Wilcox M57D1 tube limiter restoration

[abbey road d enfer]

From RDH3: The 6SA7 is somewhat similar in construction to the 6L7-G but No. 3 grid has a super control characteristic whereas in the 6L7-G the first grid has a super control characteristic.

Weird. According to datasheet, G3 is the RF input; I wonder why signal should go to an input that has less linear transconductance...  ???

 

[Heikki]

Weird. According to datasheet, G3 is the RF input; I wonder why signal should go to an input that has less linear transconductance...  ???

That seems to be the way it was usually done in compressors and expanders.

https://worldradiohistory.com/hd2/IDX-Site-Early-Radio/Archive-Radio-Engineering-IDX/IDX/35/Radio-Engineering-1935-12-OCR-Page-0026.pdf

 

[abbey road d enfer]

That seems to be the way it was usually done in compressors and expanders.

https://worldradiohistory.com/hd2/IDX-Site-Early-Radio/Archive-Radio-Engineering-IDX/IDX/35/Radio-Engineering-1935-12-OCR-Page-0026.pdf

Well, since there is no technical justification for this choice, I have to take it at face value.
I still feel it strange, particularly considering it's a feed-forward arrangement.

 

[hereforever]

I've had a little more time to work on this...

200uA meter is working well strapped across a 220 resistor at the junction of R5 & R6/R7

I moved the sidechain signal tap to the 6V6 plates - each side through a 220K resistor & .047uf cap.  Otherwise the amp was getting too distorted before I could get the threshold into a useful range.

Next I will replace the carbon comps with metal film and then fine tune a bit more

It sounds pretty good

 

[abbey road d enfer]

I moved the sidechain signal tap to the 6V6 plates - each side through a 220K resistor & .047uf cap.  Otherwise the amp was getting too distorted before I could get the threshold into a useful range.

Not surprizing since you had one less stage in the side-chain after your mod.

Next I will replace the carbon comps with metal film

Don't throw them away. Put them on ebay. They're full of mojo.  8)

 

[hereforever]

Not surprizing since you had one less stage in the side-chain after your mod.

Not really of consequence, but I'm not sure this was modified before I got it - I could be wrong but the wiring all looked original.  Perhaps an earlier revision because the date codes pre-date the schematic I'm working from by about 8 years

 

[hereforever]

When I inject sub 60hz signal into the unit I'm getting some motorboating that occurs only while limiting. I didn't notice this when the sidechain was fed from the 6SN7. Increasing C6 helps somewhat but it doesn't entirely disappear. I haven't tried increasing C1 yet.

Before fine tuning I will wait to replace resistors and try to make sure the input tubes balance is OK - In particular, V1's plate resistor (R6) has really drifted in value. I thought of also adding a balance trimpot on the cathodes of V1/V2.

 

[abbey road d enfer]

Not really of consequence, but I'm not sure this was modified before I got it - I could be wrong but the wiring all looked original.  Perhaps an earlier revision because the date codes pre-date the schematic I'm working from by about 8 years

  My understanding of your first posts, you took one 12BH7 and put a 6V6 instead...?

 

[hereforever]

  My understanding of your first posts, you took one 12BH7 and put a 6V6 instead...?

No, I redrew the circuit as it was when I got the unit.  The Army manual I have is from 1952 but the pot codes are 1946. It seems possible I have an earlier version with a pair of 6V6 on the output rather than a 6SN7 and minus the 6SN7 side chain gain stage

 

[abbey road d enfer]

OK, got it.

 

[hereforever]

I've been working on this a bit more, here and there, and re-drew the schematic:




The sidechain is now fed from the 6V6 plates.
I split V3 cathode resistor from 2K to a pair of 3.9K, although I will probably change this to a pair into a shared resistor.
Negative feedback also added from the 6V6 plates to V3's cathodes. I measured about 60db gain before adding the negative feedback. Haven't measured again yet. The negative feedback made a big difference with lowering distortion and lowered hum a bit too.
I've made C6 and R17 variable, on rotary switches. Also trying a 25K pot for now between R17 and the CT of T4 as attack control
200uA meter strapped across a 470 resistor between C1/R5 and R6/R7 - I'm also using a 1K resistor across the meter terminals to zero

It is sounding good- fairly clean with modest limiting and gritty in a good way with heavier action

One thing I'm curious about, is there any advantage to T4? I'm not familiar with other tube compressors or limiters that use an interstage transformer in this position

I thought I might add a balance control for good measure, although thump hasn't been noticeable. Any suggestions on best placement? I thought maybe between R6 & R7 on V1/V2 plates

 

[abbey road d enfer]

One thing I'm curious about, is there any advantage to T4? I'm not familiar with other tube compressors or limiters that use an interstage transformer in this position

Are you positive the resistors that drive T4 are 270k?
That would imply the transformer primary had an inductance of 1500+H...
Have you evaluated the ratio of this xfmr?
EDIT: I checked the original schemo. So you drive this xfmr, that was designed for being driven by about 12-15 kohm, with a drive impedance of 540k. This has two consequences. The LF response of the side-chain will probably roll-off at low-mid frequencies, and the attack time will suffer considerably.
Rolling-off LF in the side-chain may very well be a "feature" since many commercial units include it as an option, that avoids pumping with bass and kick, but the slow attack is rarely welcome.

 

[emrr]

Yeah the ratio of T4 is of interest, as well you would expect to see a load resistor to fix transformer response to some degree in a case like this, like from 1-3.

 

[mjrippe]

T4 would more often have been a tertiary winding on T3, so as the guys have said there must be a reason for the extra expense.

 

[abbey road d enfer]

T4 would more often have been a tertiary winding on T3, so as the guys have said there must be a reason for the extra expense.

The original schemo shows an additional 6SN7. Since the original output tube was also a 6SN7, the designer may have considered that combining the two tasks on one tube was too much, and thus separated the functions. We're talking about a unit that's been significantly modded, and a schemo of a different version...

 

[MaxDM]

Check the ratio of the GR Transformer. It would make more sense as a step down.

Schematics often had mistakes

 

[hereforever]

Thanks all for the analyses.

Check the ratio of the GR Transformer. It would make more sense as a step down.

Schematics often had mistakes

Will check. As pictured it seems to be wired 100k:10k

The original schemo shows an additional 6SN7. Since the original output tube was also a 6SN7, the designer may have considered that combining the two tasks on one tube was too much, and thus separated the functions. We're talking about a unit taht's been significantly modded, and a schemo of a different version...

May wire it up with the 3 6SN7's on the schematic and see how it performs

 

[MaxDM]

yes, if it's a step down, then the added gain from the 'extra' tube is going to be reduced, so that in fact it's not increasing much in voltage, but it does increase current, so that it can drive the GR cap faster.

I have my suspicions about R16 as well, which seems to be way too high in value.

I think r16 should be connected to the center tap of the GR transformer and the cap to the diodes, in a normal fashion. Then the transformer with the extra tube is actually an improvement over other typical designs, which take the signal for GR unbuffered from the output, such as the 176 etc.

 

[hereforever]

I've finally had a couple of weekends to return to this project. In the meantime I came across another for sale and now I have a pair. The second unit has Thordarson input & interstage transformers and the same Stancor output transformer as the first.

The new unit is wired same as the print schematic with a minor addition- the T3 & T4 center taps are fed from a 10K resistor off the B+ line. I've wired my original unit to match this and I've been testing some modifications in the side chain. I made attack and release both variable and added a 'Speed' control inspired by the Sta-Level / Collins 356E / GE BA-9 circuits. I've been testing values for the capacitors and resistors by ear and how far I can push it until I get unstable behavior.

It's working well in this current configuration.

I also had a front panel made, mounted the units to rack shelves and used Molex connections to bring the circuit points I need out to the panel.

In measurements taken in Room EQ Wizard it seems to be fairly full frequency with a few db hump up around 12 or 15 kHz.

 

[hereforever]

I've been continuing to test time constant values and what I am finding is that the release is relatively fast. A 5M linear pot wired as a variable resistor from points 7 & 8 on the rev2 schematic in the last post isn't doing much of anything to vary this. Is this due to R18/R19 on the sidechain transformer? I'm unclear on the best way around this without a considerable redesign of the sidechain.

I'm also interested in adding a balance control on the cathodes and/or plates of V1/V2. I thought I'd try replacing R2 with a 200ohm resistor wiper to ground and similarly a 1K resistor between R6/R7. To set the balance could I tap a signal off one leg of the heaters through a 1uF cap and inject it at T4 center tap? I've seen something like this being done on other tube compressors.