The best time to replace Rifa capacitors is immediately. ;)

On 9/10/22 19:46, vk2bea via groups.io wrote:

I opened up an HP counter of mine and saw the dreaded RIFA X2 & Y2 caps.
Out of an over abundance of caution I replaced them.
Looking carefully at them they appear to be in reasonable shape. There is some minor plastic cracking but it seems at odds with the horror stories you hear.

Those are the problematic ones. They will fail. The cracking that you see is the beginning of the failure. We've replaced dozens of them at LSSM, some having exploded while others had gotten close. They aren't "horror stories", it's a well-known and understood failure mode.

-Dave

--
Dave McGuire, AK4HZ
New Kensington, PA

I looked at the manual and I think my assumption is correct.
At least the HP8510 has a command to read the list segments...? I had to jump through hoops to do it on the 8753 8-( ...?

Michael

I opened up an HP counter of mine and saw the dreaded RIFA X2 & Y2 caps.
Out of an over abundance of caution I replaced them.

Looking carefully at them they appear to be in reasonable shape. There is some minor plastic cracking but it seems at odds with the horror stories you hear. (of course I have no idea on how the instrument I have was used before I got it. It may have had a good low use life)
Does any one know if there are good guide lines as to when to check the equipment for these problematic caps? You can decode the date from the HP S/Ns. What years were instruments built with the bad RIFA caps?
Are there particular models that, as a general rule, are likely to eventually exhibit a failure ?

Michael

That usually indicate there is no data (as you found from the error explanation).
Are you sure the command is doing what you think it's doing?
I've not used the 8510 but with the 8753 you cannot get the discrete source (frequency) values for each point. You must interpolate it from the start/stop (or center/span) data. (or use marker of course)
There is a 'list frequency sweep' mode, where you can enter either discrete frequencies or many start/stop values. This allows you to use your 201 (1601 or whatever) points economically if you have non contiguous areas of interest.
Maybe the 8510?OUTPFREL command is similarly targeted. (all I could find in a quick Google search is "OUTPFREL - Output frequency list".

Michael

Swapping C1 (100uF) for what I had on hand, 470uF, brought the rails back to somewhere near nominal (+7V is now +7.3V which is out of spec) but it brought the meter back to life so that I can now coarse zero it, etc.

I will need to get an axial to replace the radial I've tacked in but it looks like the unit is on track to being repaired.

TonyG

And as soon as I posted this, I went back and checked C1:



And previously I swear it tested correctly but now it is testing at 650nF instead of 100uF - Clearly I did something wrong.

Please ignore.

TonyG

Hey All,

I'm currently working on fixing up a non-working HP 432A unit - I've identified that there are several issues with the A2 board that I need to address before moving forward:

• Power rails outside of the spec, and
• 5KHz Multi-vibrator is bouncing around.

I'm taking the approach of addressing the rails first. The DC rails should be:

• +7V nominal - Anywhere between +6.5V and +7.1V - Broken unit 6.4V
• -13V nominal - Anywhere between -12V and -14V - Broken unit -11.1V

The broken unit I have is outside of those values - I started looking at it and eventually settled on looking at A2CR6 (6.2V Zener) as everything is referenced to the +7V rail and sure enough, it is out of spec with only a 5.8V drop.

Before changing that though, I put the scope on the rails of the non-working unit and looked at them - These images show what I saw:

7V Rail

-13V Rail


You can see there is quite a bit of ripple on the rail. I have a working A2 board and swapping that in shows nice clean rails:

7V Rail

-13V Rail


Has anyone had this problem before and could provide any thoughts/insights/comments on what might be causing the ripple? I know that the issue is contained in the A2 board because I've been using the same chassis, A1 board, and Thermistor mount to test so the only difference between the two cases is the working A2 board and the non-working one.

Look forward to any comments the group has.

TonyG

I have this one 5340A microwave counter that just won't work right despite lots of effort. I have four of these altogether, two Nixie style and two LED. Three work beautifully.. The problem one (LED) would be number two in the lineup, so I'd really like to get it going right. The number four unit (Nixie) is the crappiest in all respects except that it works great - I had to do some major kludge repairs in its power supply many years ago, it has no provision for OCXO, and is worst cosmetically.

The problem one at first couldn't lock onto anything above around 10 GHz. After much swapping of boards from unit four, I thought I had it made. It looked like A7 had problems. After some study of the manual, I put the original A7 back and did some tweaking, and it started working, until it quit again. It seemed like a thermal or intermittent issue somewhere, but after much study, I still can't tell. I even tried swapping all the boards between the units, and ended up with neither working. Swapping them all back put unit four immediately back to excellent working order, while the problem one remained a problem. Then today fooling with the tweaks on A7 and jiggling things around, it suddenly took off working fine again - then not at all. Lots of board jiggling and poking around in the bad one seems to mostly make it worse, even if it happens to start working for a few moments. The same action on unit four has virtually no effect - it just keeps going and going just fine.

This part of the system is quite complicated, with a number of interactive control and phase locked loops that are automatically tweaked and activated according to the counting conditions (mostly frequency), so I think I'll have to build a test thingy like the "K05-5340A Feed-Forward Simulator" described in the manual. I think it will help to isolate the functions somewhat, to figure out WTH is going on in there. One issue with this is that the procedure in the manual is about using it, but some of the steps have ambiguity. For instance, there's a test point marked "DC" on the K05-5340A circuit, where I think you'd take the DC measurement, but the text says get it from the pin 10 on the connector. These two points differ by a 51 ohm resistor, which can result in a 5 percent difference in reported voltage (the load is about 980 ohms). Maybe it's not that critical, but so far it looks like everything is quite sensitive to the adjustments being just right. Has anyone been through a full CAL of a 5340A, and knows the ropes on this procedure?

Anyway, it could be a bad part or connection somewhere, but very hard to tell when everything goes unstable. There are a lot of solid-Ta caps in the analog circuits that potentially could be leaky, but at least HP used the nice hermetically sealed can types, which tend to be quite reliable.

Ed

On 9/10/22 18:15, Ozan wrote:

Looking at it again I think you are right, this unit seems to be a prototype based on mix of two existing products. Board on the top is bolted only at 3 places to two bars with large portions unsupported, not production-HP like. I am not next to the unit but I will check the serial number and report. It is possible the box and bottom board is an existing distribution amplifier, top board is an existing GPS receiver, ?likely?59551A. ?ROMs on the top board have stickers that say 59551-8000x. I will check if any of HP distribution amplifiers match to bottom board.
I see GPS antennas that advertise 3V-5V operation with 28dB gain. Is there anything particular I should look for?
I attached the picture of what I think is the GPS board.

That's a Motorola Oncore GPS receiver module.

-Dave

--
Dave McGuire, AK4HZ
New Kensington, PA

On Sat, Sep 10, 2022 at 08:38 AM, paulswed wrote:

One additional comment. This is a prerelease widget. Or its one heck of a hack. Notice the GPSDO LEDs are simply drilled through the front cover. No labels. Also and this may have been someone else. The RS232 hack. A nice unit no matter. But a odd one. Whats the serial number??

Regards

Paul

WB8TSL

----------------
Hi Paul,
Looking at it again I think you are right, this unit seems to be a prototype based on mix of two existing products. Board on the top is bolted only at 3 places to two bars with large portions unsupported, not production-HP like. I am not next to the unit but I will check the serial number and report. It is possible the box and bottom board is an existing distribution amplifier, top board is an existing GPS receiver, ?likely?59551A. ?ROMs on the top board have stickers that say 59551-8000x. I will check if any of HP distribution amplifiers match to bottom board.

I see GPS antennas that advertise 3V-5V operation with 28dB gain. Is there anything particular I should look for?

I attached the picture of what I think is the GPS board.?

Thank you for all your comments.
Ozan

开云体育

"Bernd Laengerich" <bernd@...> writes:

Am 09.09.2022 um 08:27 schrieb Sven Schnelle:

Thanks, but i already ordered a HxC emulator to replace the floppy
drive. I don't even know whether the floppy drive does even work. Were

Is this one of the Sony MP-F52W drives too? I (now) know how to
emulate them properly with a Gotek hardware, FlashFloppy software and
HFE file. Had to dig a little bit to get all pieces working, the disk
change reset is a little tricky (and does not work on all versions of
Gotek hardware).

I think the MP-F52W drive is the 3.5" version? My HP4145B (which
contains the 3.5" floppy) was already modified to a HxC, so i don't know
what the original floppy was.

The Analyzer i'm converting is the 4145A, which still has the 5.25"
floppy drive.

"Sven Schnelle" <svens@...> writes:

The HP4155/56 seem to be a bit more problematic - both have broken power
supplies. One is not starting at all, the other one is starting for a
second and shuts down afterwards.

I haven't found a schematic/service guide for them - does anyone know
any details about the power supplies? One is a Voltex D2160, the other
one i don't know yet. But they look quite sophisticated, with a lot of
different voltages and several PCBs.

I've spent way to much time figuring out the problems, so don't laugh
:-)

The HP4156A just had an open thermal fuse. That one is located below (and
glued to) three ceramic resistors. I wonder what the best glue is for
that task? It has to cope with heat + high voltage. The original was a
white and hard glue.

In the other one (the HP4155A) it was actually just the fan - someone
had replaced the fan, but with a different type. While the fan was
spinning, it somehow had a different current consumption which made
the safeguard switch of the supply. I took the Fan from the 4156A,
and it worked. I spent way to much time with investigating things
like over voltage protection... I was expecting the PS was shutting
down because of some regulation problem.

So i'm wondering which one to keep - the newer ones are likely overkill
for home use. Does anyone have experience with Triax cables? There are
some chinese ones on ebay, but i have no idea about the quality. Getting
cables from somewhere else seems to be hard.

Regards
Sven

I have ran the unit all morning without any fail codes. Maybe this thing has sat for a long time and it's getting juiced back up? I think I will go with it as is and if the code comes back I'm gonna read and store the data with a programmer.? Cleaning the cards has helped a lot of issues. Thanks for all the help and suggestions from everybody.?

开云体育

I have same issue in tek 2235, failure of the little buggers, same series bourns 3329....not looking fw to getting in there to replace them on the main board input amps.
is there anything more reliable? ...only thing I found to fit.
搁别苍é别

Hello group -
Recently acquired a 3458 DMM, great condition, but reported errors (I don't remember what they were now...)
Dug the battery out of the cal NVRAM and found that it was dead.
So replaced the three NVRAMS, and it boots up fine, with errors that say it needs calibration.

I just want to be sure, before I send it to the cal lab, was there any other magic that died with the NVRAMs? Like any OS code that needs to be loaded somehow?
Or will the cal lab be able to bring it back?

Thanks -

Dan