How best to repair broken tang on keypad switch (HP3330B)

[intabits]

So ThunderThumbs was removing the keytops from a recently acquired HP3330B Automatic Synthesizer to wash them, when he broke two of them.

One is just a cracked open keytop, and should be quite easy to repair.

But on the other, the "tang" on the keyswitch that clips into the keytop has snapped off.
The switch's sliding shaft reduces from a wider shoulder to the thinner tang that enters and holds the keytop.

So the tang needs to be reconnected to the shoulder, but it's very thin and weak. To allow for that, the repair would probably involve some widening of the shoulder and/or tang to provide some extra strength.

But less than 1mm of the tang is outside the keytop, and when fully depressed, less than 2mm of the shoulder is outside the switch. So any method of rejoining the tang onto the shoulder must involve less than 3mm of material wider than the shoulder, or else the key either won't depress fully, or sit higher than the other keys.

I'm hoping someone has a clever way of rejoining these two. Thanks for any help or suggestions you can provide.

The cracked keytop and the broken tang, and a good keytop that the tang would mate with.


The broken keyswitch (top left) missing the tang.

[intabits]

After posting, I had a thought. Could this work?
 
The switch will probably need to be removed from the keyboard, to make it accessible for repair.

1/ Slip a loop of (say 3mm) wire around the shaft/shoulder to account for travel.
2/ Push a bit of paper or thin plastic over the shoulder, making a hole in it, and so that it sits snuggly around the shaft. (yes, I know)
3/ Slip a largish 2.5mm thick washer over the shoulder. (Washer hole is big enogh to fit over the shoulder)
   And washer is thick enough to extend past the top of the shoulder.
4/ Stick/melt/whatever the tang onto the shoulder
   Now there is a "pool" space around the tang and shouder which can be filled with epoxy resin.
5/ When hardened, remove wire loop. The keytop hides the washer and repair.

No?

[Zucca]

After learning this:



I never break stuff anymore.

I would take a little piece of wood and with a lot of patience... make a support for putting back the keypad....

[tooki]

Just a tip, if you google using the correct terminology (keyswitch “post”, not “tang”, and “keycap”, not “keytop”), you’ll find a lot more discussions on such repairs, since it’s a fairly common thing to break in vintage computers.

[KL27x]

For the tang thing, I would cross pin it. Tiny drill bit, tiny brass or steel pins. At oblique angles from either side and all around, so they can't come out. Maybe thin piano wire. 

If it's abs or styrene, I'd put a drop of acetone in the seam and see if that helps. If not abs/styrene, then I'd super glue it before drilling and pinning.

I doubt this would work very well. It's just what I would do. Or try.

 

If you have the ability to make a new tang out of aluminum or brass, you could maybe extend the base a bit and cut out a few teeth/spikes in it. Drill undersized holes, then bed it in there. Maybe even heat the brass/alum with a torch and press it in. Pull it back out after it cools, then superglue it in.

[intabits]

Thanks for the replies.

Just a tip, if you google using the correct terminology (keyswitch “post”, not “tang”, and “keycap”, not “keytop”), you’ll find a lot more discussions on such repairs, since it’s a fairly common thing to break in vintage computers.

I couldn't find much for these type of switches, most hits were for the more common and more sensible "cross" profile posts (which btw, seem to be called "stems").
(For my switch type, I'll continue to use the term "tang" though, as I think it's useful to distinguish it from the thicker stem.)

The only usable hit I found was for HP9830 calculators, where someone used a loop of piano wire as the tang, and melted it into the stem.

...If you have the ability to make a new tang out of aluminum or brass...

As described just above, a similar idea to yours but with a wire loop.
I was skeptical of it's strength against sideways bending, but I just tried a loop of stiff wire in a keycap, and it's fine.
So I think I'll give that a shot. Pics to come if it works...

I've removed the keyboard, and can now see that the situation is tighter than I first thought. When depressed, there is only a mm between the keycap and the switch body. So my earlier idea would be harder, but the wire loop thing is still very viable.

[Rasz]

3d printer is always an option
EDIT:melting nails into plastic + epoxy is the way I usually go about fixing those types of defects

[intabits]

Thanks, interesting video. But different keyswitches though.

I've gone with the wire loop fix:-

A little messy but it works well. (If I had to do a few of these, I reckon I'd be doing it very nicely by the 3rd or 4th one)
Also shown, the broken keycaps. I super-glued them, plus some heatshrink around the post to keep it all together.

All done now, just like a bought one!

The key sits about 1mm further out than the others. I could mess with the wire loop to fix that, but I think I'll just leave it as is for now.

[coromonadalix]

nice repair trick,  instead of the shrink tube, maybe an epoxy filling would be okay

[precaud]

Your 3330B has a beautiful front panel! Most of the paint has flaked off of mine.

[intabits]

Are they prone to that?
Mine isn't bad, but not perfect. There are a few nicks, and a large patch not very visible in the photo, left of the left most buttons.

[precaud]

Are they prone to that?

I'd say yes, they used the same paint as they do for metal surfaces and it doesn't adhere as well.

Mine isn't bad, but not perfect. There are a few nicks, and a large patch not very visible in the photo, left of the left most buttons.

Yours looks very nice. I bet if you remove that sticker, the paint will come off with it, though...

Do you use yours with a 3570 or 3571 ?

[intabits]

... they used the same paint as they do for metal surfaces and it doesn't adhere as well.
... I bet if you remove that sticker, the paint will come off with it, though...

Do you use yours with a 3570 or 3571 ?

OK, I'll leave the sticker there for the moment. (It serves as a reminder that the unit is 30 years overdue for calibration!)
And I'll also take a high quality photo of the text around the slide switches, in case I one day it can be used to make a replica.

I bought this unit on ebay, just to have a sort of precise frequency reference, not knowing much about it.
But when I learned that it was part of a system, I also got its companion 3570. But I know very little about that. Not even sure if I've turned it on.
I plan on playing with them at some time, to better understand Network Analyzers and their usage. For the moment though, the 3570 is pretty much wasted on me...

Speaking of calibration, I currently have limited means of checking the 3330B, but what I have done seems to indicate that it's output is always around 20Hz high - a constant frequency offset. (The picture shows the scope saying 400Hz too high, but I think it's less than that)  I haven't investigated whether that sort of thing is possible and a normal, expected point of calibration, or a fault. Do you know about that, just off the top of your head?

[precaud]

Speaking of calibration, I currently have limited means of checking the 3330B, but what I have done seems to indicate that it's output is always around 20Hz high - a constant frequency offset. (The picture shows the scope saying 400Hz too high, but I think it's less than that)  I haven't investigated whether that sort of thing is possible and a normal, expected point of calibration, or a fault. Do you know about that, just off the top of your head?

I've worked on a few over the years, and I haven't encountered anything like that. Definitely check all the clock freqs. IIRC, these units are pretty straightforward to calibrate, no exotic instruments required...

[andy2000]

If the key cap breaks again, I would just fill the whole thing in with epoxy.