HP-21 - Crazy numbers

[BusyCalc]

Hello,

I've purchased an HP-21, which shows a strange behavior when operating it.

If the calc was not used for several hours, it shows 0.00 after switch on and I can key in. For example if 1/0 is entered, it shows correctly the expected "Error" word on display.

But after few seconds, the calc becomes crazy and it starts showing odd numbers (zeros, commas), usually endind with a random number like "0185934170".

After this, if I switch off and on the calc, the wrong operation starts from the very beginning, I don't get the 0.00 and I can't key in anything. Instead of it, crazy numbers are showed.

If I let the device off for hours, then again it starts properly for few seconds.

Background of the calc: I opened it because the old battery pack had a small leak, mainly under the Batt Plus contactor. Some adjacent printed circuits were light corroded with white, green and blue stuff. The 4 button contactor was also light corroded and didn't contact.

All this was cleaned and it is now in good visual shape.

I also checked several points of the ACT:

- Pin 1 Vss: around 6V.
- Pin 2 Vgg: around -12V.
- Pins 16 and 17 Phi2/Phi1: around 180kHz rate.
- Pin 20 Sync: around 1,6kHz rate.

Could it be related to the tantalum caps? I find strange that after a "nap", it starts properly for few seconds, but afterthat it becomes crazy and a further reset doesn't help. I was wondering if any component remains loaded and it blocks the normal operation.

Could you please help me with tips to check further?

Thanks!

Oscar

[Maximilian Hohmann]

Hello!

To me that sounds typical for a poor battery connection. Some residue of corrosion on the battery contacts maybe. It is able to conduct the "nominal" operating current but whenever the calculator needs more the voltage drops below some value and the calculator operates erratically.

The easiest way to verify that would be to connect a regulated power supply (2.5V 1A) to the battery contacts, using either clamps or little pieces of wire soldered to the PCB traces leading to the battery terminals. I prefer the latter variant.

Regards
Max

[BusyCalc]

Hello Max,

thank you for your reply!

I did as you said: I soldered two wires to the lines leading to the battery terminals and I connected 2 AA Ni-Mh batteries, showing around 2.55V. Unfortunately, the behaviour was the same as before, going crazy after about 1 minute of operation.

I decided to clean deeper the corrosion seen in the traces above the ACT, around the 2.2 uF capacitor to the POR (image). The initial green corrosion was easy to remove using water and vinegar and then isopropyl alcohol, but the blue corrosion is really hard to dissolve.

I also cleaned better the Button 4 connector, because when it wasn't pressed, I could measure about 5-10 MOhm instead of infinite. Also, apparently some rests of the battery leakage was closing slightly the circuit at this point.

After this, I switched the calculator on and it is till now working properly. I cross the fingers!

I have hypothesis for this:

- It was a problem due to the green and blue corrosion, which was joining slightly two lines and introducing noise to the ACT.

- A capacitor was somehow "lazy" after many years without voltage and after several switch on and off it has recovered its properties. I say this because before cleaning better the PCB, I had the impression that the calc was working properly longer after every switch on.

I also had this impression when I restored an HP-65 and tried recording magnetic cards. The first 10-20 attempts were unsuccessful but suddenly, it started to record. I read then that a weak capacitor wouldn't filter the noise of the amplifier and make impossible to record a card.

I'm not sure if this hypothesis of a "dormant" capacitor is feasible.

Regards.

Oscar

[Geoff Quickfall]

As discovered, corrosion and copper sulphate will change internal electrical specs. Vinegar is a great remover of sulphites followed by a good rinse. Of corse after you have removed 5h3 led block and keyboard.

I have soaked a 25 PCA in a glass of vinegar over night without detriment to the circuits.

Also, Electrolytics caps can reform when in use. Others like the button and ceramics a very stable and do not leak or short as the Electrolytics can.

Good job,

Nice to no another ACT Woodstock has survived.

[BusyCalc]

Hello,

thanks for the tips!

Is it then possible to immerse completely the main PCB in vinegar? Is it not hazardous for the chips or the coil?

How do you rinse it afterwards? With distilated water?

Apart from this, I've noticed that the back of the LED PCB was dirty with something similar to black dust, like carbon particles. I could clean it a little bit with isopropyl alcohol, but it remained quite dark. I saw also that the pins that are inserted in it have black parts, specially the parts that get in contact with the holes of the PCB.

I tried to clean the pins but in this case I saw no improvement, they are still partly black. I don't know if they are burnt or simply I'm not using an appropriate cleaner.

Good is, that the calculator works perfectly, but I'd like to clean everything properly and increase the posibilities that it works for many years. I have a bad feeling when I know that there are still some rests of corrosion and dirt inside.

Kind regards.

Oscar

[Maximilian Hohmann]

Hello!

(05-02-2019 08:56 AM)BusyCalc Wrote:  Is it then possible to immerse completely the main PCB in vinegar? Is it not hazardous for the chips or the coil?

How do you rinse it afterwards? With distilated water?

Yes, you can immerse the entire PCB in vinegar (no need to dilute either). It won't do any more harm than the leaked battery liquid already caused. I just rinse it with normal tap water. Then I use either compressed air (from the compressor in the garage) or a hairdrier to dry it. Or place it in the sun for the afternoon.

Some of the corrosion needs alkaline chemicals to dissolve. I have had some success with baking soda (plain "Backpulver" from the supermarket here in Germany). One bath in vinegar, one bath in baking soda and a final rinse. A bit like it used to be when one developed film in the dark room...

Regards
Max

[BusyCalc]

Hello!

OK, thanks.

I'll do it so the next time I open it. I think it'll be a good solution because I could see some green corrosion on the PCB under the ACT, where there is only about 1 mm gap.

One idea was to use dental floss to remove it, but it'll be easier to immerse it in vinegar and let the chemistry work.

I bought some time ago a special lack in spray to protect electronic circuits against moisture. Do you think it's a good idea to spray it after cleaning?

Regards!

Oscar

[albertofenini]

(04-29-2019 06:16 AM)BusyCalc Wrote:  I also had this impression when I restored an HP-65 and tried recording magnetic cards. The first 10-20 attempts were unsuccessful but suddenly, it started to record. I read then that a weak capacitor wouldn't filter the noise of the amplifier and make impossible to record a card.


Oscar

I'm having the same problem on an HP-65, waves for channels RA and RB are clean and in fact it reads every card
Waves for channel WA is clean, but wave for channel WB is dirty and it doesn't record a single card
Other than keep trying to record did you experiment anything that is worth ?
Thanks for help !!!

[BusyCalc]

Hello Alberto,

in my case I didn't have an oscilloscope to check the signals. I only could see that reading cards wasn't a problem (after restoring the gummy wheel, of course), but recording wasn't possible.

I cleaned the cards with isopropyl alcohol, I inserted the cards as smooth as possible, etc. with no result.

I read that the calc needs more power recording than reading and that a fully charged battery pack is desirable to do this. So, I fully charged the batteries in an external charger (I use 3 Ni-MH AA) and I got then about 4.3V.

After 2-3 unsuccessful attempts of recording, it started then doing it properly. I suppose that the caps were reformed after some use.

Were you able to check the caps of the reader/writer amplifier?

Regards.

Oscar

[albertofenini]

Hello Oscar
We do have an oscilloscope and the waves are very good and clean
When this happens usually the capacitors are good
We have not checked yet that the switches recognize the fact that the card is write enabled
We will do this soon
I have also opened a post asking the meaning of the WE pin on card reader
Do you know what it means ?
Thanks for help Alberto

[BusyCalc]

Hello,

I've immersed the main PCB in white vinegar and the results have been very good. The first minutes I could even see small bubbles coming out the rests of corrosion.

After rinsing and drying there are no blue rests anymore, and the calculator continues working, hehe.

Thanks for the support!

Regards.

Oscar