Gates Sta-Level Stereo Link

[gg85]

Hey guys, I've finally got my drip dual gates sta-level build at about 95% completion. I'm really stoked with how it sounds! There's a bit of a 50Hz hum issue that I need to trace down on both channels, one more than the other. I've balanced the 6V6's so i'm pretty sure that's not where the hum is, but anyways.

I've built it up with stepped attenuators with 0.5% and 1% resistors and they're measuring exactly the same on each channel with a DDM.

One thing that is strange is that before both channels start compressing, the left channel is about 1.7dB hotter. This follows up until the compression threshold. Once they both get to about 7dB compression, both channels are within 0.1/0.2dB of each other all the way up the range of the in and out attenuators. This is with the unit in Dual Mono. I am not 100% sure on the best way to stereo link them and at the moment I am just linking the sidechains together with a switch. The same thing happens even with the SC linked, although it does induce much more hum. Maybe there is a better way to link them and keep the unit quiet, but at this stage the difference in stereo balance at low to no compression is the main issue I am stumped with.

And to everyone that has helped me with the build so far, thank you very much. It's been a serious learning curve!!

 

[emrr]

Look at the Collins 26U-2 schematic for stereo linking that is isolated from the time constants. 

I've found with any old simple vari-mu circuit, you have to decide how much compression you want, THEN trim gains to match.  If it must be precise, dump everything to mono, or use a test tone once the basic settings are dialed in, then go back to stereo input.  If you deviate much from the one amount of compression, one side will always drift off from the other in gain, unless you find those virgin unicorns known as perfectly matched tubes (in all specs).  The better the tube match (all constants), the better the gain/compression tracking.  Truth is these aren't precision devices in the first place. 

 

[gg85]

Man I completely missed the whole Tech Drawing section of this forum. wow. thank you guys!

Doug, to me it looks like the main difference between the 26U-2 and how I currently have my dual sta linked is the 1N459 diode. I've got the single/double/triple mod on my build, so would the best way to add this just be a diode across the S/D/T switch? I also noticed the 47k series resistor between the two CV's as well. Do you think this would be necessary for me, or is it just a way to keep the two CV's at a reasonable level instead of summing too much?

Regarding the difference in L/R between the two channels, do you know how Manley seem to combat this (other than having stock they can sort through and find matches)? I know they aren't using the 6386's anymore, but if you were to retube one with JJ's or even old GE's I would assume you'd run into the same problem I am having, although I can't seem to find much reading on people having this issue in a Manley, so I'm just trying to see if there's something I'm missing, or a mod I can implement in my build.

The schematic shows 1.2V at no compression. In my build i'm sitting a little higher than that, more like 1.45V, but not exactly the same in each channel. I don't know if replacing R8 and R9 with trimpots to bring that idle voltage down to match the other channel would be something worth exploring. I also have slightly different voltages at R12 where the schematic says 105V. I'm sitting a bit higher, more like 117V but a few volts difference between channels. Maybe a trimmer after the 20k resistor could bring it down to 105V on both channels. I know this is all within acceptable range of this circuit, and they weren't designed for the precision I'm trying for, but if it is at all possible then I definitely would like to try and see how good I can get it

I also noticed a difference between the 12AT7 cathode voltage in each channel, which follows if I swap the tubes, but to be honest I don't completely follow what that tube is doing in the circuit and whether it would affect L/R balance.

Thanks!

 

[XAXAU]

@gg85

Did you ever solve the stereo link dilemma?

Looking to build a drip dual sta  myself

 

[lassoharp]

The Manley Vari-Mu implements a stereo link simply by taking a feed from one channel just after the attack resistor (junction of attack resistor and the shunt timing cap) and sends it to other unit at the same place. There are no diodes, xtra caps or iso resistors indicated on the schematic.

 

[emrr]

The Manley Vari-Mu implements a stereo link simply by taking a feed from one channel just after the attack resistor (junction of attack resistor and the shunt timing cap) and sends it to other unit at the same place. There are no diodes, xtra caps or iso resistors indicated on the schematic.

That changes the timing since the caps become parallel.

@gg85:
The 26U-2, the diodes isolate the caps from one another so the timing doesn't change half/double depending on coupling. The 10M bleeder around the diode is interesting. The 47K series resistor isolates the input trans secondaries and also introduces some slight decoupling of the link. The diode iso for timing decoupling is the important part, otherwise mono and stereo are not the same compressor.

 

[gg85]

That changes the timing since the caps become parallel.

@gg85:
The 26U-2, the diodes isolate the caps from one another so the timing doesn't change half/double depending on coupling. The 10M bleeder around the diode is interesting. The 47K series resistor isolates the input trans secondaries and also introduces some slight decoupling of the link. The diode iso for timing decoupling is the important part, otherwise mono and stereo are not the same compressor.

Got it!

The 10M in the STA seems to be the difference between Single and Double modes. I'm assuming in the 26U-2 there was so switch for that and the Double time constants were chosen as default?

Do you think it would be worth experimenting with the 47k resistor and trying different values there? Does the 47k correspond to anything in particular with the secondaries?

 

[emrr]

I took the 47k out wasn’t obvious to the ear. This 10M isn’t a single/double resistor.

 

[gg85]

I took the 47k out wasn’t obvious to the ear. This 10M isn’t a single/double resistor.

Great, I'll just stick with the diodes then. Thanks Doug!!

 

[Script]

Implemented diode iso successfully on a PRR varimu unit using gernanium many years ago, after having read above explanation for the first time Also tried resistor like 47k and it indeed decouples the channels. I put it +different value though+ into a filter across the channels.

But I never bothered about .1uf and 10M, which makes me wonder why. Is that cap there as reserve to overcome the diode drop ? Or does it form a filter with the 10M ?

 

[gg85]

Implemented diode iso successfully on a PRR varimu unit using gernanium many years ago, after having read above explanation for the first time Also tried resistor like 47k and it indeed decouples the channels. I put it +different value though+ into a filter across the channels.

But I never bothered about .1uf and 10M, which makes me wonder why. Is that cap there as reserve to overcome the diode drop ? Or does it form a filter with the 10M ?

In the case of the STA would you suggest putting the diode in parallel with R37? I'm unsure if the single/double switch will cancel it out in this arrangement...

 

[Script]

I might be wrong but I have a feeling that the 10M in the 26U-2 mentioned above and the 10M in your STA schematic (IMG8988) only happen to be of the same value.