HP Forums

Full Version: Cards for the HP9810
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

As you've probably guessed by now, I am repairing an HP9810. The next part to do is the card reader. The first problem is that I don't have any cards for it.
As I understand it, the cards for this series of machines (9810,9820) came in 2 lengths. 6" ones, which worked in both the 9810 and 9820, and 10" ones which were designed for the 9820 only. But a friend of mine in the UK has said that he's used the longer cards in the 9810.
Does anybody know the real facts? Can I use the longer cards in the 9810 (I need to be _sure_ if I am going to attempt to use them, so that I can be sure that any errors I get are due to real problems and not just due to using the wrong cards. Note, I don't care if the
machine only puts 6" worth of data on the card.
Thanks in advance for any help

Yes, both length cards work. I have a few of each and there are no problems reading/writing either. Furthermore, some of these cards were recorded way back when and they read just fine with complete programs on both the longer and the shorter cards.

If you have a deteriorated drive roller on the card reader, take a look at the article that I posted here on how to fix it. Also, I have a way to repair a deteriated platten on the printer, but it's quite a lot of work to rebuild that.

Thanks for the information. So I can try out the longer cards and if/when I have problems I know it's a real fault to be diagnosed. That's useful.
I've seen your article on repairing the roller (how _did_ you guess I would have to do that :-)). It looks fairly easy (I've had the reader mechanism apart already -- it's relatively simple compared to most things I work on). I am a little
worried as to how well the O-rings will stay on the original hubs -- I would have thought that they might slip off sideways. I will probably make new (larger) hubs) with a groove for an O-ring, like I did for the 9100 (Yes, I do have a lathe).

I have used both in the 9810 with no problems. Did you locate some cards to use? If not, I may have one or two to spare.

Also, I found the card readers very forgiving. I had a reader that would slip and I occasionally had to help pull the card through and it still read fine.

Thanks for the offer. I think I have managed to locate a few in the UK, but if that doesn't work out, I'll get back in touch with you.
With regard to the reader being 'forgiving'. I think, that like the 9100 reader, the 4 tracks (in each direction) on a 9810 card are 3 data tracks (again, like the 9100, the 9810 uses 6 bit keycodes) and a clock track. I haev no idea if the format is identical to that onf the 9100, though.
But the presence of the clock track means that the reader should work no matter what the card speed is (within reason). So I am not too suprised that you could get it to work by helping the card through by hand.
Getting it to write reliably might be harder -- if the card is pulled too slowly then the data bits (and clock pulses) will be too close together on the card, and will not be able to be read. Still, we shall see.
The motor in the reeader is a 24V permanent magnet DC motor and as far as I can see there's no speed control circuitry, either on the interface board or inside the motor. So the speed can't be _too_ critical.

I am fairly sure the cards were not preformatted (I seem to remember a guy mentioning he bulk erased some), so the clock track is probably written along with the data.

The o-ring repair that I did has held up surprisingly well without having to cut a v-groove or gluing them on. I thought that they'd slip off too over time, but so far they haven't.

Also, the R/W head that I had originally was bad -- one open coil -- I managed to track down a place in California that had a spare head and would mount it. (This place speciallizes in repairing magnetic heads for all types of equipment. I sent them the head in the mount and they found that they coudn't fix the coil but that they had an exact replacement. I ended up not buying from them however, a well known collector friend ccame to my rescue with a spare 9810 reader assembly first) Let me know if you need their services and I'll dig up the contact information. But I think that bad r/w heads are not a common problem.

I am _sure_ the cards are not preformatted. The card reader hardware writes the clock track as it writes the data (same on the 9100, BTW, except that the 9100 does rather more in hardware than the 9810 does).
That's why the card speed is relatively critical for writing. If the card is fed too slowly, the bits will be written too close together, and will be unable to be read).
To digress slightly, the HP71 and HP75 cards _are_ pre-formatted -- the clock track is written at the factory. When you write on one of those cards, the card reader _reads_ the clock track to determine when to write the next bit. That's why those machines can get away with hand-pulled cards for writing. The older machines (9100, 9810, 9820 which have a clock track that's written along with the data, and the HP65, HP67, HP97, HP41, which have a self-clocking data encoding scheme) don't have a pre-written clock track, which is why you have to
have the card speed about right for writing.

Something else to worry about. I'd not even considered an open-circuit head -- I've never seen that fault in an old disk drive, tape drive or card reader (but I guess it can happen). I will have to check this, and then demagnetise the head afterwards (the ohmmeter current can be enough to magnetise the head).
Is this head anything close to standard, I wonder. It doesn't look particularly custom. Incidentally, there's little shim under one end of the head mounting block (between the block and the side plate of the reader). Is this standard practice from HP, or is it a kludge to correct for misalignment later? Anyone seen that sort of thing before?

My guess is that it's a head missallignment correction kludge. The replacement head that I installed (from my friend's spare reader assembly) and was able to read the cards written with the original head without any special allignment -- not that you can adjust the alginment anyway. I can't belive that HP would have done this, perhaps it was some past owner of the calc that was trying to adjust the head alignment in order to read cards from another calc that was misaligned. I used to do exactly that on my PET computer.

I've never seen a problem with magnetizing a tape head with the tiny current (~1 ma) from a multimeter, but I suppose if it were a really sensitive head, like in a disc drive it could be an issue.

I think that the 9810/20 head was made by Amperex for use in 4-track recorders, but I don't have a model number on it, just a recollection of my phone conversation with these repair specialists. Although I have never come across an open head coil either, they told me that it's a quite common occurance and is often repariable if the coil break happens at the junction of the input pin and the coil wire. They dig through the epoxy fill and repair it, amazingly! The head that I had sent them had the coil break deep down and was not fixable.

It would appear that HP aligned the head into the diecast support frame, then epoxied it in place. And therefore any head assembly should be a drop-in replacement in any reader.
But I guess if the side plates were slightly distorted then the head might not fit in quite the right position, so it would need an alignment shim. But I wouldn't have thought that alignment or head position where _that_ critical -- the data density is pretty low, and the card can slip slightly sideways anyway.
Be careful with the current from an ohmmeter. Some of the older analogue meters (which I perfer for most work) will supply 50-100mA on the ohms *1 range. Easily enough to magnetise a head or damage some components (I recall a story of an idiot who tested an entire box of 50mA fuses and claimed they were all open-circuit. Well, they were after he tested them...)
DIgital meters tend to use a lower current -- for most types, the currnet is chosen so that the voltage across the highest resistor that can be measured on a particular range is 200mV. Which would give a 1mA test current on a 200 Ohm range, agreed.
I thought the head might be standard, for the simple reason that it doesn't come close to using half the card width. One day when I've got nothing better to do I'll measure (with a travelling microscope) the core positions on the head face relative to the mounting flange. I suspect I'll find that the track spacxing is something fairly standard.
And finally. corrosion of fine wire windings is fairly common. It's often thought to be due to contamination of the wire (e.g. with acids from the skin of the person who wound the coil) when the coil was made. I've found transformers that have failed in that way. I've never found a magnetic head that'suffered from it, but I guess it can happen.

I know some people who CAN fix a broken head coil. Problem is they are the only people on the planet who can (or will) do it, they know it, and they charge accordingly... it takes about a kilobuck just to get them to answer the phone.

They also do unholy things with obsolete and degraded media. One client spent well over half a million dollars to have a reader built for some ancient tape format. The tapes were HIGHLY degraded and were destroyed in the recovery process but every bit of data was recovered.