Interesting new pHEMTs

So Broadcom and CEL axed all their 3-GHz class pHEMTs, but CEL has introduced some interesting new 12 GHz ones, e.g. .

That one has a typical 0.3 dB noise figure at 12 GHz! I'll have to get a few and measure their 1/f noise. The SKY65050 (which you can still get) has about a 30 MHz 1/f corner, but is otherwise very nice. So hopefully these ones don't have 100 MHz corners.

As with most pHEMTs, it has a horribly low drain impedance of about 200 ohms, and like most SS microwave parts, it has a low V_DSmax (4V).

It would make a dynamite cascode with a BFP640 though, if I can keep it from oscillating. The 3-GHz pHEMTs are very well behaved by themselves.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs
Loading thread data ...

Nice, but too low current for a lot of uses, like laser drivers. We mourn the passing of ATF-50189.

0.3 dB NF sounds physically impossible somehow. That's a resistor at 21 Kelvin.
--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

I ordered 25 of the cheaper SOT343 ones, the CE3514M4-C2, 0.42 dB NF. I'll try cascoding them and see how it works. Could potentially make a fairly fantastic scope probe, for one thing--6X faster than an ATF34143 and a bit cheaper even.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Some of the Tek and LeCroy fet probes cost insane amounts, like $6K or something.

We'd like to get into the scope probe business (especially differential) but we've been too busy to get serious about it.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

I don't know if I dare trying to make the upper stage a bootstrap.

Probably so. ;)

The first time I tried prototyping one, it oscillated at about 12 GHz, and that was with an ATF38143. That was the "manual wavemeter" one, where the amplitude went up and down as I moved my hand towards it. The envelope went through a full cycle in about half an inch, which is how I know it was 12 GHz. One of those nice Murata BLM18BB ferrite beads in the base cleaned it right up. Trashed the bandwidth of course, but I only needed 100 MHz for that job--the reason for the 40 GHz BJT was its insane Early voltage, which let me run it at a huge DC gain.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

The fast probes are probably wirebond hybrids or full custom ICs.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Parenthetically, I benchmarked a bunch of beads a while back including that one, or at least one in the same family (BLM18BB121SN1D). It's a nice part but it falls short of the all-time champ (HZ0603C651R-10):

formatting link

The current rating on the HZ part is somewhat lower, but it's still good for powering MMICs in the +15 dBm range.

-- john, KE5FX

Reply to
John Miles, KE5FX

a

Nice, thanks. The one I like for stabilizing those 40-GHz BJTs is the 10-oh m one, BLM18BB100.

Cheers

Phil Hobbs

Reply to
pcdhobbs

Cool, but what do they do under bias?

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

I have some beads that would resonate acoustically after having seen DC bias. Applying just the right amount of inverse DC would remove the resonance again.

Jeroen Belleman

Reply to
Jeroen Belleman

I once cracked a toroid into about twelve bits, when I was testing it with a pulse saturation circuit, and hit a mechanical resonance of the core. Was going along normal, then "tick" and the waveform changed (much more current all of a sudden) -- poked at the coil and the core-bits fell out of the winding!

Do you mean not so much "resonate" as "produce sound at all"? That would make sense if you'd magnetized the core (or some of it), so that its magnetostriction becomes biased, and AC causes a ~proportionate change in dimension (rather than a squared change).

I've also had a case where, a laminated iron toroid runs quietly under normal conditions, but once it's been subjected to hard saturation, it's always noisy. The noise varies with time, like it's flux-walking, all by itself (varying with a long time constant, thousands of cycles), even though the circuit is AC coupled (with a big fat film capacitor).

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

Dunno. I never run beads past 1/3 of their rated current, and even that's probably pushing it.

I should put that fixture back on the 8753 and run a couple hundred mA through it...

-- john, KE5FX

Reply to
John Miles, KE5FX

The resonances were at several hundred kHz and very narrow. There were three modes, present only after having passed a fair amount of DC in either direction and could be removed by passing just the right amount of DC in the opposite direction. That certainly looks like it was caused by residual magnetization. I could see no trace of it in a de-magnetized core: 0dBm is probably not enough excitation.

I concluded it must be acoustic because touching the core with a plastic rod damped it.

I recorded the observation in , near the end.

Jeroen Belleman

Reply to
Jeroen Belleman

Maybe you can make a fluxgate magnetometer from a ferrite bead.

--
John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
 Click to see the full signature
Reply to
John Larkin

It'd be a terrible performer, because ferrites are brittle; any microcrack would throw the whole calibration curve off.

Hall devices, too, are excessively susceptible to mechanical stress. A good reluctor, though, can work in full battle conditions. Anything that has ductility, is a lot more fault-tolerant than hard materials. That's why earthquake-safe concrete has lots of low-ksi steel in it.

Reply to
whit3rd

A good magnetometer operates in zero-flux mode. The nonlinear magnetic element is a null field detector. The measured field is cancelled by a DC feedback field, which could be a PCB trace or a nearby air-core inductor or something

Or just buy the TI fluxgate chip for around $4.

--
John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
 Click to see the full signature
Reply to
John Larkin

Flux gates work by flopping back and forth between saturation in two directions.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Sure. But the 2nd harmonic thing is usually the null detector, with a feedback loop and field feedback coil around it.

--
John Larkin   Highland Technology, Inc   trk 

jlarkin att highlandtechnology dott com 
 Click to see the full signature
Reply to
John Larkin

Well, zero _average_ magnetization, sure. But in this case it's like the guy who drowned in a river whose average depth was two inches. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Good plan. It's rarely seen from mfgs, with Steward/Laird being the standouts with a full catalog of plots. Fair-Rite has them here and there. Presumably, similar chips (size, impedance, frequency) will have similar performance from other mfgs. But who knows, maybe everyone else cuts their ferrites with sand? (Ironically, sand would make them much better, because air gap.)

Typical values are saturation more like 1/10th of rating, give or take what impedance the part is, and how much saturation you can handle. That is, a -30% impedance threshold might be 0.1, or even 0.05 times rated current. But you might still be fine with -90%.

Oh and, don't forget to de-embed the inductors you use to bias the thing. :) Proper power inductors, around 10uH or more, should be fine. Consider stacking a few (and of progressively smaller values) in series to keep capacitance low.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
 Click to see the full signature
Reply to
Tim Williams

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.