Jump to content


New Members
  • Content Count

  • Joined

  • Last visited

Community Reputation

54 Excellent

About Nelno

  • Rank
    Space Invader

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Not quite. I didn't have a connector (i.e. the receptacle part) at all on the motherboard -- it had shattered and been removed back in 1987. Neither did I have an original transformer, cable or plug. I didn't "have to replace both the connector in the computer and the one on the cable" because I designed a receptacle that didn't fit. I had to replace all of that because I had none of it, and I designed a receptacle that doesn't fit the original because I have no references or specs for the original plugs aside from the pinout voltages. Like I said, this would only be useful to someone who's lost both the power supply and receptacle on their motherboard. If I did have those specs, I'd consider designing a full replacement that matches the original parts. That gets tricky to do without the physical parts to test and validate with, though. Even with measurements it would be a guess as to the tolerances and whether or not it would really match the originals.
  2. I've posted a project for a power plug / socket replacement I did for my own SV-328 in this forum here. Thanks to everyone who contributed to this thread. It helped me tremendously!
  3. Back in 1987 my Spectravideo SV-328 got knocked off a desk (power cord was tripped over or something, I can't remember) and the plug receptacle on the motherboard was broken. Undaunted, my 16 year-old self cut off the plug and soldered the power cables directly into the motherboard. Unfortunately, I left a little too much wire on the bottom and it contacted the shielding when the case was closed. This fried the transformer. Fast forward to 2020 and I got a 3D printer. I've kept my SV-328 all of these years and had promised myself I'd fix it one day. This is the result. This is a plug and socket made to replace the plug and socket in an SV-328 computer. WARNING: BY USING THIS PLUG AND SOCKET YOU ARE DOING SO AT YOUR OWN RISK. YOU SHOULD VERIFY THAT THIS SOCKET MEETS GUIDELINES FOR POWER PLUGS WHERE IT WILL BE USED. THE SV-328 HAS AC POWER COMING DIRECTLY INTO THE MOTHERBOARD THROUGH THE POWER RECEPTACLE SO THERE IS A REAL RISK OF ELECTROCUTION IF YOU DON'T KNOW WHAT YOU'RE DOING. This will only be useful to you if you no longer have a working original SV-328 power supply, receptacle and plug. If you have the original plug and socket you may be much better off reusing that with another power supply vs. printing this. The plug is not compatible with the original SV-328 power receptacle on the computer. I did not have the original (that is the first piece that broke) to even attempt to make it fit. There is another project out there somewhere (sorry, I don't know the link) that is a replacement plug for the original socket. This plug is designed to be used with 0.062 inch Molex connectors. The male connectors go in the receptacle and the females in the plug. This is important because YOU DO NOT WANT LIVE AC VOLTAGE ON EXPOSED PINS ON THE PLUG (POWER) SIDE!! YOU MUST ABSOLUTELY VERIFY YOU HAVE THE PINOUTS CORRECT BEFORE PLUGGING IN. DO THIS WITH A MULTI-METER TO VERIFY THE CORRECT VOLTAGES ARE ON THE CORRECT PINS OR YOU CAN DESTROY YOUR SV-328. Molex connectors I opted for two separate wall transformers because these were easy to source on Amazon: 16VAC, 1000 mA wall transformer 9VAC, 1500 mA wall transforme I've released all of the files, STL + FreeCAD, on Thingiverse here. I'm releasing this to the public domain in case it's useful to someone else later. This thread on AtariAge discussing Spectravideo power supplies was of great use to me in this project. Anecdote: When I first powered on my machine on with these transformers (and an a hacked up connector using a 4-pin DIN connector, which I later decided was a bad idea because it had the AC power on semi-exposed pins) it worked fine until I typed in the following program: 10 print "Hello, world!" 20 goto 10 Upon typing "run" the computer's video signal promptly died. I probably should have made it "Goodbye, world!" instead. This turned out to be a bad transistor in the video out logic that was almost certainly an age-related failure and nothing to do with the new power supplies or connectors as nothing else on the board was damaged, and the power all checked out fine. After finding and fixing that it's worked flawlessly. This just goes to show you -- you never know what will happen if you decide to do this and it may just expose something else that's wrong with a 30+ year old computer that is completely unrelated, so do it at your own risk. I take no responsibility for what may happen should you decide to use or misuse of this project. Be careful not to fry your Spectravideo, as there aren't a lot of them around! 3D models: Pinouts, as well as the installed receptacle: \ Final plug: The transformers I used:
  4. The possibility of breaking up the PCB into 5 rows did occur to me, but this would: - make it more expensive to produce - make the structure less rigid - make it more difficult to secure in the case (some sort of additional support framework would be necessary, I think) Because of that I think an adapter that allows mounting the keys at a slight angle (I believe it is 15 degrees) seems like an easier option, as those could almost certainly be 3D printed. An adapter will also be necessary to mount any original keys to modern switches. To be clear, I haven't tried that yet, and mounting to the Kailh low profile switches is probably going to be more difficult than mounting to regular Cherry or Kailh switches because of the dual prongs used by the low profiles. To allow for regular, non-low-profile switches I think I can take the basic design, deepen the switch substrate, and adjust the mounting brackets for the PCB by simply shortening them. I'm not sure if the regular profile switches use the same pattern for the connectors but that's probably just a minor inconvenience since everything in the CAD files are parametric and use arrays for key placement. But my hope here is that this will end up being only a minor modification to the low-profile design.
  5. Thanks. KiCad is what I decided to start with after trying a couple of other programs. I've been able to import the outline of the PCB into KiCad from FreeCAD, but I didn't know there was a library for Kailh switches, so I'll look into that! I only use FreeCAD only for the parametric modeling of 3D printed parts. It doesn't have any native PCB tools but it does have an open source addon that will allow importing from KiCad. However, in this case I'm trying to go the other direction. I've mocked everything in FreeCAD and used 3D printed stand-ins to ensure everything fits to my satisfaction and now I just need to design the PCB to exactly match the shape and dimensions from the 3D printed prototype in KiCad, which is what I've done here: Now I've just got to reverse-engineer the TI's traces and make my own, along with adding the connector, which I hope is some standard part still in production. I did some reverse-engineering of the traces already in DesignSpark a while back, but I wasn't using the final board shape or anything and I intended it to be throwaway.
  6. Excellent. I thought that was the case, but it's good to have it confirmed.
  7. I've been working intermittently on a project to create a replacement for the TI-99/4A keyboard using modern keyboard switches. So far I have the PCB shape and key placement working well enough to slot into the case. In the attached photos you can see the 3D printed part representing both the PCB (bottom layer) and the keyboard switch substrate that the Kailh low-profile switches plug into. In its final form there will be a PCB on the bottom layer and a 3D-printable layer on top of that which the switches plug into. Some additional clips that are screwed into the original screw holes on the case should hold it in place, though I haven't designed those yet which is why you see a bunch of tape holding the prototype onto the case. This is an earlier prototype. It's printed in two pieces and joined because my 3D printer isn't large enough to print the 24 cm wide piece in one go. The raised area is the substrate for inserting the keyboard switches. They just snap in. The lower flat area would be the separate PCB, secured to the substrate with some screws. I chose the Kailh low-profile brown switches because: a) I had some of them b) the low-profile switches with low-profile keycaps match up nearly perfectly to the height of the original keys with just the PCB + switch substrate. Using full-size switches would require a much thicker substrate, which is not the end of the world, but would take a lot longer for me to print prototypes. c) I think browns are probably the closes modern approximation to the TI's Hi-Tek keyboard. I've also designed 3D-printable low-profile keycaps for these Kaihl low-profile switches, but I am expecting that some injection-molded keycaps would be used in the final build. One issue with this is the tilt of the keyboard in the case doesn't leave much option for adjusting the angle of the top of the keys. In the original Hi-Tek keyboards the tops of the keycaps are level. In order to do that with modern key switches, the only real option is to make the keycaps themselves attach to the switches at an angle, because the switches themselves must all plug into the single, slanted PCB. This would mean custom keycaps, or some sort of adapter for original keycaps. I'm not too worried about this, though, because in modern keyboards its natural for the keycaps to not be level when the keyboard is tilted at an angle (i.e. by the keyboards back feet). Ideally anyone building this could just use standard low-profile Kailh keycaps. In retrospect, using the Kailh low-profile switches may not be the best choice because keycaps are much harder to find than they are for Cherry or Kailh standard switches. The main challenge I will have is PCB design. I have exactly zero experience with PCB design, so I'm learning as I go. Currently trying to get a good workflow for FreeCAD to PCB design software. Fusion360 + Eagle is no good because it only allows 80 cm^2 PCBs in the free version and I'm not paying Autodesk's insane software rental fees for a project I intend to open source. At any rate, the PCB for a keyboard seems like one of the easiest things to start with and a good way to learn. I started this due to my experience replacing the plungers on my Hi-Tek keyboard and realizing these keyboards are relatively rare compared to the Mitsumi membrane keyboards. I don't know if this would fit in a case that takes Misumi keyboards, since both TI-99/4As I own have Hi-Tek keyboards. Are the various keyboards interchangeable? My end goal is to provide a replacement keyboard with a modern feel, using modern parts that are easier to source, but hopefully something that looks close to and has the option to be, at a glance, difficult to distinguish from the original keyboards. Interested to hear any feedback. Am I wasting my time with this? Would anyone else find it useful? Has this already been done? (I couldn't find anything indicating it has when I was working on my keyboard plungers about 6 month ago).
  8. This 1 mm extra lip is probably an improvement over my design (and the original pieces). First, it makes the part of the plunger holding the key cap, which is a weak point in the original pieces, somewhat stronger. Second, it allows the hole at the top to be beveled, which will make it a bit easier to insert the key caps. Third, it will also act a bit like the optional brim I added to help the pieces stick to the printing surface. Finally, it may prevent the pieces from falling all the way into the waffle frame if you are testing for fit without the spring, though that shouldn't happen if the nubs are large enough, but did happen to me a few times during iteration of my design.
  9. Turns out that a Tech Tangents YouTube channel else has already gone through this process in 2018, made a YouTube video about it, and even released the model files on Thingiverse: Comparing these to the ones I designed: - The X/Y inner and outer dimensions are exactly the same. - The nubs on the bottom are less pronounced, which would make them easier to pull out than mine. - The inner spring rest is about 1.7 mm lower than mine. This may affect the rebound way the keys feel. I measured this again on the original plungers and it's 10.5 mm from the bottom, the same as my model. - There's no beveling on the outer corners (I included beveled and unbeleved because it might make difference depending on how your printer slicing software works, and whether you use PLA, PETG or some other material) - There's a lip on the top that doesn't exist on the original pieces. - I maintained the split in the pieces, though for PLA parts this can be brittle depending on the quality of your filament. For PETG it helps with flexibility. - I angled the inner spring rests at 45 degrees on one side so that they print without any supports on any printer than can handle 45 degree overhangs. The Thingiverse version has horizontal overhangs which are unlikely to work at all on my printer. You definitely don't want to have to add and clean up supports on a piece this small. I haven't tried these in a TI, but the Thingiverse link (https://www.thingiverse.com/thing:2871775) does say they are also for TI. I searched all over for something like this before starting, including on Thingiverse, and never saw this, but from what I have been able to tell, Thingiverse has a terrible search engine. Anyway, if you're looking to replace your key plungers, you now have two options to try. One or the other may yield better results for your printer.
  10. Thanks. I have a 4K resin (SLA) printer on the way that, supposedly, will print with an accuracy down to 0.01 mm. Even though I've got the PETG ones working well, they still need a bit of sanding. I am hoping with the SLA printer I will be able to simply print them and be done without any sanding. I was able to also fabricate the copper contacts used for each key to replace the one that I broke. I purchased a 0.15 mm thick copper roll and basically cut these pieces out with a tin snipper and bent them into the shape of the one original that I removed without breaking it. I actually received 0.25 mm copper roll, but it worked perfectly well. The original contacts are 0.16mm thick and that is extremely hard to source. I only found it one place after hours of searching, but I went with 0.15 mm that had a somewhat sooner arrival time (still took about a month to get here). Attached are some pictures of the contacts. Probably the most useful is the profile pics that show how the contacts need to be bent to actually work correctly (touching when the key is depressed and separated when it is not). You can see from the pictures that I didn't bother with the tabs on the outside, or the retaining clip that is punched in the middle of the original. I couldn't find any way to reproduce the clip with the tools I have, and I tried the tabs but accidentally cut them off. My tin snips are very unwieldy. I don't think this will be a problem, however, once I solder the copper strips in. The clip really only keeps the strips secure when the solder is removed, but it doesn't prevent them from wiggling around like the solder does. The tabs on the side just prevent the copper strip from going too deep, but the same can be achieved by careful sizing of the part and the notch that is cut on the bottom to make the solder pin. To be really useful I need to figure out a way to reproduce these with more precision, though I would recommend not breaking them as the best solution. My current thinking is to make a 3D model of a form that can be printed, then the copper strips can be bent to conform to that. Another possibility would be to make a 3D model of the contacts, make a mold from that and then poor in molten copper to make the contacts.
  11. Update: Printing these with PETG works much better: - it is more flexible, which helps when inserting the plungers, which are designed to flex a bit to allow the nubs to go in. - it seems to print more accurately and though it's more "stringy" it ultimately requires much less sanding an prep to get the pieces to fit. I've attached a slightly updated file with the "nubs" fixed to match the original (affects depth after insert) and made minor adjustments to the wall width to accommodate (what I found to be) more accurate PETG printing. hi-tek_key_plunger_B_3.zip
  12. I've been wondering... has anyone tried to fabricate a replacement stainless steel exterior for the TI-99/4A? I have two 99/4A's and my original one is missing the lower metal exterior piece. I have no idea what happened to this. I seem to remember it was slightly bent and was always coming up and getting in the way so I just removed it. I was probably < 12 years old then and didn't care what happened to it. It seems to me these could be fabricated from some sheet metal and a form, though I have no actual experience doing that. Another option would be creating entirely new pieces from plastic, sort of like the custom XBox 360 face plates you used to be able to buy, to customize your 99. Also considering getting back into TI-99/4A assembly programming and making a game. I spent 18 years in the AAA video game industry and there's an argument to be made that I owe my career to my TI-99/4A. I could never afford the expansion bus when I was younger, so my only option then was the mini-memory, which had some pretty terrible limitations, iirc.
  13. Replaced the 10K ohm thermistor (temperature sensor) in my spa heater, made some minor changes to the Hi-Tek keyboard plunger and uploaded it here!
  14. I was considering the entirely new PCB approach, too, but that would involve a fair bit more learning on my part. The PCB for these keyboards is pretty simple, though, so I don't expect designing a new one with the same functionality would be very difficult. However, since the modern switches are quite a bit shorter than these Hi-Tech High Profile switches, and because you want the switches to be held in place by more than just the soldered contacts, some sort of support structure to snap the switches into would still be required. In the end my guess is I wouldn't save anything by designing a new PCB. It might make such a project easier overall by removing some restrictions and limitations, but the extra freedom in key placement might also require a lot more prototyping to get things right. Still, I am intrigued by the possibilities here. I just doubt I would have the time for such a project right now and I'm not sure it's really how I should spend my spare time, considering my work and some of the other projects I have in flight. But honestly, the things I could learn doing it do make it attractive...
  15. I feel like the plunger is in a pretty good place to release the source files now. I decreased the width of the internal bar from 1.6 to 1.5 mm and I changed the lower cut-outs to have a tapered top which I think provides a little extra strength. In printing with some gray filament I picked up I had issues with these snapping right at the top of those bottom cutouts, so I decided to try and reinforce that a bit. I haven't had any of the ones I printed in white snap, though, so I think that other filament is just poor quality. I don't think I have the notches quite right. After inserting a polished one, these sit maybe 1 mm lower than the original and I believe it's because I've got the side notches a little too high. I've attached a zip file with the FreeCAD source file, plus two STL files, one with filleted edges and one without. The walls are so thin that slicing may cause some artifacting with the filleted version, so I included the non-filleted one just in case. It should be rotated to the correct angle for printing. If not, make sure the protuding nubs are on the top. All of the overhangs are angled from that direction so they should be printable. If it fails on overhangs it's probably rotated the wrong way for printing. Any feedback you can give would be helpful. See the readme file in the zip for more details. hi-tek_key_plunger_A_17.zip
  • Create New...