Read?Write HP 67 card reader

[Geoff Quickfall]

After:

1) replacing two key snaps with a donor snaps from a dead 55 harvested board.

2) replacing the dampening couple with a silicone and insulation repair.

3) Replacing the pinch roller with a milspec silicone tubing.

4) cleaning the keyboard bezel, re-lacquering it in flat clear.

5) cleaning the keyboard and logic PCA connections.

The calc writes. Card tested in working 67. The calc does not read it’s own cards or any others producing the standard Error on each read.

The current draw on the card reader is correct. The card reader head reads and writes with a known functional spare board swapped for the suspect board which is pictured.

Apparently, with searches, these two caps are suspect and may not have the correct ESR. Before diving into a de-solder and etc. what are the replacements. Are they polarized, would smd mounted tantalums be viable or electrolytic.

At the moment looking at 16v 47uF and 22u F to replace. See picture.

Any help would be appreciated. This is for the BOOK!

[Geoff Quickfall]

Problem solved.

These two tantalum caps smooth out the power supply to the read/write IC when the card reader motor initiates with a card inserted.

If they fail in smoothing the circuit you will get a spike and the READ function will fail.

Will do some probing of voltages and scope this to see which has failed.

But will replace both with like tantalum polarized 16v through hole caps

[John Garza (3665)]

Hi Geoff,

Good job!

The problem you describe is usually caused by capacitor A (the 47uF one) in your picture.
But as usual, it's good to replace them ALL once you're in there as preventive maintenance.
And HP solder lands have little squares for the positive leads, just to make it easier. The old caps themselves usually have a little dot painted on the positive side.

Also, I know it's practically canon to use wire insulator for the coupler replacement. But I've found an easier way. Just dig out all the remains in the cylinder sleeve. Pack it with RTV gasket maker compound and re-assemble the unit forcing the motor axle down through the goo. Then use needle tweezers to clean up any spillage. Let dry overnight. The result is a rubber-like replacement. Voila!

-John

[Geoff Quickfall]

That works but how do you get it to be dead centre? Otherwise it will induce a wobble.

But if it works, it works and will be mentioned with a few other fixes. Me, I have purchased some rolls of thin copper wire sheathed in hi temp resistant silicone. It is my got to wire now as it does not offend my OCD with melted plasitc.

But the silicone best mimics the original rubber for dampening. And I can center the motor pivot in it. With the aluminum sleeve and the outer wire insulation it compresses onto the motor pivot and the worm gear.

But if it works it works.

Further, I have that fix in a bunch of these from 2009 so it looks like it lasts at least as long as the decomposed rubber did with HP.

Cheers, Geoff

[John Garza (3665)]

If you look at the worm gear end that fits on the coupler, you will see hole in it. Push the motor axle into it and it will be straight. When the compound dries, you can ease the shaft out a bit and it will be in a straight line.

And this stuff lasts a long time. I've used this since the early 2000's as well and never had to replace one. I've used the silicone insulation approach as well, its very low friction material good for this application with plastic on plastic movement. It works too but is more tedious. The gasket compound is easier to use and drastically reduces problems with friction and heat (heck, the stuff is used to make gaskets in automobile engines). The only negative with the gasket compound is that it has to dry overnight to solidify.

All the Best,

-J

[Geoff Quickfall]

Well, will try it tomorrow. Got to try before it goes in the book!

[AnnoyedOne]

(10-04-2024 09:43 PM)Geoff Quickfall Wrote:  Any help would be appreciated.

Those two parts look like tantalum capacitors to me. Tantalums are always polarised. There are non-polarised electrolytics but not tantalums. They appear to be part of a switch mode power supply.

https://en.wikipedia.org/wiki/Tantalum_capacitor

Unlike electrolytics tantalums don't obviously "dry out" with age. Perhaps faulty anyway. Can't hurt to replace them (don't insert them backwards--they don't like that!). Not costly.

A1

PS: I see two other (smaller) tantalums on the (top center of the) PCB. I'd replace those as well.

PPS: Looking at the date codes on the IC's I see

7805FC
7912
7902CV
7844
7842
7915AL
7909AD

so my guess is that this board was made in 1980/81--42+ years ago. I don't know what the failure mode(s) of tantalum capacitors is/are but I'm sure they're old and tired They need to be "retired".

[Geoff Quickfall]

Hello Annoyed ONE!

Yes, tantalum, my capacitor recognition level is low but learning.

These are now ~ 50 years old and are pretty reliable BUT, these two specifically will take a load from the switching on of the card reader motor when it is in the READ mode.

They are supposed to smooth the flow to the IC and when they age (apparently) they allow a voltage spike into the IC disprupting the read command.

The symptoms are, after the pinch roller and dampening couple are replaced and the pinch roller height is correctly set, the ability to WRITE but the READ produces intermittently and then in the worse case an ERROR.

The others are not involved in the spike smoothing but may affect the calc in another manner.

I will remove them from the circuit, replace them and test the old ones for ESR and capacitance plus leakage.

Cheers and thanks.

Parts ordered from digikey.



Thanks to Anthony Nixon (TEENIX) for this extremely helpful circuit diagram which can be found at his site (PLEASE check it out):

TEENIX site!

[AnnoyedOne]

(10-05-2024 03:53 PM)Geoff Quickfall Wrote:  These are now ~ 50 years old and are pretty reliable BUT, these two specifically will take a load from the switching on of the card reader motor when it is in the READ mode.

From Wikipedia tantalums can fail from

"Field crystallization"

https://en.wikipedia.org/wiki/Tantalum_c..._mechanism

New to me.

A1

[Geoff Quickfall]

From years of radio work, that is new to me too!

So they crystalize and electrolytics dry out.

Well, one more addendum to the book!

Cheers

Willdo some research also.

[AnnoyedOne]

(10-05-2024 03:53 PM)Geoff Quickfall Wrote:  ...test the old ones for ESR and capacitance plus leakage.

Geoff,

I've never bothered testing old components. In the waste they go!

Unlike you I don't have a "lab". I used to do any personal stuff at work where I had access to some pretty nice equipment. Not now. I have a old soldering iron, solder sucker, and solder wick. Wires are pretty much it these days The brain is willing but the body is not.

Some months ago a 13 year old consumer electronics item failed for the 3rd time. I didn't even open it up. Out it went! I'd already purchased a replacement so it was all good. However the manufacturer had changed the product but kept same model number so I had to "tweak" some stuff to get back where I used to be.

So it is not just old calculator stuff that fails.

A1

[AnnoyedOne]

(10-05-2024 03:53 PM)Geoff Quickfall Wrote:  ...circuit diagram which can be found at his site (PLEASE check it out)

I've looked. It looks like a -Ve voltage (VGG) is being generated for some of chips. Note the polarity of C4 (2.2uF).

The motor (MOT) is driven by the "CRC" chip (U3). VGG also goes to this chip.

I don't see any value for VGG but 3 of the 4 tantalum caps are in that power circuit.

A1

[Geoff Quickfall]

Ha,

I hear you, for me, just a curiosity. Would be interesting if they test good and yet the replacements allow the system to function.


I like detective stories too!

Cheers, Geoff

[teenix]

The 47uF is directly across the battery when the unit is switched on. All of the LED models had this cap, although the Classics and Woodstocks had a 60uF type.

It is likely used for power supply decoupling and smooth out small load changes placed on the power supply. The sense IC is also connected directly to the battery when switched on, and when the CRC tells the sense IC to start the motor, a brief short is placed across the supply and draws a lot of battery current. As the motor speed ramps up, this load will drop and at max speed it will be a fairly constant as determined by the speed setting resistor which tells the sense IC how much voltage to feed the motor - around 2.4V. DC motors can also generate a nasty negative spike when switched off, but this might be handled by the sense IC.

The 47uF cap could act to continue to supply voltage to the power circuit while this motor spike occurs, but it is not large enough to absorb a long drop in supply voltage. The LEDs are also connected to the battery, (Classics via driving coils) and this will create supply noise as well, so the cap can help reduce this. The 22uF is acting similar for the 6V Vss supply for the ICs.

Low ESR tantalums might be of use here, I don't know if the originals were.

The battery itself is the biggest absorber for the motor spikes, so all connections should be kept clean, including the brass connector on the charger terminals and the battery should be charged and in good order.

There are 1 or 2 smaller caps on the power supply (2.2uF each), depending on the model. These are used to smooth the Vgg rail (-12V), and for the Classics, a supply for the cathode driver IC (I think around 8V). One was used for the ROM0 led driver in the Woodstocks.

There is another small 2.2uF cap that is used to time a power on reset signal for the ACT.

For the sense IC, there is a 6.8uF capacitor connected to the power input pin and 2 x 22uF caps used for the head circuit, plus a 3.3uF on the motor terminals to reduce motor brush noise. These might be due for replacement as well. The 97 has an extra 6.8uF cap which I think was used to stop the motor from briefly running when the calculator was turned on/off.

Any of this circuit noise that finds its way into the inputs to the sense IC (basically a high gain amplifier) could affect its ability to read the head signals properly. It was reported that dirt/corrosion around the head input pins could also affect the reading ability.

cheers

Tony

[AnnoyedOne]

(10-05-2024 09:29 PM)teenix Wrote:  DC motors can also generate a nasty negative spike when switched off, but this might be handled by the sense IC.

Yeah. Anything inductive (DC motors, relays, etc.) generates "back EMF" when the current flow ceases. Usually limited by a diode (reversed) across the coil. There could be one in the IC.

Quote:Low ESR tantalums might be of use here, I don't know if the originals were.

I don't think they existed in the mid-1970's. They shouldn't hurt anything, if used, though.

A1