Hello,

Over at my household, I started the day thinking I would clean up an HP-34C. Now its in pieces all over the place.
Right next to the computer here, piles of little keys, clips, plastic case parts, its starting to look pathetic.

Anybody want to start a betting pool for whether I get it back together ?? If anybody bets against me and I get it together, they have to mail payment within 5 days.

I paid a high price for this HP-34C, and the guy said it was perfect.

Well, no sooner do I try operating it a few minutes, and its quite unreliable. Just flaky, acting real flaky.

Now, this is in keeping with the early SPICE calculators. Are we really sure that Carly wasn't working there earlier, maybe as a junior engineer? I mean, this thing about chips stuck into the unit without any soldering. That's dumb, real dumb. It's Carly-dumb.

I stripped it all the way down, so that the IC's all fell loose. I see oxidation on pins of the 3 memory chips. I tried soldering to them, and you can't solder to them, not unless you use the tip of the solder iron to rub each pin for awhile. Obviously this is the failure mode for the early SPICE units, the oxide on the pins prevents contact.

So I am experimenting, folks. I'm converting this UNSOLDERED product to a SOLDERED product. I'm about 60% done. It looks difficult, but promising.

There was a fork in the road, I decided that I would accept an inferior cosmetic appearance, for a more reliable exterior performance. I mean, I can't realistically even remove the solder flux afterwards, so it wont look real pretty, but who cares, you have to tear it all the way down to see the solder and the flux.

The Leica self-illuminating low-power stereo-zoom microscope is real handy. There's a certain model that can be had brand-new for $850 , and its zoom, x7 to x 30. It helped me make sure that each connection was solid, esp. with the solder-resistant pins on the 3 memory chips.

WHAT'S KILLING me is I want to replace the 22uF electrolytic caps, figuring they'll go dry soon. I drove to the usual watering holes,here in Seattle. One folded, the other moved. Kind of pathetic, really, a country where there's only one or 2 stores that you could go to buy tantalum capacitors, in a city of 3 million.

No, Radio Shack doesn't count.

I would appreciate any input from those who have tried, or refused to try, converting an unsoldered unit to soldered.

...about a dozen Spices like yours in the past: I let them spread around the table and recalled them back together. To be perfectly honest with you, I lost the count of how many HP's and TI's I disassembled, washed case and keys, and successfully reassembled. My last challenge was getting to life two "franken"-HP42-"Stein" from three diceased ones. A few days ago I did exactly the same you did with.. an HP34C, my HP34C! It's happily working, all operational. With unsoldered IC's, you dam bet!

I use metal polisher in all IC's terminals and over the connection surface. I call the connection surface the flex-PCB, metal (not golden) coated, right? After letting all termiinals and connection surfaces shinning, I use micro oil (Tutti; seems to come from Italy, with this name) and allow a tinny coat (contacts shine a bit more) to prevent oxidation. When you "sandwich" it back, preassure will be enough to connect all terminals together again.

Amongst the Spices I had in hands, at least five of them had missing segments and two returned [ Error 9 ] after [STO][ENTER] self-test execution. All of them were successfull recovered and worked well. No one needed to be resoldered. In fact, there is an HP38C with me that has the keyboard frozen and if you switch it OF and ON a few times, [ Pr Error ] is shown even with fully charged batteries or with AC adapter plugged in, and this is a soldered-type Spice, not a "sandwich" type. I suspect there is noise from power supply and it is interfering with calculator operation, but I have no scope to track signals. 8^(

I would not solder components; I think these "sandwiched" Spices are versatile to allow you to maintain them without the need of a soldering iron, unless you need to repair the power suply unit. Even the flex connector to the batteries may be easily cleaned with metal polisher.

Wanna run the risk? Bag it all and sent it to me. I'll do whatever needed (except send my HP34C to replace it) and whatever I can to put it back to work. It's too bad Captain Zener is not allowed to cross country borders, because he could simply fly here and bring it in a few minutes, tops.

If you actualy want it, Norm, it will be my pleasure trying to help you. Send me ane-mail and I'll pass my actual address in Brazil. Mail delivery is from one to two weeks for each side. Anyway, I'm almost you'll find others close to you that would gladly help you faster. Anyway, be my guest.

Success.

Luiz C. Vieira - Brazil

I have somewhat the same problem. Dumb old me did not realise that the three chips were not permanently attached and they fell off. My problem is that I don't know where each one goes or which way it is to be placed on the card. Could somebody point me to a post that has a photograph of these three chips in the correct position. Of course somebody couls alwayd pull their working 34C apart and then let me know. LOL
Thank you in advance.

Hello, Howard;

the IC's area lways oriented in the same direction. If you position the calculator with the display in the upper position, you'll see there are three places (square holes)at the left for the ROM and RAM IC's and the big IC (Processor) rests in the right side. All of the IC's are oriented to top, meaning their pin #1 is at the top-left side of each place; there is always a smal mark in IC's: a painted spot or a single recessed dot. Any mark is close to pin #1, and looking from top, pin #1 is always at the first one at top-left.

Now have a look at the flex-PCB where the IC's are mounted; let's call it the main assy, O.K?. Looking at the side the IC's are mounted, find the location of the three small IC's and place the main assy so these locations are at your left side. Now check th econnections from one IC to the other: they are exactly the same, meaning pin one of each of them are connected together, pin two, three, etc... The only exception is for the top-right pin (terminal 8) of the topmost IC: it is not connected to the others. This is what we were looking for: the RAM IC place. I was informed at this very forum that this is the position where pin 8 is fed by the batteries while calculator is OFF.

In my HP34C the topmost IC is 1LB902. As it's aRAM IC, maybe you'll find something closer

The other two are:
mid position - 1MA10014
lowest position: 1MA11(?) - cannot read, it's too faded; would you check for me, please?

As these two IC's are in parallel, I checked if it works when positions are inverted and it runs smoothly.

I hope this helps you rebuilding your HP34C.

NORM: the main guy has a paper seal over it where it reads:
34C/B2102
00034-69110. But I was wrong: as I told you, I recovered many Spices for the last days and I confused the HP33C forthe HP34C. The one that is "sandwiched" is the HP33C; my HP34C is one with soldred IC's, I mean, the mainboard is conventional and the plastic support has bigger square holes. It is intended solely to lock components together and not to keep IC's terminals touching the flex circuits. I'm sorry I confused both calculators (in the begining of this month I had nothing less than 11 calculators opened in my workbench; there are only two, now) and gave you wrong information at first.

Success to you all.

Luiz C. Vieira - Brazil

Hello Luiz,
I checked and the ICs are:
1LB902
1MA114
1MA115
I take it from your message that I should assemble them in that order top to bottom. Only the middle one has a painted dot on it but they all have a HP emblem on them at one end. So I feel that all those emblems should be positioned in the same way.
Thank you for your help.
But my main problem and the reason why I pulled it to piecesis that this calculator had a bad battery leak and it has corroded and totally destroyed the flexible connecting strap from the battery terminals. Totally disintergrated!!
I have traced things out and I have found where the leads (that might not be the correct term) - the traces within the flexible connector go to on the PC board. I feel that I could make up some ordinary wires connecting from the battery terminals (and the charger terminals) to this board where they make their first real connection on that board.
Do you think that this will work?
I really want to save this calculator as it was initially working after I had cleaned off all the corrosion but too much damage had been done to the connecting strip and it gave out after a little while. I have a 33C and I was going to use the connecting strip and PC board from it but the internals are a little different as the ICs and the connecting strip seem to be soldered on rather than the clamp type connections of the 34C.
I hope you can give me some advice.
All the best.
Howard. (I can be contacted by e-mail but take the "spam" out of my name).

Hello, Howard;

I'm glad you could idnetify the IC's and their positions. Go for it!

About wiring the batteries' connector with ordinary wires: it has already been done, and I remember reading about this at this Forum. One thing to notice is that at least one of the cooper trails that come from batteries connector goes straight to the power supply board.

There are three cooper trails comming from the connector: one from (+) batt pole, a common one from (-) batt pole and one of the AC adapter connections and a third one, from the other AC adapter connection. This last one (AC) goes straight and only to the power supply, and the other two connect to the mainboard and the power supply. After identifying them well, you should cut the corroded portion of the flex circuit (think of it as a cancer that must be removed before spreading around the rest of the calculator; show no merci!) and connect wires at their place. If you find a pice of flat cable, it will easen you job, look better and do the same.

Do you have the skills to try? It's not hard. If you need more precise advice about the connections, I'll open the 34C again and guide you as deep as possible.

I decided to post (and I try posting as much as I can) because maybe others can use the information as well. If there are some private statements, no matter at all: I'll e-mail you directly. I have done this with others.

Best regards, Howard. And success.

Luiz C. Vieira - Brazil

Hi Howard,

I tried writing you direct but your "spam big pond" address did not work for me.

Here is what I tried sendign to you direct, then:

Hello Howard,
My name is Norm, I am enthusiastic about the HP-34C.

Sorry to hear about the mistake with the chips. Guess you
didnt know that they are not soldered. This information is
available if you carefully read the discussion of 34C at
the HP Museum.

It is the heart of the problem with the "SPICE" series
of calculators, because it makes them unreliable. The pins
oxidize and then the connections go intermittent because
there is no solder. THEY CHANGED THE DESIGN
TOWARDS THE END OF PRODUCTION, so the
later version has a standard normal circuit board.
I think those are much less common than the original
unreliable version. As you saw on my post,
I experimented with soldering the chips. I have mixed
feelings about it. It is hard to do it with a level of
workmanship that is acceptable, although I did so.
It required a microscope to check the pins. The pins
did not want to accept solder sometimes due to
the oxidation.

NOTICE that the pins on the little chips may be taking
on a blue-ish color that may suggest they wont conduct
electricity properly. I found that the little chips did NOT
ACCEPT SOLDER unless each pin was "rubbed" with
the solder iron for a significant period of time, to break
thru the oxide and get fresh tin-lead thru them. This
requires professional methods, including to de-flux
afterwards.

I have a professional temperature controlled iron.
Flux removal is important. You are also playing with
fire due to ESD (electrostatic discharge) and dealing
with ESD is important and can't be typed down in
just a sentence. I have no idea
what is your best strategy, for you personally. Maybe
you would just re-assemble and hope for the best.

Luiz wipes everything with a thin layer of electronics
compatible oil, then reassembles it. He also mentioned
apply a bit of refinishing to the pins on the 3 IC's.

HERE ARE MY TOP 2 SUGGESTIONS FOR YOU:

(a) Go to the office supply store, and ask to buy a
"silver marker". They put out paint that is silver, and
it dries. The brand is clearly called out as "PILOT"
It says "PILOT SILVER MARKER" right on the side.
Before I took apart the 34C, I put dots on the chips
with the silver marker, and let them dry. You know,
enough so that they were clearly identified.
One dot on chip #1, two dots on chip #2,
three dots on chip #3.

(b) Simply acquire additional specimens of the calculator.
If you like the calculator enough to service it, then you could
have more than one unit. Then it is very easy to cross-check
the type of question you are asking.

For this reason, I am refraining right now from an
extended discussion of "where the wires go". I could
type it to death and you wouldn't want to read thru it.
It's easier just to see a real unit, that you could acquire.
I also do not have one apart right now.

Luiz may already be describing it for you.

HEY, did you get that good deal about 3 weeks ago, on
eBay for a 34C that was in the box ???? From the
seller in Ontario? It went cheap about $70 .
I was going to snipe it, but never got back to the
computer after running an errand.

Good luck ! - Norm

Yes, replacing the battery flex cable with wires works very well. I have done it to dozens of these nasty, vile, wretched little machines.

I suspect that eventually all of them will need it someday. I have seen cracks in the flex circuits of machines that were new-in-box and never had batteries in them.

I used wires once. Next thing I will try is a flex circuit I found in a electronics store. It is for a microwave owen panel, but it looks like it will work ok, and appearance will be much better than wire. Did not have time for it yet, but as i paid one dollar for this circuit, I thought you might want to give a try.

Renato

in the manner of doing a camera test I took some really bad low light shots of a disassembled 32E I'm repairing- if you want to use the photos as props for discussion.

http://www.koyote.cx/static/photos/handhelds/HP/SPICE/

Note the 32E hasn't got the third IC......

-C

Maybe this should be a classified (sorry) but while we are here, anybody got a big chip for an HP-34C they could sell me? I also need the rubber key membrane.

Luiz and I are studying something: display blanking on spice calculators.

Say you type in sin(45). Do the commas blink while it does the computation?????? They would on the older HP-25C.
Luiz definitely has some SPICE units where the commas blink, and others where they don't (same model number, different results).

Luiz says it must be a firmware upgrade that they did at a certain date forwards. I am wondering if they sort of "degrade" into this condition when the connections all start getting loose (spice chips not soldered down). That is, maybe they start out with display blanking, but don't finish with it, and its evidence of some deterioration?

I perceive that ALL HP-34C have display blanking, but maybe I'm wrong? If somebody clearly recalls that they bought a brand-new HP-34c and the commas blinked on sin(45) from day #1, plz let us know.

Another thing to watch for. Type in:
2.2222222 and you will get an extra flash of all digits,
each time you push "2" if there is not display blanking.

Certainly display blanking is superior. I find it deeply unacceptable to get this extra material in the display while I'm trying to run my calculations. I have an HP-34C here right now that commas blink while it types sin(45) .
It has a low serial number by the way. I consider the
'personality' of the calculator sufficiently inferior with the extra flashing, that it would be worth replacing
the big chip to get the display blanking back.

HP-25C I am more accepting of the flashing, because maybe that WAS its personality. But not the 34C, it had a cleaner behavior. Anybody got my parts?

- Norm

I use a small wire brush to clean dull IC legs before soldering, thinking the metal bristles being generally all in contact with each other, it won't generate any static - am I deluding myself? I know that just because I haven't actually killed any IC's this way doesn't mean I haven't punctured some oxide layers and damaged them.

Also I clean tin plated PCB's this way. Looking at Christof's pictures, it looks like that flex circuit has been cleaned, it is so shiny.

Have you used low-residue liquid flux? That works great for hand assembly of surface mount components along with very fine gage rosin core solder. Apply the liquid flux to the PCB pads before placing the component. It dries right away but still does its job. You can generally tack solder the corner pins to hold the part in place without any additional solder, as long as the PBC is tin plated, then go back and solder all the pins with some more solder. I think the low-residue flux is quite poisonous, though - possibly just in the liquid form? It is water-clear and has very little odor.

If I get a no-solder Spice, I think I will try cleaning it as Luiz described and keep it no-solder. Maybe that expensive magical contact cleaner would work well here.

Hello Norm, folks;

I checked again the HP33E and the HP34C and some other Spices (HP31E, HP37E and HP38E/C). The "not blanking", or "comma-blinking" (plus minus signal) happens when entering any nuber that contains a "2" digit elsewhere. If you type in a number whitout a "2" digit, no flicks at all. But at the moment a "2" is entered, every comma that follows every digit after the entered "2" blinks as you type the rest of the number.

with [STO][ENTER] the display remains quiet; [ENTER] will flick if the number has a digit "2" even in the hidden part, say, after the fifth place with [FIX]4; commas will blik only where is a hidden "2" when entering; all number functions that cause internal computation cause blinking commas (is there a chance any of these functions does not use the "2" digit internaly?); and the last interesting fact: [Sigma+] (for statistics) will cause a blinking display if any (or both) X- an Y-register contents have a "2"; if both numbers do not have a "2" digit, [Sigma+] executes with a silent display.

Because of this "2"-related selectivity I thought about firmware-related bug. And I observed it in some "sandwich"-mounted Spices.

I agree with Norm that this is somewhat weird. And his concerns about internal damage have theoretycal foundations for as long as I have the knowledge to technicaly judge them. Silicon-based microchips, MOS-technology, are susceptible to ESD, no doubts about.

I like this sort of discussion and I'd like having more knowledge and data to go further.

Best regards.

Luiz C. Vieira - Brazil

Could this be related to the behavior of the 25 while a program is running - where the digits that flash are related to the angle mode and/or display mode? BTW, I ran across an article in the PPC Journals about that but I don't know if I bookmarked it. I'll try to find it again. Maybe I'll run into an explanation of this Spice behavior!

Here's another way to do some soldering... This is shown for SMT devices, but the overall technique and discussion of solder pastes, etc. may stimulate some thought within the Spice Fix-It Group.

Best Regards,

Dan M
Bellows Falls, Vermont
USA

Toaster Oven as Soldering Device

That was really good stuff, Luiz.

I am trying to draw attention to it with this post.

You are right, I sort of thought it related largely
to pushing the "2" button. But I didn't prove it like you did.

It just somehow seems like a deterioration in progress,
or a low-serial-number bug. Something I dont like anyway.
Would really like to know if its a factory thing, or a deterioration.

The missing link here is somebody who REMEMBERS how these low-serial-number units behaved on day #1, esp. 34C.

One guy said 'use toaster oven for solder paste for soldering down the chips'. The problem there is I think the whole assembly might melt (keyboard etc) so I'd be favoring a temperature-controlled professional-soldering-iron.

My 25C does not blink or show commas doing any of the things you describe. Serial 1802S.....

tm

Hello Norm, guys;

The "bliking" comma+minus_signal may also be observed clearly when in PRGM mode. If any of the digits - program step, program instruction - contains a "2" digit, single-stepping or back-stepping over that particular instruction, i.e., when "leaving" the struction, will cause the "blinking" behavior.

The fact seems to happen when any processed data (in a general form) has the ASCII code for digit "2", that is '32'hex or '0011 0010'bin. I wonder if there is any particular internal routine where '32'hex code is part of display control or display status. If first models' firmware did not observe this fact and it became a "bug", corrected in later models.

Just my US$ 0.05 contribution.

Luiz C. Vieira - Brazil

A KeyNotes decribed the HP65 display system which may (or may not) be similar to the Spice machines. One inetrnal register contains the number to display and another one contains formatting codes for each digit (number, blank, sign, decimal, etc). By doing certain invalid operations you could hose up the formatting register and get weird displays. I would suspect that there is a hardware bug and that the "2" display code is interacting with the formatting register to give the funky display.

Also during program execution the display blanks and unblanks with each program step. By stopping execution on a certain (even/odd) step after doing certain invalid operations you got even werider displays.

that's an important tidbit of information.
In the end we can only guess the truth, by
observing actual units. That means even with 25C
some do, some don't. I have definitely seen
25C that blink on sin(45) etc.

It seems now that we have a wide palette of HP
calculators where some do, some don't, keep
a clean and blanked display w/o "2 artifacts".

Guess what, for my specimens I definitely want the
ones with a clean display and no artifacts.

Hey, Norm;

Congrats, after all!

You have a collectible item, so far. The bugged ones are like coins with the same stamp on each side... not so rare, indeed, but they are collectible items. Keep it and try to find an HP34C that does not blink, so you'll have both. Unless a collector chooses to buy your or offer to trade your "blinking" one for a "blank" one ;-)

Hey, I'm sending you an e-mail right now. Do not answer to this post before reading it.

Best regards.

Luiz C. Vieira - Brazil

I want to try that!

One thing my company learned when they first started doing surface mount is that one reason for the temperature profile is to allow all the moisture in the solder paste to evaporate before soldering temperature, otherwise little balls of solder splatter all over.

My 34C (S/N 2051S) doesn't flash commas or anything else while executing sin(45) or typing 2's. Running a program containing 2's, all digits, commas and minus sign flash. Looks about the same in FIX, SCI, ENG, DEG, RAD, GRAD modes. It's hard to write a program that loops without any 2's because you need GTO, (h LBL), or I with keys on the second row. The best I could do was a series of 16 repetitions of "5 +". This program flashed everything as described before.

Wires work just fine and are much more robust and easier to work with than flex cables. I doubt that you could get a flex cable routed to the proper places and keep it working.

Routing is not too hard - I once built a flex circuit using scotch tape and chomerics copper foil, and used it on a 34C (soldered ok on both ends). Advantages are: cosmetics and, as it is thinner than wires, it fits well in the case.

Ooooh, way too much work for us lazy slobs. There is plenty of clearance for wires. I usually use three conductor ribbon cable salvaged from old PC LED harnesses.