[time-nuts] Upgrade an HP 5342A microwave frequency counter to have an oven oscillator.
Dr. David Kirkby (Kirkby Microwave Ltd)
drkirkby at kirkbymicrowave.co.uk
Sat Nov 29 12:20:13 EST 2014
On 27 November 2014 at 22:38, Dr. David Kirkby (Kirkby Microwave Ltd)
<drkirkby at kirkbymicrowave.co.uk> wrote:
> The HP 5342A has an optional oven oscillator. I just bought one of
> those counters, but mine has a TCXO and is about 50 kHz off at 10 GHz.
> I'm sure I can trim it closer than that, but if possible I'd like to
> stick an oven oscillator in it. Does anyone know what is involved? I
> have at least one 10811A oscillator here that I could stick inside,
> but does it need any more, or just the oscillator?
> Someone said the oscillator should be an 10544-60011, but a 10811A is
> backwards compatible.
It was very easy to fit the 10811A oven (HP 10811-60111) and get the
oven working, although the oven is not powered when the counter is not
switched on, even if there is power at the AC mains input. I can live
with that. In fact, I prefer it to be honest.
The procedure was
1) Remove top and bottom covers, which means the two front feet, plus
4 rear feet/protectors need to be removed.
2) Undo one screw at the top that held a board with a TCXO on it. The
board plugs into a 15-pin edge connector.
3) Pull out the board (HP P/N 05341-60047) with the TCXO. The TCXO on
the board is a marked DALE, TCXO-22-1, 0960-0394, 10.0MHz, SET 1.0Hz.
(normally I would put a space between a number and the units, but I've
written what is actually on the TCXO).
4) Attempt to insert the 10811A oscillator. This failed, as there were
some wires around the optional board for GPIB which were restricting
the space too much. So I had to cut a wire tie, and move the wires out
5) Fitted 10811A at the top.
6) Invert the counter, and screw in the two screws which secure the
10811A to the chassis. For this I needed to temporarily move a ribbon
cable, as the screw was below it.
7) Powered it up, and it worked. It shows "OVN" in the right of the
LED display. Once that went out, it still took a minute or two for the
readings to become pretty stable, although no doubt it will take
months to become as good as it will get.
I've not adjusted it yet, as I don't have any accurate frequency
reference. But whilst the actual frequency indicated on the counter is
different from what my signal generators are supposed to be producing,
the last few digits (100, 10 and 1 Hz), are not all jumping around
when seeing 10 GHz.
The frequency indicated on the counter when connected to two different
signal generators, which both have ovens of unknown type, are:
1) HP 83623A 20 GHz sweeper set to 10.0 GHz, fed into high frequency
input of the frequency counter.
HP 5342A counter indicates 10,000,000,690 Hz (relative difference = +6.9 10^-8)
2) HP 8656A set to 100 MHz,
HP 5342A counter indicates 99,999,987 Hz (relative difference = -1.3 x 10^-7)
With the old TCXO in the frequency counter, the indicated frequency of
the 10 GHz signal was about 48 kHz off, but it moved around a KHz or
so. In contrast, now the oven is installed, the reading is a *lot*
more stable, with it shifted about 15 Hz.
I don't currently know the absolute accurate any of the references in
the test equipment are, but certainly the readings are a lot more
stable after fitting the oven.
I will need to get a GPSDO before adjusting any, but if nothing else,
the short term stability of the oven is clearly superior to the TCXO.
Long term should be too, but I can't determine that from what I have.
More information about the time-nuts