The Museum of HP Calculators

HP Forum Archive 20

[ Return to Index | Top of Index ]

HP 28S: shucking the clamshell, further. Book?
Message #1 Posted by Geoff Quickfall on 1 June 2011, 5:11 p.m.

Hello all!

I should wait for the relevant chapter of the book to be finished but I had to share the preliminaries of the Clamshell section and the book.

The book

Firstly, the book! Just a quick note, every calculator family up to the stretch pioneers has been disassembled, repaired and restored. Mostly based on my collection, but with some help from two loaners, the 10C and the 19C printing calculators.

Most of the chapters are complete, but I did not realize how big an undertaking this was at first. Especially considering the ease of repair the early calculators are versus the later calculators. Also the HP 41C up to the 41CL now, almost deserves a book on its own!

Today and the 28s

Case in point, I have a 28S that has a bad connection between the keyboard and the PCA. I assumed this was due to a decomposing compression foam connection between the keyboard and the PCA. After looking up info and emailing I decided to remove the keypad overlay and enter the calculator via the heat stakes in the front. This will entail shaving a minor amount off the top of the heat stakes so that reassembly will be facilitated. Enough info on the pry method leaving the keypad overlay in place indicated to much damage can occur to the plethora of heat stakes present.

Well, I did not apply heat to the top label as I tried to remove it from the calc. This resulted in half the label remaining on the calc and half staying with the clear plastic overlay. By heating the remaining half of the overlay the label came off cleanly. I can make up a reasonable facsimile with the printer reusing the clear plastic of the overlay.

Using this knowledge, I heated the lower overlay and gently removed it in perfect condition. Much easier then the overlay removal of the pioneers. The overlay is composed of three parts:

1.  clear plastic cover
2.  reverse applied label
3.  glue layer.

If the label is not heated gently, then the glue and label separate from the clear plastic cover and is not salvageable at that point. Not a problem on the top, but a big problem on the lower keyboard overlay.

Back to the calculator, WHICH, I thought I would disassemble to the nth degree for documentation. This was due to the ASS-U-MED need to get to the deteriorated foam. In my defence the problem with the calculator keypad; non functional ENTER, 7,4, 1 keys manifested itself exactly like a Pioneer/Stretch Pioneer with a deteriorated compression foam.

From the pictures below you can see the actual problem with the keyboard PCA connection and the importance of heat stakes. Due to a manufacture flaw OR severe flexing of the calculator, some of the heat stakes separated from their bases.

In retrospect, I did notice that the side with the weak heat stakes was sitting prouder to the back shell then the rest of the keypad. Also that pressing gently on that corner facilitated a functioning calculator!

Okay, now to edit in light of a careful examination and help from Randy (the stuff in italics is wrong). I thought I would leave the wrong stuff in but italicize it to show how, without manuals or fore knowledge, as system can be mis-interprete; at least by me


I thought the loose parts were replaceable. In fact they are broken stakes. I will try Randys suggestion of using a styrene post affixed from the back, and protruding through the now vacant hole in the keyboard side. I will have to drill a hole through the back to facilitate this. Once the base is securely glued in place I will melt the top into position. I guess one could also sew with wire throught the holes and back down the side with the broken stakes. Not pretty though as the wire would be visible from the back.

Now the dilemma, I have a very easy repair that does not require further disassembly of the good heat stakes. Simply a glue fix with clamps overnight and then reapplication of the keyboard bezel.

Fortunately for me and the book, I have a second functional 28S with a dropped LCD column in the display. Something I could live with but don't have to now. I also have a, according to the seller a completely defunct 19B Business Consultant II on its way which is a candidate for complete disassembly.

So I thought I would have all the documentation done today for the Clamshells but it has now been delayed for the arrival of styrene rods.

In any case I thought you would appreciated some preliminary photos of the 28S disassembly, to the point where the fault was found.

Cheers, Geoff

The calculator as it arrived:

The fault label that succinctly describes the problems:

How not to remove the label as explained in the text! The fix will be to photoshop a new label and use the original plastic overlay to protect it. If done correctly it should almost be identical.

The keyboard label removed intact, with the loose heat stake caps. These mushroom caps ends of the stakes are heated to form a single piece rivet. The flaw in the flex of the keyboard and the fact that this is where the latch is located indicate a flaw in the design. Randys posting also reinforces the idea that this is due to a bad flex in the design. They are actually stuck to the adhesive back of the keyboard label:

The location of all the heat stake tops including the broken ones highlighted with translucent yellow circles.

Functional calculator as evidenced by the display when gently pressing down where the broken heatstakes are missing:


on closer inspection, these are broken posts, as Randy discusses. So opening the calculator up will now go ahead as it is not a case of simply gluing these into position

Edited: 1 June 2011, 7:20 p.m. after one or more responses were posted

Re: HP 28S: shucking the clamshell, further. Book?
Message #2 Posted by Randy on 1 June 2011, 6:17 p.m.,
in response to message #1 by Geoff Quickfall

Due to a manufacture flaw OR severe flexing of the calculator, some of the heat stakes separated from their tubes.

I had about 50 dead 19Bii's from a purchase many years ago. I sorted out the units that did not turn on and proceeded to trouble-shoot about 20 units. It became very quickly evident that the sole problem was exactly what you found, heat stakes that had sheared off, primarily from the case back, not the mushroomed heads. I suspect it is the physical force from opening the case that stresses the heat stakes in the latch area that causes them to fail. Once that happens, there is a zipper effect to the left that eventually leads to failure of the smaller diameter stakes under the display that causes the classic loss of keyboard to logic board contact. Same symptom but a different cause.

The only idea I had for repair was to drill holes through the back of the case and solvent weld new posts in. The only issue was that I could only find white styrene rod stock so it would not be a "pretty" repair but it would certainly work.

IMO, if you add up the time required, the horrible battery door design and the lack of supply for replacement doors, I decided it was wasn't worth the effort to undertake commercially.

Re: HP 28S: shucking the clamshell, further. Book?
Message #3 Posted by Geoff Quickfall on 1 June 2011, 6:27 p.m.,
in response to message #2 by Randy


You are absolutely correct, the time involved versus the outcome does leave alot to be desired.

For the hobbiest though, with free time, a little guide to get in and get out might be helpful.

To bad the casio, sharp and other cheapies from the early era and many screws wasn't the case here.

In any case, these have to be the most tedious to repair as you forced to deal with so many heat stakes and the styrene post repair. I guess your styrene post could be sanded and the back of the calc painted a close match, but again, just how much work for the outcome. I will try your suggestion though after I open it up.

Cheers, Geoff

Edited: 1 June 2011, 7:22 p.m.

I think I'll stick to the repair of LED models only
Message #4 Posted by Michael de Estrada on 1 June 2011, 7:25 p.m.,
in response to message #3 by Geoff Quickfall

It seems to me like repair of most of the LCD models is terribly involved and time consuming with a high likelihood of disappointment and frustration. Fortunately, all of my LCD models, even my HP-28S, are still in good working order. I've had great success repairing Classics, Woodstocks and Spices due mainly to their ease of entry and soldered connections (Spice first versions excepted), but think I wouldn't have the patience or skill to attack one of the newer LCD models. Some day no doubt my HP-28S will no longer fire up, at which time I will simply remove the batteries and bid it adieu.

Re: I think I'll stick to the repair of LED models only
Message #5 Posted by Geoff Quickfall on 1 June 2011, 8:17 p.m.,
in response to message #4 by Michael de Estrada

I would have to agree with you Michael!

But I have to do this for the book. I shure wouldn't take these on for repair!

In any case, they may become rare enough one day that someone will want to fix them!

I have a 19C comming my way (29C cousin) and even if I can't get it working (it isn't) it is still fun to work on!


Quick update on the 28S and thanks Randy!
Message #6 Posted by Geoff Quickfall on 2 June 2011, 8:09 p.m.,
in response to message #1 by Geoff Quickfall


Well I drilled out the offending broken stakes, I thought. I did this without opening the calculator. Selected the correct sized bit and drilled straight through the body of the calculator starting at the keyboard side.

Purchased some styrene piping from a hobby store of the correct diameter. The holes were slightly smaller then the styrene creating a very tight fit which did not need glue.

When I cut these down and glued the backs of the piping as it protruded from the back of the calculator. Let them dry and then shaved them flush.

This corrected the faulty 7,8,9, key response but left the x,+, -, / keys unresponsive. Pressing firmly under the lower edge of the display reactivated the keys so now its time for major surgery.

I drilled out the top of ALL the stakes leaving my new ones in place. Pried open the calculator and found that 15 more of the stakes in the numeric part of the keyboard were also defunct!

Well, the culprit for unresponsive keys, the PCA to keyboard foam compression. I cut two new ones, (over sized) and fitted them in place. Dry assembled the calculator and IT WORKS.

All the remaining heat stakes have more then enough left to melt them into position plus a drop of styrene glue solvent. That combined with my stake fix (idea suggested by "RANDY") and the calculator will have a 100% recovery after surgery.

Some shots (for more detailed explanation with labeled pictures you will have to get the book ;-)

The drill press set so the depth is controlled by the drill press lever. This allows for just the right amount of stake shaving in a controlled motion and consistent manor:

The drill press was used to bore through the broken stakes and to shave the good stakes:

You can see the new stakes are in place in the above photos and here is a slant view:

All the heat stake tops are shaved in this view and the calculator is ready to be gently pried apart.

Under the keys once the top is removed with the keys (they fall out so be ready to catch them!) are four layers sandwiched above the keyboard circuitry. Here they are starting at the top rubber layer:

Next the snap layer that gives the keys their feel:

Next a thin paper layer:

Under the paper and covering the keyboard circuitry is a clear film, not visible unless removed so once again here are all four layers:

Of course all this was to get to the compressed foam which I have replaced with new foam cut to a thicker dimension. The foam lies underneath the vertical striped ribbon:

Here is a shot of the back of the calculator before the stakes are shaved flush. Next step will be a permanent reassembly with the rear case cleaned up and the stakes painted to match. When done the fix should be almost invisible from behind and seamless from the keyboard side.

That is for tomorrow when the glue has set and the paint has dried.

Stay tuned:


p.s. thanks Randy for suggesting the styrene piping.

Edited: 3 June 2011, 12:29 a.m.

Completed HP28S pictures!
Message #7 Posted by Geoff Quickfall on 3 June 2011, 1:55 p.m.,
in response to message #6 by Geoff Quickfall

Well, what started as a restaking of the 28S without taking it apart ended up being a complete dissection of the right hand side to the PCA.

This was due to two problems, misinterpreted as one problem, a faulty keyboard PCA connection due to broken stakes at the hatch lock side of the numeric keyboard.

In fact, after replacing the stakes, described above, most of the play observed before restaking was removed and a lot of the faulty keys were rectified. However this left the = - X / key still unservicable.

So diving in the culprit was found to be the foam that is used to press the keyboard ribbon into the contacts of the PCA. After replacing the foam and a dry assembly the calculator performed flawlessly.

Here it is assembled with a keyboard test done.

New stakes cut flush and sanded flush. I might touch them up or paint the back completely in the appropriate colour. New rubber pads and it is done:

Looking at the keyboard side, the stakes are flush, and all the remaining original stakes have been remelted with a soldering iron. I also placed a tiny drop of styrene glue on each remaining rivet:

The keyboard test started (you can see the tops of the new stakes):

The keyboard test finished:

The finished product, I am still working on the top label since I destroyed the original. Just printed one off to cover the rivets:

I will compile the photos for the article section, but there will be much more detail including the left hand side and hinge for the book.

Cheers, Geoff

Re: Completed HP28S pictures!
Message #8 Posted by Gerson W. Barbosa on 3 June 2011, 3:09 p.m.,
in response to message #7 by Geoff Quickfall

The finished product, I am still working on the top label since I destroyed the original.

I might send you one if you tell me how to remove it safely (from a dismantled non-working 28S).



Re: Completed HP28S pictures!
Message #9 Posted by Geoff Quickfall on 3 June 2011, 3:42 p.m.,
in response to message #8 by Gerson W. Barbosa

Hello Gerson,

With a dissembled calc the removal should be very easy. Heat it up a hair dryer, and since you are not concerned with overheating the internals, heat it up enough to be very warm (not hot) to the touch.

Slip a sharp blade underneath on corner and lift, apply heat when it feels like it is adhering to much. You don't want to crease as you lift it.

And thanks for the offer, I would greatly appreciate the label.

I will pm you with an address.

Thanks and cheers, Geoff

main problem with my removal of the top label is that I did not apply ANY heat, just peeled it off cold. Don't know what I was thinking ;-(

Edited: 3 June 2011, 4:02 p.m.

[ Return to Index | Top of Index ]

Go back to the main exhibit hall