[HP48GX] Testing repaired keyboard

[fnas]

Hello,
I need advice to do the testing on a HP48GX. I have opened the calculator because the ON key wasn't working. So I placed under the flatpack a piece of rubbery scotch tape to be thick. But before closing the calculator I want to test if everything works. I connected the calculator to terminals of a 4.5V stack. Now as I can see if the PCB and flatpack makes contact? The negative pole is connected to the PCB only through the twists hooks? I attach same photos
Tanks,
fnas

[mfleming]

Hi,

Here's a link to an article on the keyboard schematic of the Pioneer style calculator. Although not the same as that of the HP48 series, the key matrix arrangement follows the same pattern.

http://www.hpmuseum.org/cgi-sys/cgiwrap/...i?read=451

Have a look through some of the old articles in the link below to see if you might find the same information for the HP48. I think your testing approach would be to identify a column or row trace, then measure the drop in resistance on the other traces as you press keys. As the above note says, resistance should drop from 10 Mohm to under 5 Kohm.

http://www.hpmuseum.org/cgi-sys/cgiwrap/...ticles.cgi

Getting a good contact with the flex circuit is the tough part - you don't want to hold two probes against the circuit while trying to press a key with a third hand! To arrange semi-permanent contacts with the flex circuit, try taking an appropriately sized double-row header and bending the pins on the long side together to form something like a clamp. Then you can slide the flex circuit between the pairs of pins that are touching each other.

The old Pioneer flex traces were 0.1 inches apart, so you could use double-row headers like those for an Arduino. Not sure of the HP48 series though, they could have a metric pitch.

Good luck!
~ Mark

[TomC]

Keep in mind that in the 48S and 48G series, the 'ON' circuitry is NOT part of the regular keyboard matrix.

The 'ON' key provides a direct battery connection to a pin on the main processor IC.

Schematic here:

http://www.hpcalc.org/details.php?id=3543

TomC

[TomC]

Also, you may be interested in some pics posted here:

https://sites.google.com/site/tchrapkiewicz2/hp48

TomC

[fnas]

Hi Guys,
to test the contact between keyboard and PCB I need to give the negative pole of stack to the PCB. Can you tell me if the PCB take the negative pole from the contacts between the twists hooks and the copper pad A, B, C, D in the picture?

http://imgur.com/a/8txGX

Thanks,
fnas

[mfleming]

Yes indeed, ground is supplied through those twist posts to the main board. Be sure to twist them back enough to reestablish electrical contact (one of my newbie mistakes when doing my first repair job).

It sounds as though TomC's observation about the ON key wiring should guide your testing. There could also be a mechanical root cause for your problem in the form of dirty key contacts. There are more posts addressing that problem if you eliminate an electrical fault as the root cause.

[fnas]

(07-25-2016 01:58 AM)mfleming Wrote:  Be sure to twist them back enough to reestablish electrical contact (one of my newbie mistakes when doing my first repair job).

Hi Mike,

This is my main concern in the testing phase. So I wouldn't twists the hooks only to test the contact between the flatpack and PCB. I fear that the hooks can break and the copper on the pads can be scrape. Therefore I'm thinking to develop a specific tool to obtain the electrical contact between the A, B, C e D pads and the ground.

The tool I have in mind consist of four pieces of a copper pipe with 5mm internal diameter and 20mm tall. Then I will put the four cylinders over the A, B, C, D pads and I will connect electrically its to the ground.

With this tools I can give the ground to the A,B,C,D pads and can push the PCB against the keyboard to control the contact PCB/flatpack.
What do you think about?

[mfleming]

Hi fnas,

You should need only a single connection to a pad to supply ground to the CPU board. Alternatively, you might use a jumper wire from a ground point on the board to the case ground. Placing the board over the twist ties and pressing down with sufficient force would be enough to test the electrical connection.

As one member reported, the twist ties may be good for five or six open/close sequences, but once they break you're in trouble

~ Mark

[fnas]

(07-25-2016 02:32 PM)mfleming Wrote:  You should need only a single connection to a pad to supply ground to the CPU board.

may be I don't understand well: if I connect one, and only one, pad to the case ground to supply ground to the CPU board, the other 3 pad connection are unnecessary?

If I'm not mistaken I see that the A, B, C and D pads are independents and each is connected to the ground only through its twist tab. So each components connected to the A, B, C, and D pads, have its ground connection. Therefore I'm thinking to connect the 4 pieces of pipe together and then one to ground on the case.

In other words: the PCB needs 4 independent connections to the ground, isn't it?

Tanks,
fnas

[mfleming]

One way to determine if the four pads are independent of one another would be to use a continuity tester on pairs of pads. As a practical matter, I would expect the PCB to have a single ground plane that uses one or more twist ties as a return path to the batteries.

~ Mark

[fnas]

Hi Mark,
I have checked and your are right:

Quote: the PCB to have a single ground plane that uses one or more twist ties as a return path to the batteries.

the only pad that isn't connected to the ground plane is that I marked in the pic as A.

Now I have twisted the tabs but I have anyway a bad connection between the PCB and flatpack keyboard. So I need to reopen the twist tabs :-( and I need to find something to put in place of the rubber. I took the measurements of the existing rubber and its section is about 3x4 mm. I noticed that the original rubber is flexible enough to be crushed but also stiff enough to push the flatpack against PCB and now I'm looking for something on the market that can be good for size and consistency.

I had another problem when I closed the unit: when I powered the unit almost the whole screen was switched on but I don't know why. (it could be dust or overpressed the PCB against the flatpack?)

http://imgur.com/a/vj8ne

[mfleming]

Hi fnas,

The sort of material you are looking for is a stiff but compressible foam rubber. One suggested source is a thin mouse pad. You'll see it has the right consistency to match the rubber strip you are replacing. By sheer luck, the anti-static pad I have on my work surface is just the right thickness to match the height of the rubber strip. The pad now has a number of rectangular notches along the bottom

You will need to take some care in removing the old strip, as it is glued to the case and will tear easily if mishandled. The approach I have used is to take the thin blade of an Xacto knife and insert it between the top of the strip and the plastic case, then press down to pull the strip away from the case and break the glue seal. Do the same again by placing the blade along the bottom of the strip and press upward. If you repeat this once or twice, you will loosen enough of the seal so that you can peel the strip from the case without tearing it or leaving small bits if it behind.

Have a look at the article "Calculator Restorations" by Geoff Quickfall in the #19 issue of HP Solve for some excellent guidance. You can find the issue here:

http://h20331.www2.hp.com/hpsub/download...e_2010.pdf

When you replace the CPU board, press down on it before closing the twist ties so you get a good contact between the keyboard flex circuit and the PCB, and the same between the LCD zebra stripes and PCB. The twist should return to a 45 degree angle or better. Use your multimeter to verify continuity between the metal case the the PCB pad that the twist should touch.

Good luck!
~ Mark