BloodyCactus wrote:

except PCB cost$ is by dimension. adding another 1/4" by whatever height * number of boards would be a huge expense.

Yes, it would have cost more, but I'm not buying that it would have been a huge expense. While it may have been rounded up and cost disproportionately more than just the quarter inch than it is, the extra cost to the PCB is dwarfed by what it would cost to fabricate custom brackets. I would have happily paid a buck or two or even five more. If the board house they've chosen can't produce the board for a reasonable price, then there are others that can. The move to texelec seems to be consistent with further commercializing these boards as viable products. If it's a commercial product, then it needs a practical solution to attach a compliant ISA bracket. But anyway, this is just nitpicking. I'm glad to have the boards back so I could build one.

For what it's worth, it failed again last night, passes 123 and 124 again. At least the behavior is deterministic and probably wouldn't be overly difficult to get to the bottom of. I wish the source code to checkit was available (is it?) so that I could see exactly what checkit was doing when it failed.

For what it's worth, it failed again last night, passes 123 and 124 again. At least the behavior is deterministic and probably wouldn't be overly difficult to get to the bottom of. I wish the source code to checkit was available (is it?) so that I could see exactly what checkit was doing when it failed.

I do wonder if checkit has a bug.. exact same 123/124 passes? .... sooo many things would have to line up for it to be so exact but without error all the other times...

Actually, doesn't CheckIt DEFINITELY have a bug? I know it logs results to memory, and running a FULL memory test with many iterations results in an error due to running out of RAM.

Or was that another piece of software I was using. I need to double-check.

keenerb wrote:

Actually, doesn't CheckIt DEFINITELY have a bug? I know it logs results to memory, and running a FULL memory test with many iterations results in an error due to running out of RAM.

Or was that another piece of software I was using. I need to double-check.

That is check-it you're remembering. I'm not sure if that's a bug or a feature (it stops gracefully rather than just blowing up).

This is different in the respect that it's the memory test itself that's failing, and it's failing in the EMS region, not the main memory region. I've been running a test for several days now, this time using "full" memory test rather than "quick" so each pass takes much longer. I wish checkit had a pass counter so I knew where it was currently at.

so I replaced my VGA card (triden 9000B) with a ACUMOS AVGA1. Stability went from 1% to 20%! So very strange. Now I need to get that testing program and run a mem test on it.

ftp and other memory/bus intensive things will lock it up. ftp with a large file locks up 100%

anyway. getting closer. might look at one of those lotech ram cards to try.

smbaker wrote:

I built the 2MB EMS board today. It's working fine in my Xi 8088... at least so far.

One thing that's annoying about these Lo-Tech boards is that they seem to be a nonstandard size -- the standard keystone bracket won't fix it, so I have no ISA bracket attached to it. Seems like it would have been insignificant cost to just make the thing a tiny bit longer. I'll have to see if I can custom fab some kind of bracket to go with it.

Scott

Well I finally got my CF Card assembled and installed. Everything is working great. Not too sure what version of dos to install so I went with 6.22, does anyone have any opinions on what is the best version for the 8088?

Scott, How does the 2MB EMS board function in the 8088 system? I'm ordering parts to assemble mine soon. Just wondering if you need a memory driver to allow the system utilize it?

I'm interested in setting up a master slave config with multiple CF cards. Does anyone know how to set a particular board to master or slave?

Sir Isaac wrote:

Scott, How does the 2MB EMS board function in the 8088 system? I'm ordering parts to assemble mine soon. Just wondering if you need a memory driver to allow the system utilize it?

There is a lo-tech EMS driver that you install into your config.sys. It was a piece of cake for me to get working.

BTW, I did recently design my own through-hole Flash / EMS board:

I originally intended this to be a flash drive using 39SF040 flash chips, but it turned out that it was really simple to make it so the board can function as an EMS board (move a couple of jumpers, use ASC6C4008 RAM instead of 39SF040 Flash). No video or blog post on it yet, but soon.

In other homebuilt 8088 news, Sergey now has his "Micro 8088" boards available for sale. There's a thread describing them on the retrobrew forum.

I've got one arriving Friday. I think I have everything I need already to put it together. I am optimistic the FE2010A chipset will address the stability issue in the xi8088...

Next stop, 286?

That reminds me, does anyone know what bios extensions the Tandy 1000/SX uses to initialize the Tandy specific 286 accelerator card?

it'd be neat to add those to the micro 8088 bios and see if that card works.

keenerb wrote:

I've got one arriving Friday. I think I have everything I need already to put it together. I am optimistic the FE2010A chipset will address the stability issue in the xi8088...

Built mine over the weekend. My PC is every bit as unstable. ;(

This makes me think there's something else going on. We've eliminated all of the wonky old x86 chips except the CPU. I'm using all F series logic. The board is small enough that I can do a pretty thorough job of verifying that I built it correctly. I'm think if the problem isn't the Xi 8088 / Micro 8088, then it must be something else wrong with my build -- the backplane, the video card, or the power supply. Going to spend some time fooling around with the other components of the system.

gdjacobs wrote:

Next stop, 286?

I have a Tiny Turbo 286 sitting here waiting to be tried out with either the Xi 8088 or the Micro 8088 when I get a chance. I have no idea if it will be compatible.

I have a 286 sbc.. I looked at the datasheets of the PCchips SCAT + F80806(its some wonky chip#), these two would be super easy to wire up a 286 sbc, but each one sells for like 30$ a chip on ebay. which makes me kinda go hmmmmm dont think I want to try.

One would think with the micro8088, its so few chips that making it work should be a piece of cake. all the timings and bus handling is done by the one chip...

smbaker wrote:

Built mine over the weekend. My PC is every bit as unstable. ;( […]
Show full quote

keenerb wrote:

I've got one arriving Friday. I think I have everything I need already to put it together. I am optimistic the FE2010A chipset will address the stability issue in the xi8088...

Built mine over the weekend. My PC is every bit as unstable. ;(

This makes me think there's something else going on. We've eliminated all of the wonky old x86 chips except the CPU. I'm using all F series logic. The board is small enough that I can do a pretty thorough job of verifying that I built it correctly. I'm think if the problem isn't the Xi 8088 / Micro 8088, then it must be something else wrong with my build -- the backplane, the video card, or the power supply. Going to spend some time fooling around with the other components of the system.

gdjacobs wrote:

Next stop, 286?

I have a Tiny Turbo 286 sitting here waiting to be tried out with either the Xi 8088 or the Micro 8088 when I get a chance. I have no idea if it will be compatible.

Maybe it's a BIOS problem. Maybe try the Wildcard bios, if you have a real 8088 laying around? Sergey said it worked for him, more or less.

smbaker wrote:

I'm think if the problem isn't the Xi 8088 / Micro 8088, then it must be something else wrong with my build -- the backplane, the video card, or the power supply. Going to spend some time fooling around with the other components of the system.

but we all have different backplane, different psu, video card etc. but we all see similar problems. I dunno what the answer is.

smbaker wrote:

Built mine over the weekend. My PC is every bit as unstable. ;( […]
Show full quote

keenerb wrote:

I've got one arriving Friday. I think I have everything I need already to put it together. I am optimistic the FE2010A chipset will address the stability issue in the xi8088...

Built mine over the weekend. My PC is every bit as unstable. ;(

This makes me think there's something else going on. We've eliminated all of the wonky old x86 chips except the CPU. I'm using all F series logic. The board is small enough that I can do a pretty thorough job of verifying that I built it correctly. I'm think if the problem isn't the Xi 8088 / Micro 8088, then it must be something else wrong with my build -- the backplane, the video card, or the power supply. Going to spend some time fooling around with the other components of the system.

gdjacobs wrote:

Next stop, 286?

I have a Tiny Turbo 286 sitting here waiting to be tried out with either the Xi 8088 or the Micro 8088 when I get a chance. I have no idea if it will be compatible.

Are you using the UMB feature?

BloodyCactus wrote:

but we all have different backplane, different psu, video card etc. but we all see similar problems. I dunno what the answer is.

It could be we're all suffering from different issues. For example, if I recall correctly, my bus terminator board improved my stability, but did not improve keenerb's stability when he tried it.

I don't have an 8088 on hand, but I did order one from eBay yesterday.

I'm not using any UMB. I've tried the Micro 8088 both with an without the SPLD and had similar results.

Could even be an iffy FE2010 chip. I've got everything I need for a non-UMB build, and I ordered two boards and I have two FE2010 and two FE2010A chips at this point to test, and several known-good 8088 chips, so I can try a bunch of things.

I got rid of the compactflash adapter and I'm seeing improved stability. Anyone else try that? To early to draw any concrete conclusions at this time.

I notice Sergey's compactflash adapter doesn't have a data buffer, whereas some of the other compactflash adapters (current Lo-Tech, etc) do.

Scott