Dbx Vca Construction

[Michael_Joly]

David Blackmer (for whom I worked under for 15 years) must be chuckling in his grave at the thought of someone wanting to resurect his circa 1972 discrete VCA and / or RMS modules. I say this because David was always about pushing the envelope in the quest for better sound. And were he alive today he would not be puttering around with work he did 40 years ago. But have fun!

But speaking of old times...anybody remember reading Davids article in Recording Engineer Producer about his new dbx noise reduction system?

 

[JohnRoberts]

hmmm....The only reasons, I would even care to want to try to build my own VCA are:

1) Main Reason for me is SOUND - I haven't had the chance to discrimintely listen to a THAT chip vs. a Discrete VCA. I have much to learn and the experience of the people on this forum is priceless. If it is  wasted effort, as it sounds it could be sonically (if the discrete just sounds BAD) then it might be  a waste of time

2)  second is just to learn. This is an end in itself

Thanks
AC

Trying not to repeat myself, and trying to avoid subjective judgements (like this kind of distortion sounds good and that one doesn't), there are clear objective improvements between even the best early VCA hand built by David himself, and the newest generation IC technology. There is a pretty long AES paper from one of the THAT VCA design engineers a few years back, counting the ways. 

Not making a judgement, but there are obvious subjective perceived differences between different compressor side chains. The control path would color the sound even if using a perfect VCA that doesn't exist. This IMO justifies the popularity of some classic legacy designs. So even with a near perfect VCA there will be X% deviation caused by the gain manipulation and Y% from the VCA errors. With a lesser performance VCA the X% will still be there but the Y% contribution will be larger. .

I don't want to damp your enthusiasm or inquiry. The open question (to me) is it the X% sound that you want to replicate or the X+Y? Does the larger Y% error added by the old discrete VCA add some useful euphonious coloration, or detract from the goodness of the X%.

Perhaps build them up each way and listen.

It doesn't matter what I think... it's your time and money.  Just know what the variables are.

JR



 

 

[damnyankee]

Trying not to repeat myself, and trying to avoid subjective judgements (like this kind of distortion sounds good and that one doesn't), there are clear objective improvements between even the best early VCA hand built by David himself, and the newest generation IC technology. There is a pretty long AES paper from one of the THAT VCA design engineers a few years back, counting the ways. 

Not making a judgement, but there are obvious subjective perceived differences between different compressor side chains. The control path would color the sound even if using a perfect VCA that doesn't exist. This IMO justifies the popularity of some classic legacy designs. So even with a near perfect VCA there will be X% deviation caused by the gain manipulation and Y% from the VCA errors. With a lesser performance VCA the X% will still be there but the Y% contribution will be larger. .

I don't want to damp your enthusiasm or inquiry. The open question (to me) is it the X% sound that you want to replicate or the X+Y? Does the larger Y% error added by the old discrete VCA add some useful euphonious coloration, or detract from the goodness of the X%.

Perhaps build them up each way and listen.

It doesn't matter what I think... it's your time and money.  Just know what the variables are.

JR

Hiya John,

AC may have a side project in mind with these VCAs, but he's spearheading the dbx 160VU cloning project here.  He's completed the pcb and has provided footprints on the board to accomodate a 200 VCA OR a THAT 2180.  Some on the build will likely build both on the board and toggle to between the two VCAs. 

As for the old opamps: the original BC-109C/BC-263 opamps have newer NTE equivalents which we will be using.  The tricky part will be hfe and Vbe matching AND keep them thermally neutral (glue???). 

Now I have several dbx 119s that I can harvest the 200 VCA and 208 RMS out of, but since this is my first foray into building something that's not a guitar pedal, I'm chalking this up as a cheap education and it will be a great learning experience for me.  The cost of the entire clone project - including meter and rack will be under $200.

DY

 

[damnyankee]

Trying not to repeat myself, and trying to avoid subjective judgements (like this kind of distortion sounds good and that one doesn't), there are clear objective improvements between even the best early VCA hand built by David himself, and the newest generation IC technology. There is a pretty long AES paper from one of the THAT VCA design engineers a few years back, counting the ways. 

Not making a judgement, but there are obvious subjective perceived differences between different compressor side chains. The control path would color the sound even if using a perfect VCA that doesn't exist. This IMO justifies the popularity of some classic legacy designs. So even with a near perfect VCA there will be X% deviation caused by the gain manipulation and Y% from the VCA errors. With a lesser performance VCA the X% will still be there but the Y% contribution will be larger. .

I don't want to damp your enthusiasm or inquiry. The open question (to me) is it the X% sound that you want to replicate or the X+Y? Does the larger Y% error added by the old discrete VCA add some useful euphonious coloration, or detract from the goodness of the X%.

Perhaps build them up each way and listen.

It doesn't matter what I think... it's your time and money.  Just know what the variables are.

JR

Hiya John,

AC may have a side project in mind with these VCAs, but he's spearheading the dbx 160VU cloning project here.  He's completed the pcb and has provided footprints on the board to accomodate a 200 VCA OR a THAT 2180.  Some on the build will likely build both on the board and toggle to between the two VCAs. 

As for the old BC-109C/BC-263 in the original 200 VCA: there are newer NTE equivalents which we will be using.  The tricky part will be hfe and Vbe matching AND keep them thermally neutral (glue???). 

I have several dbx 119s which I can harvest 200 VCAs and 208 RMS's out of, but since this is my first foray into building something that's not a guitar pedal, I'm chalking this up as a cheap education and it will be a great learning experience for me.  The cost of the entire clone project - including meter and rack will be under $200.

hfe matching seems pretty straight forward.  Have you ever done Vbe matching?  Is it difficult?

DY

 

[abechap024]

    Yes I agree on pushing the envelope and pursuing the best sound. But sometimes doesn't that mean searching the history to find the future? I mean Tubes are an old technology and technically hugely inferior. But If I was offered the choice to have a studio filled with old tube equipment or a studio filled with clean, noise free, no distortion equipment - I would have to say at this point in my life and personal tastes I would quickly choose the Tube equipment.

Its good to know that this clearly is Not going to be a walk in the park. Its going to be a lot of work and frustration with almost guaranteed inferior technical specs....

Still into the fan my hand goes  :-\

I'm thinking the matched transistor array is a great idea... then a bunch of trimmers to optimize the resistor values.
I mean its not going to be easy, but at the same time - (here comes my niaveity) I'm confident its not going to be *impossible* !

I need to make it obvious for people looking into building the 160 that the discrete VCA is going to be strictly a DO AT YOUR OWN RISK sort of thing. I am going to get it a try when I get all my parts....and see how it goes...
So the question isn't really how and why NOT to build a discrete VCA unit

But how to build a vca unit ;D

even though this thread has very good real advice, and I do respect the voices saying that it will be a PITA with possible inferior results. but I still want to move ahead with it, with out seeming that I'm disregarding the advice, because I have taken it to heart and I hope anyone else attempting to build a discrete VCA will do the same.

AC

 

[Biasrocks]

David Blackmer (for whom I worked under for 15 years) must be chuckling in his grave at the thought of someone wanting to resurect his circa 1972 discrete VCA and / or RMS modules. I say this because David was always about pushing the envelope in the quest for better sound. And were he alive today he would not be puttering around with work he did 40 years ago. But have fun!

Chuckling like Rein Narma did when CJ mentioned that people were still using his Fairchild limiter 50 years later.

Most designers are looking for the next best thing, nature of the beast.

It certainly doesn't mean that something they designed isn't still useful. As with many recording tools that have been around awhile and are still in use, there's a good reason.

Mark

 

[JohnRoberts]

Hiya John,

AC may have a side project in mind with these VCAs, but he's spearheading the dbx 160VU cloning project here.  He's completed the pcb and has provided footprints on the board to accomodate a 200 VCA OR a THAT 2180.  Some on the build will likely build both on the board and toggle to between the two VCAs. 

As for the old opamps: the original BC-109C/BC-263 opamps have newer NTE equivalents which we will be using.  The tricky part will be hfe and Vbe matching AND keep them thermally neutral (glue???). 

Now I have several dbx 119s that I can harvest the 200 VCA and 208 RMS out of, but since this is my first foray into building something that's not a guitar pedal, I'm chalking this up as a cheap education and it will be a great learning experience for me.  The cost of the entire clone project - including meter and rack will be under $200.

DY

The transistors will need to be held at a constant temperature while you are matching them since Vbe and beta(?) also changes with temperature. Perhaps clamp a bunch of them to a metal bar to force them to the same temp before measuring. Match the beta and Vbe at the roughly the same current they will be operating at in the final circuit to avoid surprises.

I think the Vbe needs to be matched from NPN to other NPN devices, and from PNP to other PNPs,  Ideally the beta could also match between NPNs and PNPs to help hold down a different distortion mechanism, but don't kill yourself.

The BC109 and 263 are relatively high beta for small signal transistors, so when subbing look for something similar.

JR 

PS: Under $200 not counting your time?

 

[damnyankee]

Hi John,

I planned on using a room air fan...and the metal bar is a great idea. 

I think the plan is to sub the BC109s and 263s with BC550s & BC560s.  These should work, yes?

I have read in another forum someone tried to recreate a dbx VCA without Vbe matching and the results were awful as the builder hfe-matched the transistors but didn't Vbe-match them.  I learned a lesson at someone else's expense.

And you are correct: the $200 doesn't include my time.  But ya know...like gardening, this is a hobby for me.  I can go to a quiet place and enjoy some quiet time with a soldering iron. 

DY

 

[abechap024]

Thats a good point. there is a lot of talk about "time". some people do EE/equipment construction repairs as their fulltime job others just as a relaxing hobby. I feel I'm in between. So the time spent trying to make a usable VCA will not be ill spent. The bar is a great Idea...and also maybe a bucket of mineral oil if you don't mind getting a little messy. Just a thought.
AC

edit: also I think for sure going to change the PCB on the 160 to accept a matched transistor chip.

 

[livingnote]

I really like that PCB layout they have

Of course back when I was at Volker's I was thinking about cloning these things too, but the
more I got talking about VCAs the more obvious it was that "ummm...no". But if you're seriously
gonna get into these, I'm game.

 

[damnyankee]

I really like that PCB layout they have

Of course back when I was at Volker's I was thinking about cloning these things too, but the
more I got talking about VCAs the more obvious it was that "ummm...no". But if you're seriously
gonna get into these, I'm game.

Abe worked his rearend off to create that pcb with input from Matthias, Arksa, and a few others. 

Not only does the board offer THAT2180 and 200 VCA footprints, but the electronics for the original VU has been altered so a non-dbx VU will look/behave just like the dbx original.  Speaking of the non-dbx VU: another helpful individual took apart their original VU and scanned the VU label so one can photo edit out the blue "dbx" and text edit whatever they choose.

For the 3 button switch: Mike/Hairball Audio can provide his 1176 switch (yeah, it's 4 buttons not the original dbx 3 button) and the plan is to utilize the 4th button as a power off button.

It's coming along - we're just awaiting the boards to come in.  I'm going to harvest my 200 VCA from one of my old dbx 119's as well as purchase a THAT 2180; I'll probably insert a toggle in between to see which VCA I prefer.  ;D

 

[livingnote]

Aha - ok, that's great. I was just thinking that if anyone's gonna go on a quest for excalibur
i might be able to help with a free redraw, but if you're already that far into it then I'll join
the curious ones

 

[emrr]

   
Its good to know that this clearly is Not going to be a walk in the park. Its going to be a lot of work and frustration with almost guaranteed inferior technical specs....

Still into the fan my hand goes  :-\

Good man; have fun.

 

[JohnRoberts]

Hi John,

I planned on using a room air fan...and the metal bar is a great idea. 

I think the plan is to sub the BC109s and 263s with BC550s & BC560s.  These should work, yes?

I have read in another forum someone tried to recreate a dbx VCA without Vbe matching and the results were awful as the builder hfe-matched the transistors but didn't Vbe-match them.  I learned a lesson at someone else's expense.

And you are correct: the $200 doesn't include my time.  But ya know...like gardening, this is a hobby for me.  I can go to a quiet place and enjoy some quiet time with a soldering iron. 

DY

The fairchild data sheet shows that these are available graded for beta, I would consider the highest beta version "C". Since they are not available in metal cans, you may need to thermally bond them together some other way.

Ideally you would try some compared to 109 and see how they act... They look similar on paper.

JR

 

[abechap024]

I really like that PCB layout they have

Haha, Yes It isn't the most pretty design, of that is what your hinting at. But then again its my first design and no one else seemed interested in doing it the 1+ year the 160 thread was hanging around. So I figured I would give it a go. I apologize for the design not being better.
Livingnote some of your designs are very great, I like how you take the creative approach and turn your pcbs into real works of art. Me, at this point, just want to make a PCB that works. And yes  you can probably tell it was mostly all auto routed, and personally don't really have a problem with it. Just as long as it works. (Ahh! I hope it works!)
Now that I have seen the PCBS in hand though, I think I'm going to move some vias around for te next batch.
Kind regards
Abe


Sorry for the Edit: I realized I could just be a little paranoid....haha are you talking about the 200 vca pcb that is floating around? or the 160vu i have...
Thanks
AC

 

[abechap024]

I'll probably insert a toggle in between to see which VCA I prefer.   ;D

Hi DY,
Just a Note, due to the parts that need to be changed around the VCA there can be no easy way to insert a toggle switch to switch between different VCAs. On these boards, you just have to choose what VCA you want in the beginning, there is something like 3-4 resistors that change if you are going with the that or  discrete.
You should be getting your board soon!
Abe

 

[abechap024]

Hey,
You know what would be very helpful...if someone out there could open a 200 unit to measure some of those select resistors... to see how widely they vary.

 

[damnyankee]

Hey,
You know what would be very helpful...if someone out there could open a 200 unit to measure some of those select resistors... to see how widely they vary.

Is that a hint, AC?

I have my dbx 119 already opened up so I'll see if I can pry the can apart.  Check yer PM...I have questions for you.

DY

 

[damnyankee]

I'll probably insert a toggle in between to see which VCA I prefer.   ;D

Hi DY,
Just a Note, due to the parts that need to be changed around the VCA there can be no easy way to insert a toggle switch to switch between different VCAs. On these boards, you just have to choose what VCA you want in the beginning, there is something like 3-4 resistors that change if you are going with the that or  discrete.
You should be getting your board soon!
Abe

That's why on the 8th day God created perfboards - and I'm sure there's a way to mount one above the main PCB.  If there's a will, there's a way.

And thank you for the board - I'm frothing at the mouth!!!

DY

 

[abechap024]

Hey,
You know what would be very helpful...if someone out there could open a 200 unit to measure some of those select resistors... to see how widely they vary.

Is that a hint, AC?

I have my dbx 119 already opened up so I'll see if I can pry the can apart.  Check yer PM...I have questions for you.

DY

Haha, Yes it was a hint if you could check those we would be that much closer to figuring this out. Hopefully those metal cans aren't too hard to pry off. If they are then don't worry about it. Also the only PM i have from you is the one I replied to a couple days ago...am  I missing something?
AC