Original TI30 circa 1977 key bounce problem

[Thomas Plummer]

I have four TI30 calculators from 1977 era. Some have the key bounce problem. Over the years I have tried to analyze and fix this. I thought I had a permanent fix which was based on the premise that the keyboard pulse and sense wire welds were bad.

I was wrong. I destructively tested a keyboard and found that the welds welds were so good in fact that I had to use a lot of force to rip them apart. I have lately discovered that the key bounce problem seems to be greatly affected by the foam sheet between keys and dome switch assembly, for instance if I remove the sheet, key bounce really increases. So I am left pondering the effect of 46 years on the physical properties of this elastomeric foam sheet as well as the properties of the adhesive plastic sheet which encapsulates the metal dome strips. For now I plan to try other sheet material to see if I can eliminate the bounce.

I think the TI engineers debounced the keys in the main chip based on the properties of new keyboard material and that they were very close to the edge of bouncing which the passing years made evident.

I have researched this problem extensively on the web without finding a good fix.

Would like to hear from anyone who has worked on this.

Best regards,

Thomas

[osa...@gmail.com]

This keyboard behaviour got worse with later TI models, in fact some of the LCD models were so bad that they became useless within just a few years! I would be intrigued by what you find out. I did read somewhere that the behaviour was improved by spraying Kontakt contact cleaner into every hole on the keyboard. This would imply that the debounce code performed worse as the wires tarnished or veneered. The domes and adhesive sheet were common on models without the wire keyboards and I believe that these don't tend to suffer from the same issue.

To replace the old foam, which I have had no luck in finding, I use the thinnest Neoprene sheet I can find, or a piece of that rubber exercise band that you may get issued by a hospital physio department.

If the keyboard is just a series of X and Y wires, and you have the means, would constructing a simple PCB to replace it be an option?

[Thomas Plummer]

Good morning and thanks for your thoughts. Doesn't everybody get up at 4:30am to talk about calculators?

I think oxidation may not be involved since there were reports of new TI30 with the bounce problem and the wires and domes appear to be gold plated. The foam sheet is not precisely located and can shift quite a bit so that simply taking it apart and putting back together will bring a new section of foam under the button nibs or flipping it over. 

I may try your neoprene suggestion. Curiously absent is any discussion of using capacitors on the wires or any oscope signal investigations.

Best regards 

Thomad  

Thomas C Plummer

Whaling and Technology Historian

   

--
You received this message because you are subscribed to a topic in the Google Groups "OldCalculatorForum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/oldcalculatorforum/nuQPfwXUt0Y/unsubscribe.
To unsubscribe from this group and all its topics, send an email to oldcalculatorfo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/71b32194-47e6-4769-a7bf-5ff8f43c3c19n%40googlegroups.com.

[Thomas Plummer]

I have some fresh research and restoration tips for the TI-55-II calculator

16 FEB 2024

TI-55-II restoration

I have recently completed the restoration of a TI-55-II calculator circa 1984

My research showed that this was the worse model for key bounce problems and so I decided to focus on this as a challenge to restore. Based on other restorations I had done on TI30 and TI55 calculators circa 1977, I used my acquired skill on the TI-55-II

As received I could not enter the digits 1-8 without bounce after ten attempts so this one really had the defect.

GENERAL NOTES

On disassembly I noticed that the foam sheet was perforated where it touched the keys. It was also a bit sticky. I folded it over to 4x thickness and squeezed it and observed that it simple crushed to a thin layer and stuck together, the layers sticking to each other. Over 43 years the resilient foam had become an adhesive.

I had a TI30 foam sheet that was resilient and non sticky so I used it.
The process was as follows

Remove old foam and clean the key nibs with alcohol
Create and install 9 small 10 mil plastic pressure pads over welded spots on dome switch assemble
Put TI30 foam sheet back in, a little short but still covers keys
Cut a 5 mil mylar sheet the same size as the keyboard white plastic and place over foam
Reassemble

TESTING

Exercise the keyboard by firm pressing all keys ten times
Lay calculator on hard surface
Enter the numbers 1-8 ten times and look for bounce, repeat this 10 times
Also enter 1-0-1-0-1-0-1-0, and 9-6-9-6-9-6-9-6 ten times per set to get 0 and 9 coverage

In my case no bounce was seen in 800 consecutive key strokes.

Note on welds

It has been observed by another researcher that welds have been defective in his TI calculator project, that is to say they were no longer welded together but just touching. Now this could have occurred at the factory or over time. I have destructively tested two TI30 keyboards and found solid and strong welds requiring a lot of force to break, so I cannot confirm his observations in my limited testing.

Now I found that the behavior of this TI-55-II without pressure pads was different when held in the hand or laying flat on a hard surface. In the hand occasional bounce was seen, and almost none when laying flat. I concluded that slight twisting and bending on the dome switch assemble was occurring. The unit is slim and thus unable to resist bending and twisting by hand use.
I surmised that this change in behavior could be do to some defective welds, just touching.
To reduce this effect I wanted to apply some pressure over the welded spots to force the two metal contacts together. This was done with 10 mil small plastic shims and a little superglue to hold them in place. The pads focus pressure on the welds. The layers in order are: a sheet of foam, then a sheet of 5 mil mylar pressing against the dome switch assembly and leaning on nine pressure pads.
Bad welds may be the reason but it could also be slight changes in the way the snap dome switch hits the wire contact. I have a method of directly testing welds that I have not yet put into practice, this would settle the issue by direct observation.

The pressure pads greatly reduced the bounce when hand held.
I have not been able to completely eliminate the hand held rare bounce but I will say that the unit is very usable for all tasks.

For 40 years this problem has not been solved. I seem to have a talent for doing these kinds of things, the proof being that this restoration was completely successful.

I hope these notes help you fix your own TI calculators.

Best regards,

Thomas Plummer

New Hampshire

[osa...@gmail.com]

A great description and well done in doing this. I got a little lost in trying to imagine the layers, would it be possible to post some pictures. It reads like you are adding a spacer layer on top of the dome and foam and supplying extra plastic beneath each of the keys to reach through this new layer. Am I completely wrong?!

[Thomas Plummer]

Thanks for reading my post. 

The foam sheet lays on the keys as in the original, then the mylar sheet lays on the foam. If you were to look at the dome switch assembly you would see 9 small rectangular pieces of 10 mil plastic and these are located directly above the welds. The whole purpose of these is to focus a bit more pressure from the foam/mylar system on the weld contacts in the case that a weld has failed and the contacts just touching. They do not affect the dome switches at all. 

I will put together a photo layout so all this becomes clear

Best regards,

Thomas

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f8799606-3f0e-4abc-b703-0951406d7734n%40googlegroups.com.

[Thomas Plummer]

I have examined the welds and they are fine. Careful examination revealed that two welds had forced the wire to be bent up and shorting to nearby dome strips.

This caused shorts to adjacent dome strips when the keyboard was slightly flexed or held by hand or RST key pressed.

There is very little clearance between the wires and the surface of the keyboard, they are recessed a tiny bit and each wire has a bump formed in it to just touch the dome strip for welding. I noticed that the ears of the dome strips at the welding location were pushed up above flush and thus creating a short.

The solution was simply to push the weld back down until the dome strip ears became flush, thus providing clearance.

The reason the pressure pads worked is not because of bad welds but because they leaned on the bent up wires and cleared the shorting.

Here is a link to the project pictures

https://drive.google.com/drive/folders/1PYO9XYEaUalSiuvjJHhaIrKklbj3W4Oc?usp=drive_link

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f8799606-3f0e-4abc-b703-0951406d7734n%40googlegroups.com.

--

Thomas C Plummer

603-479-7946

Whaling and Technology Historian

thomasp...@gmail.com

[Thomas Plummer]

TI-55-II extra notes

In my last post I showed two bent keyboard wires and how to fix them. I am pretty sure that I caused this damage.

While prying the keyboard loose with a small screwdriver I unknowingly forced a couple of wires back into the assembly a small bit.

This force was relieved by the wires bending up a little.

These wires stick out a little at the keyboard end and the two in question were the closest to that end.

Also the calculator showed new operational defects that were not there before such as the RST key performing reciprocal operation as though the key above it had been pressed, a shorting defect.

So now I am very careful to avoid the wires when removing the keyboard.

On Fri, Feb 16, 2024 at 11:33 AM osa...@gmail.com <osa...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f8799606-3f0e-4abc-b703-0951406d7734n%40googlegroups.com.

[Thomas Plummer]

TI-55-II materials suggestion

I realize that mylar sheet and original TI30 foam sheet are not available to most calculator restorers.

I have found a substitute material that seems to work even better.

It is a single 9x12 sheet of 0.030 inch polyester (fake felt)

Cut to the size of the white keyboard and install, no need for mylar or original TI30 foam.

I bought this at Michaels craft store for 0.49 dollars.

The thickness of these sheets is highly variable so bring your calipers and get the one closest to 0.030 inch

So far it has given superior performance even when handheld

/media/thomas/CANON_DC/DCIM/100CANON/IMG_0549.JPG
/media/thomas/CANON_DC/DCIM/100CANON/IMG_0548.JPG

Best regards,

Thomas

On Fri, Feb 16, 2024 at 11:33 AM osa...@gmail.com <osa...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f8799606-3f0e-4abc-b703-0951406d7734n%40googlegroups.com.

[Thomas Plummer]

There was a person inquiring how to take apart a TI-30 and TI-55 circa 1977 .

I'll show you the easy way I do it in this Youtube video

https://youtu.be/BClGoV0PfkM?feature=shared

On Fri, Feb 16, 2024 at 11:33 AM osa...@gmail.com <osa...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f8799606-3f0e-4abc-b703-0951406d7734n%40googlegroups.com.

[Thomas Plummer]

Please see my new brief report on the restoration of 9 TI calculators.

https://drive.google.com/drive/folders/1UEQ5tfRO4JZ9rIkQG5jjufNEpwZ7Dljw?usp=drive_link

Thanks,

Thomas

On Fri, Feb 16, 2024 at 11:33 AM osa...@gmail.com <osa...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/f8799606-3f0e-4abc-b703-0951406d7734n%40googlegroups.com.

[osa...@gmail.com]

Brilliant, many thanks Thomas. 

While all this has been going on I also had a look at a TI-51-III that was in very poor cosmetic condition with a keyboard that barely worked at all. There is no battery corrosion and the PCB is in good condition. The foam is still quite good, considering its age. Not powdering and still quite squishy. Emboldened by your previous mails I found that several of the keys were not making any circuit at all when pressed. I started peeling the cellophane layer that covers the domes. Horrid stuff, looked like is was slightly rubberised, left behind a nasty black sticky layer that took quite an effort to remove. The domes are all linked together in strips. About half these strips just fell away, the crimped joints to the up / down wires having failed. I decided to remove all the strips. They didn't put up much of a fight! I then relaid the dome strips, but this time I soldered the crimp joints. I then applied strips of plain old electrical insulation tape to stick the domes down. Reassembled with the original foam layer. Keyboard works perfectly, absolutely no double presses and they work first time. There is some LED display ghosting, which is almost certainly a slight short somewhere in the keyboard. If the calculator was in better condition I'd investigate further, but the exercise to repair and improve the keyboard has worked better than I expected.

[Thomas Plummer]

Wonderful to read your report of successful repair. You clearly had bad welds, something I haven't yet encountered in the nine units I have.

I know first hand what a difficult and technically challenging repair you have achieved. I have destructively examined about three keyboards and found the black sticky tape as well. I believe this is mylar with black adhesive which resisted five different solvents and grudgingly could be removed with acetone. The clear dome switch assemblies are much nicer to work with. 

I have also recently found a hard button press for #5, /, and = key was needed on one TI55. This defect was new and I cannot account for it. A little surgery and tiny amount of DeOxIt under the dome with immediate blowout with air and excess squeezed out with paper towels soon put it right. First time I have needed to do this.

I hope my advise is not merely anecdotal and only good for my nine units. In your case you have proved that the foam was not causing any of the problems.

Well done on a hard repair.

Best regards,

Thomas

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/001c84b4-ba68-42fd-adba-7213ded5e5ddn%40googlegroups.com.

[Thomas Plummer]

TI-55-II exhibits bounce problem again when handheld

What I am finding is that 800 key strokes without a single bounce is no predictor of future performance, so what is going on here??

I picked up the TI-55-II today for a keyboard check and it bounced right away three times. I am unable to explain this change.

Now this unit has been exercised a great deal. Is it possible that this clears away snap dome/wire oxides for several days, only to return?

This also has the 0.030 inch felt sheet in it as a replacement for the original foam.

I wanted to try DeOxit on this but its a bit tricky. I applied DeOxit under each dome (45 in all) with a very small syringe and sharp needle taking care that I didn't DeOxit myself in the process. A little pressure on the plunger provided a tiny amount of fluid under each dome. The dome switch is now drying out under a light to drive out any remaining DeOxit.

I will report on the results of this technique but I am hesitant to say what works and what does not work having had to reverse myself several times in this dialog.

Here are pictures of small syringe and tiny needle for applying DeOxit under the domes.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/001c84b4-ba68-42fd-adba-7213ded5e5ddn%40googlegroups.com.

[osa...@gmail.com]

I like your technique. Are you based in the US? Here in the UK syringes and needles are quite difficult to get hold of. The US has a better supply of old calculators and at more modest prices, in the UK about the only source is eBay, we don't have estate sales nor yard sales in the same way as the US, and the prices are generally too high!

[Thomas Plummer]

Thanks for your note. Yes New Hampshire US. DeOxit had no effect on problems. The subtle and perverse behavior I have seen during hours of working on these TI calculators tells me that I may not have a permanent fix.

The test of time is time itself so I plan to periodically test the lot and take notes.

I do have a nondestructive way to test each weld for a solid connection. I am considering a video to demonstrate this. It requires very close observation of the dome strip ears at each weld looking through the holes in the back of the keyboard while direct pressure is applied to the ears with a needle-like tool and seeing if it moves away from the wire it is supposed to be welded to. I used this method to confirm that all 9 welds in the TI-55-II were sound.

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/b356341b-4b2d-40bd-a568-98ce6fe74250n%40googlegroups.com.

[osa...@gmail.com]

I no longer consider these welds as sound at all. Since detaching and soldering each weld point I have had no issues with the keyboard at all. I have slight ghosting on the LED display which is almost certainly due to a rogue contact somewhere. I shall endeavour to find out which dome strip is responsible.

[Thomas Plummer]

Yes, you did have bad welds. As to ghosting I've not seen it yet with my units.

The LED display is heavily multiplexed and may be using the same timer and sequencing as the keyboard.  Your thought of a partial short in the keyboard may be spot on. The sense wires are only recessed 0.006 inch below the surface and clearances are tight all around. If you used rosin flux for soldering then all is well, if acid then a complete washdown with water should remove all sneak paths for current. The keyboard is not hermetically sealed as some have stated. If you were to direct a stream of water at the holes in the back of the keyboard I think you would do a thorough job, then a blowout with low pressure air followed by a mild heating for 6 hours to completely remove any water .

Good luck,

Thomas

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/9bb65601-8b42-459e-92ef-002aaf59c630n%40googlegroups.com.

[osa...@gmail.com]

I took the tape off the top and noted that the ghosting appeared as I pressed anywhere around the second row, second key from the right. The edge of the dome appeared to be making a very slight contact with the wire as it came in to the dip for the key.  I applied an iron and moved the strip a tiny fraction and the problem went away. The keyboard and display now work perfectly every time. I'll leave it for a while and see what it's like. I haven't cleaned the wires at all, but the underside of the domes may have benefitted from the sticky removal. If this fix is persistent I think it's fairly conclusive that the crimped joints are a weak point in the design.

[Thomas Plummer]

Well done. You have good observation and troubleshooting skill. 

To view this discussion on the web visit https://groups.google.com/d/msgid/oldcalculatorforum/00c76bda-f529-49ff-b2a9-0e43c0a8a353n%40googlegroups.com.