wierd_w

Since this a new unit for you, I expect that you will be in need of both a speech synth enclosure, and the tipi+32k enclosure. I expected that 2 days ago, which is why I started printing the speech top to go with the speech bottom I had printed just before it. With any luck, the printing will be done by either the end of tonight, or tomorrow. (currently at about 70% printed, according to the LCD on the front of my cheap printer.) Paradoxically, I have found that warpage is reduced if I turn off the heated bed, at least for the desert tan. (the PVA glue on the bed seems to be sufficiently adhesive.) We will see if my modifications to reduce the need for sanding bear fruit. Also, I think your image is the only one I have seen that is clear enough to see that the beige systems have that "textured" plastic finish. (the kind that harbors filth.) Perhaps my lacklustre sanding job I did on the "engineering sample" I mailed to arcade shopper was not far off then. Good to know. Anyhoo... The speech top should come off the printer "soon." Then it is time to do the "without door" versions of the base and top. (I think I will barely have enough left over for that, but I am unsure.) I should get the new roll of material by then. Amusingly, when it comes to the idea of embellishments, I have considered monochrome black sprite image cameos, done with some horizontal bars framing above and below them, from popular game titles. Things like the parsec ship and some swoopers, the wumpus, munch-man, etc.. I dont have time to doodle the idea here at the moment (getting ready for work), but if I am in a light duty unit tonight, perhaps I can doodle then, and get some feedback.

Wow, you actually found one? Great! You should try converting some BASIC programs for it.

Nuh uh! it's Personal Computer Memory Card International Association.

So, it's been like 10 days, have you taken the plunge?

In my case, I was the "child computer prodigy" (ahem), and had an early summer job at a local slave-pit (erm, small operator computer store), doing grueling benchwork and phone support. From the early to mid 90s, through the early 2000s, I was there in the trenches, getting my hands cut the hell up by those abominations. Seriously, when you have to wear welding gloves and go over all the metal edges with the blade of a screwdriver to take all the burrs off, JUST so you can safely work inside it, there is a problem from the factory. Throw into that, the obscenely nasty habit that manufacturer had of putting SRAM sockets on the boards for external CPU cache, and then NEVER POPULATING THEM, and setting the jumper to "No cache", (in the days when SRAMS were still pretty expensive)-- it got real tiresome, real fast, trying to explain to the customer why the fancy new overdrive CPU they had installed, still ran like it had a concrete butt plug holding back a dumptruck load of wet cement. "What do you mean I need to add 256kb of cache memory? It's got 4mb installed!!" and other such glorious groaners-- and then they would see the prices for SRAM and about soil themselves. The sad part, is that these same customers could have been vastly better served had they just bought a cheap assed MSI, TDK, KDS, or Shuttle motherboard, and a modular case. The price difference was not worth the pain. It really wasn't. Those systems were subsidized on pure pain and suffering, for both the user, and the support people they INEVITABLY took them to, while those generic board makers actually installed at least a token amount of cache memory, and the boards were actually rather solid (most of the time anyway) (And the people making modular cases knew what a deburr tool was! I am serious, I have MF-ing scars on my hands from those PB bastards!) You have no idea how happy I was when that company went out of business. No idea at all.

I admit that I am most familiar with MS Quickbasic. It has some useful features, but it is sorely lacking in others. (No microsoft, PSET based PUT is not a substitute for having a proper sprite routine!!. I DO appreciate having MODE X though.) "Which you like best" or worse yet, "which is best?", is very subjective. Best for what purpose? Because each of those console BASIC implementations were just as "intended role" biased as the computers they ran on. As a pixel artist, (who does not really find pleasure in having to beat his head against walls to overcome engineering choices of yesteryear) I do not favor most 8bit BASIC implementations' graphic routines. I need at least 4 simultaneous colors to make anything really nice (without going down a pyrrhic pit of despair to accomplish anyway), and most 8bit BASIC implementations did not want to offer that. the holy grail in a BASIC implementation for me, would be one that offers true 2, 4, 8, 16, 64, and 256 color modes, with hardware sprites (or at the very least, quality software abstraction that is fast), and a decent sound generator. However, if you ask somebody that did math heavy things, they would favor some other basic. (maybe one that does better recursion.)

I think it is meant in "Subjective 'I start workload, and system does not drop into a coma trying to do it' metrics of power. Some of the 68k processors were hands-down better designs than their x86 peers, for instance-- as regards how they accessed memory, did memory mapping and protection, etc--- but the complex instructions inside actual x86 chips of the time could accomplish certain kinds of compute-heavy workloads in fewer cycles. Like anything in the real world, the devil is in the details. Early amigas for instance, were highly timing-cycle optimized, with lots of support chips doing all kinds of heavy lifting. This was both beneficial, and detrimental, depending on how you look at it. (Beneficial in that the CPU was not having to slog through pedestrian tasks, like graphics data IO, but on the other hand, it meant that future upgrade paths were limited and hindered, as later series chips were developed.) All those older systems had areas where they shined, but they did it by accepting warts somewhere else. It was the era of compromise all around-- you had to decide where you wanted to compromise, and then buy appropriately.

I dunno.. I ... (and my poor, scarred hands) remember Packard Bell. I swear to the unholy god of technology, those people who made those things had no fucking clue what a deburr tool was. (or that external cache on anything newer than a 386 is NOT OPTIONAL, regardless of what the spec sheet says.) So many painful support memories from Packard Bells. I would not wish that on somebody who is young and unsuspecting. Give them a nice period Dell instead.

EISA lived mostly in 386s, but also in some early 486s. Mid 486 had VLB, and "VERY LATE" 486 had some early PCI implementations. (though those are semi-rare to find.)

There's hardware devices that service this niche, IIRC. https://www.simulant.uk/shop/retro-vintage-computer-wifi-modem-rs232-serial-hayes-compatible (just an example, there are quite a few such devices.) If the price was not in the stratosphere, I would just suggest a Xircom Pocket Ethernet 3 though. (It plugs into an LPT port. Slow, but gives you a real ethernet port you could attach to whatever switch, bridge, or other appliance you want to get online with.)

I think I am gonna have to order more desert tan. The top parts of the enclosures use a lot of material. (This is fine, I personally hate flimsy enclosures; these things are paired up with systems that have withstood multiple decades of ill-treatment, more often than not, and weathered it with aplomb. I think any peripheral enclosures need to meet that same standard. You need to be able to play baseball with the things, in my opinion; and that's how I designed these.) I am gonna have to do test prints of both "No door" and "With door" models, and I don't think I have enough tan left. (lots of wood color, but the desert tan is in high demand.) AND 2 rolls ordered. That should get me through this project. When christmas rolls around dont forget me ok?

The LED does not protrude. It is not long enough to protrude. One COULD fill the small cavity partially with epoxy, so that it acts like a light tube. I think that is what I will do.

And, an "initial design" that has the door.

OK, how's this look for a "No Door" version? Note the tiny hole for LED activity light to shine through. (It's a swept hole, that is bigger on the inside so that the LED can fit, but this allows the foot/wall to retain more integrity.) Here's the back.

OK-- Investigating the fitting of a TIPI+32k in this enclosure-- the top will DEFINITELY need revision. There is collision with a functional component with one of the ICs. this can be addressed fairly easily. I will just lop off that part of the retainer foot. It should still have sufficient meat to hold the stand-off, with that part of the lobe hacked off. I will need to dog the corners so that it overlaps the board (eg, presses down on the PCB once top shell is screwed down), to have positive retention of the board. Otherwise, the 32k board will "just barely" fit in the bottom component of the enclosure. These are the corners I will "dog". I will also cut holes for the LED and for the USB power for the 32k board. Then it's on to figuring out how I want to deal with the door.

The recess will make it difficult to sand after backfill... If we forgo the recess, not hard at all. Just need the font.

It would be grossly inconvenient to have to disassemble the enclosure whenever you want to futz with the SDCard. J-Data also just responded a few minutes ago to my STP file request, and delivered. I am investigating the data now.

SWEET

Good news Omega, I found a suitable vector image of the TI logo I can (ab)use. A little prodding with inkscape, and it becomes a suitable .dxf, and VOILA, it's in my cad software and I can use it to make a cutout in that shape. Looks like tonight I am gonna be wrangling the blueprint drawing of the 32k+tipi board for speech synth, and cobbling together my own digital mockup, then doing modifications to the existing enclosure data to give it a print-in-place door. I wasn't really intending to embark down this road just yet, but I guess that's what will happen.

You haz rezin printerz? Why no print?

He has the individual bits and bobs it is made from in STP format, but not the whole assemblage.

No it's not. The speech synth has 3 screw holes that I am using to mechanically restrain the board with. This device has no such provision. I will have to mechanically restrain it with an overhang in the top half of the shell, and I would like to avoid colliding with anything when I design such provision. A STP file will give me all I need.

The way I am holding the speech synth down is mechanically with screws. There are no such screws, nor provisions for screws, in the 32k board. I have requested a STP file in the appropriate dev thread for the board. I can import that into my cad software, and see about coming up with a clever solution to holding it down. Vertical clearance inside should be no problem. There's a LOT of it in there. The issue is preventing it from just being loose and rattling around in there. I might "dog" the corners of the top part of the inclosure in so that they overlap the corners of the 32k board, and thus there would be no wiggle room once you screw down the top of the enclosure. But I want a model of the board for digital fitting all the same.

Hey Yo, I have had a request to make a beige enclosure that can fit this-- Can you get me this 3D object in STP format? I dont need pretty colors or anything, I just need an accurate volumetric model.

Hmm.. I am unsure that my enclosure will properly hold a tipi+32k board. All the pictures I have seen, seem to suggest it has different screw hole patterns. I would need such a board to properly test fitting. EDIT Actually.. wait.. It's open hardware right? Is there a CAD model of the board?