Perhaps. If I want a bad stomack. I will go for that pizza with 40 slices of jalapeno pepper and extra chilli dressing instead of sugar. 😁

orinoko wrote:

jarreboum wrote:

Turns out my dead clicky springy IBM keyboard wasn't so dead after all! After an extensive repainting of most of the membrane tracks, every contact were finally working. I took some time to clean all the keys and mount everything back, and while all keys are working ( I'm typing with it right now), some keys seem to have a faulty spring as they activate at the gentlest touch, not a normal keypress. Ugh, it means I'll have to open it again to change them. Removing keys and reseating them is only fun the first time, after the fourth time you just want to be done with it.

Please tell me your repaired a Model M2!! I tried to redraw the membrane traces using a liquid silver conductive pen but the resistance looked way too high and putting it all back together just to test it is a pain...

I did! some of the corroded traces were near the board and inaccessible without taking it out, which made me decide I would be better of buying one that works. Unforeseen expenses made me rethink that goal, and after a long struggle I managed to extract the board off its plastic cradle without breaking anything. I can see why people say they break easily though, everything is so tight you need to put a lot of force on the legs and breaking them is very likely. I think I was lucky on that one.

Working with silver ink is tedious. The conductivity goes up as the ink dries, and I found no difference in resistance over similar distances compared to the original traces. There were some instances where I had to scrub away the protective varnish, again very carefully as not to remove the underlying trace. The solvent of the silver ink dissolves the traces too, so it plays a role in the initial lack of conductivity. I found this out as I spilled some ink over the neighbouring traces (which would connect everything upon drying): when I wiped it out, I wiped the traces with it! Luckily they were all straight lines so I could redraw them easily. I waited 24h after each application before testing, as the smallest wet spot can prevent conductivity.

Finding the corroded part is key. It's often a small bit that spoils everything, and when that bit is under varnish, strategic scrubbing needs to be done to pinpoint the fault and repair it.

I also thoroughly cleaned the contacts on the board (it's a simple edge connector), and tested it by half-inserting it and holding it tight while pressing on the membrane with my finger. This was mainly to test if there was any short and if the inputs were somewhat reliable, though the unseated board can add to the errors. Using a dedicated keyboard testing software helps a lot.

I did break some of the plastic legs holding the keyboard together though, which made me experiment with new ways of gluing plastics. I had bad experience with traditional super glue on plastics, so I used tough but elastic super glue (the Loctite gel control has rubber particles in it to make it less brittle), then applied epoxy around the fractured bit to consolidate it further (with 24h drying period between every step). So far it's holding well.

I learned a lot from that experience, in the unlikely event that I will have to repair another keyboard (in which case I will say suck it, I'll buy a new one.) Mastering silver ink is cool though, I might use it now instead of rewiring faulty traces on boards.

So I finally got around to fixing my 386 machine today... I've rebuilt it 3 times already. Initially it was a 33MHz soldered SX, then a 33MHz socketed DX, then a 40MHz soldered DX with cache and finally I switched to a slightly faster socketed DX with VLB.

Initially it had a 1MB Cirrus Logic ISA card, a GUS GF1 (Primax Soundstorm M16C) witch ended up in my K6, a 3Com ISA lan card, a winbond multi I/O card and an Adaptec 1542C SCSI card + double height 1GB Toshiba 3.5" drive. Before switching motherboards I I replaced the GuS with a SB16 for better SB compatibility and the Cirrus Logic with a VLB Trident 9420. Removing the GUS and replacing the ISA cirrus logic card with the Trident messed up win 3.11 so I tried to reinstall it - to no avail. Turns out the old 1GB HDD developed bad blocks and attempting to fix them would cause the machine to hang - so I removed the SCSI setup and replaced them with a 1.2GB WD HDD, on witch I'm currently installing DOS and win 3.11.

This is the board I settled on:

I'm going to bed but I hope to finish setting it up tomorrow if nothing breaks again.

Fiddeling around with the connector for external devices on my Adaptech 1542c-ISA controller.
I have a centronics 50 pin cable, and I have this 50 pin cable with a centronics plug.

So I rigged up some drives with the right device id, connected them on that cable and connected
it to the conputer, using the centronics cable.
I went into the controller BIOS and disabled the adaptor terminator and enabled it on the Zip-100.
I do not have any Zip-100 media, so it will only function as a terminator.

Installed the EZ SCSI suite in Dos and rebooted. Loaded up the format tool from the suite and
made a small 800 megabyte test-partition on that IBM 10.000 rpm drive.
Everything worked beautifully.
Once I learned how to jumper everything right, I think SCSI is somewhat faster and easier to set up.

The plan now, is to build my own external drive case from a small ATX case.
Need lots and lots of cooling, and I can use the ATX psu for powering the drives.

Was really fun to learn. 😀

Kanecvr:

Looking good. Great job. 😜

SCSI is definitely easier to set up, and the controllers have less CPU overhead - witch is why I wanted to use a SCSI drive in my 386 - but eh... I think I still have some SCSI drives but I don't know is I can use them with the AHA1512c or how large they are...

kanecvr wrote:

SCSI is definitely easier to set up, and the controllers have less CPU overhead

Speaking of - Is there a good piece of benchmarking software for observing I/O and CPU performance from one controller to another? A way to benchmark different I/O cards, basically. I may have an opportunity to compare a few in the somewhat near future.

kanecvr wrote:

SCSI is definitely easier to set up, and the controllers have less CPU overhead - witch is why I wanted to use a SCSI drive in my 386 - but eh... I think I still have some SCSI drives but I don't know is I can use them with the AHA1512c or how large they are...

What size and what interface are they?

If you look closely on my IBM drive, that are attached externally. I have used a 80-pin to 50 pin adaptor.
The controller detects that drive without problems. I just had to use the format util from the EZ suite.
It can not make partitions larger than a certain size. I don't know exactly how large.
As of now. The drive has a partition that are smaller than one GB.
Anyway... It is quite possible to get a 9gb 80-pin drive running on a 1542c controller.

Did some more benchmarking around 7800 GTX SLI and some other small things here and there. Also created my personal timeline, this will help me plan a little bit better ahead with my videos. Big picture decisions so to speak. Tweaked the benchmark graphs a bit as well.

Quite a few items arrived from eBay, but I haven't gotten around to test any of it yet 😵

Lots of small bits. Cache chips, BIOS chips, adapters...

brostenen wrote:

What size and what interface are they? […]
Show full quote

kanecvr wrote:

SCSI is definitely easier to set up, and the controllers have less CPU overhead - witch is why I wanted to use a SCSI drive in my 386 - but eh... I think I still have some SCSI drives but I don't know is I can use them with the AHA1512c or how large they are...

What size and what interface are they?

If you look closely on my IBM drive, that are attached externally. I have used a 80-pin to 50 pin adaptor.
The controller detects that drive without problems. I just had to use the format util from the EZ suite.
It can not make partitions larger than a certain size. I don't know exactly how large.
As of now. The drive has a partition that are smaller than one GB.
Anyway... It is quite possible to get a 9gb 80-pin drive running on a 1542c controller.

Dunno - here:

It seems the 386 still won't cooperate... I got everything up and running except for networking. I think something is interfering with the LAN card's base address... possibly the sound card? The LAN card was on 300h, but maybe for some reason it wants to stay on 330h - same address as the midi port on the SB16... I get "cannot get address from DHCP server" when starting windows... I set the gateway and DNS correctly, I double checked...

Is there any way to check what base address the lan card is using? Like a utility that enumerates hardware / address / irq / dma?

Hmmm... It's a 1gb drive, according to this:

http://www.4drives.com/4drives/MK537FB.htm

My SCSI controller is set up to IRQ-11, DMA-6 and I/O-340.
Have you checked if there is an INI/CONF file in the directory, were you have the network drivers?

EDIT:
I missed that you are running Windows. Assumed that you were running MS-Dos 6.22 and prior.
Try testing this out in Dos, and see if if can function together, using the SCSI drive.

Try this setup:
- Soundcard: IRQ-7, DMA-1+5, I/O-330
- Network: IRQ-5, DMA-3, I/O-300
- SCSI: IRQ-11, DMA-6, I/O-340

brostenen wrote:

Hmmm... It's a 1gb drive, according to this: […]
Show full quote

Hmmm... It's a 1gb drive, according to this:

http://www.4drives.com/4drives/MK537FB.htm

My SCSI controller is set up to IRQ-11, DMA-6 and I/O-340.
Have you checked if there is an INI/CONF file in the directory, were you have the network drivers?

Yup, 1GB drive. Unfortunately it's dying. I/O cuts off while reading or scanning the drive for defects. I replaced the whole setup with an IDE HDD.

The LAN card seems to work fine but for some reason my router will no longer assign it an IP address.

kanecvr wrote:

Yup, 1GB drive. Unfortunately it's dying. I/O cuts off while reading or scanning the drive for defects.

Have you tried a Low Level Format (found in the controllers BIOS)? Just for the fun of it. 😁

brostenen wrote:

kanecvr wrote:

Yup, 1GB drive. Unfortunately it's dying. I/O cuts off while reading or scanning the drive for defects.

Have you tried a Low Level Format (found in the controllers BIOS)? Just for the fun of it. 😁

^^Yes^^

Worth a shot. A few times I've seen SMART errors clear up after I DBAN a drive. When you touch every sector intentionally, you give the drive controller a status update and it can make adjustments accordingly.

TechNoirMK wrote:

Got an AcerPower/M Ultima PC along with an Acer 6311 K9R/RU keyboard for free. Now for the question - how the heck do I open this case up?
It seems to be somehow screwed inside the front panel, so it's impossible for now without getting thin pliers or just breaking it, which I don't want to do.

http://imgur.com/a/DGJ2D

You unscrew the back and there are big clamps on each side of the front panel you squeeze and pull

I'm sweating and swearing my ass off putting together a shelving system to contain (part of) my retro collection. Damn Whalen Systems for selling kits with bent parts in the boxes that I have to spend hours straightening out with a mallet.

Well, I WAS going to be swimming in PCI Powermac sweetness today. Responded to a listing on cragislist but when I was about to go pick it up they told me it had just sold.

Think I'll be cataloging what parts I have left and deciding exactly what I want to build and start selling off the remainder. Wife is really on me about how much clutter there is.

Out of sight, out of mind. If you can pack things away so they're compact and tidy, she probably won't care how much stuff there is. Maybe sacrifice some cases but keep components, that sort of thing.

Funny, I was just looking on ebay for some ATX cases to rebuild my older systems into. Would be nice to have matching cases that would stack nicely in a corner or something when not in use. Thinking about these:

http://www.newegg.com/Product/Product.aspx?It … 051&ignorebbr=1

With the promo code it comes out to $33 shipped plus a $10 rebate card after the fact. All of my older motherboards are ATX. These cases aren't too hideous so it should be ok?

senrew wrote:

Well, I WAS going to be swimming in PCI Powermac sweetness today.

Which one? I had over the years a number of them (though I'm not exactly a mac guy...)
Still have a G3 (beige), G4 (sawtooth) and G5. Systems I don't have anymore were a 7300 and a 8100. And a bunch of pre-PPC systems also which I gave away.