CA3046 VCA

[jorgen.bergfors at idg.se]

PC:chrisc at zetnet.co.uk on 98-09-09 00.42.32
To:  Jorgen Bergfors at IDGSE
cc:  
Subject: Re: CA3046 VCA 

Hi Jorgen you wrote....

>>
>>1. Terrible CV bleedthru (ca 0,5 V) and no way to trim it. I used av 
CA3086. 
>>Maybe it is better with a CA3046 but it still ought to be trimmable

>Hmmmm... I haven't had that problem with mine and neitehr has anyone 
else 
>repoted it.... in fact CV bleedthrough ought to be stopped by the output
caps...

Well, even if you have an output cap, a CV bleedthru will cause a pop if the CV 
is fed a step, like quick attack ADSR. And if you use it to modify control 
voltages you can't have an output cap.

>>2. Negative control voltages cause a DC voltage on the output

>Granted, don't use them on the Modulus system (there is no way to 
develop a
-ve 
CV....

An LFO usually swings both positive and negative. Of course yo usually offset 
it so it is all positive, but maybe you get some cool effect if you don't. It 
ought to be possible, at least.

>>3. Input is AC-coupled, which makes the VCA unsuitable for control 
voltages

>Granted, front panel say <Quad Audio VCA>

I want a general purpose VCA, that can be used for both audio and control 
voltages. I'm going to make my own PCB anyway.

>>4. Some ringing with faster op-amps

>Hmm... not been a problem with TL074's and they're as fast as I ever 
use. That

>is a solveable problem with slower op-amps though....

There was ringing with a TL072. You have to know how to look for it, to see it 
though.

>>5. No headroom above 10V p-p

>Eh?....Oh? you mean signal rather than CV.... ummmm....don't overdrive 
your 
>VCA's, isn't 10V P-P native enough for you?

If the signals are 10V p-p you definitely want some safety margin.


>>6. Too much gain

>No such thing, you can always turn it down....

If I send it full control voltage I want it to pass the signal as it is. If it 
is 10V p-p at the input it should be at the output too.

>>7. Linear control response is a problem if you want log.

>Granted, uncomitted log converter in the pipeline, never needed it 
myself, 
always considerd builting expo envelopes instead as a satisfactory 
solution......

>>
>>1. Is fixed by adding a 10k trimmer. The ends are connected to the 47k 
>>resistors that used to be the control voltage input. The input signal 
is 
>>connected to the wiper. CV bleedthru can now be adjusted to as low as 
4 mV.
>>

>Ok, granted it works, like I said, not had  a problem before....


>>2. A diode is added to the CV input. This does not mean that nothing 
happens 
>>below 0,6 V CV, as you might suspect. The VCA starts to output signal 
at 0,15
V 
>>CV. That is quite acceptable in my opinion.

>OK, I can dig it, can be added by users as a mod ATM I might integrate 
>it on the  next version of the board, don't hold your beath waiting for it, 
I've 
>got other  fish to fry ATM and relayiong will involve a new software 
package....


>>
>>3. Corrected by simply removing the cap on the input. This does not 
>> disturb the function of the VCA.
>>

>Hmmm...debateable, I thought the whole thing about Gilbert Multipliers 
>was the lack of DC coming through the signal path....

It must be DC-coupled if you want to use it for control voltages. But of course 
if the input signal has a DC offset, it might cause problems later in the 
signal chain.

>>4. Added a 22 pF cap in the feedback loop of the output op-amp. An even
>>smaller cap might be sufficent, but I had no smaller. This of course is 
>>dependent on what op-amp you use. I use a  TL072.
>>

>Hmmm will integrate onto new board if and when....once again current 
>circuits easily field modified...


>>5. Increased the input resistor to 68k
>>
>>6. Lowered the feedback resistor in the output op-amp to 56k
>

>OK... once again field fittabel if desired, will experiment sometime....

>>7. Mine is still linear but I plan on making a log version too.
>

>Yeah, but I doubt i can get four log Ip circuits onto a Eurocard. The 
>whole point of this circuit was a very clean (noise and distortion wise) 
>audio VCA. 
>There were any number of multipurpose, but rather noisy/distorty (????) 
transconductance amp designs, this was an alternative....

>chrisc 

I think it is too good to limit it's usefulness by saying it's just for audio. 
With my modifications it will be a general purpose VCA, without sacrifying any 
of it's performance for audio,

Jorgen