HP-28S Battery Compartment Door(teawea.com)
teawea.com
HP-28S Battery Compartment Door
https://www.teawea.com/hp-28s-battery-compartment-door/
43 comments
Disagree, and I used to run a company that did FDM printing. If you look at this design it's packed with tons of overhangs regardless of printing orientation, and a particularly nasty undercut ending in a sharp point (see right side view). This is actually a great example of where FDM is a bad solution and the part should be redesigned (ex. split it into multiple pieces or find a way to re-engineer it so it can be printed in some good orientation) or you should pick a different method. Yes, you could FDM print this, but the end quality would be garbage, the tolerances would be poor, and you'd end up with a lot of support material cleanup and probably failed/unacceptable prints.
If you print this angled so the curved side is down and the two sides are at 45 degrees to horizontal, it should print more or less fine without any support whatsoever.
Agreed. I’ve printed dishwasher replacement parts using ABS that hold up perfectly in high temps soaking in water. PLA is also great for certain applications and can look and feel very nice when done right.
The design itself could also probably be made more FDM-friendly.
This seems like the wrong part of the door to fix.
This brings a whole new angle to HP calculators' motto: "By Engineers, for Engineers".
I actually love this little projects which extend the life of useful tools for a very, very long time.
I actually love this little projects which extend the life of useful tools for a very, very long time.
I have this calculator and for me, the case cracked around the battery door. I haven't put batteries in it in a long time, but IIRC the springs are very powerful and that plastic case is very strained there. My calculator is 30 years old now and I'm guessing the plastic is less elastic now than it used to be...
Plastics have an ingredient called plasticizers to add elasticity and reduce brittleness. This materials sometimes evaporates or degrades over time, hence plastics become so-called dry and become brittle.
This is why things crack over time and become more brittle. Some modern formulas are much more stable and can protect their properties for a longer time.
Of course, we see the effects of this resiliency as pollution nowadays. Looks like we made some wrong choices while selecting the items to make more resilient.
This is why things crack over time and become more brittle. Some modern formulas are much more stable and can protect their properties for a longer time.
Of course, we see the effects of this resiliency as pollution nowadays. Looks like we made some wrong choices while selecting the items to make more resilient.
My old HP 55 calculator is still looking like new, after what, nearly 50 years? They must have changed the case plastic at some point, for calculators after that one.
Looks like the unit used for this picture on Wikipedia https://en.wikipedia.org/wiki/HP-28_series#/media/File:HP_28... could use one of these too (I assume the battery compartment door is in the upper right corner, where you can see some duct tape)
How about ordering to somewhere running HP MJF printers? The PA11/PA12 materials they use is much like ABS or PLA but has no visible layer lines.
Isn't MJF essentially the same process as SLS that was used in this project?
No. Microjet printers use a binder (like a glue/epoxy) with powders (usually ceramic) and sometimes add colors as well to achieve the end product. SLS uses a laser to melt together (sinter) the powder (in this case nylon, but could also be metal). They are both powder-based, but the method of turning the powder into the end result is quite different.
Needs some acetone vapor smoothing.
https://www.youtube.com/watch?v=NiG1jDEG0kA
https://www.youtube.com/watch?v=NiG1jDEG0kA
I am intrigued by the support for the 28S these days. I still have my 28S (with working battery door) from when I was at University. But, while I was impressed by it's potential, I never loved it as a device to use and I've stuck with my older FX-7000G as the calculator on my desk.
As explained here (edit - not my site, but I agree with the POV): http://www.ganjatron.net/retrocomputing/hp-28s/index.html
"I must say I have mixed feelings about this calculator. All this sophistication comes at the expense of ease of use. I find the HP-28S aggravatingly clumsy to use, and it's not because of RPN. For starters, the clamshell design, while ideal protection, is awkward -- you'll find yourself constantly flipping back and forth because keys on both sides are frequently used in typical operation.
Another major problem is that the HP-28 is, for the most part, not keystroke programmable; most commands are indirectly accessible through menus (over 20 of them)."
Not having the ability to import from exteral storage (the IR link is out-only) is also a big minus.
As explained here (edit - not my site, but I agree with the POV): http://www.ganjatron.net/retrocomputing/hp-28s/index.html
"I must say I have mixed feelings about this calculator. All this sophistication comes at the expense of ease of use. I find the HP-28S aggravatingly clumsy to use, and it's not because of RPN. For starters, the clamshell design, while ideal protection, is awkward -- you'll find yourself constantly flipping back and forth because keys on both sides are frequently used in typical operation.
Another major problem is that the HP-28 is, for the most part, not keystroke programmable; most commands are indirectly accessible through menus (over 20 of them)."
Not having the ability to import from exteral storage (the IR link is out-only) is also a big minus.
That’s a good review of the 28S. I had the same issues with mine, To the extent that I’m not sure where it is now!
I later got a 32Sii which is simpler, more focused, and better as a desk calculator. I had a 41cv for many years and loved it even more.
I later got a 32Sii which is simpler, more focused, and better as a desk calculator. I had a 41cv for many years and loved it even more.
I was gifted an HP 28S that was given as a bit of swag when our lab picked up an HP mass spectrometer. The odd battery size was a challenge back in those days. That silly thing absolutely saved me on a physical chemistry exam where I could get something close enough that the calculator could do the integration. My trusty 48G is still on my desk. Incredible how solid these were compared to the HP of today.
You can get a modern HP-41 from Swissmicros which just launched its DM-41X. I just received mine and it is a very nice device.
I have one too, it’s superb. Love that a business exists that is doing this.
32sii gang represent
I have the 41CV app on my iphone and it's wonderful.
> I intrigued by the support for the 28S these days. I still have my 28S (with working battery door) from when I was at University. But, while I was impressed by it's potential, I never loved it as a device to use
I came to the higher end HP calculators with the 48SX, which was essentially v2.0 of the 28S. Aside from the 48's larger display and single keyboard, the two machines operated essentially the same, and my experience tends to align with yours. (Even if the 48 addresses the specific concerns you mention about direct access to common features, bidirectional I/O, and the split keyboard.)
For me, the I way I'd put it is that these machines were better pocket computers than they were pocket calculators. Everything was just a touch too slow and overwrought to be as quick as you want a pocketable tool to be.
In HP's defense, these machines were designed in the mid-1980's as a sort of next generation calculator architecture. They developed a higher level programming language (RPL) and a suite of tools for the purpose of getting themselves a solid foundation for continued work after the HP41 series (which was pushing the limits of HP's traditional SW/HW processes).
So the The first RPL calculator in this new series was a business model (19B) very similar to the slight later 28C and 28S scientifics. Notably the 28S was sold in parallel with the 41CX, largely, I think, because HP recognized the appeal of the 41 and the fact that the 28 wasn't in a place to be a full replacement. The 48 was the combination of the two lines, learned from the issues of the 28, and was intended to replace both the 28 and 41 series.
So like I mentioned above, the 28 really the first try at a high end scientific based on RPL and the 48 is the second. And for better or for worse, this is where funding for further development became highly marginal. There was a 48GX that doubled the clock rate and bundled the formula library, but my understanding is that that was really the last major development of the platform by the original engineering group, and they didn't ever really do a third version. (Which is probably where they'd have gotten it fully right, based on experience with other HW/SW products over the years.)
One place people might argue with this narrative is that there were, in fact, two follow ons in the form of the HP49 and HP50. These bundled in additional open source libraries developed by the community (MetaKernal, IIRC) and at least one used an ARM CPU running emulation of the old HP Saturn CPU used previously. Neither were built to the same mechanical standards of the previous models, and I don't believe there was much continuity in the engineering teams. (The follow on work was done at least in part by members of the open source community that HP hired to briefly form a new calculator engineering group.)
> and I've stuck with my older FX-7000G as the calculator on my desk.
I use an emulated HP42S.
I came to the higher end HP calculators with the 48SX, which was essentially v2.0 of the 28S. Aside from the 48's larger display and single keyboard, the two machines operated essentially the same, and my experience tends to align with yours. (Even if the 48 addresses the specific concerns you mention about direct access to common features, bidirectional I/O, and the split keyboard.)
For me, the I way I'd put it is that these machines were better pocket computers than they were pocket calculators. Everything was just a touch too slow and overwrought to be as quick as you want a pocketable tool to be.
In HP's defense, these machines were designed in the mid-1980's as a sort of next generation calculator architecture. They developed a higher level programming language (RPL) and a suite of tools for the purpose of getting themselves a solid foundation for continued work after the HP41 series (which was pushing the limits of HP's traditional SW/HW processes).
So the The first RPL calculator in this new series was a business model (19B) very similar to the slight later 28C and 28S scientifics. Notably the 28S was sold in parallel with the 41CX, largely, I think, because HP recognized the appeal of the 41 and the fact that the 28 wasn't in a place to be a full replacement. The 48 was the combination of the two lines, learned from the issues of the 28, and was intended to replace both the 28 and 41 series.
So like I mentioned above, the 28 really the first try at a high end scientific based on RPL and the 48 is the second. And for better or for worse, this is where funding for further development became highly marginal. There was a 48GX that doubled the clock rate and bundled the formula library, but my understanding is that that was really the last major development of the platform by the original engineering group, and they didn't ever really do a third version. (Which is probably where they'd have gotten it fully right, based on experience with other HW/SW products over the years.)
One place people might argue with this narrative is that there were, in fact, two follow ons in the form of the HP49 and HP50. These bundled in additional open source libraries developed by the community (MetaKernal, IIRC) and at least one used an ARM CPU running emulation of the old HP Saturn CPU used previously. Neither were built to the same mechanical standards of the previous models, and I don't believe there was much continuity in the engineering teams. (The follow on work was done at least in part by members of the open source community that HP hired to briefly form a new calculator engineering group.)
> and I've stuck with my older FX-7000G as the calculator on my desk.
I use an emulated HP42S.
I, too, have a 28S that rarely sees the light of day. Instead it is my older, and much less advanced, HP-10C that gets occasional exercise.
The 28 and 48 series sat in a kind of awkward spot between pocket calculators and "real" computers. The programming language is essentially Lisp turned backwards, but for any task where you actually need that capacity, it must have been easier, even at the time, to just use a PC (in the wider sense, not just IBM PC and clones). The 48 had better communications facilities, so I imagine it might have been useful where pocket carry was actually necessary (it seems to have been popular with surveyors, for example, and I have seen them hooked up with a cable to a total station).
The 28 and 48 series sat in a kind of awkward spot between pocket calculators and "real" computers. The programming language is essentially Lisp turned backwards, but for any task where you actually need that capacity, it must have been easier, even at the time, to just use a PC (in the wider sense, not just IBM PC and clones). The 48 had better communications facilities, so I imagine it might have been useful where pocket carry was actually necessary (it seems to have been popular with surveyors, for example, and I have seen them hooked up with a cable to a total station).
The language is basically a hybrid forth-lisp. I liked it, sadly my HP-48G bit the dust a while ago and was not recoverable (water damage, lost a lot of things).
And it may have been easier to use a PC, but PCs aren't portable. The calculator could easily move with you. This is not a trivial matter when moving between desk and lab and the field. With today's laptops and smartphones a calculator like that is less useful compared to the alternatives. But in the 90s and 00s it was definitely handy.
And it may have been easier to use a PC, but PCs aren't portable. The calculator could easily move with you. This is not a trivial matter when moving between desk and lab and the field. With today's laptops and smartphones a calculator like that is less useful compared to the alternatives. But in the 90s and 00s it was definitely handy.
> it must have been easier, even at the time, to just use a PC (in the wider sense, not just IBM PC and clones).
Agreed.
It was probably a tradition dating back to the fact that HP did the original pocket calculator, but even through to the 48's, HP did literally all of the engineering. Everything from the physical design of the device down to design and fabrication of the custom CPU and other semiconductors. There was just no way that business model was going to stay competitive with PCs over time.
(While I'm reminiscing, however, I would've loved to see the RPL software running on a machine based on the HP100LX chassis and a faster CPU... Much bigger display, a workable clamshell design, and RS232/IRDA/PCMCIA expansion.)
Agreed.
It was probably a tradition dating back to the fact that HP did the original pocket calculator, but even through to the 48's, HP did literally all of the engineering. Everything from the physical design of the device down to design and fabrication of the custom CPU and other semiconductors. There was just no way that business model was going to stay competitive with PCs over time.
(While I'm reminiscing, however, I would've loved to see the RPL software running on a machine based on the HP100LX chassis and a faster CPU... Much bigger display, a workable clamshell design, and RS232/IRDA/PCMCIA expansion.)
I use a physical HP42S ~1989. I've lost count of the number of batteries replaced over the years. Doesn't have an excess of functionality beyond what I need and fits easily in the hand. Only piece of hardware that I would grieve.
Edit: Reminds me of about 10 years back when a box of HP calculators was found in a local warehouse. Somebody went around taking orders and high-tailed it down there.
Edit: Reminds me of about 10 years back when a box of HP calculators was found in a local warehouse. Somebody went around taking orders and high-tailed it down there.
I have a 28S and a 50G. While the 50G is more powerful (newer), the 28S is my favorite.
For programming custom functions the full keyboard is a lot more convenient. The 50G is too cramped, so it was hard to get into programming it, just too inconvenient whereas the 28S was a joy.
Love the clamshell. When using it for programming I lay it flat on the desk fully open. When using it purely as a calculator, flip it backwards and use handheld.
The build quality (except the battery door of course) also feels better. The 50G keys don't have the classic HP feel anymore.
The only real complaint I had of the 28S was not being able to save code between battery changes. I wish it had a backup battery to keep memory while changing the main batteries.
Still have my 28S and still use it most days, although sadly don't find much reason to program it anymore.
For programming custom functions the full keyboard is a lot more convenient. The 50G is too cramped, so it was hard to get into programming it, just too inconvenient whereas the 28S was a joy.
Love the clamshell. When using it for programming I lay it flat on the desk fully open. When using it purely as a calculator, flip it backwards and use handheld.
The build quality (except the battery door of course) also feels better. The 50G keys don't have the classic HP feel anymore.
The only real complaint I had of the 28S was not being able to save code between battery changes. I wish it had a backup battery to keep memory while changing the main batteries.
Still have my 28S and still use it most days, although sadly don't find much reason to program it anymore.
I also have 15C from my early University days and my 28S from graduate work, both still working. I got the 15C used from a junior EEE student that was quitting the program. I think it's on the 3rd set of batteries in 35 years. It got used several hours a day for the first three years.
The 28S burns through three N batteries every two years, even when it doesn't get used much. That may have contributed to the battery doors wearing out on the 28S. Mine is still in great shape however, as spent most of it's life on a desk rather than thrown in a backpack with heavy books that the 15C endured with aplomb in it's case.
The 28S was at work and the 15C at home. Now that I'm working from home and they both are on the same desk, I find I pull out the 15C unless it's complex enough that I want the four line display on the 28S. And if it's slightly more complex than that, a spreadsheet is probably the better tool.
The 28S burns through three N batteries every two years, even when it doesn't get used much. That may have contributed to the battery doors wearing out on the 28S. Mine is still in great shape however, as spent most of it's life on a desk rather than thrown in a backpack with heavy books that the 15C endured with aplomb in it's case.
The 28S was at work and the 15C at home. Now that I'm working from home and they both are on the same desk, I find I pull out the 15C unless it's complex enough that I want the four line display on the 28S. And if it's slightly more complex than that, a spreadsheet is probably the better tool.
Every time I see a 3D printing project like this, I experience a mix of disappointment and hope for further progress. If the finished result is going to look so rough and unfinished, it's just not for me yet. If, in order for it to look good, I need to carefully sand the result and account for that sanding in the initial design, I might as well take up woodworking. Sanding is laborious.
The PLA print isn't any good, but the SLS print looks great IMHO. I don't think they needed to do much sanding on it at all.
Yea screw scanding, but I tthink you've got to accept that this is always going to be prevalant purely down to how 3d printing works.
Acetone vapor chambers are where it's at, closest you'll get to a injection moulded piece without doing the injection moulding.
Acetone vapor chambers are where it's at, closest you'll get to a injection moulded piece without doing the injection moulding.
Acetone will only work with ABS which has another set of issues. Reducing the layer height and a coat of paint will accomplish almost the same thing using PLA.
This is a bad example. The layer heights are too high. Look at thingiverse.com for examples of what you can do with a 3D printer.
That's is my take as well, the PLA print is of poor quality.
There are better and worse examples of FFF printing. The current heavyweight printers from Creality, while being some of the cheapest you can buy, have been dialed in to the point they can make amazing output.
The problem here is that you're zooming in on a 1x2 CM part...It's like taking a 12 Mpixel photo, then looking at a 320x200 size part of it...it'll still look pixelated.
The problem here is that you're zooming in on a 1x2 CM part...It's like taking a 12 Mpixel photo, then looking at a 320x200 size part of it...it'll still look pixelated.
I just bought a 3D printer, partly for fun, partly as I want to print some mechanical keyboard cases.
I've only printed a few things so far. Some board game tiles are perfectly smooth on the "top" (they were printed face down), and the very slight texture on the bottom makes a nice pattern when it reflects the light. The bottom texture is as coarse as a typical cotton fabric, but hard not soft. In other words, I can feel the texture, but it's even, like rubbing a pebble.
I also printed some chess pieces, which are essentially cylinders. The texture is visible, but completely consistent all the way up each piece.
If this battery door really does show the "limitation of an entry level 3D printer" then I'm very glad I spent a bit more money. I still know very little on this subject, but while choosing what to buy I found tutorials on building a printer from parts for €100 or so. This reminds me of the first print from something like that, before the expert showed how to correct the worst problems.
I've only printed a few things so far. Some board game tiles are perfectly smooth on the "top" (they were printed face down), and the very slight texture on the bottom makes a nice pattern when it reflects the light. The bottom texture is as coarse as a typical cotton fabric, but hard not soft. In other words, I can feel the texture, but it's even, like rubbing a pebble.
I also printed some chess pieces, which are essentially cylinders. The texture is visible, but completely consistent all the way up each piece.
If this battery door really does show the "limitation of an entry level 3D printer" then I'm very glad I spent a bit more money. I still know very little on this subject, but while choosing what to buy I found tutorials on building a printer from parts for €100 or so. This reminds me of the first print from something like that, before the expert showed how to correct the worst problems.
I've had the same objections and was waiting for prints with resolution so high you can't see layers. But this year I had some time and money to spend so I bought a cheap FDM 3d printer and honestly it's good enough for most uses.
Depending on the material, color, lightning and size of the part - usually you don't even notice the layers, and in many cases even if you do it doesn't matter. For example I've printed big boxes for our bikes out of black PLA and they look fine - you wouldn't see it's 3d-printed unless you stare at it from 1 meter under good lightning.
And when you need a different finish you can always paint it. I've printed about a dozen plant pots of different sizes because my wife bought 10 ceramic pots and liked them very much and couldn't find similar ones anywhere. We painted them and they look identical and you don't see the layers under the paint.
I don't even print with the smallest layer size possible because it takes too much time and difference between 0.1 mm and 0.2 mm or even 0.3 mm is negligible in practice, but the printing time scales almost linearly.
Depending on what you use it for YMMV.
Depending on the material, color, lightning and size of the part - usually you don't even notice the layers, and in many cases even if you do it doesn't matter. For example I've printed big boxes for our bikes out of black PLA and they look fine - you wouldn't see it's 3d-printed unless you stare at it from 1 meter under good lightning.
And when you need a different finish you can always paint it. I've printed about a dozen plant pots of different sizes because my wife bought 10 ceramic pots and liked them very much and couldn't find similar ones anywhere. We painted them and they look identical and you don't see the layers under the paint.
I don't even print with the smallest layer size possible because it takes too much time and difference between 0.1 mm and 0.2 mm or even 0.3 mm is negligible in practice, but the printing time scales almost linearly.
Depending on what you use it for YMMV.
These aren't very good examples of what is currently possible using a decent FDM printer. (Prusa Mk3s for example) While not beautiful, PLA prints are pretty darn smooth and surprisingly strong if you orient the print in the right direction. (you don't want to stress the part on the axis of the layer stacking)
I bought one a year or so ago and while definitely still in the "toy" category, you can do some neat things with them. That said, woodworking is going to let you build way nicer things, but you don't get to just click a button and have it reproduced, nor share plans with others. But ya, FDM doesn't print things that are "nice" really, just functional.
I bought one a year or so ago and while definitely still in the "toy" category, you can do some neat things with them. That said, woodworking is going to let you build way nicer things, but you don't get to just click a button and have it reproduced, nor share plans with others. But ya, FDM doesn't print things that are "nice" really, just functional.
FDM 3D printing like you can do at home, is a toy/hobby. It's great if you want to spend the time fine-tuning your printer and dealing with all kinds of brokenness, and get a mediocre result at the end.
To get good results, upload the files to a service like Shapeways and use their hugely-expensive SLS printers. You'll get your printed stuff in several days and the quality will be great.
To get good results, upload the files to a service like Shapeways and use their hugely-expensive SLS printers. You'll get your printed stuff in several days and the quality will be great.
This is great work... not specifically for 28S, but exactly what 3D printing should enable, more reuse and repair of otherwise working systems.
He needs to realign how the part is printed, the layer height, and possibly nozzle size. He can also anneal the parts for added strength.