My experience replacing the battery on the HP 3457A

[mrnuke]

I recently scored a HP 3457A for an amazing price. Unfortunately, the battery on the unit was dead, and as I am trying to replace it, I found the process to be a confusing, and I would like to share my experience, along with the information that I discovered along my journey.

The meter was constantly popping up with error code "9 NOT CAL'D". CALNUM showed 0, and the readings were a few thousand counts out of spec. It was still good enough for 4 1/2 digits. I only realized it was out of spec by comparing the readings to a calibrated 33401A. I opened up the 3457A, and, as you might have guessed, the cal battery had leaked, and was measuring 19 mV.



I was worried that the leak might have damaged the PCB, so I immediately proceeded to removing the battery. I took the digital board out, then snipped the pins off the battery. Not knowing the extent of the chemical damage, I was really worried I might strip a trace, or rip out a via.

I heated the battery pins from the back side of the PCB, then carefully pulled them out. I then proceeded to carefully remove any excess solder with some solder wick, then clean the residue off with 99% isopropyl alcohol. PHEW! The PCB is in mint condition. The corrosion was contained to the positive baterry pin.



Even having the dead battery, I had difficulties figuring out what to get as a replacement. I started with the inscriptions on the cell:

ETERNACELL SDX
3 VOLTS No. B9511
H.P. P/N 1420-0278



Searches for those part numbers turned out empty. I spent the next few hours measuring the dead cell, and comparing its dimensions to specs for other lithium batteries. Turns out that this is a:

2/3 A size (not 2/3 AA)

At this point, I also found some eevblog posts about this, and people were replacing them with either 3 V or 3.6 V cells. I was able to find the service manual, and the schematic eased some of my fears.

The NVRAM rail, NVVCC, is a 5V rail. The battery feeds this rail through a diode, CR504. Alright, so anythng between 3V and 5V will work. Now it's time to find the right battery. The only battery that I could find on digikey that had the right PCB mounting pins was the Panasonic P226-ND.

I decided to order the battery, but to first attempt installing a battery holder. Someone else on the eevblog forum already tried the battery holder, and they couldn't get it to fit quite right. I decided to order every 2/3 A holder that Digikey had to sell in the hope that one will fit.

I'm still waiting on the parts, and I will update this as I go along. Please excuse the crudity of the images. I only had a phone available to take the pictures.

[mrnuke]

First part of the Digikey order arrived. I didn't get the batteries yet, because they had to "ship ground" due to "shipping restrictions".

Back to the repair. First order of business was to create a jig to keep the calibration RAM powered on during the repair. Since I had calibrated the DC voltage ranges, I didn't want to lose that data. I decided to use three AA batteries and connect them to the board with Pomona 3925 clips. I used CR504 to hook up the positive lead, and a nearby capacitor connected to GND.



I thought that was a neat way of connecting things. Hooked it up with the system running, then I was able to remove power and work on the board. The grabbers stayed in place nicely, and I had about 20 cm of slack wire for the temporary baterries, so that wasn't a problem.

Now, to finding the best fitting holder. Sadly, none of the holders I had would fit. The mounting pins were spaced too far off to align with the holes in the board. Bummer! My best hope was an MPD 6S-2-3A holder. I was able to bend its pins inwards, then up so that it aligns with the footprint.




I put it in, I soldered it. It all felt to fit nicely. It was only after I was done soldering it that I realized it was crooked. It looks like the "very conveniently" placed resistor nearby pushes the holder just slightly. What's actually happening is that the battery mounting holes are not aligned, so the entire holder is naturally rotated, and it still sits flush with the board. I think the other person here who tried this got the same result.



If anyone is curious, I think one of the holes in the holder can be used to anchor it to the PCB more securely. There's no trace underneath the mounting hole close to the positive terminal. I think it's nice, but I'm not going to drill a hole in the PCB for this purpose. I think the holder is already securely mounted.



That's all I can do until I get the batteries. I've put the unit back together, and the calibrations for the DCV ranges were not lost. Hopefully, others can find this procedure useful. I really liked using the Pomona grabbers because they had a surprisingly strong grip. I should get the "shipping-restricted" batteries next week and finish off this project.

[mrnuke]

Alright. The new battery is in. I decided to go with the Panasonic BR-2/3AG. My reason was that the same series of batteries is used on the HP 3468A, so they should work. I also considered using the Tadiran TL-5955/S, but that one is a 2/3 AA size, so it's a little thinner than the others.

All in all, a fresh battery is in place, and I did not lose any calibration data. Now to take some measurements...

[uncle_bob]

Hi

A "commonly seen" solution:

The battery is a DC beast so short leads are not an issue. You do not replace it often so accessibility to the holder is also not a high priority.

Stuff a holder plus battery "in line" with some wires. Cover the whole thing with heat shrink. You now have a battery with really long wires hanging off of it. Mount the "blob" anyplace you can with a wire tie. Solder the leads into the pc board. In 20 years somebody will have to clip the wire tie and replace the heat shrink.

One small advantage of this approach: If the battery oozes a bit, the plastic heat shrink isn't going to be bothered.

Bob

[mrnuke]

I thought about that. I have a 34970A data acquisition system with a dead battery, but not enough space to mount a holder on the PCB. I might try this trick there.

[GRFixedGear]

What the bloody hell is going on. I have an 1989 unit with all the 600-series reference designators on the common electronics (main board). But unlike the schematic for the earlier version, the new one from the Edition 3 service manual does NOT seem to show cross-references for where the NVVCC signal goes, nor does it depict the power/ground for the U603 SRAM chip that apparently now does double duty as standard SRAM and CAL SRAM.

Am I being a dunce and missing that information somewhere? I see the main analog side had a table with supplies/grounds for the different IC's. Thats not present either!

Hoping I get to learn something new here and haven't stumbled across a situation where "well DUH of course thats the way it was done" applies.

Thanks
Jason