VOGONS


Toshiba T5200 mods and upgrades

Topic actions

Reply 320 of 534, by OldCat

User metadata
Rank Member
Rank
Member
yourepicfailure wrote:

I am making a request here. I am looking to see if someone is willing to disassemble the plasma panel assembly (or LCD if you have that version) of a T3200SX and T5200 for pictures. Preferably individuals who have faulty panels (so as not to chance ruining a working display).

T3200SX is indeed complicated, az jaZz_KCS mentioned, T5200 is trivial, as you can unhook the whole screen/monitor lid from the base.

For T3100SX. I strongly suggest cehcking Maintenance Manual here: http://www.minuszerodegrees.net/manuals/Toshi … ce%20Manual.pdf

Reply 321 of 534, by yourepicfailure

User metadata
Rank Newbie
Rank
Newbie

Thanks for the suggestion, however:
The maintenance manual is incorrect.

Notably it labels pins 33 and 34 on machine side connector as ground, when they are actually +5v supply lines. Verified with probing and testing.
But I already have figured out the addressing scheme, I just needed pictures of other units to verify compatibility. I actually managed to power on the display and got 162vdc from the left side (view from component side of panel) inverter. Right side inverter (which generates the bias to actually "light" the panel) will not fire up without clock pulses, so I'm figuring that one out. Also got 3.3v logic on the I/O pins which is exceptional.
I will get the example pictures up tonight for folks who will take it on.

Reply 322 of 534, by henryVK

User metadata
Rank Member
Rank
Member

Not sure if these laptops have a big chunky shield in the back where the ports are. If they don't, you could probably low-key mod the case to have a ribbon vga cable go between the case and the port shield. This could connect a lcd driver board to the vga-out via with a low profile vga connector; reversible and you don't "sacrifice" the vga-out in case you want to connect a CRT monitor.

Reply 323 of 534, by OldCat

User metadata
Rank Member
Rank
Member
yourepicfailure wrote:

Thanks for the suggestion, however: the maintenance manual is incorrect.

Notably it labels pins 33 and 34 on machine side connector as ground, when they are actually +5v supply lines. Verified with probing and testing.
But I already have figured out the addressing scheme, I just needed pictures of other units to verify compatibility..

I have found my photos from 2017 when I was dismantling my T3200SX - won't have time to disassemble it now, but maybe these can help.

Attachments

  • 20171109_112204.jpg
    Filename
    20171109_112204.jpg
    File size
    1.58 MiB
    Views
    1537 views
    File license
    Fair use/fair dealing exception
  • 20171109_112129.jpg
    Filename
    20171109_112129.jpg
    File size
    1.08 MiB
    Views
    1537 views
    File license
    Fair use/fair dealing exception
  • 20171108_145239.jpg
    Filename
    20171108_145239.jpg
    File size
    1.22 MiB
    Views
    1537 views
    File license
    Fair use/fair dealing exception
  • 20171108_145105.jpg
    Filename
    20171108_145105.jpg
    File size
    1.13 MiB
    Views
    1537 views
    File license
    Fair use/fair dealing exception
  • 20171108_143951.jpg
    Filename
    20171108_143951.jpg
    File size
    1.28 MiB
    Views
    1537 views
    File license
    Fair use/fair dealing exception

Reply 325 of 534, by yourepicfailure

User metadata
Rank Newbie
Rank
Newbie

Apologies for not getting the images out sooner. Had things get in the way.

Anyways, here are examples of images. Don't care the format, so long as you don't have JPEG compression that limits the clarity of a zoomed in image. ~70 quality in photoshop save for web, and 8+ in save as jpeg.
For the component side of the panel, good enough I can zoom in and make out component names. There are components mounted behind the visible board- please don't remove the board. As seen, the name of the component behind the board is silkscreened in.
For the display panel, a good enough picture I can count the pins will do.
For machine-side connector, again good enough I can count pins.

These are high res images that won't fit with the image uploader, so used an external.
https://mega.nz/#!GTJ03YjS!z_xR7v7_2ZdKsnUnT2 … Vzp-fOTmCzI0KDU

Thanks OldCat, it would appear as though your display connector on the panel-side is 16 pins. However, similar interface in terms of addressing. Looking at the maintenance manual for an idea, Toshiba termed the connections differently, but they are similar. The same three clocks to be exact. So similar logic, but made simpler (4 ttl I/O pins vs 8).

I wish I had the funds to purchase a T3200SX and T5200 to research the panels on each, but this will have to do.

Reply 326 of 534, by my03

User metadata
Rank Member
Rank
Member

Hi guys,

so i ended up contacting the gentleman on the german ebay and bought the "matrix" T5200 from him, to use as a parts-donor for my own T5200 with the parity error 2 message. While doing this, as i wrote earlier, i found one morning when i tested my machine that the parity error was not there anymore (it counted nicely to 8MB of ram totally). So now i have this new "matrix" machine at home, and should my own machine behave, then this one do not really need to be a parts donor.

But this crazy "blockiness" got me intrigued so i started out with some experimentation. First of all, i made sure that all ROMs are actually in place and in contact with their respective sockets. Then i pulled the "casette" from the back (with controller + vga card) from my working machine and inserted it into the "matrix" machine. The result: Still playing scenes from "The Matrix" even with my "working" VGA card. Vice versa, my working machine still works fine with the "matrix" grahpics board inside.

I guess this leaves me with the advice given here earlier by a fellow member, namely to recap the PSU (i should probably try to measure it first using my voltmeter). If that will not work, then i assume its either dry-joints (could it be that??), something fishy with the plasma screen itself (recap?) or the ROM thats there right now (??)

Behind these "blocks", i actually do see some text output from the boot process (the usual bios battery message and the option to press F1). Not chrystal clear, but its there alright.

Inside the bios, i can see that this machine differs from my other, working, one in that it has (i believe) the award bios instead of the phoenix one (can tell from comparing images uploaded in this thread earlier). Another odd thing is that the matrix-machine has got a 500MB hard-drive inside it (which i guess either means that the bios is old enough to not have the restricting checks in it, or that the bios is modded somehow).

What would you gut feeling be? Recap first or put new ROM in it?

Reply 327 of 534, by yourepicfailure

User metadata
Rank Newbie
Rank
Newbie

Without seeing the machine, it's hard to say.
Dry joints are a bit on the lower chance. These units used leaded solder and through-holes, with a few plcc and qfp chips.
That said, not impossible, but not very likely.

Also, one that I've found examining the panels is that the ribbons that connect to the panel drives are always on the verge of coming out. Take a look at my uploaded images you'll see that the contacts of the drive ribbons are showing.

I'd suggest not recapping, but examining the PSU first. Check voltages with and without load, look inside for leaks or bulges. Nonetheless a recapping is still to be on the list.

Also, a rom dump and a difference check is a good idea. I know versions of the bios have been posted.
My offer still stands, I would very much like one of those to study especially for the panel.

First priority should be inspecting that PSU.

Reply 328 of 534, by my03

User metadata
Rank Member
Rank
Member

Update: The PSU is fine. I stripped the machine completely and opened up the PSU to see if anything seemed off, which it was not. Then i "transplanted" the working machines PSU over to this machine (and vice versa) to see if that changed anything, but it did not change anything at all (so, my healthy machine works fine with the "faulty" machines PSU, but the "faulty" one still shows the blockiness).

To summarize i have swapped:

- screen assembly (no diff)
- I/O & VGA cards (no diff)
- disconnecting floppy & HD (which is a 1.3GB Toshiba drive w. 3.5->2.5 adapter) (no diff)

  • This is kind of peculiar since it says something about read and write errors on the vga ram at adress <so and so>. But since the VGA memory is located on the discrete card (which i swapped), it does not makes sense.

- visually inspected the screen cable for anything odd (nope, looks fine)

And since the few capacitors on the system board looks fine, then it must be something else on the system board causing this.

I have since also bought (still waiting) one pair of bios roms (one each with the latest Award and Phoenix bios + AT-ide from this thread). At the same time, i also bought the VGA rom (32kb variety). It will probably not help at all for this machine, but at least that will hopefully allow me to use something else than the 100MB drive on the working one.

Reply 330 of 534, by my03

User metadata
Rank Member
Rank
Member

It is consistent. This is what it says:

vram error
read data 00FF00H
write data 000000H

It hangs on this output for some 10 seconds and then moves on to counting the RAM etc.

Googling a bit (presumably?) gives me that some kind of flash i/o could be the issue and i guess it could have something to do with the CMOS chip itself (?) or probably more likely something that it tries to communicate with using some port mapped IO? Question is what...

after the RAM count, it either says that the configuration is incorrect (9 out of 10 times) and i can press the function-key to enter the bios. 1/10 it gives HDC error.

While it does all this, the HDD led is constantly on (but i have of course tried to disconnect both the power and hdd cable to rule them out, but it still results in the same "blocky" craze).

Reply 331 of 534, by dj_pirtu

User metadata
Rank Member
Rank
Member

Just bought myself a T5200, had one over 25 years ago. I have 486DLC/40 and 486SXL2/40 CPU's and want to put the latter one on Toshiba. Will it work with cache enabled out-of-the-box or will it need some soldering? Is it possible to swap oscillator-chip from 40MHz -> 80Mhz? That 486SXL2 has internal clock-doubling so it'll work @40Mhz anyway so oscillator swap not necessary.

I have IBM P70 portable too and those 486 upgrades didn't work with cache enabled so good. Corrupted data on hard disk.

EDIT: found it 😍 Re: Toshiba T5200 mods and upgrades

Reply 332 of 534, by my03

User metadata
Rank Member
Rank
Member

Hey guys,

on a sidenote: when you start using the xt/at-ide instead of the native dito and it detects the HD alright and boots half way, but ends up complaining that the command can not be found, what might be the cause of this?

I got my modified roms yesterday (from Ians files, burned on suitable roms) and tested both the Phoenix one as well as the Award one (both seemed to work just great btw). They both showed the at-ide/xt-ide banner but stopped half-way through the boot-process. When i restored to my original bios again, it worked normally.

what is the difference in how they handle the HD?

Reply 333 of 534, by dj_pirtu

User metadata
Rank Member
Rank
Member

Ok, did the a20 and cache flush -mods. I dont have modded bioses yet but i assume that the cyrix.exe works the same, maybe. So with cyrix.exe i enable flush and a20 pins and cache. But in wolf3d it scrambles graphics if cache is enabled. It did the same thing before those mods on the motherboard.

Or do i need the modded bios this work correctly?

Edit: if i disable caching 640k -> 1Mb area then there's no graphics corruption and it doesnt affect speed at all.

Reply 334 of 534, by my03

User metadata
Rank Member
Rank
Member
my03 wrote on 2020-01-15, 21:03:

I got my modified roms yesterday (from Ians files, burned on suitable roms) and tested both the Phoenix one as well as the Award one (both seemed to work just great btw). They both showed the at-ide/xt-ide banner but stopped half-way through the boot-process. When i restored to my original bios again, it worked normally.

what is the difference in how they handle the HD?

Just to close this one. After exchanging the original Conner 100MB hd for a DOM module (4GB. Overkill i know, but thats what i had available) coupled to a "pin-switch" adapter, now everything works perfectly well. The machine is much snappier (with the Cyrix 20/40mhz drx2) and load times are very reduced.

I had some issues with the Phoenix rom so i exchanged it for the Award one and all problems went away (which is weird considering i had a Phoenix ROM in there to start with).

Anyway, have anyone here been successful in starting up the dos-browser Arachne on your Toshibas? When i try to select the vga setting (256kb, grayscale and vesa graphics) it just won't start up, but rather goes back to the settings screen again.

Reply 335 of 534, by dj_pirtu

User metadata
Rank Member
Rank
Member

Little update from here:

Got eprom-chips and updated both system bios and vga bios. Swapped the original 100Mb Conner to 10Gb Seagate (taken off from original XBOX long time ago).

Next I'll put bigger heatsink on CPU. Got some oscillator-chips too, maybe some overclocking with 50Mhz, 60MHz, 66MHz or 80MHz pieces. I think 50Mhz FSB could achievable so CPU becomes 40MHz -> 50MHz (because it has doubleclocking).

Need more memory too, 4Mb just isn't enough for Doom. Death Rally doesn't start at all.

Reply 336 of 534, by my03

User metadata
Rank Member
Rank
Member
dj_pirtu wrote on 2020-02-18, 09:28:
Little update from here: […]
Show full quote

Little update from here:

Got eprom-chips and updated both system bios and vga bios. Swapped the original 100Mb Conner to 10Gb Seagate (taken off from original XBOX long time ago).

Next I'll put bigger heatsink on CPU. Got some oscillator-chips too, maybe some overclocking with 50Mhz, 60MHz, 66MHz or 80MHz pieces. I think 50Mhz FSB could achievable so CPU becomes 40MHz -> 50MHz (because it has doubleclocking).

Need more memory too, 4Mb just isn't enough for Doom. Death Rally doesn't start at all.

Hey pirtu,

i guess (?) we're the only ones still watching this thread 😁

a few pages back, i read that it was next to impossible to overclock this thing by just swapping the chrystal since the rest of the machine could not keep up. Did you try to increase the speed?

Reply 337 of 534, by henryVK

User metadata
Rank Member
Rank
Member

You guys, I'm still watching this one!

I keep seeing T3100s and T5200s on local ads but I have my other computer and an increasing number of vintage cameras to deal with 😁

Reply 338 of 534, by my03

User metadata
Rank Member
Rank
Member
henryVK wrote on 2020-03-13, 08:31:

You guys, I'm still watching this one!

I keep seeing T3100s and T5200s on local ads but I have my other computer and an increasing number of vintage cameras to deal with 😁

Oh, over here (Nordics) you seldom see them for sale. Consider picking them up imho 😀

I bought a Cyrix CX486DRx2 20/40 and installed it in my 5200 meanwhie (plus i got the Arachne browser to work after all). Its by no means perfect, but the difference is astounding relatively speaking. Just love this thing 😀

Reply 339 of 534, by Hamby

User metadata
Rank Member
Rank
Member

I'm still following the thread somewhat.
A bit scared, atm...

last week my water heater ruptured and flooded my apartment. When I came home, everyplace was a wading pool.
And I had had tucked under one of my desks, still in its case, my T5200...
I haven't had a chance to examine it yet, but the flooding didn't seem to have extended quite under my desk; the case felt slightly damp under one side of the back (the case had been sitting back-down, not flat, thank God).
My "weekend" starts tomorrow, and on my list of things still to be checked out (including a C128D and a 1084S monitor) will be the T5200.

The thing is, it was under my desk because I wanted to preserve it as much as possible; if it survived this, now I'm going to make it one of my main computing machines. I'm going to enjoy it while I can.