Clean, modern and inexpensive VCA?

[Rob Flinn]

The noise performance of these chips is heavily dependent on the Rin and Rout resistors.

Using 30k for Rin/Rout (the largest recommended value for the 2164) gives the worst noise performance, while using 7k5 reduces output noise by up to 8dB in Class AB mode or up to 12dB in Class A mode.

Further still, those resistor values need to be divided by the number of VCAs if you're using multiple sections in parallel, in order to actually realise the noise improvements of that arrangement.

Unfortunately, the measurement thread linked above doesn't seem to mention which value of resistors were chosen for Rin/Rout in those tests.

The attached images are cropped from the SSI2161 and SSI2164 datasheets.

True, but I'm wondering if the SSI2164 data sheet necessarily applies to the Cool Audio V2164 that was the VCA actually used to get these figures.

 

[JohnRoberts]

I don't know if this applies but back when I was still competing in the mixer market last century, Behringer copied a new small mixer of mine. Their trade show literature copied my specifications verbatim.

Trust but verify...

JR

 

[cyrano]

I'm glad this forum exists. There's so much knowledge here. And so little in the outside world. It seems there aren't very many manufacturers out there you can trust. At least for specs...

I've learned a lot about VCAs in this thread. I've got two Behringer Cybermix boxes to play with. One has seemingly original dbx chips. The other one has Coolsound chips. I think. I'll try to compare those, but it'll take a while as I'm currently moving.

 

[Lee_M]

True, but I'm wondering if the SSI2164 data sheet necessarily applies to the Cool Audio V2164 that was the VCA actually used to get these figures.

Probably not exactly, but the value of those resistors should affect the V2164 in the same way that they affect the SSI2164, so I'd still expect to see an improvement in noise if lower values were used.

 

[abbey road d enfer]

Probably not exactly, but the value of those resistors should affect the V2164 in the same way that they affect the SSI2164, so I'd still expect to see an improvement in noise if lower values were used.

Do you mean the input resistor? The improvement with reducing the value comes at the detriment of distortion. Indeed, using 4 cells in parallels allow using a 5k resistor instead of 20k, so noise would decrease by 6dB for identical distortion.
CoolAudio claim having improving the max operating range from 0.5 to 1mA, which is a 6dB improvement over the SSM part.

 

[Lee_M]

Do you mean the input resistor? The improvement with reducing the value comes at the detriment of distortion. Indeed, using 4 cells in parallels allow using a 5k resistor instead of 20k, so noise would decrease by 6dB for identical distortion.

That's right, the input resistor, as well as the "Rout" feedback resistor on the I/V converter.

CoolAudio claim having improving the max operating range from 0.5 to 1mA, which is a 6dB improvement over the SSM part.

Yeah, SSI have the same improvement listed on their datasheet. It's interesting that Cool Audio decided to improve that aspect, but neglected to fix the "missing rail" problem.

 

[porkyc]

With regard to the comparison table.
I used a Neutrik A2D test set with the noise figures verified by a Sennheiser UPM550.
I had the VCA daughter boards on a console module powered at ±15v on a bench in a shed/workshop. All done on the same evening, the same way.
I used the SSI 4-wide VCA schematic as in their Datasheet/Application notes, but to be fair, I used a CoolAudio2164 because it was DIL package.
This table was posted on another thread that started the same way; are these 2164s any use?
I also tried to use the That 218x specified test conditions to give a fair comparison.
I was not looking for absolute values, mine was a real-world application; would this VCA chip be of any use to replace black dBx202s in a console that was going to need 48 of them.
Thus my arbiter was VFM and be better than the existing units and be closer to the That product.
Since the answer to that arbiter was No, and I felt it unlikely that the SSI chip would be wildly superior, (How would I know, sweeping generalisation and totally unfair!), I did not bother to mess around with I/O values. The only value I changed was the Class A/AB determining resistor in an effort to get it vaguely better.
So the fact that the CoolAudio2164 (in particular, noise figures) did not meet the SSI predicted specs in the SSI schematic did not worry me, because the distortion was completely unacceptable.
I think I threw everything to do with the 2164 stuff out afterwards.....stick with the That product. Might have the gerbers of a 202 format DIL somewhere, though.
Hope this clarifies things.
PC

 

[cyrano]

Some interesting info here:

https://www.njohnson.co.uk/index.php?menu=2&submenu=2&subsubmenu=3

 

[neil.johnson]

Some interesting info here:

https://www.njohnson.co.uk/index.php?menu=2&submenu=2&subsubmenu=3

Thanks

Neil

 

[cyrano]

Silly me. Should've known Neil was a member here

 

[scott2000]

pedalpcb.com has the 4305 to 4301 adapters back in stock...

https://www.pedalpcb.com/product/ic-that4301p20-u/