Increasing noise performance by replacing mechanical pots

I really like the idea of using VCAs, feel kinda stupid that i didn't get there myself, thanks. I'll probably use one of those. Thanks a lot to Nico for that suggestion!

What's your problem? Ever considered that not everyone has english as first language? I ask here because I don't know better, not to show off.

Adrian

Wow, I never heard of them. They sell matched transistors too! Are they on the same die?

Thanks

George H.

Yep, same die. I did some VCA designs for a start-up in San Jose in the '70's that ultimately joined up with THAT. I need to make a trip to my off-site archive facility to remember the company name :-) ...Jim Thompson

Yes. They aren't as good as the MAT14, but they're better than the CA3046.

Cheers

Phil Hobbs

Nor as expensive.

For matching? I don't know. Those large geometry devices match pretty good. (Although I'll have to look... the CA3046 may have been laid out for tightness rather than match.)

...Jim Thompson

"Phil Hobbs"

.... Phil

MAT14s are only a few bucks--when they reintroduced the MAT04 and renamed it, they also dumped a whole lot of the production tests that made it so expensive. They *gutted* the datasheet, anyway.

The THAT devices seem to have a large Rbb' and relatively poor beta linearity compared with the ADI ones. They don't make nearly as good laser noise cancellers, anyway, as we both have some reason to remember.

The CA3046 devices have low beta and poor beta linearity, and low beta at low current usually means a lot of 1/f noise.

Cheers

Phil Hobbs

It is a LOT worse than that.

.... Phil

Did you happen to design their differential receivers with the nifty tricks for getting low CMRR out to high audio frequencies?

Those things are gorgeous.

His problem is he is a bodily orifice.

TI has ap notes on how to roll your own with just a bit of glue. I guess it depends on your budget. My recollection is the ADI parts are full blown Gilbert multipliers, while TI app notes do the VCA with OTAs. Best bet is to check both websites. Obviously you don't need four quadrants.

Formerly Elantec, but that is Milipitas. Probably not your design.

SSM was the other company. God was their stuff shit cubed. However, PMI bought them. Those were Derek Bowers designs. Eh, I said too much already.

Glue is for sniffing, not for smoking.

chips

=A0Are

t -

Hi Phil, which are the cheaper ones mat14 or mat04's?

In the hundreds of pieces range I see mat14's for $7-$10 each.

The mat04's seeem even more outrageous. Except for 20 pieces at Quest which I can get for ~$6 each.

George H.

chips

They discontinued the MAT04s for a year, then brought them back in a de-spec'd version as the MAT14. I'm pretty sure they're the same part, but I haven't done an A-B test to make sure. (Some enthusiast might have thrown away the masks when the part was discontinued.)

Cheers

Phil Hobbs

Tooling gets scrapped periodically. Stuff happens. It isn't unheard of that (speaking in a Mafioso voice) "something unfortunate" happens to a mask, allowing a different project to take its slot in the queue, enabling bonus money or some other goodness at the expense of screwing the company and another engineering team.

Note the "THAT corporation" parts are on DI, while the ADI parts are bulk. DI should be cheap these days, but my reality distortion field is not as good as the late Steve Jobs.

Engineering or production costs? An appnote is exactly what it is: an appnote, not a production ready design!

Possible solution: a digital pot - which can add "zipper" noise and spikes when being changed; else that gives reasonable isolation.