|
On 3/3/25 12:41, ed breya wrote: That's an interesting picture of the failed cap. I'd guess the molten blob on the can end is just some electrical solder that ended up there by coincidence sometime in the past, or maybe from inside the part. Try hitting it with a soldering iron - it should melt. If it doesn't, then I guess it could be Ta.
Metallic tantalum does not normally look like that, all silvery and shiny, because oxides form on the surface with a range of iridescent colors. It's more of a refractory metal that would not exist molten at a temperature that wouldn't have turned the rest of the part to ash. Besides that, the "solid" Ta caps are mostly Ta oxides and very little metal, sintered into solid slugs, and I can't picture there being that much actual Ta metal present (in proportion) inside a cap of that size. There could have been that much electrical solder in there though. It's used to make internal contacts, and especially to solder the electrical leads and glass/metal hermetic sealing disc into the can. Ordinarily I would agree, but I've seen this "blob of metal on the side of a dipped Ta capacitor" failure many times, and it's always shiny and looks a lot like solder. -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
I forgot to comment on hermetic solid Ta caps. I have seen only a few failures of these over the years, and I consider them among the best of parts for reliability. I have seen lots and lots of failures in all kinds of non-hermetic Ta caps, especially certain brands of dipped style. My experience here is with the older parts in the same vintage as most of my gear and such, so not so much about more modern parts which should be better. For instance, I have seen very few failures in surface mount style solid Ta types that are so common now. These are non-hermetic but seem to survive OK. Or, maybe after these rack up many years they may degrade like the others. Only time will tell.
?
Ed
|
That's an interesting picture of the failed cap. I'd guess the molten blob on the can end is just some electrical solder that ended up there by coincidence sometime in the past, or maybe from inside the part. Try hitting it with a soldering iron - it should melt. If it doesn't, then I guess it could be Ta.
?
Metallic tantalum does not normally look like that, all silvery and shiny, because oxides form on the surface with a range of iridescent colors. It's more of a refractory metal that would not exist molten at a temperature that wouldn't have turned the rest of the part to ash. Besides that, the "solid" Ta caps are mostly Ta oxides and very little metal, sintered into solid slugs, and I can't picture there being that much actual Ta metal present (in proportion) inside a cap of that size. There could have been that much electrical solder in there though. It's used to make internal contacts, and especially to solder the electrical leads and glass/metal hermetic sealing disc into the can.
?
Anyway, I think it's just solder on there, but if it is actual Ta metal, then it was quite a feat to nearly instantaneously melt down, collect a decent glob of liquid Ta inside the core, then burst the can end and eject the Ta gently into a nice external glob. All without a hint of discoloration or deformation of the can. I'm very curious about it. If you still have the part, it would be interesting to see what that actually is. Please check it out if possible.
?
Ed
|
The other one ooks just as weird but is healthy according to the chinese component tester.
?
Are there guidelines on HP capacitor replacements? Such as caps that never/very rarely fail and those who absolutely always do?
I want to recap this unit to keep it. It's not calibrated anymore of course but it's so great. Quality long life capacitor are expensive, I don't want to replace what does not need to be.
?
Thanks,
Arnaud.
|
On 3/2/25 18:17, AArnaud via groups.io wrote: It is indeed a Tantalum capacitor, from the Kemet T110 series. <> From my research that would be T110B226K015AS. It's expensive.
Here's how it looks:
My manual <> is definitely wrong on that part naming. It is indeed a Tantalum. Wow. I have *never* known one of those capacitors to fail. I replace dipped tantalum capacitors all the time, but never one of those! -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
Hey Guys,
?
It is indeed a Tantalum capacitor, from the From my research that would be T110B226K015AS. It's expensive.
Here's how it looks:
?
is definitely wrong on that part naming. It is indeed a Tantalum.
?
Arnaud.
|
On 3/2/25 09:06, factory via groups.io wrote: The manual on the web gives type 150D for C1 & C2 on A7, these are hermetic sealed solid tants, of course they might have been replaced with a bead or the disguised bead in an axial or radial block package in the past, or HP (or someone else) might have decided to use the non-hermetic axial ones in a repair, or later build using them. Yes, that's why I wanted to see a picture. I've never seen the ball-of-molten-tantalum failure on anything other than a dipped unit. And I notice they are listed as 15V parts on 12V rails, not much derating, if using non-hermetic tants make sure to use at least double the voltage rating for replacements. Seconded. -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
The manual on the web gives type 150D for C1 & C2 on A7, these are hermetic sealed solid tants, of course they might have been replaced with a bead or the disguised bead in an axial or radial block package in the past, or HP (or someone else) might have decided to use the non-hermetic axial ones in a repair, or later build using them.
?
And I notice they are listed as 15V parts on 12V rails, not much derating, if using non-hermetic tants make sure to use at least double the voltage rating for replacements.
?
David
|
On March 1, 2025 4:43:06 PM "AArnaud via groups.io" <a.cabiyaud@...> wrote: According to the manual they're actually both electrolytics.
Thanks for you kind word! Can you post a picture of these capacitors? -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
Hey Dave,
?
According to the manual they're actually both electrolytics.
Thanks for you kind word!
?
Arnaud.
|
On 3/1/25 11:57, AArnaud via groups.io wrote: Hello everyone,
I'm glad to report the HP 3403C works again!
I've replaced the duos of LDs A/D chips, but this was not enough to revive the unit. It kept blinking +1 although it had lost its tendency to sometime display garbage.
I turns out I lost the +12V supply rail which I previously had. On the test point of the supply had a dead short. Taking boards out I found that it was on A7 (the board between the converter and the motherboard). Looking at it I found that both C1 and C2 looked really weird has they had a ball of solder on them and you could see inside of them. That solder-looking ball is metallic tantalum that melted and burst through the outer shell of the capacitor. This is a classic failure mode for a shorted tantalum capacitor, usually the "dipped" variety. I've taken both them out and although C2 is apparently healthy the C1, which is the smoothing cap of +12V, was shorted. Good thing it was regulated.
With them out the unit works again and is not too far off reality aswell. Very happy about it.
My beginner mistake was that I made the power supply measurements with basically all of the boards out. However I think the new chips were necessary as the display was not working even when A7 was out.
I'll order new capacitor for that board, and i'll also recap the power supply aswell.
Here it is reading the probe test of my oscilloscope together with its friend the Racal Dana. An excellent recovery of an excellent instrument. Nice work! -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
Hello everyone,
?
I'm glad to report the HP 3403C works again!
?
I've replaced the duos of LDs A/D chips, but this was not enough to revive the unit. It kept blinking +1 although it had lost its tendency to sometime display garbage.
I turns out I lost the +12V supply rail which I previously had. On the test point of the supply had a dead short. Taking boards out I found that it was on A7 (the board between the converter and the motherboard). Looking at it I found that both C1 and C2 looked really weird has they had a ball of solder on them and you could see inside of them. I've taken both them out and although C2 is apparently healthy the C1, which is the smoothing cap of +12V, was shorted. Good thing it was regulated.
With them out the unit works again and is not too far off reality aswell. Very happy about it.
?
My beginner mistake was that I made the power supply measurements with basically all of the boards out. However I think the new chips were necessary as the display was not working even when A7 was out.
I'll order new capacitor for that board, and i'll also recap the power supply aswell.
?
Here it is reading the probe test of my oscilloscope together with its friend the Racal Dana.
?
 Thanks everyone!
|
It happens! I'm okay with that, especially since they were a high chance they were defective anyway. I'll update soon, at least for the supplies and clock.?
|
There are days when Murphy's Law just seems to run in overdrive mode, but the fun will resume in a few day and I hope to have more news soon !
Joel
|
Well, It did not go well lol.
This is actually not that badly packed. It is somewhat dense but this is really well designed. The boards can be freed from their metal case easily, and A22 (AD board) is socketed from the rest, and the LDs themselves are socketed too. Really easy to work on actually. Nothing to complain about, great job from the good fellas at HP.
Expect I wanted to reseat the chips in case of moisture but my hand slipped when doing LD110 and it badly bent a pin which broke on straightening. Meh. I guess the replacement is mandatory now. I have ordered the pair. I'll still take some measurements later but there it goes. At least the chips are easy to find on Ebay, and wasn't even crazy expensive nor far away.
|
My guess is that the clock source must be OK, otherwise the blinking function would not work or the display would blink at a "wrong" rate.
I think I would probe all pins around both the LD110 and LD111 with a voltmeter and an oscilloscope, and check for correct supply voltages, reference voltage, "illegal" logic levels, proper multiplex signals etc ... I know the 3403 insides are "densely packed" and the probing may be difficult, but it may give very useful clues.
... And of course please keep us posted !
Jo?l
|
Hello,
For Joel:
1. Yes, the blinking +1 indication remains on DC. Actually, it remains +1 on any setting, with or without the converter inside. It does not seem to care.?
2&3. I will report the results.
?
For Jeff:? The clock is actually a good old NE555. This is what HP used but also the solution used by Intersil in the datasheet. I'll check the clock output.?
?
|
Actually, if the display is goofy, not counting or "wiggling", maybe the clock oscillator that drives the system or display is not running. Probably a simple oscillator crystal with a pair of gates.
Look around with a scope and see if you have gating action anywhere. Jeff Kruth
|
Hello Arnaud,
I think there's a few things you may do to locate the source of the problem.
1 - Try operating the 3403C in DC, linear mode. This will bypass all the RMS to DC converter and the LOG circuitry, and will let you measure different DC voltages, positive or negative, and see what value is displayed. Is it always +1 ? This will be a first indication.
2 - Check the voltages which are fed to the LD110 and LD111 : power supplies and reference voltage, and check the clock input.
3 - With a known voltage at the 3403 input (for example +0.5 VDC), check the input voltage on pin 15 of the LD111.
And of course, please let me know !
Sincerely,
Jo?l Setton
|
I kept looking up on the LDs chips, on the Tek groups and here with a thread on the 8558. It uses the same duo for the AD converter of its frequency display it seems, with a very similar failure as it was stuck on "+1".? At this point i'll probably blindly replace the pair as they seem to have a high failure rate with time. They seem particularly vunerable to overload, i'll check the power.
I still don't totally understand how the digital section work but I don't think I need to for now.?
|
Dave McGuire
factory