Fairchild 670 Solid State Concept

[abbey road d enfer]

The anamod principle - as I understood it from talking to Greg Gualtieri a couple of years ago - is about analoguely simulating a digital simulation of the analogue - i.e. back-translating dsp-functions into analogue domain.

That's a smokescreen (with a capital B and a capital S). Emulating a piece of analog equipment in the digital domain involves convolution. It's impossibly complicated to do a convolution in analog. They use analog building blocks that they tweak to do a fac simile of other analog building blocks.

Greg and Dave wouldn't say that it's "analoguely simulating a digital simulation."

Basically, it's the same principle as an analog computer (which is what op-amps were first invented to implement). A continuous-time mathematical model of the device in question is developed, and this model is implemented entirely in analog electronics.

-a

So, that's what I said; it's an analog processor. You can't implement DSP algorithms (except maybe the basic four operations and some logic functions) in analog computers. Convolution, table-based processing and recursive operations are essentially out of reach.
I have no problem with the notion of using an analog computer to perform some processing, but there is very little, if any, correlation with DSP processing.

 

[bockaudio]

To me, tape simulation is to recording what putting cards to the spokes of a bicycle is to motocycling.

;D Just had a coffee-through-the-nose-laughing experience here. Thanks.

;D

 

[abbey road d enfer]

To me, tape simulation is to recording what putting cards to the spokes of a bicycle is to motocycling.

;D Just had a coffee-through-the-nose-laughing experience here. Thanks.

Sorry for spoiling your coffee break 

 

[abbey road d enfer]

To me, tape simulation is to recording what putting cards to the spokes of a bicycle is to motocycling.

;D Just had a coffee-through-the-nose-laughing experience here. Thanks.

;D

The story of my life...people laugh, must be something I said (or wrote)...

 

[jackies]

I found out that the transformer in front of the tc network is also needed for a specific reason.

What was the specific reason though?

 

[CJ]

probably a few reasons,  i would guess step up and isolation would be high on the list.

Rein told me that if he were designing the 660 today, that he would use opamps, circuit boards, etc.


then hopefully he would revert back to tubes and transformers, 

 

[hop.sing]

I found out that the transformer in front of the tc network is also needed for a specific reason.

What was the specific reason though?

The main reason is the way the fairchild is biased. You need a floating voltage in front of the timing network.
It is not easy to explain for me, but if you look at the cathode biasing arrangement of the 660, you see that it consists of a mixture of a negative voltage and a connection to ground (through a resistance)
As soon as you decrease the current through the tubes, which happens while applying a negative voltage to the grids, the cathodes are tied into the dirction of the negative voltage, too. The solid state SC would run out of voltage very fast, I had like 4 dB gainreduction max. without the transformer. With the transformer the grid voltage is floating, so it is indepedend from the +-18v in the sidechain amp and I can achieve 25dB GR with the same sidechain.
To illustrate the case: In my pc189 compressor with 0v on the grids (no GR) I have ca. +7.5v on the cathodes. With -10v on the grids I have something like 0v on the cathodes. going into deep gain reduction leads to something like -30v on the grids and -10v on the cathodes. And to achieve that with a +-18v Sidechain you need that transformer.
If you bias the cathodes straight to ground with a balance pot and a resistance, you would not need the transformer, but then it is no fairchild any more. I found that the compression characteristics are different that way.
I hope it is possible to follow my explanations...

 

[abbey road d enfer]

I found out that the transformer in front of the tc network is also needed for a specific reason.

What was the specific reason though?

The main reason is the way the fairchild is biased. You need a floating voltage in front of the timing network.
It is not easy to explain for me, but if you look at the cathode biasing arrangement of the 660, you see that it consists of a mixture of a negative voltage and a connection to ground (through a resistance)
As soon as you decrease the current through the tubes, which happens while applying a negative voltage to the grids, the cathodes are tied into the dirction of the negative voltage, too. The solid state SC would run out of voltage very fast, I had like 4 dB gainreduction max. without the transformer. With the transformer the grid voltage is floating, so it is indepedend from the +-18v in the sidechain amp and I can achieve 25dB GR with the same sidechain.
To illustrate the case: In my pc189 compressor with 0v on the grids (no GR) I have ca. +7.5v on the cathodes. With -10v on the grids I have something like 0v on the cathodes. going into deep gain reduction leads to something like -30v on the grids and -10v on the cathodes. And to achieve that with a +-18v Sidechain you need that transformer.
If you bias the cathodes straight to ground with a balance pot and a resistance, you would not need the transformer, but then it is no fairchild any more. I found that the compression characteristics are different that way.
I hope it is possible to follow my explanations...

The grids are biased at about -9V at idle, and the rectified voltage is also negative. The bias voltage varies from about -9 to -70V from idle to max GR, so definitely, you can't do it from +/-18V rails. It has to be high voltage, probably with about -80-100V negative rail.

 

[hop.sing]

I found out that the transformer in front of the tc network is also needed for a specific reason.

What was the specific reason though?

The main reason is the way the fairchild is biased. You need a floating voltage in front of the timing network.
It is not easy to explain for me, but if you look at the cathode biasing arrangement of the 660, you see that it consists of a mixture of a negative voltage and a connection to ground (through a resistance)
As soon as you decrease the current through the tubes, which happens while applying a negative voltage to the grids, the cathodes are tied into the dirction of the negative voltage, too. The solid state SC would run out of voltage very fast, I had like 4 dB gainreduction max. without the transformer. With the transformer the grid voltage is floating, so it is indepedend from the +-18v in the sidechain amp and I can achieve 25dB GR with the same sidechain.
To illustrate the case: In my pc189 compressor with 0v on the grids (no GR) I have ca. +7.5v on the cathodes. With -10v on the grids I have something like 0v on the cathodes. going into deep gain reduction leads to something like -30v on the grids and -10v on the cathodes. And to achieve that with a +-18v Sidechain you need that transformer.
If you bias the cathodes straight to ground with a balance pot and a resistance, you would not need the transformer, but then it is no fairchild any more. I found that the compression characteristics are different that way.
I hope it is possible to follow my explanations...

The grids are biased at about -9V at idle, and the rectified voltage is also negative. The bias voltage varies from about -9 to -70V from idle to max GR, so definitely, you can't do it from +/-18V rails. It has to be high voltage, probably with about -80-100V negative rail.

I took the 660 as an example, because the grid is at 0V and that is easier to think through.
See, without the transformer you just have the  -18V rail, which is not sufficient at all, as I wrote, but with the transformer you have the full +-18V Voltage swing available which may translate to maybe -40V DC in use. I used a 1:2 transformer (But a 1:1 tranny works nearly as well) so it is even more. Since the grid is coupled with the cathode voltage (That is what I think and wrote up there, at least, somebody correct me if I am wrong) it goes even lower in voltage. That might not be -70v but pretty close.
I use tubes with a B+ of 130V which is half the B+ in the fairchild, so when I am nearly in cutoff mode, and I can do that, the fairchild would probably still deliver some signal, but should be in heavy gainreduction already, probably more than ever needed.When I have the time to open my compressor again, I will take some measurements and see whats possible

 

[bockaudio]

even with a SS side chain you need the xfmr. Look at the EAR660 which has both those elements.

 

[abbey road d enfer]

even with a SS side chain you need the xfmr. Look at the EAR660 which has both those elements.

Why would it be necessary, as long as the SS side-chain can produce the high-negative voltage needed? EAR may have decided that it was a better compromise, but certainly not the only solution.

 

[hop.sing]

even with a SS side chain you need the xfmr. Look at the EAR660 which has both those elements.

Why would it be necessary, as long as the SS side-chain can produce the high-negative voltage needed? EAR may have decided that it was a better compromise, but certainly not the only solution.

I think with the way the cathode is set up, it is necessary, since the grid voltage follows not only the sidechain voltage but also the cathode voltage, I am pretty sure the behaviour of the compression would be different if the rectified sc voltage would not be floating. Even if that would not be the case, the transformer solution is very elegant, since it is probably difficult to get to -50v or more with solid state components and no transformer alone.

 

[analag]

The Fairchild 670 is an all tube compressor, dumping solid state stages into it makes it something else. Personally I have found that the sonic signature of the CV amp plays a large role in the compression characteristic. So in essence it can't be done by short cuts and cost cuts. Call it a tube compressor but leave the "Fairchild 670" out of it.
For an idea I have a bag of remote cut off pencil tubes...that would be nice in conjunction with the sand state chips and dips, but it won't be a 660, 670 at all would it. Gimmick 670 .

 

[hop.sing]

The Fairchild 670 is an all tube compressor, dumping solid state stages into it makes it something else. Personally I have found that the sonic signature of the CV amp plays a large role in the compression characteristic. So in essence it can't be done by short cuts and cost cuts. Call it a tube compressor but leave the "Fairchild 670" out of it.
For an idea I have a bag of remote cut off pencil tubes...that would be nice in conjunction with the sand state chips and dips, but it won't be a 660, 670 at all would it. Gimmick 670 .

I have no experience with the fairchild tube sidechain amp, but I always thought of it as a powerful clean amp that of course distorts ugly when you misbias the 12ax7s for dc threshhold, but you can mimic that behaviour with transistors, as done in the EAR. But it could be that I underestimate the role of tubes in there. Imo the way of biasing the cathodes is the key to those specific compression characteristics I like so much.

 

[abbey road d enfer]

I think with the way the cathode is set up, it is necessary, since the grid voltage follows not only the sidechain voltage but also the cathode voltage,

Where do you see that?

 

[hop.sing]

I think with the way the cathode is set up, it is necessary, since the grid voltage follows not only the sidechain voltage but also the cathode voltage,

Where do you see that?

In reply #26 I explained how cathode voltage changes with current flow through the tubes. And the grid voltage follows accordingly.
I observed that while wondering how the grid voltage went from 0v to something like -20v with just a few volts of ac on the transformer primary.

 

[abbey road d enfer]

I think with the way the cathode is set up, it is necessary, since the grid voltage follows not only the sidechain voltage but also the cathode voltage,

Where do you see that?

In reply #26 I explained how cathode voltage changes with current flow through the tubes. And the grid voltage follows accordingly.
I observed that while wondering how the grid voltage went from 0v to something like -20v with just a few volts of ac on the transformer primary.

I think you got it wrong. It's the cathode voltage that follows the grid voltage, like any other triode. A few volts of AC on the transformer primary is a quite high audio voltage, which in turn is amplified by the side chain and producs a negative voltage at the rectifier output.

 

[hop.sing]

I think you got it wrong. It's the cathode voltage that follows the grid voltage, like any other triode. A few volts of AC on the transformer primary is a quite high audio voltage, which in turn is amplified by the side chain and producs a negative voltage at the rectifier output.

A few volts on the primary of the sidechaintransformer, that is what I meant.
Maybe I am not able to explain myself ,
but if you ever have a chance to measure that circuit, measure cathode voltage and grid voltage during gain reduction and you will understand what I mean 
Tobias

 

[abbey road d enfer]

I think you got it wrong. It's the cathode voltage that follows the grid voltage, like any other triode. A few volts of AC on the transformer primary is a quite high audio voltage, which in turn is amplified by the side chain and producs a negative voltage at the rectifier output.

A few volts on the primary of the sidechaintransformer, that is what I meant.
Maybe I am not able to explain myself ,
but if you ever have a chance to measure that circuit, measure cathode voltage and grid voltage during gain reduction and you will understand what I mean 
Tobias

Unfortunately, it's very unlikely someone would lend me his $30k pice for satisfying my curiosity.
But I can assure you that it's the cathode that follows the grid, not the contrary - unless the input stage was bootstrapped, which it isn't.

 

[hop.sing]

Unfortunately, it's very unlikely someone would lend me his $30k pice for satisfying my curiosity.
But I can assure you that it's the cathode that follows the grid, not the contrary - unless the input stage was bootstrapped, which it isn't.

Well, ok, its interdependent, but as I see it, it is the negative bias voltage on the cathodes, that forces them more into negative voltage the less current is flowing through the tubes. The negative gridvoltage during GR is of course responsible for less currentflow, but the cathodes are tied into negative voltage through their bias network and the way I see and (as long as I did interpret it right) measured it, the gridvoltage followed the cathodevoltage.
But as usual, I might be wrong.