VCA82x series of variable gain amps, experience?

Too vague ;-) How many "few dB", how much slew rate and how much swing?... numbers please.

...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     |

3dB control range required, 4-5dB range would be nice. Granularity 1dB is ok.

Slew rate should be 1500V/usec or more. 3rd order at 5Vpp -75dBc would be nice and should not be much worse then -65dBc (that's an issue with the VCA). The amp will be fed a fixed positive voltage to the IN- so we have a symmetrical output swing of +/-2.5Vpp with a unipolar (positive) input. The load for that case would be 100ohms. We can also do half that swing into 50ohms, giving up source termination.

I am also looking into RF muxes like this ...

... but paltry datasheet, misses the DC signal range and such. Cajoling that out of AD (they bought them) is IME time-consuming.

Since I can make sure the intermediate signal swing is 0V to 2V (or possibly even away from 0V) I can also use 4:1 muxes or maybe even RF muxes. But those either have Rdson in the tens of ohms or a lot of capacitance. Might work though, I am looking at some right now. That would allow me to use fast amps where I already know the performance.

[...]

--
Regards, Joerg 

http://www.analogconsultants.com/

Any multiplying D-to-A's with dB steps? I've done that on custom RF chips but don't know if they exist off-the-shelf. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     |

I've also done that but on the board level. Then the client would be forced to a uC interface though. Ideally I should provide jumpers (or a potmeter) and simple interfacing. The latter would be external logic lines or an external control voltage.

RF muxes are a pain. No DC specs yet many of them bluntly state DC - xGHz on the datasheet. Then in the fine print "decoupling caps required".

4052 style muxes have marginal figures of merit (Rdson versus capacitance).

Maybe I should just have a divider with several resistors and RF FETs to ground and call it a day. It's one of those urgent pre-Christmas projects. Calendar is a bit full and will also interfere with my home brewing :-)

No kidding, so far I did Pale Ale, Koelsch, IPA, Cream Ale, Belgian Witbier, Stout, Irish Red and Autumn Amber. Some of them more than once.

--
Regards, Joerg 

http://www.analogconsultants.com/

JFETs are particularly well suited to small gain range adjustments like this. Though they may not be suitable/desirable for other reasons (low impedance, negative bias, inconsistent Vpo, >= 2nd order distortion, capacitance).

You may've thought about this already, and just didn't mention it; so, just to put it out here...

Or, you could always make your own OTA, but I'm guessing it would be as bad or worse distortion (including 2nd order because of device imbalance), and all the other fiddly things that an IC solves handily...

...On that note, you could make a MOSFET OTA, which operates in the quadratic range, so has lower distortion (and less adjustable gain range within that region) than with exponential devices (subth. MOS, BJT). The signal could be split between a complementary pair of stages, and their outputs combined, to cancel 2nd order products, presumably/hopefully leaving less distortion than the same design with exponential devices. (But again, fiddly discretes, and double of them, no less...)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: http://seventransistorlabs.com

I just decided to do it the usual way, redneck style. Series resistor from one stage to next. Then two resistors in R2R fashion to ground but via small MOSFETs. These get switched as needed for gain adjust. NXP makes some that are only very few pF in Cds but well under 10ohms Rdson.

More importantly, the new brew supplies just arrived! Including stuff for a "Subduction Cascadian Dark Ale". Now I'll have to find the time to brew it.

--
Regards, Joerg 

http://www.analogconsultants.com/

A high frequency MOSFET OTA or bjt/mos hybrid OTA still sounds pretty cool, though. It's like a thing they should make...

Yeah, that's a real pet peeve. "Any DC voltage you like, as long as it's 0." And then you're lucky if the graphs in the data sheet go below 100 MHz.

They just started shipping the MEMS switches they announced a couple of years ago:

These offer true DC support to +/- 6 volts, but they don't appear to be any cheaper than a true-DC RF switch (PE42020). Being MEMS- based there's a list of caveats as long as your arm.

-- john, KE5FX

For a better FOM, look at the nmos-only ones intended for digital muxing, like the FST3125 and similar. With creative power supply rails you can often accomodate the swing you need. I would mostly use them to short nodes to ground, or to steer current from a high impedance output (e.g. collector of a transistor) to taps on an attenuator. If you are sneaky you might even be able to use the switch nmos as a cascode. I would not suggest you use them as a switch in a low impedance signal path if you need low distortion, as the Ron will be modulated with the drain and source voltage.

The FST3125 became popular with radio amateurs in the "H-mode mixer" circuit. Then they went and tested a bunch of other switches, so if you google "H-mode mixer" you will find a lot of similar parts too. e.g.:

I have looked at logic switches but their FOM isn't better than analog ones from NXP where I don't need creative supplies. They must have a secret sauce in their process.

--
Regards, Joerg 

http://www.analogconsultants.com/