chopper transistors

Can't you use a mosfet, or an analog switch?

John

JFETs are quite popular for that as well, on account of their low capacitances.

Tim

-- Deep Friar: a very philosophical monk. Website:

John,

relevant point on Mosfets, however my design is tailored around having Vbe's closely matched. I believe this is more difficult to do with Mosfets, although not to sure about Jfets.

Tim,

Thanks for that, will be giving them a try as they're readily available from RS.

Matching Rds(on) should not be too hard (at same drive level).

Zetex makes some rather nice low sat dual transistors..try (in this order): ZXT12N50DXTA, ZXT09N50DE6TA, and ZDT1053TA. As far as singles go, the ZTX1048A, ZXT1049A, ZTX1051A, or FZT1053ATA. You will notice that i purposely picked high beta transistors; for lower on saturation voltage, run them inverted. Zetex seems to make a goodly bunch of high inverse beta transistors - which makes them better as choppers.

JFETS are good as low level choppers. Bipolars suck, all sorts of offset problems to design out.

I've never tried MOSFETS. Should be as good as / better than JFETS, AFAICS.

JFETs rock in this domain. The only MOSFET that I like in this area was a quad, the SD5400. But it has changed ownership and to me it's now too much of a boutique part. Digikey doesn't have it, expensive.

CMOS analog switches work. You just have to be cautious with regard to charge injection. ...Jim Thompson

On a sunny day (Fri, 22 Oct 2010 13:03:00 -0700) it happened Jim Thompson wrote in :

Conrad here no longer seems to carry the 74HC4053, they do have the 74HC4051. What is happening? I find 74HC4053 one of the most useful chips ever.

For us non-chip guys the main problem in the last 2-3 years is a whole different one: You have to be able to obtain the parts without a 40+ week leadtime. Seriously, that is driving a lot of design decisions these days.

Just got off a conf call with an IC design house and my client. Man, you guys have it easy. "Can we change the Rdson for the n-channel device over yonder to about 20% or less?" ... "Yep, sure, consider it done". All they do is type another number into the channel width attribute field. Now try that in the discrete world in the year 2010. Type in your part number ... out-of-stock ... type in another ... 22wk leadtime with no guarantees ...

Stick to ancient parts ;-) ...Jim Thompson

I usually do anyhow :-)

All semiconductor parts are hard to get these days. We have lead times from 3 months to over a year. Hopefully, won't be like the late 90s when resistors and capacitors were hard to get.

As for CMOS switches, I've found Fairchild's FSA66 and FSA266 to be super duper. Somewhere around 0.05pc of charge injection.

And status = obsolete!

Most of those logic-family analog switches have huge charge injections.

John

Fairchild calls this a "datasheet"!

Most of their analog switches are documented about this well.

John

That is amazing, the best I found for my application was 0.5pC - the DG636 (and some others in the same ballpark).

And then there is the LTC6943/LTC1043, as featured in AoE and around half of all LTC application notes, I guess they're pretty proud of it.

Anyone know of any significantly better (non-obselete) ones?

Of course. It's "eco-green", so obviously not intended for use by anyone with any technical competence.

Thanks! Rich