OT: Ping Jim Weather report

Kinda hot out there eh?

Cheers

Reply to
Martine Riddle
Loading thread data ...

Not really. It's only 107°F at the airport, 121° on my north patio (indirectly heated from reflections off the concrete).

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

About the same here in Mesa, a cold snap!

Luhan

Reply to
Luhan

It's beastly here, 66F at 5 PM. I'm at work, downtown, playing with step-recovery diodes. I'm sure it's a lot cooler at home, about 2 miles away, in the Alemany Gap, which is a break in the coast hills, sort of a wind tunnel between the ocean and the bay.

None of these srd's seem to be as fast as advertised... maybe I'm doing something wrong.

John

Reply to
John Larkin

Isn't it dI/dt that matters, after recovering from high I.

--
 Thanks,
    - Win
Reply to
Winfield Hill

I'm playing with regular SRDs, Metelics and M-Pulse, the small-signal kind. These are normally forward-biased with lowish DC currents, 10 mA ballpark, then slammed reverse at 100 mA or so. The surface-mount packages seem to add a lot more capacitance than the datasheets specify... maybe they are spec'ing bare chips. I have some flea-market parts that are pretty good, but I can't go into production on those!

I tried using pulsed forward bias, not DC, to see if that would speed things up. HP did this in their classic 1430 sampling head, about 4 ns of forward bias just before the snap, as I recall. With these diodes (better control of doping profiles these days?) it doesn't seem to matter.

John

Reply to
John Larkin

Hello John,

106F right now, only about 100 miles east from S.F. Going to dash into the pool in a half hour since it is only 100ft from the office.

Somehow I always get the feeling that SRDs are a niche product. No chance to avalanche a really hot transistor?

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

On a good summer day, it's chilly in SF and the gradient is about 1 degree F per mile, going north, south, or east. As in 60 in SF, 80 in Berkeley, 100 in Walnut Creek.

Nope, too slow. I want to build a fast 2-diode signal sampler, and I need complementary impulses, + and - 3 volts maybe, 100 ps wide and preferably a lot less.

One could start with a fast logic level - 35 ps edges are affordable - differentiate, and amplify with some fast mmics, but that's a lot of hassle. A fast mosfet or a slow gaasfet driving a single SRD should be able to do it.

It's looking like I have drive problems. The trick is to get lots current into the SRD fast, but the integrated current is burning up the stored charge in the diode, and the faster the diode, the less charge it stores. Lots of physical and numerical tradeoffs.

They are indeed a niche business. Practically nobody stocks packaged parts... they have a few wafers around, and package to order. So it's hard to talk them out of samples, especially the third or fourth time around. Only MaCom has distributor stock parts, in SOT-23, but they're damned slow.

Somebody (was it you?) told me that some varicaps snap nicely. I have the spiffy Skyworks kit full of varicaps, so maybe I'll try a few.

John

Reply to
John Larkin

Hello John,

I wonder if this could be done with a couple hot rod transistors a la BFP540 plus a wee strip line or coax to set the pulse width. They make even faster ones that could happily oscillate above 10GHz but then the max Uce won't be enough for a 3V swing.

That can be a problem for a small run. Also for a big run unless you can be sure the manufacturer will have enough defense contracts and be around for a while.

Not me but IIRC it was Jorgen Lund-Nielsen and he meant tuner switching diodes. The ones that are sold a lot by Philips and Infineon. With the latter the parts procurement for a prototype series can be a nightmare though. It probably wouldn't be quite fast enough for your case anyway.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

I've never had luck getting "fast" bipolars to make fast edges, even the 45 GHz SiGe parts. Don't exactly know why.

I just tried a Skyworks varicap, SMV1405-079. It snaps like gangbusters. The best one in their kit has a V0 capacitance of 2.1 pF, so it's slow, risetime about 270 ps, but the concept looks good. I think the hyperabrupt varactor doping profile is about the same as an SRD. So all I've got to do now is locate some lower capacitance parts that use a similar process.

But I'd sure like to meet the guy who invented the SC-79 package. At a scenic overlook in, say, the Grand Canyon. At night.

John

Reply to
John Larkin

FWIW, there's always Bill's tip from May, Agilent's app note about using PIN diodes as SRDs: AN 1054, Low Cost Frequency Multipliers Using Surface Mount PIN Diodes, 5966-4998E.pdf.

"PIN switching diodes with low values of transition time can multiply frequencies up to C-band similar to step-recovery diodes (SRD)." ... "None of the multipliers was as efficient as a multiplier made with an expensive, high-quality SRD, but then these examples were fabricated using inexpensive PIN diodes."

10dB loss for a 600 MHz-to-1.8 GHz tripler, 15dB loss going from 1.8 to 5.4 GHz, for whatever that tells you about the edge rate.

James

Reply to
dagmargoodboat

Hello John,

I have, but it was not SiGe and eons ago at an RF institute. IIRC it was the ones they made themselves and were in the old style "satellite" package where you had to drill a large hole into the board to seat it correctly. The layout of the board was really critical. A fraction of a pF too much or an input mismatch and performance was gone.

When I have time (and Infineon figures out to get samples to me) I'll try that with a BFP740F. That thing really looks like a hot rod. Very low capacitances. An alternative would be the BFP640F which Digikey has in stock.

I'd really be surprised if you could get a Varicap under 1pF that isn't a boutique part. I guess Win would know.

Now, now...

The Golden Gate would be more en vogue, and there are sharks down there.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

For a page of PDFs paste 5966-4998E.pdf. into google - take the 3rd hit down, ignore the XLS button and hit " view as HTML" the text is in Russian but it looks as if the PDFs are probably in English.

Reply to
ian field

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.