The Museum of HP Calculators

HP Forum Archive 08

[ Return to Index | Top of Index ]

When is a 42S truly dead?
Message #1 Posted by Randy Sloyer on 3 June 2002, 8:48 p.m.

I have my trusty 12yr young 42s which got soaked with rain water about a month ago. It sat for 2 days with the batteries in it while being waterlogged before I found it. I let it dry out for 3-4 days and when I put in new batteries it was doa, no display. After a proper mourning period, I read Paul Brogger article on "Pioneer Internals" and popped it apart.

I found two vcc (bat+) traces that had been somewhat damaged by electrolysis from what I assume was the water between the pcb and the rear metal shield. The traces were not broken electrically, but I resoldered. Still nothing (not that I was expecting a miracle at that point!).

So it was scope time. Absolutely nothing. No sign of anything AC - only +4.5 vdc on the CPU and memory chip. So my question is:

1. When in idle mode, is the CPU clock running? Should I be able to see anything on any CPU pins?

2. If the clock is not running when off, does the "ON" button apply power to a CPU pin to wake it up? I can't seem to find any connection.

My last hope is that something happened to the keypad, but I hold little hope at this stage. I was hoping to resurrect this unit so I could add the 32k chip....

Any ideas?

PS - Are there any schematics out there? Even partials from perhaps some of Pauls work?

Thanks All!

      
Re: When is a 42S truly dead?
Message #2 Posted by Paul Brogger on 4 June 2002, 12:28 p.m.,
in response to message #1 by Randy Sloyer

I did fix one calculator recently that had a keypad problem. But it turned on, at least -- it showed faulty keypad behavior, and after replacing the "sandwich" layers, the thing works fine.

If, by chance, it is the "on" key that isn't working, then perhaps a keyboard fix would help. The best way to test is to get 'hold of a 17BII and replace its circuit board with your 42s board. If you get a business calc with scientific behavior, you're on to something. Otherwise, it's probably dead. If it works, you may wish to go on and repair the -42s keyboard.

To fix the keyboard, I cut off (but saved!) all the mushroomed plastic heat stakes with an Exacto knife, lifted up the metal plate, took out the rubber and mylar layers, and replaced 'em with some from another calculator. (I think all the Pioneer keyboards are the same under the skin.)

To re-fasten the keyboard sandwich, I held it all together by hand, and melted one of the decapitated mushroom heads onto each heat stake stub with a small soldering iron. I swirled each around a bit, and pressed the still-hot plastic with my finger to flatten and merge the pieces. I don't think it's nearly as strong as it used to be, but it seems to work fine.

I don't have a 'scope, and so can't advise you on the EEG characteristics of a -42s in REM sleep . . . (Maybe someone else could chime in?)

            
Re: When is a 42S truly dead? ( . . . and may I add ?)
Message #3 Posted by Paul Brogger on 4 June 2002, 1:21 p.m.,
in response to message #2 by Paul Brogger

If the 17BII/42s hybrid works, you might wish to paint over the keyboard and attach a fold-out, laminated key map. I've got a couple of these, -42s innards in algebraic business calc cases with kludgy keyboard guides. If functionality is what you're after, it might suffice.

Which brings me to a related angle -- I've got two of these, upgraded to 32K, and working fine. If your PCB is toast, I'd like to either buy your case, or sell you a working PCB . . . (We gotta get these working pieces together!)

                  
Re: When is a 42S truly dead? ( . . . and may I add ?)
Message #4 Posted by Randy Sloyer on 4 June 2002, 1:58 p.m.,
in response to message #3 by Paul Brogger

Since you've had many pioneers apart and the keyboards are the same, is there any chance you ever (in a fit of insanity) decoded the keyboard matrix? If so, I could then determine if the ON/OFF button works with a resistance check.

As for the parts swap, the aluminum bezel is a bit battered from years on road, but keyboard is in good shape cosmetically. I agree we should try and make one good from two bad if at all possible.

                        
Re: When is a 42S truly dead? (Keyboard schematic?)
Message #5 Posted by Paul Brogger on 4 June 2002, 2:07 p.m.,
in response to message #4 by Randy Sloyer

Yes, in fact, I did once. It's in my project box (I think) on a piece of crumpled graph paper. (You asked this in your first post -- I should have indicated such.)

I'll look, and either retrieve it, or do a new one. (BTW, I've not been wonderfully consistent in my follow-through of late, so I won't be offended if you choose to remind me in a few days . . . )

      
Re: When is a 42S truly dead?
Message #6 Posted by Tony Duell (UK) on 4 June 2002, 5:23 p.m.,
in response to message #1 by Randy Sloyer

A lot of the below is from memory, and I've not worked on a Pioneer for a while, so I might have misremembered something. I can check up the details if you have problems. Firstly, I think the CPU buses are stopped in the power-down mode (so don't expect to see activity on the pins of the RAM chip), but that the clock is still running (don't these machines have a hardware realtime clock, even if it's not used by the 42's firmware). So you should see activity on the pins of the Xtal even in power-down mode. But isn't there a deep sleep mode that stops the clock as well? Now for the keyboard. The first trap for the unwary is that the keys measure about 10k Ohms when pressed. They are not a dead short. They can't be. The keyboard lines are also the address lines to the RAM (and to external ROM if present), so they can't simply be shorted together. If you don't realise this you can think the keyboard is not making contact properly even when it is. The ON key is not in the matrix. It's 2 connections are brought out separately, to the 2 keyboard connections on the same side of the machine as the ON key. One of them is connected to battery +ve on the logic board, the other goes to a special pin on the CPU (which is some kind of power on signal). Again the 'on resistance' of this key is about 10k, but this time a dead short won't matter. You can trace the connections on the processor board (IIRC there's a handy via on the trace you can use for testing) and try shorting that point to battery _+ve (with the board in place, connected to the LCD) to see if that makes it power up. If not, try again, but this time with an isulating sheet between the logic board and the keyboard connections in the top case, so the keyboard is not connected to the logic. If it works then, suspect a short in the keyboard matrix that's shorting 2 address lines together (I've seen this in a 48, which uses a similar keyboard design). I do have a hand-drawn 42S keyboard matrix diagram, but as I said, the ON key isn't in the matrix, so you probably don't need that yet.

            
Re: When is a 42S truly dead?
Message #7 Posted by Randy Sloyer on 4 June 2002, 8:58 p.m.,
in response to message #6 by Tony Duell (UK)

Thanks Tony and Paul.

I'll have a go at the ON/OFF switch this weekend. Gotta go do work for a few days. It seems to always get in the way of the stuff I think is more important!

Thanks for the 10k info. That would have gotten the keyboard cracked open in a flash had I not known. And I'll try with the whole keyboard isolated and see if it will power up. Good stuff indeed.

Also, I think I'll look around for a 17b to try Pauls logic tree of kbd/pca test.

Thanks to all. I'll report later.

                  
It's alive!!!!!!!
Message #8 Posted by Randy Sloyer on 13 June 2002, 9:00 p.m.,
in response to message #7 by Randy Sloyer

Thanks to Tony and Pauls help, I discovered I have only a bad "ON" key!!!. Shorting the keys traces turned it on and only that key is dead. The logic board is a-okay. It must have been in hibernation mode since I saw no AC activity.

Since I have some plastic industry tools around (I work for an extruder manufacturer!) I think I'm going to try to reform the kbd heat stake mushrooms up and back into a straight pin then disassemble the kbd since its dead anyway. I'll let you guys know how I make out. Gotta 17B kbd around just in case.

                        
Re: It's alive!!!!!!! after thoughts
Message #9 Posted by Randy Sloyer on 13 June 2002, 9:07 p.m.,
in response to message #8 by Randy Sloyer

Does anybody know if a 20S kbd is the same matrix? It sure looks the same. If I need parts for this puppy, I'd much rather sacrifice a 20s then a pristine 17B (not that I need an AOS biz calc...)

Just wondering.

                              
Re: after thoughts -- at least 27S is the same . . .
Message #10 Posted by Paul Brogger on 13 June 2002, 10:02 p.m.,
in response to message #9 by Randy Sloyer

Glad to hear of your progress.

I believe all Pioneer keyboards are indeed the same.

I have replaced a 42s keyboard circuit with that from a defunct 27s (but those models are of the same vintage). You should be able to make that schematic that I promised you (ahem!) and test the 20s keyboard to verify.

I would be VERY surprised if they're at all different (but I can't honestly say that I KNOW they're identical).

Let us know!

                                    
My monster lives!
Message #11 Posted by Randy Sloyer on 14 June 2002, 9:57 p.m.,
in response to message #10 by Paul Brogger

Hooray!

I cracked open a 20S and it's the same. Part number that is. Yeah, beleive it or not, the part number and rev of the pcb are the same but the trace layout is different!(5180-3049 rev A -2905). Go figure.

After getting the 4 layer mylar sandwich out, I plopped into my 42S and voila! a working, 100% functional 42S back from the dead. Kinda fun Paul! You never mentioned how the stakes are smaller in diameter than the holes in the metal plate. Especially the two doosies in the middle with the smaller holes! It really does make it tough to get it all back together and not have a soft key since when HP put the 4 layer sandwich in, it had some scoot room around the stakes since the holes are larger until the stakes are fused. Gotta really work slow and carefully.

Rather than use Pauls method of cutting the mushrooms off with an X-Acto knife, I used a !Woodburning! tool (from my local Sears Hardware) with a knife edge to melt the mushrooms back up into a stake . I limited the temperature to about 200-250 using a powerstat so that the plastic didn't melt, only soften. If the material melts and sticks to the tip, its too hot! Once reassembled, I used the tool to remelt the stakes while pressing the metal plate down with my other hand. One the plastic softened enough, I used a small (1/4") aluminum rod that had a counter-sunk hole to press down to make the mushroom again. Had to repeat on a few to get the whole thing tight with no loose keys.

In closing, many thanks to all for the help and background work that made this possible. WELL worth the time and effort.

                                          
Re: My monster lives!
Message #12 Posted by Ellis Easley on 15 June 2002, 8:40 a.m.,
in response to message #11 by Randy Sloyer

Randy, Thanks for suggesting a new way to deal with heat staked sandwiches! Your "insider information" is enlightening: first, the appropriate temperature, and then the aluminum tool for re-staking. Thanks!

BTW, do you think HP used an ultrasonic welding technique to stake these in the first place? I've seen that in use at a factory where I worked. The design of the tool that contacts the plastic is critical but the results are amazing. Only thing is, exposure to high level ultrasonic energy can damage hearing, as I recall.

                                                
Re: My monster lives!
Message #13 Posted by Randy Sloyer on 15 June 2002, 5:02 p.m.,
in response to message #12 by Ellis Easley

As I said in previous post, I really wanted to thank all those who helped. I think this is a truly great site and I would like to "give back" best I can.

As for the ultrasonic welding, I kind of doubt it. I think this is used more when trying to fuse two separate pieces together. I have always likened it to inertia welding without the motion. All that needs to be done here is soften the material and force it down in a controlled fashion. A short blast of cold air and its solid again. Here's my reasoning:

First, I think the case is some type of styrenic material. Probably ABS with some impact modifiers or other "engineering grade" resin with good impact properties. ABS is the same stuff on the inside of your refrigerator (the food liner and door liner) - it handles the cold well, takes quite an impact without cracking, bends before fracture, and most of all: It's easy to scribe your initials into! I'm sure that was in the spec somewhere.

Assuming the case is of some material in this family, its softening point is somewhere around 150F. It never really truly melts, it only gets less viscous with higher temp. Since heating it past 450 or so only degrades it, which lowers its strength, I wanted to keep the temperature as low as possible, hence the powerstat. I didn't actual measure the temp at 200-250, I was only guessing. I could do this with an accurate thermocouple if anyone is interested. I tried my IR laser pyrometer, but its spot size was too big to get a good reading. The strong broad tip of the tool was easy to manipulate and roll the mushrooms back into a stubby pin. I found it better to roll the edge up, then back over onto itself. Trying to pull the sides up to form a tall pin resulted in the "ring" of the mushroom falling off after becoming a tall thin cylinder.

After now doing two units (I put the 20S back together first before the 42S!) I find the real problem is getting enough pressure with my fingers on the front plastic and rear metal plates to get the proper key "feel". I have been daydreaming of perhaps milling a female plate that has keytop recesses to drop the front of the calculator into and use several toggle clamps to get enough squeeze on the back plate to make a good keyboard sandwich. Then both hands are free to do the reheat and form. Since the heating element end of woodburning tool has a #8 or 10/32 female thread, I could machine a female mushroom tip and have the final product be pretty much indistinguishable from the original. Maybe....

Just a curious question: Has anyone here every offered a "no guaranties" repair service? I know Paul Bragger has repaired some, perhaps privately, but I don't know if it exists as a service to our community at large. I for one, would be willing to give time to this cause since "They don't make 'em anymore!" and my fingers will be cold and stiff before I use an AOS calculator! Okay - so I'm biased! Why else would I spend 15 to 20 hours to fix a throw-away calculator? And kill a working AOS unit in the process.

                                                      
Re: My monster lives!
Message #14 Posted by Christof (Davis, CA) on 16 June 2002, 4:26 a.m.,
in response to message #13 by Randy Sloyer

I haven't got any broken calculators, but....

I don't trust my soldering skills quite enough to do a memory upgrade on a 42s, but would love the upgrade. So I'd say it's a possible interest. (the other possibility being getting the tooling and learning smd on my old 48s that needs a ram upgrade- if I can find the info for the s version, instead of the g version.)

-Christof

                                                            
Re: My monster lives!
Message #15 Posted by Randy Sloyer on 16 June 2002, 8:02 a.m.,
in response to message #14 by Christof (Davis, CA)

Sure, I could do the ram upgrade. I have a full smd rework station at work. The hardest part is on the fingers, getting the case open. I think if this looks like something more people want, I'll make a tool to pop the lower 4 stakes apart from inside the unit.

Post a message here or send me an email if you want to do this.

PS: I'll check Pauls instructions to get the ram part number, see where I can find them and how much $ is involed.

                                                                  
42S ram upgrade
Message #16 Posted by Randy Sloyer on 16 June 2002, 8:29 a.m.,
in response to message #15 by Randy Sloyer

The 32K low power static rams are available at Digikey for $2.48 ea qty 1. Sounds like I need to buy some and give this a try!

                                                                        
Re: 42S ram upgrade
Message #17 Posted by Christof (Davis, CA) on 16 June 2002, 2:45 p.m.,
in response to message #16 by Randy Sloyer

yeah- it sound like a plan :)

                                                                        
Re: 42S ram upgrade
Message #18 Posted by Randy Sloyer on 22 June 2002, 8:17 a.m.,
in response to message #16 by Randy Sloyer

I did the mod yesterday will Paul Broggers instructions. It was a snap. I found super fine (0.035") solder wick a real help in preping the pads for the new chip. The regular stuff I had (about 0.100") didn't wick well on such small amounts of solder.

A suggestion that I have on the proceedure is (obviously?) to remove the batteries and then short the battery contacts with a jumper while working on the unit. This prevents the backup caps from building a charge and having stray voltage on the vcc line when you install the new memory. Perhaps putting the circuit board on an anti-static mat would be better. I also covered the top of the cpu with thin tape to prevent anything from falling into the leads of the chip carrier while doing the job.

The memory test reports 31550 bytes free!

Thanks to all for the infomation that made this possible.

                                                      
Re: My monster lives!
Message #19 Posted by Ellis Easley on 17 June 2002, 6:30 a.m.,
in response to message #13 by Randy Sloyer

I've never opened up a Pioneer but I have taken apart two Voyagers: I put my 15C back together by putting thick superglue on one post at a time, holding a couple minutes till set. My 12C I never completely fixed and the post tops are cut off. I just filled the space between the case back and the PCB with a pad of stacked duct tape. I have thought of making a fixture to hold the sandwich together while glueing (or now, re-heat staking!)

BTW, Tony Duell says it is fairly easy to take the keyboard bezel off a Pioneer to get access to the staked end of the bottom posts, instead of just pulling apart till they pop.

I've looked at the top posts and I see they are hollow, and as I have been looking into Plastite screws lately, I wonder if it might be possible to use screws to reassemble Pioneers. What I have learned about Plastite screws can be summarized this way: to use a regular sheet metal screw (circular thread, 60 degrees) in a plastic post, the post wall thickness needs to be greater than the screw diameter (or original hole size - I'm not sure exactly which) but with Plastite screws (tri-lobed thread, 48 degrees) the post wall thickness can be less than the diameter (of whichever it is) - the main point being the post wall thickness can be less than is needed with a sheet metal screw.

                                          
Re: My monster lives! -- Congratulations!
Message #20 Posted by Paul Brogger on 17 June 2002, 11:55 a.m.,
in response to message #11 by Randy Sloyer

Sounds like you did a superior job. And I agree, WELL worth the investment!


[ Return to Index | Top of Index ]

Go back to the main exhibit hall