Products > Test Equipment
HP3048A phase noise measurement set
(1/2) > >>
Leo Bodnar:
Does anybody use HP 3048 phase noise measurement kit in anger?
I have put one together even though it wasn't an easy ride.  I have ended up with several HP9000 series workstation and HP35650 DSAs as a result before I have realised that its software makes such a massive chunk of its IP.  It's 40,000 lines of HP BASIC code.

Between four applications that are available for it (original HP9000 suite, original MS-DOS suite, modified HP BASIC suite to run on HTBasic emulator and PN3048) I have settled on using HTBasic emulator to generate calibration files and PN3048 for actual measurements.  It's a killer kit when it works.

I have a few questions on both HP11848 PN interface and HP3561 DSA.  HP11848A, amazing as it is (despite having HP-IB interface it does not have a single clock source or even MCU), is full of analogue circuits.  Do electrolytics in signal path need replacing or are they better left alone if it passes calibration tests?

Can its internal sources 10MHz and 400MHz be improved, upgraded or replaced?  They have quite a few spurii that I'd like to get rid of.  I have to use external Wenzel, Spectracom and Vectron OCXOs that would make sense putting inside 11848.

Can the tests be automated?  I'd like to perform few thousand repetitive test on a DUT while modifying DUT setup in an automated loop.

Overall, it is a valuable tool.  I am working on improving spurii performance in our mini GPS clock and got good results so far - thanks to HP and their HP3048.

Leo

Wolfgang:
Chapeau !

This huge block of machinery is well worth to go into a vintage hardware museum. At the time, it was an absolute milestone.
Nowadays, the pace is set by Rohde & Schwarz FSWPs or Holzworth equipment. Did you try to compare the measurements you got
on your vintage machine to newer equipment ?

BTW: I bought your GPS clock and I like it.
pnv57:
Apologies for the bump, but I am curious about the 3048 test set and the state of PN3038 and would basically end up making the same thread. What modes in the counter are used that restrict it to the following counters?: HP5384A, HP5385A, HP5386A, HP5343A, HP5316A, and HP5316B
S57UUU:
You can do measurements without a counter, it is used only to speed up beat check, VCO slope measure and PLL lock, which can take quite long, especially with low tuning slope VCOs.
When I measure HP10811s with 0.3Hz/V slope, it can take 20 minutes until the measurement starts, but with "normal" VCOs, a couple of minutes is more typical, on my counter-less HP3048 setup with DOS software.

Otherwise, I find the 3048 a quite good instrument. It may be a bit slow, but I have yet to encounter a oscillator, that would need deeper than the 3048's -180dBC/Hz noise floor. The lowest I ever measured was -175, an OCXO followed by a crystal filter. The newer boxes advertise a -200 floor, but for that, you need to feed them with +23dBm signals, etc, You can't beat physics, 300K means -174dBm/Hz!

One problem I encountered lately on my setup, is that the "line porcupine" is slowly growing. I guess the caps in the 11848 power supply need to be replaced.

Marko Cebokli
G0HZU:

--- Quote ---Can its internal sources 10MHz and 400MHz be improved, upgraded or replaced?
--- End quote ---

I know this is an old thread but a while back I had a go at designing a low phase noise crystal oscillator at 10MHz. The aim was to get low phase noise rather than good stability and ageing and I did it using slightly unconventional means. I used a low noise class A amplifier and ran the crystal quite hard.

I didn't bother trimming it to exactly 10MHz, this was just an experiment but I managed to get the same results as a basic simulation using Leeson's equation.

I also tested a 102.4MHz VCXO circuit and the aim was to triple this to use as a DDS clock. However, I never got around to it. I'm still hoping to buy one of the E5052 SSA analysers from the company I work for.

The spurious signals seen on the 10MHz trace are caused by pickup in the busy RF lab. I didn't have the oscillator in a screened test enclosure. otherwise I think it would have been spurious free.

The aim was to try for -160dBc/Hz at 1kHz offset with a noise floor at about -170dBc/Hz and I managed to get quite close to the phase noise prediction from the simulation. I think the phase noise of the E5052A itself may be limiting the result slightly at around 1kHz.


Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod