The Museum of HP Calculators

HP-65 HP-67 HP-97 "Clutch" Fix

This article is specifically aimed at users of HP-65/67/97 calculators who have had the defective gummy wheel replaced, either by the Silicon Tubing or O-Ring fix, and are now experiencing card jamming and misreads. Those who have a calculator with a gummy wheel fix that was performed by others are highly encouraged to read through the documentation describing both solution procedures before attempting any "clutch" fix.

A brief history. Faced with that gummy stuff sticking to program cards, and with a small amount of research into the problem, I was, thanks in most part to hpmuseum.org, presented with two disparate yet equally elegant solutions. One is Mike Davis' Silicon Tubing solution (http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/articles.cgi?read=179), the other Mark Hoskins' O-Ring solution (http://www.hpmuseum.org/guest/hoskins/67crfix.pdf). At first I opted for the O-Ring solution, as Mark (waterhosko@aol.com) offered not only a detailed step-by-step procedure, but the actual O-Ring material as well, and for an entirely reasonable 4 or 5 bucks. His documentation also included some excellent diagnostic material. When my HP-67 read its first card after all of those years of sitting dormant, Mark became my instant hero.

His instructions also included a reference to the slippery "clutch" problem, which I promptly ignored, even as the cards became more and more twitchy to get to run smoothly through the gauntlet.

Well, not content with simply accepting a single solution, I opted to attempt the Silicon Tubing Solution. Everything was perfect, everything fit, and yet, the thing stalled time after time. I'd get a motor buzz, followed by an "Error" display. During reassembly, I was also greeted with detached power wires. Not the ones that plug into the power supply board, but the ones that solder to the underlying main board. Oh, joy.

So I naturally went back to the O-Ring solution, which worked great for a while, before cards eventually began to stick every time. This is when I revisited Mark Hoskins' clutch slippage solution advice, which he graciously supplied with his O-Ring fix documentation.

The advice was to use a little super glue to tie things together. This left me feeling a tad apprehensive.

So, lets look at what we have. Though we can only get at the drive wheel with complete disassembly of the calculator, access to the motor and "clutch" can be attained with removal of the back panel. I removed the two slotted screws holding the motor to the frame, holding the calculator body on edge to allow gravity to keep the tiny slotted screws in their case holes (as suggested in Bernd Schmeling's wire insulation clutch solution http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/articles.cgi?read=262), and, remembering where they connect, unplugged the red and black motor wires from the power supply board. I was faced with the following...

A motor with a thin 5/64 spindle extending into a 3/32 I.D aluminum sleeve, and, from the other side of the sleeve, a brass screw shaft that meshes with the grooves of the card reader's plastic drive wheel. I used rubbing alcohol to remove EVERY trace of gummy "clutch" material from the motor shaft, the screw shaft, and from inside the aluminum sleeve.

Well, here's the thing. The motor shaft engages loosely with an indentation in the screw shaft, and, originally, some kind of gummy material adhered the two pieces within the oversized 3/32 I.D aluminum sleeve.

I tried everything but the super glue: the wire insulation fix, minute silicon 3/32" O-Rings I purchased from smallparts.com, candles and incantations, nothing worked. Things jammed completely, or either the motor shaft or the screw shaft would spin freely.

Until I came up with this. It was like an epiphany.

The thing isn't really a "clutch" at all. It's a start-up torque inhibitor and vibration suppressor, all meant to make screw gear and wheel engagement not only be kind to each other, but to be mutually forgiving as well. If the two little nylon ball switches that let the calculator know that a card is in place but not moving ever fail, and the assembly does indeed act as a last resort clutch, it would probably have to be replaced after a short time anyhow, and the main board electronics would have to be deemed questionable.

Assuming that the electronics and nylon ball switching are operational, the following fix, though a tad tedious, works perfectly.

Please read the entire thing through before commencing. Once all of your parts are cleaned of all residues, there's no going back.

Materials:

A small tube of DAP 100% SILICONE Household Adhesive, hopefully fresh. Don't scrimp on this, its only, like, two or three bucks a tube.
Some scraps of 24 lb glossy paper, or equivalent.
Tissue paper.
3/32 metal bit (won't be used for any drilling).
Magnifying glass (usually handled, used for reading small print).
Rubbing alcohol.
Q-Tips.

Procedure:

Ignore how drastic this fix seems. When done successfully, your card reader "clutch" will work like new. But please be cautioned. After reading through these instructions, if you are at all unsure, please seek out other solutions.
Remove the back cover, hold the case to the side, release the power supply connector to the main board and flip it up in order to access the two tiny slotted screws holding the motor in place.
Remove the motor and shaft assembly, remember where the red and black wires are connected to the power supply board, and gently pull on them to disconnect. With the magnifying glass, note the tiny brass clips that facilitate reconnection.
Being careful not to exert enough pressure to warp the aluminum sleeve, separate the screw shaft from the motor...it will pull off. The aluminum sleeve will usually remain attached to the screw shaft rather than the motor shaft.
Make sure that a tiny nylon washer is still surrounding the motor shaft, and push it up against the motor body.
Separate the aluminum sleeve from the screw shaft. After separation, use a small knife, if necessary, to carefully remove the hardened gummy stuff surrounding the end of the screw shaft.
Using rubbing alcohol and Q-Tips, clean every bit of residue from the motor shaft, the now separated screw shaft, and the inside of the 3/32" ID aluminum sleeve. Poise these three items on your desk and wonder, as I did, how you will ever get them to work again.
If the aluminum sleeve is warped, due to handling during removal, use the blunt end of a 3/32 bit to reform it, by setting the sleeve on end, and, gripping the bit with needle nose pliers, force the blunt end into the sleeve; then work the sleeve back and forth with your fingers until it can easily be removed.
Note that the end of the screw shaft has an indentation that accommodates the motor shaft. The motor shaft and screw shaft will meet here, while entombed within the oversized aluminum sleeve. Practice inserting the motor shaft into the screw shaft a couple of times so that you can get used to the feel.
Now, practice propping the motor against something stable so that the shaft is pointing straight up...a small soft-edged bench vise would work well for this...I just kind of prop the motor up against the inside edge of the calculator case from where the thing had been removed.
Again make sure that the tiny nylon washer is on the motor shaft and is snug against the motor body.
The DAP 100% SILICONE Household Adhesive will be acting as a filler and torque/vibration inhibitor between the inside of the oversized 3/32 I.D. aluminum sleeve and the motor shaft/screw shaft connection.
Put a big old glob of your DAP 100% SILICONE Household Adhesive on a scrap of glossy paper. Have a tissue handy. Have the drive motor standing by within easy grasp.
Taking that little aluminum 3/32" ID aluminum sleeve in hand, plop it onto the silicone blob from both ends, packing it up. Don't try to be precise. Just glom both sides.
Grab the tissue and, without trying to be too exact, wipe around the outside of the miniscule sleeve to remove access silicon material. Wipe across the open ends as well, the silicon will retreat into the sleeve, which is just what you want, as it will be displaced by the motor shaft/screw shaft assembly.
OK. Here we go. Place the silicon filled sleeve over the motor shaft, and, while working to keep things centered using the thumb and forefinger of the hand that's holding the motor, use the thumb and forefinger of your other hand to insert the screw shaft into the other end and, by feel, mate the screw shaft and motor spindle, within your silicon filled aluminum sleeve.
If step 16 finds you with silicon everywhere and nothing meshing, or even with the tiny sleeve sticking to your fingers or falling on the floor, take a breath. Work through the steps. Clean everything thoroughly with rubbing alcohol and begin again at step 10. Clean things quickly...silicon cures over 24 hours, but begins to set in minutes.
You've meshed the motor shaft and the screw shaft, all within the silicon filled aluminum sleeve. Prop the motor up so that the screw shaft is pointing up, and over the next several minutes adjust the outer sleeve by using your thumb and forefinger to gently turn the shaft, and keep readjusting the sleeve so that it looks centered.
Also, while keeping pressure on the screw shaft to keep it engaged to the motor shaft within the silicon-filled aluminum sleeve, pull back up on the sleeve a bit to keep it just from touching that nylon washer. This is easier than it sounds, and will make the sleeve seat up against the thicker part of the screw shaft. You know you're right when you can see a slight separation between the sleeve and the nylon washer that is sitting against the motor body.
Once adjustment is difficult, due to the silicon within the sleeve beginning to set, its time to throw caution to the wind.
Let the silicon set for about three hours.
Holding the case on edge, reassemble the motor into the frame, over the drive wheel, with the two tiny slotted screws. Reinsert the red and black wires into the power supply board, and reset the power supply connector into the main board.
Inspect all wire connections, and, holding a battery pack against the battery compartment contacts, turn on the calculator and make sure everything works.
If no display, switch things off, remove the battery, and carefully check all power supply connections to the motor, R/W head, battery leads, etc. Although all wires attach with tiny brass clips to the power supply board, everything is soldered to the main board underneath. I've had to re-solder power leads a couple of times while replacing the gummy wheel. Have a solder-sucker handy, and use a low-wattage gun.
If all is well, and the display is lit, and a couple of simple calculations succeed, keep pressure on the battery pack against the contacts, take a breath and test a card. Mine went through like it was on fire. Motor noise was minimal, and the card took no coaxing at all. Smooth as silk.
Turn the calculator off, remove the battery pack, and reassemble the back of the case.

God knows how I came up with this last ditch, here-we-go, let's get Western solution. But oh, man, does it ever work! Satisfies all of the original criteria. Cards don't stall, and motor torque is buffered against the inherent frailties of the neoprene drive wheel.

Good luck!

Johnny Nestor

jnnestor@pioneernet.net

Go back to the main exhibit hall