42s upgrade



#14

Just did a RAM upgrade on my 42s and a clock speed double. Now it doesn't work :-(

It worked for a time but the '5' key produced '5' CHS and the '6' key produced '6' EEX. The multiplication key did a CLx. A bit more fiddling and now nothing works :-(

I don't know if I cooked the RAM or not. I've check continuity everywhere and nothing seems out of place. Will see if I know anyone who'll replace the RAM properly.


Pauli


#15

Hi, Paul;

have you checked the RAM chip for, say, compatibility and power consumption? I remember finding specs about some C-MOS 32K-RAM IC's that were not exactly low-power. All I remember (after upgrading some HP42S, cannot recall the actual figures) is that any 32K-RAM from HP48S/48G would work fine.

Success!

Luiz (Brazil)


#16

I ordered the part number from digikey mentioned on this site a while back so I sure hope it is correct.

I think I've either shorted from tracks or burnt out the RAM chip. I'm not great at soldering and don't have access to magnification any more :-(

My plan now it to remove the RAM and replace it with a spare chip and try to get someone who knows how to solder surface mount parts to do it.


- Pauli

Edited: 15 May 2010, 6:02 a.m.


#17

Quote:
I think I've either shorted from tracks or burnt out the RAM chip

You guessed it the first time, you've shorted some traces. The four keys you mention share data and address lines, when there is leakage between then, you get the behavior you saw.

If the board does not have the green solder mask, that is it is all gold plate showing, look not just in the ram area for shorts as it is very easy to have a small bit of solder anywhere shorting something!


#18

Thanks for the info. It has the gold tracks.

I figured I'd either shorted something or not got a good solder connection somewhere. My subsequent attempts to fix the problem have only made things worse -- now nothing works :-(

Will do a thorough check over tomorrow in good light.


- Pauli


#19

Pauli,

I would also make sure you have a fresh set of batteries. When you increase the speed, the battery voltage can't drop down as low as a normal speed calcs before it won't work. I believe that will occur before the batt warning would normally come on.

If you've completely given up, I would be willing to work on it no charge. I have upgraded 12-15 of these without problem. Just contact me off list.

Dave


#20

Having sucessfully done 7 of these I have experienced similar start up glitches. These were all fixed with a return to the chip and a thorough inspection.

Do you have a continuity tester. Try testing each contact point and the respective leg of the chip for continuity. This can be due to a bad solder joint.

I have corrected the problems in all cases by applying solder flux to each leg and re-heating the connection just confirm good contact. Then cleaning the residue with contact cleaner.

Also, to clear any previous problem that carries over in continuous memory, short out the battery +ve and -ve contacts and hold the on button down (without the batteries) after your fix and before turning the machine on.

Cheers, Geoff


#21

That for the suggestions.

I have continuity tested all pins to their test points and no adjacent pins are shorted.

I haven't tried new batteries but will.


- Pauli


#22

Hi, Paul;

before going ahead with soldering, consider doing that WITHOUT batteries in place AND providing the batteries contacts are shorted with each other. I am sure this should be well known, but if batteries are in place no soldering should be performed because the chance of shorting active lines in the circuit (RAM chips must be powered if its contents are supposed to be retained, right?) and causing damage are big enough not doing so. Also, if the batteries terminals are shorted, remaining capacitor charge should be neglected and ESD would be avoided as well.

And before using a fresh new set of batteries, remember that their current capacity is the highest expected, so if you still believe that you may have shorted terminals, try with regularly discharged batteries first so they will not drive much current if a short circuit still exists. The worst scenario of using partially discharged batteries would be dimmed LCD characters and low battery indication, meaning that the calculator is probably working fine. If so, then go with the fresh batteries.

Last but not least, if nothing goes right, repeat the mantra: 'FIXTHATCALC'... 'FIXTHATCALC'... These HP42 beasts are stronger than they seem to be.

Success!

Luiz (Brazil)


Edited: 15 May 2010, 7:17 p.m.


#23

I've done almost everything with the batteries out -- I did a few continuity tests with them in by accident which produced weird results.

I didn't know to short the battery contacts while soldering but will in the future.

The batterys are a bit flat -- dropping back to the slower crystal and the low battery indicator comes on.


- Pauli


#24

Quote:
I did a few continuity tests with them in by accident which produced weird results.
Actually, any continuity tester yields a small current in the circuit to be tested, so in terms of the calculator circuits, the test leads act as an extra current source. This would disrupt calculator operation, at least.

In terms of the tester circuit itself, continuity test applies to circuits without any signal, voltage or current source. So, the calculator batteries probably added a small amount of current that led to the abnormal, weird results you mentioned.

Still hoping for the successful recover of you HP42S!

Cheers.

Luiz (Brazil)

Edited: 15 May 2010, 10:46 p.m.

#25

Wait... I can use the 32k memory chip from my inoperable 48sx to upgrade my 42s??? This would be fantastic news!


#26

Been there, done that...

With the HP48G RAM chip it works fine, though! The IC identifier in one HP48G I have in hands right now is

SONY
CXK58257AM-70L (70L - 70nS -speed-, L - low power C-MOS)
352G14VS (series manufacturing data, perhaps?)
If your HP48SX is inoperative, should consider the chance that the RAM CHIP itself is not good, though.

Believe you me, I have also used three RAM chips I found in a multimedia sound board for PC, all of them with the 70L (maybe 60L) suffix. At least one of the two HP42S I still have with me uses one of these RAM chips. If you need I can find their identifiers.

Success!

Luiz (Brazil)


Edited: 25 May 2010, 9:14 a.m.

#27

Another session triple checking both sides of the PCB, looking for solder bridges and resoldering pins. I also dropped back to a 32.768kHz crystal. Different symptoms this time but seemingly positive.

Pressing ON shows nothing in the display, but then pressing shift lights that annunciator. I can navigate the menus (the up/down arrows appear) and I can execute BEEPs and TONEs from the programming menu. The busy annunicator lights while beep is sounding.

So, it seems I've got some annunicators but no dot matrix display. Anyone have any ideas here?


- Pauli


#28

More information. During the display self test the line patterns appear very very faintly. I cannot see anything during normal keyboard entry.

- Pauli


#29

Hi Paul,


Quote:
Another session triple checking *both* sides of the PCB...

According to this message, you've taken apart the display. Which in the end, may lead to all and every sympthom you've previously described.

Realigning the LCD with the PCB is somehow tricky as the stainless steel frame has gaps wide enough around LCD bezel for the screen to loose its alignment with the PCB's contact area.

Placing some isolating spacers (1mm wide aprox.) at every LCD side may help centering it. Tha bad news is that you 'll have to remove the frame once again (be gentle with the steel "twisted anchors" they're prone to break).

Hope this helps.

Cheers.


#30

Yeah, I too the PCB off the keyboard/display unit. Thought I'd been careful but obviously not careful enough. Will disassemble again and see what I can do... :-(


- Pauli

#31

That's it!!!! I've now got a working 32kb 42-s :-) Still only normal speed but I can live with that.

The screen is very dim but that is most likely the batteries. Being Sunday meant I didn't get out to replace them -- tomorrow I should be able to.

I couldn't noticeably move either the stainless steel surround or the LCD inside this using reasonable force, however I was able to very very slightly reposition the PCB in the slots.


Thanks to everyone for their very generous and very rapid assistance.


- Pauli


#32

Hi,

Good to know of yet another RPN wonder back in working order ;-))

Cheers.

#33

Hi, Paul;

congratulations! I'm happy for you.

And welcome (if you're not yet) to the 32K-42S Owner´s Club. These little beasts seem to feel a lot happier with this additional 24K RAM.

Cheers.

Luiz (Brazil)


#34

Quote:
These little beasts seem to feel a lot happier with this additional 24K RAM.

In what way? I have used my several HP42S units for almost 17 years. It is my favorite small HP calculator by far.

But I've never once been close to using more than about one-third of the 8K RAM. My fourth-order Runge-Kutta numerical solver (written for my HP-41C 30 years ago) for systems of simultaneous first-order differential equations barely uses up 1K of program space.

The 8K RAM of the HP42S is a tremendous advance over the 0.5K RAM of the predecessor HP-15C, but what is the practical value of an additional 24K in the HP42S, coupled with the degraded structural integrity of the result? The HP42S doesn't have the speed and I/O capability to make anything other than novelty use of increased memory, as far as my limited HP42S experience indicates.

If one is "upgrading" for the technical challenge or as a hobby project, rather than for any practical outcome, that's a totally different matter. But...I wonder how many perfectly good HP42S units have NOT survived this non-trivial operation.


#35

Hi, Mike;

I can tell by my own experience when I was testing statistics samples from many different areas and I offered to help colleagues to compute their own when I was in my master degree´s studies (about six years ago). I had my HP42S and for each sample I created a separated matrix so I ended up storing more than 20 samples with many elements each. I remember the calculator was filled with a lot of programs and we used it for a whole weekend. When I realized, we had already used more than 14KBytes amongst matrices and programs. It is less than half of the available 32K, but I'd need at least two HP42S to accomplish the task for me and the guys at the same time.

I was only worried about two things: could not save the job elsewhere outside the calculator and could not risk replacing the batteries as well... Scary, at least.

For the sake of the successful 'transplantation', I remember successfully upgrading at least 6 HP42S, not a single 'burn-out'.

Cheers.

Luiz (Brazil)


Edited: 16 May 2010, 1:43 p.m.


#36

I had a suite of programs developed for my university days. At the time the newest pc on the market and in our lab was an IBM 8086 with 2 stacked 5MB hardrives, amber Amdek screen, two floppies and 1 mb of ram, 640 for the progams.

Needless to say my HP 41C with quad memory, extended memory, printer, plotter, and stat module was my preferred machine. It was full even using synthetics.

Now I use the 42S in the cockpit with dedicated programs aircraft specific. I regressed the charts to quadratics and created software to print the answers with my infrared printer (saves time). I also have a suite of navigation, temperature as well as a large CONVERT program with 20 common conversion routines used in aviation for various countries including a meter to feet conversion routine based on ICAO (aviation) conversion. They round but the Chinese and the Russians round to meters differently! I quickly consumed the 8K.

I have configured the 42S as a PDA also with phone numbers and text search (labels) as well as door codes. These alpha progams use up memory also.

At the moment with a 32K memory chip I am running at 19K available. Sure wish it could communicate with the home computer.

So I guess it is a hobby but put to good use and a needed change. Now I wonder why HP allowed the architecture for 32K but only included 7K. I was told that it would be too expensive at 32K back in the 80's.

Cheers, Geoff


Possibly Related Threads...
Thread Author Replies Views Last Post
  Almost totally OT: Looking for Sharp EL-9900G ROM upgrade Marcus von Cube, Germany 3 910 03-03-2013, 09:35 AM
Last Post: Marcus von Cube, Germany
  HP10BII+ question (ROM upgrade is possible?!?) Csaba Tizedes (Hungary) 6 1,319 02-21-2013, 05:19 PM
Last Post: Eric Smith
  [WP-34S] Upgrade in 10 mins. now v 3.1 2988 Chris Tvergard 12 2,221 06-01-2012, 02:39 AM
Last Post: Walter B
  HP-41C pcb connector upgrade Hans Brueggemann 6 1,478 05-19-2012, 10:17 AM
Last Post: aurelio
  USB-41 status update: Upgrade available v3.0 Diego Diaz 2 822 04-24-2012, 07:50 AM
Last Post: Diego Diaz
  42s questions and 42s vs 35s snaggs 13 2,957 09-19-2011, 02:44 AM
Last Post: snaggs
  HP 50G - ROM upgrade under Linux Glenn Becker 6 1,275 07-10-2011, 06:22 PM
Last Post: Glenn Becker
  Upgrade to HP Museum DVD Set Version 7.0 Gerry Schultz 1 647 12-01-2009, 06:31 PM
Last Post: Walter B
  HP42S upgrade and opening... pictures ;-) Geoff Quickfall 3 804 10-19-2009, 04:06 AM
Last Post: nacho
  HP-42s Memory Upgrade DavidShenk 13 2,185 09-07-2009, 09:08 PM
Last Post: Paul Dale

Forum Jump: