[time-nuts] HP5065A Step Recovery Diode, what part#?
brooke at pacific.net
Wed Aug 29 14:09:34 EDT 2018
In addition to just the SRD there are things called "comb generator modules" that consist of a raw SRD diode (not in a
package) and an input matching circuit, typically for 100 MHz, all housed in a hermetically sealed cylinder about 1/4"
in diameter. For example HP 33002, 33003 and others.
Here is a packaged diode on eBay (item # 202135267844) advertised as for a comb generator for $30, but I don't read Russian.
Using the eBay search term "(srd,step) diode" shows many Russian SRDs under $10 and words like 100 GHz.
Item 201976756657 has some specs, $4 and may be the same as item # 202135267844?
The HP 8406 lab bench comb generator uses a SRD, but I don't know what lifetime it has.
Cheng Lai, founder of Herotek came up with a way to make a comb generator module that has a wide band input. It would
be interesting to learn how he did that.
1. The extent to which you can fix or improve something will be limited by how well you understand how it works.
2. Everybody, with no exceptions, holds false beliefs.
-------- Original Message --------
> Most of my SRD stock resides on old junked PLO brick boards - that way they won't get lost, and I can possibly use
> some of the associated circuitry. I have quite a few good PLOs, and have used a number of them in various projects.
> The ones that I have from the 1960s -1980s seldom run the oscillator above about 2-3 GHz. The higher frequency types
> usually have a high power oscillator around 1-2 GHz, and tap off power to - you guessed it - an SRD multiplier,
> followed by a filter. For higher multiplication or power, there may be a class-C power amp stage in between. The
> highest fundamental PLO I have is around 3 GHz. Everything above that uses SRD multipliers.
> So, I think the SRDs used in the sampler/PLL sections won't quite reach the Rb frequency, since they only need to be
> effective to around 1-2 GHz, maybe 3. But, you never know - it may be worth a try.
> The old PLO bricks are very popular I think, because they're useful, versatile, and easy to work on. They can usually
> be adapted for lots of situations and frequencies outside the nameplate. For example, about ten years ago I built a
> portable Rb standard unit, and I wanted to have decade frequencies as high as I could go. I had in stock a 9.6 GHz or
> so PLO that ran from 80 MHz reference. I took it apart and figured out the scheme. As I recall, it was an oscillator a
> little over 1 GHz, driving a power amp driving an SRD for around 8 or 9 times multiplication, then the filter. I
> modified it to run right at 1 GHz, added a tap for that output, retuned the filter for 10 GHz, and tweaked the SRD
> network to maximize output power there, for X10 multiplication. The reference drive needed no change - it uses 100 MHz
> from a VCXO locked to the Rb 5 MHz, and the 10 MHz comes from the PLL divider. So there you go - 10 MHz to 10 GHz in
> one box. I couldn't have gotten the 10 GHz in there without this two-in-one source. It's no slouch either - the 1 GHz
> is about +15 dBm, and the 10 GHz about +10 dBm.
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
More information about the time-nuts