HP Forums

Full Version: HP 32S Repair
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2
Got here late in the postings,

All great ideas, check out this HPSOLVE journal article. Scroll to page 6 for an HP 42S repair.


I found that article and really liked it, it helped me fix a 32SII which resulted in this repair-ish video:

Many thanks to Geoff!
Final Result:

I used two #3 x 1/4 brass wood screws in two of the top posts. (I found a selection of miniature screws ranging down to #0 at the local Ace Hardware) The flat head's diameter is about .005" larger than the hole in the casing, so it perfectly fits and pulls the two halves together without having to split the post or do anything else, and does not interfere with the battery cover.

All around a perfect result, thanks to the help from this forum. It is greatly appreciated.
I'm off to Ace Hardware tomorrow! Thanks for the report back.
I found a way to permanently repair the keyboard, but it's not for the faint of heart. I love the result, though. I recently repaired a dead HP20S that I bought online. The machine showed signs of liquid damage. The keyboard had one column of keys that did not work, as well as an unresponsive ON/C key. Dissasembly of the calculator showed that the kwyboard circuit sheet had broken traces--their fragility due to the fact that they are made of graphite printed on a plastic sheet. I tried using a silver conductive ink pen to reform the broken traces, but it did not work. So, I tried a radical solution: replace the keyboard with discrete tactile switches, available from Digi-Key and many other electronic component suppliers. I removed all the old plastic sheets and covered the steel keypad plate with Kapton tape cut from a 3d printer buildplate cover (3 mil polyimide adhesive-backed 10"x10" sheet from Micro-Center). Then I used double-sided tape to cover the areas where the new switches would be placed. 30 AWG magnet wire (or anything with finer gauage will also do) was cut to length (~.5" to 1.5")and soldered onto the appropriate contacts to replicate the switch matrix layout of the original keyboard. The graphite coating on the LCD diplay/processor board keyboard contact fingers was removed using a scraping tool and cleaned of any carbon residue using 70% isopropyl alcohol (IPA). Solder was then applied to the bare copper surfaces (15 fingers in all). Longer lengths of magnet wire were attached from the rows and columns to the contact fingers. The results are shown in the attached pictures. I used a switch rated at 100g of activation force (E-Switch p/n TL3315NF100Q). I found this to be a bit too responsive, so I plan to use a switch with a stiffer rating next time.[attachment=8111][attachment=8112][attachment=8113][attachment=8114][attachment=8115]
That's an impressive soldering job, and quite the creative solution. I've always disliked the conductive sheet pressed against pcb pads approach to inter-board connections. Too easy for it to fail or degrade. How did you reassemble the key/switch/plate stack?

The top sheet is the sheet containing the dome-shaped contacts. The next lower layer is the spacer sheet with the larger holes. This is the only sheet I reuse. I reassembled the key assembly (calculator top cover), upper spacer sheet (the one with the larger holes), and the steel plate with the new SMT switches in that order. The heat stake mushroom tops were cut off in order to gain access to the original plastic switch assembly. To hold the modified switch assembly together, I used a hot glue gun with a general purpose hot glue stick and put a dab on each stake from the bottom side of the steel plate. I arranged the wires so that they avoid being pinched by the reasembled stack. Using moderate finger pressure, I pinch the calculator top cover and the steel plate together and apply a little glue to each hole. Once the hot glue cools and hardens, the assembly is done. The other plastic sheets are not needed and can be put to other uses or discarded.

This experiment is actually a trial run for repairing one of the more expensive calculators, such as the HP 42S. I plan to buy a broken unit that seems to suffer from a keyboard issue and repair it using this method. The hot glue makes the unit quite solid when it is assembled. It's a semi-permanent solution, since you can remove the glue if needed. So take it easy on the glue--you don't need much, and too much can actually interfere with the fit.
(02-26-2020 05:02 PM)kriskim Wrote: [ -> ]The hot glue makes the unit quite solid when it is assembled. It's a semi-permanent solution, since you can remove the glue if needed. So take it easy on the glue--you don't need much, and too much can actually interfere with the fit.

That's a good tip to keep in mind since my earliest hp48 repair attempt damaged some heat stakes, and that affected key play. I have a recent 48GX with a damaged keyboard contact sheet like yours, and I may try your smt switch approach if other alternatives don't work out.

I wonder if making a specialized pcb to replace the whole stack, including metal plate, would work?
Pages: 1 2
Reference URL's