd for a pentode run as both a Pentode/triode and triode/inverter.

Curious, When you mentioned special tubes made for design challenges, how would one of these tubes be marked ?
Or would they just have a special silkscreened part number and no tube number designation, like ex:12at7. Kinda of like some unidentifiable transistors you find on power supplies, and other proprietary circuits with special part numbers on them that do not cross to any standard industry part numbers.

 

[KrIVIUM2323]

Kickback your saying a pentode wired as triode would be better ?? I dont dispute , want to hear more, I no theres plenty vari mu circuits that do that...

Just take a look here: http://www.manleylabs.com/techpage/TBAR.html

The fact is that pentodes tubes can be divided in 2 kind: sharp cutoff and remote cutoff. By nature the second group (remote) are well suited as varimu because they have control grids of uneven spacing (the grid is 'tight' in one end and 'more spaced' in the other). When the bias is large the transconductance of the tube is small limiting the gain of the amplifier hence varying transconductance: varimu by nature when connected as triodes!

6ba6 and 6sk7 are eg of remote cutoff tube, 6sj7 and 6au6 are eg of sharp cutoff.

Doesnt the 1K pot between the first tubes plates tune out imbalances ??

It does but for small imbalances, said 5% max. For larger ones it can't. This is the reason why pairing is so important in this purpose.


Gary O could you measure the consumption of your kind of Vk1? I would like to know how many milliamp a channel is using in pratctice. Thanks in advance.

 

[skal1]

Very very interesting, so KrIVIUM2323 you seem to now a lot about these things compressor i mean, anyway can you tell us what would be the max gain out of the 12at?* to drive the cv to -30v  just query something in my head.


skal1

 

[KrIVIUM2323]

you seem to now a lot about these things compressor

I'd love to know much but i'm like many in here: i'm learning!  

anyway can you tell us what would be the max gain out of the 12at?*

No i can't... because i don't understand the way the gain stage is working! ... it's not entirely true indeed as i understand how it work but not the math behind (to compute it theorically)  ;D
But... we've got an output attenuator of 30db so i would say 30db for the gain stage...

For the moment, what i noticed is that the arrangement (anode straight to primary of center tapped output transfo with B+ at center tap and common cathode resistor for both half of 12at7) really looks like a power amp output stage. In fact it's almost the same with others varimu designs.

And... by looking at 12AT7 datasheet it make sense: typical gain of classA with 250V anode Voltage, cathode resistor of 200 ohms (470 ohms driving 2 cathodes in // is approximately 220ohms for both no?) give an amplification factor of 60 ( 35db gain). If we take out 5db (feedback r103/102) which is aproximately 15% negative feedback it seems ok...at least for me! ;D


So :

to drive the cv to -30v  just query something in my head.

I won't be surprised that the treshold 47k log / 18k resistor// with 6al5 cathode input in both sides can take 15v each side which after rectification by 6al5 sum to give 30V. I'll try maths but here again i don't know how to make them...

But 30V for driving the gain reduction stage seems a lot to me. I was more expecting something in the 10V range...  ???

How did you get 30V Skal1? Assumption ? Math? I'm curious too as this is something i ask for myself too.  
(edit: after reading a post from PRR it seems that we could expect something in the range of 1db/volt for cv/gain reduction stage here: http://www.groupdiy.com/index.php?topic=1509 )

 

[skal1]

yes that seems right , so if i was to measure an 436c with out the 6al5 in the socket  , just using it like a line i should be see 30 dB ->
  Voltage-ratio = 31.6228
  Power-ratio = 1000
hum.

so that means just over 15v a side at 6al5 , so what does 30 Voltage-ratio equate to in ac after the blocking caps ?


after 6al5  you should expect at lease a good cv , yes  ;D well if you built that thing you would not see anything in that  :


skal1

 

[KrIVIUM2323]

so if i was to measure an 436c with out the 6al5 in the socket  , just using it like a line i should be see 30 dB

A little less as in 436C the output tube is a 6CG7 and it has an amplification factor of 20 so a gain of 26db. But are we talking about 436C or VK1?  Even if it's the same kind of topologie/schematic there is some change between the two designs... ;D

so that means just over 15v a side at 6al5 , so what does 30 Voltage-ratio equate to in ac after the blocking caps ?

In fact it's not really 15v from the ac signal: i've read somewhere that 6al5 need a minimum voltage of 45v to begin conduct so it's more 60v from any sides of the output which is 'seen' by 6al5 input. If cv is 30V it means 30db gain reduction... HIGH GR!
I never use comp with such high gr!...  ;D

after 6al5  you should expect at lease a good cv , yes   well if you built that thing you would not see anything in that  

CV is comp/threshold dependent so it depend what you do with your comp settings...
What do you mean? About Vk1? Purusha did crap we all know that... but Olaf did a pretty good job as far as i can see. Anyway if i built one i'll go P-T-P not pcb. And with the right kind of tube!

 

[skal1]

pray tell...


skal1

 

[KrIVIUM2323]

pray tell...

? Excuse me english is a foreign language for me i don't understand.