oskarzer0 wrote on 2022-10-02, 20:46:

WoW! Good work!

Only burned my thumb a little bit before I realized I had bought heat resistant gloves for this very reason. :3

I’ve burned my fingers so much at this point tho. I’m surprised I have any feeling left in them at times. :3

So today was basically soldering together the buttons with wires, which I had cut after getting an idea of length I would need. Came together a bit quicker than I thought it would, but now I have five working right angle buttons that will fit next to the SoM.

Pretty easy circuit, mostly it’s about hiding the 220 in there for the LED ring light that lights up around the button itself. I made a slight cut on the edge of the heat shrink so that it could “fit over” the connector - you can see the progression of that technique. All in all looks good though.

Next step is soldering it up to the pin header in the right spot.

So good thing that I decided to take a break after soldering, because thinking through how I was going to solder these onto the board I realized I would have no way of attaching it to the front plate. Good thing I took a break, avoided having to desolder wires.

Taking the faceplates out again, and putting the button in and wrenching it down to its final resting place sure was fun. As it turns out, the connector part actually spins freely, even after secured to the plate. Very nice touch of engineering.

Had to share pics because they just look so good. The push buttons work fantastic, too. They have a really nice curvature to the spring as you push it in and as it releases, really glad I spent some time finding something nice. They are like 5$/piece but totally worth it. Should make all the retro enthusiasts happy.

Fooled around getting this dot matrix printer working, and printed out the weeCee manual I wrote up alongside my testing. Printed it out on this really rare green bar tractor paper that I luckily got for cheap, but was only able to print 3 full copies before running out.

Oh well, at least I got as far as I did. That green bar paper, as it turns out, is heckin’ expensive! Trying to find some extra sheets of it and it’s like upwards from 80 ~ 100$ for a box of it. People ain’t kidding when they say retro hardware is so expensive.

Anyways, first few buyers of these units will get one of the print outs that I was able to make.

Just reposting this updated graphic to illustrate where the reset pin is located at, along with the main Vcc/Gnd for the SoM itself, which is where we're going to connect our LED into.

Welcome to the world, little weeCee!

SN7 fired right up, now to run diagnostics to make sure everything works.

And everything looks perfect after flashing the EEPROM for the audio. 😀

Now to slap the S/N on it, take some final pics for the listing, stick it in the encasing, and celebrate.

Edit: Apologies for the messy/dusty desk, didn't have a chance to clean it up before taking pics due to excitement of getting it turned on and such.

Nice work NachtRave!

- Alex

Alright folks! Happy to announce that SN7 is now up on eBay (direct link not provided for sake of forum rules regarding advertising, but feel free to DM if you can’t find it through eBay search).

Finally! It’s always so much work but it’s done. Now I can defrost my solder paste and get to work on SN8.

For those following, you’ll have 5 chances to get one made from me. SN7 is in silver colored encasement, SN8 will be in black colored encasement, and I’ll alternate between the two colors. The silver ones I put the square feet on, and the black ones I put the smaller rounded feet on.

All come with whatever spare parts that didn’t get used, that way I don’t have it laying around here. Usually that includes the optional metal endplates the case came with, and a SD to MicroSD converter thingie which they put with every MicroSD card.

ajacocks wrote on 2022-10-05, 15:31:

Nice work NachtRave!

- Alex

Thanks! It’s always nice when you finally see all that hard work and effort come out with a rock solid system. Very grateful for the opportunity to contribute.

SN8 progress. Coming together rather quickly. Did the SMD work last night and finished off all the remaining components today.

SN7 has a few watchers but no bids. I might do buy now on the last one or two, dunno yet. Kinda crazy cause I’m not really being paid for any of the work hardly, so whatever extra it gets bid up to is literally all I get out of it. The rest goes to just the cost of the components themselves.

I actually today lost the packet of MicroUSB connectors, and had to wind up buying some replacements only to, ugh, find the ones I was looking for stuck between one of the other plastic holders for the various ports and such. Of course fate plays out that way. :3

Oh well, I have spares now in case I ever need to replace one. And SN8 did manage to come together quite quick. Just have some bodge wires to solder and then do some testing. Pretty good soldering job this time, too. All the pins on the connectors checked under microscope and looked solid and clean.

Good enough for today though, I’ve inhaled enough toxic fumes (joking ofc, I have an exhaust fan/filter, not that it helps always the most always). Still though, a good solid day making dreams. Love getting the components in there and soldered up right good.

I couldn’t sleep so I decided to bring new life into the world.

SN8 now operational. I’ll do the enclosure tomorrow.

I hate to always dominate this thread, but here is where I’m at today. Plan on doing just the tiny SMD components on SN9. Save the pin headers and other larger stuff to Monday.

Also showing the remaining boards, and how it feels to start out. To streamline the process I have the bags pre-sorted to do tiniest parts (starting with resistors then caps) on to the larger ones. Each bag has the pad numbers listed on it. The boards have the crystal chip already pre-installed, as that is an entire affair by itself.

Honestly I turn on some tunes and point a fan at me and zone out on it. Just remember to use the right amount of solder paste, flux where needed, and if you need to even out the solder blobs on either side of any SMD component, just drag both sides over the two different pads to try and even out how much solder is on each side of the component. That will help keep it from one side having more than the other (just try not to knock against other neighboring components while doing so). And if you find you’re burning yourself, use some gloves, heat resistant ones if you can.

In case you get too much solder on any pad, you can always put some solder wick on it to “vacuum” up any extra solder, just remember to use a light touch.

With soldering it’s about getting the heat up to where you need it, doing what you need done fast, and then getting the heat off the soldering area. The time that heat is on pad is the biggest factor in burning pads, not necessarily the heat’s temperature value. So when soldering, it’s best to get the heat up, spend as little time as needed to get things set together, and then remove heat.

Anyways, it’s impossible to not mess something up (known as a production defect), as it’s impossible to not knick against the thin outer insulating coating of these on-demand manufactured PCBs. Isn’t the first time I’ve had to use some solder mask to “fix” problematic areas.

On the microscope is one such area on SN9’s audio chip, where one of the contact pads was dislodged after trying to move it back after it had been pushed too hard. It wound up tearing the contact pad slightly off the wafer (run still attached), which is fine since it’s easy to glue it back down with some solder mask.

Still, makes me sad when something like that happens, as I’ve been told by many folk that I’m looking far too into it. If it works in the end, who cares? Well, I try my best, but imperfections like this are unavoidable during hand made production. It takes a lot of years of experience to get good at this stuff, and we’re talking about making a small number of them for people’s curiosity. No way you’re going to get a perfect unit, with perfect solder work on each and every connector.

Regardless, I like to be transparent about these things. If someone asks I’ll gladly point out how it doesn’t affect function. In fact, I chose to use more copper in the main boards of these anyways (which costs extra), maybe they just needed some more time on the machine for the black coating to deposit but, who knows.

Last set of pics for tonight.

Using different manufacturer for these 4.7kohm resistors, so they’re green in color. I put my meter across one and it certainly was correct, but just highlights how the chip shortage affects this.

If you’re making your own, I suggest to get on mouser or digikey and find the right component and then use their menu to substitute as needed. The most important things to get right is package (and/or case), any resistance/capacitance/voltage/wattage values, and match the allowed deviation. Usually it isn’t too hard to find a substitute, just be smart.

You can also use larger BIOS chips for the audio device. As mentioned, I had some trouble finding 2kb chips, but you can easily sub a larger one in.

Another tip for those making these, when you are setting up the position of any component, especially these long connectors that fit on the board that seat the SoM, is to focus on getting a single side of the component correctly set first, that way you can get the orientation of the component just right. Then do the same to the other side - the solder work here doesn't need to be good, just good enough to hold it down. This is where attention to detail comes into play.

This long connector, for instance, has two through-hole guides to help position it, but I like to be extra special on these since I’ve otherwise have had issues come time to push the SoM into the finished connectors. That means I try to just get the connectors fitted as closely as possible to what they should ideally be so I hopefully don’t have to fiddle with it later on.

Also note that keeping the iron too long on one of these connectors can and will melt the plastic that encapsulates it. Be quick with the heat - don't stay on it trying to get it just right while the heat is on it, remove the heat then reposition if needed. Adding some flux here so that you can have solder on the iron beforehand so you just have to press against it is a good use of flux.

Once the connector is laid out, you can go along the length of it (on the other side from the two holding solder joints) and work your way along. I like to use a lot of flux with this connector because it helps the solder flow well, allowing the natural electrostatic/Van der Waals forces to correctly set the solder work per pin. I use a wide tip or a slant nose to do this work, as having the solder iron go between pins allows excess solder to move between the pads.

If you spot bridges between connectors, you can try moving over it again with the iron to help distribute the solder across pads. Along with some extra flux where needed, it shouldn't be too hard to get the solder where you need it to be. If you have too much solder in an area, solder wick can remove it, but it's also likely to suck up all of the solder in that area, so you may have to reflow with more solder in that area.

Also, if you spot the gold pad underneath not being filled with solder, chances are that the pin is just sitting ever so higher than the pad itself (and highly likely not making contact), and you will need to take the iron and press the pins downwards, in hopes that the solder gets into the bottom pad and not just sticks around on the top of the connector pin.

It is really challenging to not scratch the board up while doing these connectors, especially when encountering issue areas that need extra attention. Be careful not to drag the iron along the board itself, as much as you can help it. Again, I put flux across the entire row of pins in addition to the resin core solder, just because it's easier to let solder flow nicely to get nice strong looking connections.

Does anybody know if there is a way to software emulate MMX instructions?

For example, I'd like to run the heaven seven demo but it requires MMX instructions.

I'm guessing it's not possible but based on the other crazy things that exist I'm hoping there is a solution?

sofakng wrote on 2022-10-10, 19:41:

Does anybody know if there is a way to software emulate MMX instructions?

For example, I'd like to run the heaven seven demo but it requires MMX instructions.

I'm guessing it's not possible but based on the other crazy things that exist I'm hoping there is a solution?

That's a really good question, unfortunately as far as I can tell it won't run on weeCee for lack of MMX instructions. Looks like as soon as you run it it stops and says as such, alas. 😒

The only thing I can really think of is maybe to sub out the MMX instructions for the long winded x86 way, but that's probably something best done on the end of the people making that as there doesn't seem to be any sort of utility or otherwise that I've been able to quickly google for that looks promising.

Hi Rasteri, Nachtrave,

I'm curious about the IDE controller problem you guys have been having with the weecee, can you boot into XP and get me screenshots the HWID it has as well as the driver files it uses?
I'm quite sure it should be possible to get this working properly.

edit: I've already got a couple of ideas for how to get this IDE controller working, just need the Hardware IDs.

Also, can you get me the HWID of the XGI graphics chip in the weecee, I might have found a driver that could give OpenGL support in Windows98, possibly?

OK,
This is my 1st attempt at modding the driver based on what I saw looking at the driver with a disassembler. Please try it out.

If it doesn't work I may have gotten the Hardware ID's wrong (Tip: This is why no driver has worked as of yet for 98 - The Hardware ID checks are hard coded into the driver binary itself, which is annoying and entirely unnecessary)

Her

Duffman wrote on 2022-10-12, 16:07:

can you boot into XP and get me screenshots the HWID it has as well as the driver files it uses?

Here you go. The sysinfo.txt should have all the driver files listed.

I couldn't get your drivers to work on windows 98, but my install is so broken at this point that I'm not suprised. I'll have to reinstall it to know for sure.

Will do as soon as I can.

Thanks alot for the Info. should help.

Ah, I see that the Hardware ID for the IDE controller is PCI\VEN_17F3&DEV_1011 - the driver currently expects PCI\VEN_17F3&DEV_1010 since I wasn't sure which one we had in the WeeCee.
I'll get that changed, fingers crossed hopefully it'll work.

Also, here is that VGA driver I found that might give OpenGL in Windows 98, just an INF mod here, hopefully it works. Give it a try.