Gates SA-39B from scratch

[DaveP]

Front Panel now complete and connected up, so far no unaccounted for wiring!
Here's it from the back with the grid caps fitted.



The front panel has a jury rig threshold control so that I can work out the comp ratios during testing, they will then be hard wired into a 6 way switch.


Tomorrow I start all the testing.
best
DaveP

 

[MagnetoSound]

Beautiful, Dave. The wiring pics on the previous page are simply breathtaking.

Thanks for doing this. 

 

[DaveP]

Thanks Dan,
I try to do it little and often, its hopeless trying to do this kind of stuff late at night, the following morning you find all the mistakes :
best
DaveP

 

[letterbeacon]

Bravo Dave -it looks brilliant!

Can't wait to hear some samples...

 

[DaveP]

I have been some time designing the threshold Control.
This is what I've come up with:-



This gives a good spread above and below the original single curve.

The comp ratios vary from 2~3 on the knee to 12~16 at the end of the flat.

Curve 6 runs into some nice sine flattening at high inputs as it runs out on the max output of the amp.

My next job is to make some sound samples....any ideas of what would be a good test?
I've not made or posted a sound sample before, so any tips are appreciated.  I don't really want to fill my allocation with Soundcloud with samples so I need a site that would be pretty permanent for future forum users.
best
DaveP

 

[DaveP]

This is the data for the signal to noise ratios.
The six compression curves above, all allow different maximum output voltages.
The input control introduces more noise as it is turned up as you would expect, so using the lowest input setting (which implies a high drive voltage) the residual hum and noise is 9mVp-p.

Using this input level, the following S:N ratios are possible with the various compression curves.
Curve 1=72.0dB
Curve 2=73.4dB
Curve 3=74.5dB
Curve 4=75.7dB
Curve 5=76.6dB
Curve 6=77.5dB

If higher input level settings are used, then more noise is introduced, but the worst case is still 62.6dB and this would mean that the input signal was only 700mV.

All in all, I'm pleased that this WW2 design, made with 1930's metal tubes and modern components has turned out to have quite respectable figures; the layout must be in the right ball park.
best
DaveP

 

[alexc]

I would just love to have one of your creations in my studio.

Actually, I would like to have one of *each* of them 

I'm just wondering, what's next ?

 

[DaveP]

Well Alex,
this one will be up for sale later on.
The next is a Gates 28 CO with combined vari-mu and light bulb compression.
best
Dave

 

[alexc]

Alas, I'm pretty tight these days - selling myself,  more than anything  :-[

But I do hope some enterprising person with the funds that these beauties deserve can provide a good studio for them.

I have to say, if I was doing any more limiters, (which i'm not), a combo of tube limiting and opto limiting would be on my list too.

I find the vary-bias cathode-follower tube limiting with added opto-shunting combo intriguing.

There was a monster radio limiter done that way plus a whole bunch more posted some time back.
It's a veritable mine of subcircuits that all made good sense once I redrew them.

 

[DaveP]

Ok here are the sound samples, this is very brave of me I might add, showing the world what a crap guitarist, bassist and vocalist I am! :-[

Anyway, the first part is au naturelle, then it's repeated again but through the compressor.
https://www.box.com/s/rrmppc6y8awamqelfxmc
The first lick is a Telecaster with about 10dB, the second is a Jazz bass with about 5dB and the last is me with 3dB.
To my ears it seems to affect the bass tone most of all, but I hope you experts out there might hear something else.
best
DaveP

 

[alexc]

Definately cleans and clears up things in all cases. Much smoother in each.

Quite a bit more distortion in the unlimited cases for guitar and especially bass.

The bass has a bit of 'warbling' going on which is reduced in the limited case - it does sound like some kind of modulation could be happening in the unlimited case.

But for sure it is cleaning things up! The guitar and vocals for sure would be a good app. The bass could probably use less gain going in.

Cool! Thanks for sharing. One day I too may be brave enough to do the same! 

 

[emrr]

That sounds reasonably like my memory of the gain reduction quality of all the SA-39's I've ever heard.  Between original SA-39's the most differences I hear are in the transformer versions used, and this is not apparently outside of that range. 

 

[MagnetoSound]

It sounds really good!

I especially like the way it adds definition to the vocals (not a comment on the performance) ... but it sounds good on the bass and guitar as well. I agree with Alex, I wonder if the instruments were perhaps recorded a bit hot to begin with ..?

One really smooth compressor with balls, though. Great work!

 

[DaveP]

Thanks Dan, Doug and Alex,
The instruments went through a 9:1 DI box before going into my mic pre.  The threshold was on the lowest setting so it was the softest knee on the curves above.  I put the input control about halfway then tuned up the mic pre to get the compression level.
I played about with the attack and release until it sounded best, then adjusted the output control so it got the right level to my desk.

I think the guitar was a little hot, I like it on the edge of crunch, the effect on the bass really surprised me, I did it twice cos I thought I might have twisted the tone control by mistake!  Rescue Me bass line, don' it take you right back to '65 :
The 6L7's were not a matched pair but I very carefully balanced them and I could not get it to thump at all, there was no hint of it being microphonic either, so all in all, the SA-39B is a very good design that does not need an interstage, those old guys at Gates knew what they were doing, I would recommend it.
best
DaveP

 

[lassoharp]

those old guys at Gates knew what they were doing

. . . and were doing what they were told - "get rid of iron expense!" 


Excellent work as usual Dave!  You continue to inspire us all!

 

[DaveP]

I had to do some recording but now I've got around to the frequency response measurements.
The test was done with the 600 ohm output from the sig gen straight into the input.
The output was measured with the scope on a 600 ohm load.

Blue trace:      No Compression;      Input level 3;  Threshold position 6.

Red Trace:      5dB Compression;    Input level  4;  Threshold position 2.

Green Trace:  10dB Compression;  Input Level  6;  Threshold position 1.

I was amazed how flat this design is, 20Hz-40kHz within +/- 1dB well done Gates.

best
DaveP

 

[emrr]

SA-39B arrived in USA, I'm putting it through tests in the next few days.  It looks really great in person, great work Dave!

 

[Dr Gris]

Anything you would like to share with us about this amazing build?

Kind regards
Magnus

 

[emrr]

Sounding good.  Still trying to wrap my head around the comparison against the original Gates.  Response is pretty much the same at all in/out settings except max gain, which loses some lows.  You'd never use that setting anyway.  Gain reduction extends the bottom end further.  Original SA-39's response varies a good bit with input/output setting. 

 

[khel56]

Wonderful project! To do all of that from scratch is really impressive.

Khel56