Preamp recommendations

Den mandag den 30. juni 2014 01.49.03 UTC+2 skrev Dave M:

?

though probably not cheap

-Lasse

No kits. This is a design group. Build something, even if it's wrong.

You left out the IP3 (3rd order intercept point) specification.

1dB gain variation over 14 octaves of bandwidth is not a trivial exercise. Do you really need it that flat? Is your source and load impedance constant over the same frequency range? Since NF is a concern, I'll guess(tm) that you're trying to design a preamp for a radio scanner. With an upper limit of 1.5Ghz, and knowing that it has a fairly deaf receiver, I would guess an Icom PCR-1000 receiver. How am I doing? Incidentally, if it is a scanner pre-amp, you don't need 0-10dBm of power output, but you might get it for free with a suitably high IP3 spec.

The rule of thumb is: bandwidth, gain, low noise.... pick any two. You can't have all of them simultaneously.

With 1.5Ghz of bandwidth, you're going to have plenty of noise to deal with at the input (at room temperature). -174 dBm/Hz + 10log(1.5*10^9) = -174 + 91.8 = -82dBm of noise at the input. With a 4dB noise figure, that's -78dBm of noise power that your receiver will now need to handle. The noise will have some effect on the sensitivity (because the IF bandwidth will limit the amount of noise seen by the detector), but you will have to deal with mixing between all this noise, and undesired off-frequency signals to produce intermodulation products in the first mixer.

Why? Are you going to manufacture this preamp?

Instead of a list of random spec, how about a clue as to what you're trying to accomplish, what you have to work with (test equipment), where you stand in design abilities, etc.

Incidentally, if it is a receiver preamp, you should build one using MMIC from MCL. They're very easy

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Oops, I got interrupted and hit send to quickly:

Incidentally, if it is a receiver preamp, you should build one using MMIC from MCL. They're very easy to build and fairly stable. Using one as an receiver preamp will demonstrate all the problems I suggested.

However, if you just want a module: It doesn't quite make it to 1.5Ghz, but 0.1 to 1GHz might work for your unspecified application. It also has too much gain, but that can be handled with attenuators. Flat with 1dB over the frequency range.

3.0dB NF. $90/ea.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

exercise"

Makes high end audio look like a piec of cake doesn't it ? Even the REAL high end.

I've never done high or low end audio, so I wouldn't know. I guess pro audio is easier because one gets to rewrite the specs, units of measure, and test procedures as one goes along.

In RF, the problem is not getting the amplifier gain flat. It's calibrating the test equipment accurately enough to measure the gain with any degree of accuracy. Also, the source and load impedances of the connected test equipment is important. The amp might be flat, but not when connected to an antenna or non-constant impedance load.

Then, I get to make it play over the specified temperature range. The only way I've been able to make it work is with a "golden" amplifier. One unit, usually a prototype would be characterized to maximum available precision. When testing for gain flatness, the absolute gain would not be measured. Instead the differential gain between the "golden" amp and the UIT (unit under test) was measured. If the sweeps coincide, we have a miracle and a passing amplifier. If not, let the tweaking begin.

One can get really wide bandwidth and superior gain flatness with a distributed amplifier. For example, 50MHz to 50GHz (10 octaves) with

9dB gain and a claimed 1.5dB flatness (which doesn't agree with their own curves): The NF is lousy, but that's to be expected above 1GHz.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

This isn't what you're looking for, but it's in the ballpark if you don 't mind improving a design:

Clifford Heath.

Now THAT's a cascode amplifier. It is fed by a source follower and buffed by an emitter follower, but that is it.

Reason for this post is that someone called something a cascode which was not recently. (not on the group really, I think it was in a link)

I don't have equipment to measure the noise performance. If someone would care to educate me on how to do (and to interpret) an LTSpice noise sim, I'm interested.

It's designed to work with a large input impedance. I believe the high-end rolloff is due to the 20pF input capacitance. A smaller FET or PHEMT would be needed to improve it.

The 50R drain resister is not needed nor fitted - it was part of a drain bootstrap circuit per Phil Hobbs suggestions. That might improve the accuracy of the step response) by bootstrapping away Cdg), but cannot increase the high-end rolloff (which is what I initially tried it for)

Clifford Heath.