Connect with us

Woo Hoo HP5316A

Discussion in 'General Electronics Discussion' started by (*steve*), Dec 18, 2016.

Scroll to continue with content
  1. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    I've spent some time getting my HP5316A and HP5334B working. (They are both 2 channel counters AKA frequency meters).

    I am using my GPS 10MHz signal source as a reference.

    With a bit of tweaking the HP6316A's clock can be adjusted to within a few tenths of a Hz from the GPS, but it drifts a couple of Hz with temperature variation.

    The (much older) HP5316A has option 004, the ovenised oscillator. After allowing it to warm up, I am getting differences of a few parts in 10^-9 as per the HP5334B when it is set up to compare the HP5316A's clock with the GPS signal.

    Of minor interest is that the 5316A starts up with a significantly poor clock (several kHz off), but as the oven heats up it first overshoots 10MHz, then "rings" a few times before settling on so close to 10MHz that it's hardly worth worrying about.

    The 5334B, on the other hand, starts up much closer to 10MHz (a couple of hundred Hz off), settles reasonably nicely, but still drifts with ambient temperature.

    One thing I haven't done (that I should do) is look at the GPS clock vs the frequency counters' clocks and watch the drift :)
     
    davenn and Harald Kapp like this.
  2. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    For those playing along at home, the GPS receiver is an HP Z3805A.

    Oh, and the two clock signals look so nice together on a scope. :)

    Next step is to compare the GPS signal to my rubidium frequency standard (but not this weekend)
     
  3. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    This is the sort of thing I expect to see.

    When I compare the TCXO output to the GPS I see a slow drift between them.

    I'm beginning to doubt my understanding of the math function on the counter as the drift is much faster than I would expect from 1 part in 10^9. It looks more like 0.25Hz (given it takes about 4 seconds to drift 360 degrees).

    Considering it was new out of the box... 40 years ago, I'm pretty happy
     
  4. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    And of course it's an OCXO not a TCXO.

    Given that the accuracy of these is around +/-2*10^-8 over a 24 hour period, it makes more sense.

    The long term accuracy is only +/-4*10^-7, it seems lucky, but the labeling on the unit suggests it was regularly calibrated and my understanding is that drift reduces over time (let me know if you understand differently).
     
  5. shrtrnd

    shrtrnd

    3,734
    478
    Jan 15, 2010
    Does the HP5334B have the 010 Oven Osc Option?
    The HP5316B has a warm-up of +/- 5 x 10 -8 of final value after 20 minutes, which pretty darned good. (With your Option 4 Oven Osc).
    I know you're obsessed with accuracy (which is a good thing), but the specs appear to be very good
    for almost any application I can think of. And you have a couple of Counters made by the company
    that used to make the Rolls Royce of Test and Measurement gear back in the day.
    I don't know of too many modern-day equivalents that can match the specs you have.
     
    Last edited: Dec 19, 2016
  6. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    The 5334b doesn't have an oven, but if I was particularly interested I could probably fit one.

    However I have both a GPS and a rubidium standard, so I could just use them as external references.

    But the 5316 is pretty darn good on its own.

    I found that after 24 hours the difference between it and the GPS had fallen to about 1/8Hz, or well within 2*10^-8. I'm going to keep it powered up for a couple of days to see if this is consistent out just random good luck. Given that my shed is not temperature controlled it's a bit unscientific but still interesting.

    My GPS reference takes about 1 hour from a cold start before it thinks it's ready to go (it takes about 10 to 15 minutes to get a GPS lock, I think the rest is averaging it's position so it knows where it is. I've read that leaving it on for 24 hours improves is performance also. However I'm pretty sure the difference is marginal, right?

    My rubidium module is "well aged" and probably has far more hours behind it than in front of it. I read with interest about the technique of rejuvenating the lamp, but as it works fine, I'm not going to be diving into the "physics package" any time soon.
     
  7. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    At 2*10^-8 this morning with an ambient temperature around 20°C
     
  8. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    And he is a small battery powered meter I was given. It's a little dodgy, but better than I feared.

    IMG_20161220_073451.jpg

    The input is the GPS.

    Edit: looks like about 5ppm out, or an order of magnitude better than my rule of thumb for a crystal oscillator (0.005℅, 50ppm). I guess that's either lucky or it has been tweaked.
     
    Last edited: Dec 20, 2016
  9. shrtrnd

    shrtrnd

    3,734
    478
    Jan 15, 2010
    The manual on the 5316B with the oven says 20 minutes warm-up That's the factory spec.
    The 5334B with no oven, I'd wait the one hour just because I know you appreciate accuracy.
    I've read most of your posts. 'Marginal' to most people, is unacceptable to you.
    Most of the T&M gear I use, MY rule of thumb is 1 hour warm-up.
    I keep my Fluke 5700A on continuously and most of my DVM's, just because I don't want to wait for readings.
    I like my HP3458B because it's old and trustworthy, like me. I use a couple of Keithley 2001's cause it's easy for me
    to switch the AC readings to Low Freq RMS when I gotta do close work, and I have an Agilent 34401A that's accurate,
    but kind of a pain to set-up for some functions, and I don't like using it much.
    As a suggestion. Consider finding yourself a small logbook and when you remember, take readings and log 'em
    on your accuracy. The Age-Rate over time. You'll see some drift from time to time, but what I watch for is the
    long-duration declination of your readings. Every instrument is different, and I think you'll be happier if you know
    what your age-rates are.
     
  10. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,386
    2,772
    Jan 21, 2010
    Here is the difference in clock rates in the 3216 before and after a very quick and rough adjustment.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-