From - Fri Nov 15 21:37:37 1996 Path: nntp.ulg.ac.be!news.belnet.be!swsbe6.switch.ch!swidir.switch.ch!news.grnet.gr!bt net-feed2!btnet!newsfeed.internetmci.com!news.sprintlink.net!news- peer.sprintlink.net!cs.utexas.edu!ennfs.eas.asu.edu!noao!not-for-mail From: chavez@noao.edu (Enrique Chavez/) Newsgroups: comp.sys.hp48 Subject: Re: SC 32K Memory Cards Date: 14 Nov 1996 18:08:55 GMT Organization: National Optical Astronomy Observatories, Tucson, AZ, USA Lines: 251 Message-ID: <56fn7n$5at@noao.edu> References: <328aaefe.12755265@nntp.ix.netcom.com> NNTP-Posting-Host: sunset.tuc.noao.edu X-Newsreader: TIN [UNIX 1.3 BETA-950824-color PL0] Scott Vanzandt (j-vzandt@ix.netcom.com) wrote: >Has anyone tried using the Smith Corona memory cards for their word >processors in the HP? Sure have! Card is over 3 years old and works like a charm. Wish I had another card to modify. A long time ago there was an article on how to modify the SC memory cards so that they can properly work in the hp48sx. Here is the original article: -------------------------------------------------------------------- Date: 2 Jul 93 13:39:27 GMT From: Allan Petersen Newsgroups: comp.sys.hp48 Subject: How to use Smith Corona RAM cards in HP48 Here it (finally :-)) is, the article on modifying Smith Corona DataStore 32K RAM cards for use in the HP48 calculator.... For a couple of years many people have talked about using Smith Corona RAM cards in their HP48. I have for a long time tried to get one of them to see if it behaved like the Epson cards that I have modified for use in the 48. Until recently I had no succes in getting a SC card, but then suddenly I got the opportunity (thanks, Garth). I have now succesfully modified a Smith Corona 32k RAM card for use in the HP48SX. It involves opening the card, cutting a trace on the PCB, and adding one wire from a test pad to Vdd. It requires the use of a solder iron to solder a thin wire between two testpoints, but is otherwise straight forward to do... I will describe here basically how to do it and what I found. However, first this: *------------------------------------------------------------------* * ** DISCLAIMER ** * * This posting does not encourage anyone to do any physical * * manipulating with their own products. This posting may * * contain errors and any damage done to your own machine or * * to your cards is your own responsibility. * * I have no affiliation with HP other than owning some of * * their products. * * If the chance of a dead card (or calculator) after this * * modification worries you I suggest that you leave it alone.... * * * *------------------------------------------------------------------* Well, I first tested the original clamping voltage level and found the clamping to be released at 4.21 volts and to be enabled again when the voltage was lowered to 4.16 volts. That's too high for the card to be used safely unmodified. When I last measured the Vcc on my 48, I got 4.22 to 4.24 volts and that's not much higher than the clamp release voltage of the SC card. Any kind of noise or voltage tolerance may then let your calc go dead.... and that's just like what people report. If you're lucky, it may work.... otherwise not. I then opened it and tried the "one wire modification". The modification effectively disables the clamping of the inputs of the card when the supply voltage is low and makes it possible to use any of the SC cards in the 48 (you loose the ESD protection provided by the clamping, though). This modification is much easier than the modification (changing the voltage detector) that I described long time ago and seems to work. Something similar was proposed earlier from someone else on the net, but until a few weeks ago I had not been able to get a SC card to verify wether it really works..... but it does :-) Here we go... Step one: Remove the battery and using a thin, sharp blade (scalpel or similar) gently pry off the top cover from the card (the metal shutter is on the bottom side) starting at the end where the battery was. Be careful not to bend the thin aluminum cover. It has to be put back in place later. There is a small spring near one side of the card. You'll most likely damage this, but you can put a small piece of conductive rubber in there when you assemble the card, so don't worry. It's used to connect the two metal cover plates to avoid built up of static charge on the top cover. **BE CAREFUL** at this point - don't damage any of the components on the card. When opened, SC card looks almost identical to the Epson card... It looks like this (drawing provided by Garth Sweeth): +-----------------------+ Legend: | /------\ ==B== | ------------------------------------ | | A | +-----+ | A - Battery Area | | | | @@@ | | B - Write protect switch | \------/ +-----+ | C - Voltage Detector | +-------------------+ | D - Gate Array | | @@@@@@ 7:::::: | | 1-8 - Gold test points (1-3) are the | | @@@@@@ 8:::::: | | ones mentioned in article. | +-------------------+ | @@ - Epoxy coated chips | +----++-------------+ | :: - Blank chip mounting points | | ::||:: 5:::::: | | | | ::||:: 6:::::: | | * Cutout areas represent sections | +----++-------------+ | viewable through the original | +-----+ +-+ | plastic frame. | | | | |4 +----+ | | +--+ | +++-+ |@D@@| | | +-+| |C| +----+ | | ++ +-+ 3 2 1 | +-----------------------+ The test points 1-3 (and 4) are exactly what I refer to in the original Epson article. The clamp enable input is on testpoint 4. Ground is on testpoint 1 and Vdd is on testpoint 3. Step two: Again using a scalpel or a sharp, thin knife, cut the PCB trace going from the voltage detector (marked "C" on the drawing) to testpoint 4. It must be cut between the detector and the testpoint just where it comes out under the epoxy covering the detector. The best way to cut the trace is to make two cuts spaced about half a millimeter apart and then remove the the thin strip of copper between the two cuts with the tip of a pointed solder iron. Now step three: Solder a thin wire (wirewrap wire) from testpoint 4 to testpoint 3 i.e. to Vdd. The wire should be isolated between the solderings to avoid short circuits to the metal cover later. Testpoint 4 connects to the clamping disable input of the gate array (marked "D" on the drawing). Use the solder iron to melt a small grove in the plastic frame where the wire has to pass above the frame. This will prevent the wire from being pressed directly against the top cover when the card is reassembled. Step four: Using some kind of glue (I used a two component epoxy cement) fasten the cover to the frame in its original position. Remember to reinsert a spring or a piece of conductive rubber to interconnect the two metal covers.... hmm, well, I usually leave it out, but it *does* improve the resistance to statics :-) You mau also want to glue on some kind of handle on the outer side of the cover to make it easier to remove the card from your HP48 again. Place it over the greenish bar just above the Smith Corona logo. Step five: Now you are ready to test the card in your HP48, so put back the battery and insert the card in the 48. When you switch the calc ON, you get the message "Invalid Data Card". That's ok, the card hasn't been inilitialized yet. Just proceede to entering the port number, where you have inserted the card and press MERGE. Now check the amount of memory, which should be 32k more than before. The card is now ready for use. One warning regarding step three: The gate array input should never be left floating. It's very high impedant and an electric charge accumulated there may bring the input to any unpredictable state, so it's not good enough just to cut the trace. It has to be high to disable the clamping, so it can't be wired directly to ground (that would have been a nice soultion, though). Assuming it's a normal CMOS input means that it will be in a high state when the input voltage is higher than typically half the supply voltage. It could have a small hysteresis, but wiring it directly to the supply then implies that it could probably be high as soon as the supply voltage goes higher than, say, 0.8 Volt. This could then remove the clamping and everything should be ok.... but it might take a higher supply voltage to bring the clamping FET's to their non-conducting state, so the effective treshold could be much higher after all. It might also differ markedly from card to card and could in some cases be as high as 3-4-? volts. At least it's unpredictable, and that's why I first recommended changing the voltage detector. However, contrary to my expectations, after the modification I found that there was no clamping action at all, i.e. wiring the clamp control input of the gate array directly to Vdd entirely disables the clamping. Well, that's good news and that's bad news: It's good news in the sense that you can then safely use the card in the 48 as long as you keep it in there. It's bad news in the sense that the static discharge protection provided by the clamping is then gone... so be careful not to zap the input lines when handling the card. The metal shutter mechanism will greatly improve resistance to static discharge though, but don't entirely rely on it..... Well, I'll certainly use it in my calc. I'll just avoid taking the card out and carry it around in my shirt pocket. As long as people keep this in mind I think this simpler mod will be very useful to lots of people. Also, those people who got a card just to find out that their 48 didn't like it will now have a way to make use of it (or regret having thrown it out :-)). As the price for a 128k card from EduCalc is about $US 90, i.e. about $US 22 per 32k of RAM, I would consider using the modified SC cards if available for less than, say, $US 12. At $US 1.00 (or 5.00 for that matter) they're just a bargain. I didn't try to wire the clamp directly to Vcc on the Epson card back then, but as the SC card and the Epson card seems identical, I would consider this modification to be valid for Epson cards (32k and 128k) as well. Well, I have also modified a standard 512k bytes Epson card to include a manual bank switch (a Hall sensor built into the card and operated by passing a magnet under the calc :-)). But that's a totally different story :-) It was lots of fun to make it work, though.... and lots of work. It's real crowded in there. There are 16 memory chips from the start. Anyway - that's what I found - hope it proves helpful (or if not, then just informative) to someone. Have fun, and remember that you get to use this information AT YOUR OWN RISK. --- Allan *------------------------------------------------------------* ! Allan Petersen Dansk Data Elektronik A/S +45 4284 5011 ! ! ap@dde.dk Copenhagen, DENMARK ! !"They call it common sense. Then why does so few have it ?" ! *------------------------------------------------------------* -- ----------------------------------------------------------------------------- Enrique Z Chavez Global Oscillation Network Group National Solar Observatory chavez@noao.edu http://www.gong.noao.edu/~chavez -----------------------------------------------------------------------------