RF is horrrible

I think we're slightly at cross-purposes here. Yes, I'm talking about the type that come completely sealed in their own individual steel castings. The kind of standard you find inside pro-RF test equipment, for example the product ranges of companies such as Rhode & Schwartz.

Pity you didn't ask earlier, Phil. I've just re-fitted a perfect example into its caddy and thence into the chassis of another one of my high-end test gear recommissions. There's an off-chance I might have caught one in a photo I took to trace the connections back to its slot once the thing was re-assembled. I'll have a look-see and post a pic tomorrow AM if I can find one clear enough to be worthy of your consideration.

Those aren't MMICs, though, which is what we're discussing.

Sure, the expensive connectorized ones can be a lot better over bandwidths of an octave or so, and really expensive ones can do a couple of decades of bandwidth with reasonable flatness. JL is talking about DC-microwaves in a single band, though. Different kettle of fish.

Cheers

Phil Hobbs

Ah! Thanks for the clarification, Phil. I admit to jumping in with both feet without reading the thread first either. Soz!

Of course you don't, which is part of the problem.

And - like all your other efforts in that direction - failing miserably. There's little hope that you'll ever learn enough to realise this.

A little unkind, Bill; a little unkind.

We used 3dB 50 Ohm pads between MMICs and RF filters to reduce the mismatch, at Microdyne.

Yup. Helps make the response more predictable, for sure. Plus you can add the occasional small cap to dork the frequency response tilt.

Cheers

Phil Hobbs

(and one in the source, along with a string of beads and a parallel inductor/resistor combo)

Just finished the preliminary wringout.

It works better with no bead between the source and the bootstrap cap, and with only a 5-ohm bead in the gate (BLM15BB050SN1), instead of a

It did oscillate fiercely without the bead in series with the gate, but it made a huge difference in the e_N*C noise when using a 1000-pF MPPC array!

Despite the mass extinction of GHz PNPs, there sure are a lot of great parts out there these days.

Cheers

Phil Hobbs

I'm planning a diff pair of SAV541s. So I'll connect the sources hard and put small beads in the gates.

I owe you a beer or two.

As if Cursitor Doom were famous for his acts of kindness - his enthusiasm seems to be more for posting unsubstantiated slanders.

PKB, Slow man.

Cite?

Bill Sloman wrote in news:2c550dc0-7099-446d- snipped-for-privacy@googlegroups.com:

Your last post to me where you make yet more of your retarded, baseless presumptions.

Should work. Watch out for the crappy drain impedance--some years ago I tried making a Schmitt trigger out of ATF38143s, and it didn't work worth beans.

Nah, you're the one who got me using pHEMTs at baseband. Not that I'll turn down a few beers, you understand. ;)

Cheers

Phil Hobbs

Few? The offer was for one or two.

I was thinking that if the gate drive to my phemt diff pair is symmetric, namely from a very fast differential ECL source, why not bypass the source node to ground? In theory, the common source node won't move much when the tail current switches from one fet to the other.

I guess I'll include a cap or two on the layout and try it.

I'm doing a lot of PCB layout lately, some on a 1 mil grid. It's too hard to explain this fast stuff to my layout guys. I don't even know what to do until I push things around myself.

Plateau bargaining. ;)

You'll lose balance on the output though--with a high-Z tail source, the drain currents stay in balance even during transients. A small cap might speed everybody up a bit, of course.

Cheers

Phil Hobbs

I was thinking that if it oscillates, adding the caps is another possible fix.

The drains will be 50r resistors to a programmable V+, and SMA connectors, and the tail current is programmable too. Differential outputs.

You advertise your problems rather more obviously than you think.

You have accused me of inadequate reading comprehension, when the problem is that my reading comprehension is rather more acute than yours, letting me pick up implications that you didn't realise that you had posted.

But if you were as smart as you claim, Bill, you would have allowed for that, surely? Or could it be you're not that smart in truth? Best get out on the golf course, mate. Tick-tock, tick-tock....