[time-nuts] Restoring a Datum PRS-50

Tue May 7 18:11:48 UTC 2019

On Fri, May 3, 2019 at 4:00 PM Clint Jay <cjaysharp at gmail.com> wrote:

> I've just laid my hands on a Datum PRS-50 Cesium reference which 'released
> the magic smoke'.
> A common problem with these is the capacitors in the main PSU but these
> have already been replaced by the previous owner.
>
> Unfortunately when he powered it up there was smoke released.
>
> I *think* I've located the source of the smoke and it looks like I may have
> been very lucky indeed as the only burned part I can find is a 1uF tantalum
> capacitor across a power rail on the main 'logic' board so I've ordered
> replacements for all of the tantalums I can find, including the 2 47uF ones
> near the CS6158 Line drivers and will replace them all before I attempt to
> power it up fully.
>
> None of the SMD work or clean up holds any fear for me with the exception
> of the single 1uF tantalum that's in the conformally coated section of the
> board near the Burr Brown PCM54HP, the remoival and replacement should be
> fine but I am worried I might disturb something critical or upset some
> finely tuned parameter?
>
> I'm also wondering about testing the main PSU 'out of circuit', I can load
> it all up, that's fairly easy but is there any reference to what voltages
> it generates and if there are any test points?
>
>
I actually have 2 of these (which I've been meaning to sell, because they
are just lurking in my office, and I keep stubbing my toes on them) -- one
of them works perfectly, the other has a low ionizer voltage....

I just went to look at the list archives to try link to my post on this --
and couldn't find it. I then looked in my mailbox to try find the original
message, and discovered I'd sent it to time-nuts at febo.com, instead of
nuts at lists.febo.com (not sure how that ended up in my autocomplete list...)

Anyway, here is what I'd intended to send months ago:
"Hi there all,

I have 2 Datum PRS-50 "Cesium Beam Primary Reference Source" units.
One of them works perfectly (although, annoyingly the system time/date
cannot be sent to anything later than 2017 because of firmware version :-))

The other, well, not so much - I'd like to get both units working, and then
sell them -- I've also got an FTS 4060, and my wife keep reminding me that
I really don't need 3 cesium devices :-)

The units are managed through TL1 (which I always find unfriendly), but
dumping the alarms gives:
DATUM 08-12-15 14:21:49 M 1 COMPLD "*C,21,08,18,00,00";

This decodes to:
21 - ionizer supply error (>+-0.16 V of memorized nominal value)
08 - oscillator control voltage (>90% <95%)
18 - primary loop gain near limit (1 or 128)

Dumping the variables (RTRV-VARS) gives:
DATUM 08-12-15 14:22:37
M  1 COMPLD
"2,R+Z,11,21,01,08,18,00,+000,+000000,+25.83,01.0,-510,-0011,-068,-0025,-4780,0081,-0039,0131,+0332,5.00,-0006,-4123,+5.08,+24.2,+14.9,-15.1,c,05.2,14.6,0.49,18.6,08.3,002,+279";

Most of these values look within their expected ranges, other than the 5th
to last one (0.49) which is "ionizer supply - voltage (V) across the
ionizer filament", which  should be more like 1.6V, and the 2nd to last
(002) which is "IP HV supply - current (μA) being supplied by the ion pump
high voltage supply (+3500 V)" which the example shows as 025.

I've measured the voltage going to the HV supply, and it aligns with the
numbers the system is reporting (8.4V (3rd to last variable) and 25.7V
(11th variable (+25.83)). I'm not comfortable measuring the output of the
HV supply, but I'm guessing the current it is reporting (002 uA) is not
unreasonable if the ionizer filament isn't getting sufficient voltage.

Is this a known issue? This unit sat (unpowered) in storage for a while,
but I would expect the the voltage over the ionizer filament to be "normal"
if it lost vacuum (unless it completely lost it, and the filament oxidized
and burnt off?!). I've had it sitting and powered for a while and there is
no change on any of the values.

Does anyone have a schematic for the PRS-10, or the Datum 5045A package it
uses? I'd like the try measure the resistance of the ionizer filament /
measure the voltage myself and possibly trace the power section for this..."

Seeing as I didn't get a reply (duh!) I took a photo of the electrolytic
caps, and started marking the voltages from the good one, so I could
compare it with the bad one (I'm fairly sure the issue is a bad cap, but I
couldn't find a service manual). If I can unearth this I'll post it, not
sure if it will help...

W

> --
> Clint. M0UAW IO83
>
> *No trees were harmed in the sending of this mail. However, a large number
> of electrons were greatly inconvenienced.*
> _______________________________________________
> 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.
>

-- 
I don't think the execution is relevant when it was obviously a bad idea in
the first place.
This is like putting rabid weasels in your pants, and later expressing
regret at having chosen those particular rabid weasels and that pair of
pants.
   ---maf