HP3478A - Dead

[mfreymond]

Hi,

I have an HP 3478A I purchased from e-Bay a few months ago.

It arrived working.  The calibration was valid, and the battery was old.  I replaced the battery several months ago.  I've used the meter a few times since replacing the battery and all was well.  I also made a backup of the config

The line filtering and electrolytics have been replaced.

I went to use it the other day; and the fuse was blown.  I replaced the fuse with one of the correct value (0.25A - located in Canada).  The display on the meter will flash - all the segments are on for a part of a second - it then fades out to blank again.    It varies between the left 1/2 of the display flashing as described, or the top 1/2 of the display flashing; or sometimes nothing at all.

I've checked voltages from the power supply - both +5v, the +15v and the -15v are all within spec.

I've worked through the service manual - Following the steps on page 7-D-1, the ALE pin (Pin 11) is high (+5v) -  it should be a square wave (photo included in the manual).  There is a sine wave on pins 2 & 3 at 5.855Mhz.

I've searched; I can't find anything that would point me towards the solution (short of replacing parts which may not be attainable any more).  I also do not have access to a signature analyzer.

Before I give up and move on, is there something else I can try?

Thanks

[bob91343]

I suspect the power supply electrolytic capacitors.

[Specmaster]

I've seen a thread somewhere where a similar problem was traced to a bad connection on the display / display board

Sent from my POT-LX1 using Tapatalk

[MarkL]

Take a look at the processor nRESET (U501 pin 4).  See if it's low (processor held in reset).

[cur8xgo]

Look for visual anomalies. (aka caps like someone mentioned)

Especially areas where there was clearly heat dissipated (discoloration of plastic or pcb soldermask etc..), and see if those parts are now cold (thermally killed parts)

Inspect EVERY solder joint for cracks with a well lit magnifying glass (bad solder joint)

Push the PCB with a non conductive stick in random locations and see if it comes back to life (broken traces/bad solder joint)

Look at the design and ask yourself what would die first

[coromonadalix]

What would die first would be : everything that heat a lot (regulators, power resistors, xformer ...)  electrolytic capacitor, tantalum caps,  sure you have pcb traces with "big physical" parts, "mecanical stress" on the pcb ... broken traces etc...

Bad reset line ?  bad contacts on connectors / flat cable if any, badly socketed cpu / ic's  when it the case.

You're lucky a service manual exist for this meter


Some people gave you good tips to help find your problem,  take your time, work a section at the time, check re-check voltages etc ... (i know you did)

[mfreymond]

Hi,

To post a Saturday morning update...

When I 1st got the meter, I replaced most of the electrolytic caps.  A couple of the power filtering caps were cracked.  (I meant to mention this in my initial post, clearly I did not...)

There are reports of the display connector's being fragile/easily broken.  I'm a little nervous of going that direction (at least at the start).

I checked the CPU reset line (U501 pin 4).  It was being held low.    This reset line also connects to U512 pin 17.  It seems to be driven from U550 pin's 1&2.  This was also measured to be low.

I powered the meter off and did re-flow the negative solder joint for the battery.

I started poking about to document each of the pin values on U550.  When I came across pins 1/2 again, they were high.  And the meter was online.

U550 is a 1826-0138 -  This is a NTE834 Voltage Comparator (http://www.weisd.com/store2/NTE834.pdf).

I power cycled the meter, and it wouldn't start up again.  I found through experimentation that simply probing pin 4 of U550, the meter would start up. 

Page 7-F-17 of the service manual outlines the logic - U550 is supposed to hold the Reset line low until the voltage comes up on pin 4 - Once the voltage reaches a high enough level, it releases the RESET pin and allows the processor to process.  I found there was no connection between pin 4 of U550 and R765 .  For now I've installed a bodge wire between R765 and pin4 of U550.

I do not see any visible cracks/breaks in the PCB.

Many thanks to MarkL for pointing me in the right direction.

Take a look at the processor nRESET (U501 pin 4).  See if it's low (processor held in reset).

[coromonadalix]

NTE brand are substitue parts   like ECG was in the past,   an NTE834  should be an lm339n  or an hp branded part,  you should not find any NTE part in the dmm, unless someone has "played" in it.

Or you found an 1826-0138 equivalent document

[mfreymond]

Correct - In the meter the part has a Moto logo on it with a 8404 date code.  I looked up the part # to find the spec sheet.