VOGONS


Toshiba T5200 mods and upgrades

Topic actions

Reply 100 of 534, by IanB

User metadata
Rank Member
Rank
Member
Vipersan wrote:

Is this an error on my memory board and should both pin 19s be linked ?
or is this an oversight in the pin list ?

The pins that don't have an "A" or "B" designation contain the same signals but in most cases, the "A" and "B" sides have independent drivers so they are not connected together electrically.

Vipersan wrote:

Working on the assumption my hunch was correct ..I popped the pcb in ..
..and entered the bios ..
see picture for result

Congrats on getting it to work, I guess that proves the instructions are OK. I'll post the Cache support mods soon.

Reply 101 of 534, by Vipersan

User metadata
Rank Oldbie
Rank
Oldbie

Thanks for explaining Ian ..
I guess I want to try the cpu mods next ..
looking forward to the details.

..and it really helps during the test phase (buzzing it out) to have the contacts for the new Simm and the original slots both topside..
rgds
VS

Reply 102 of 534, by IanB

User metadata
Rank Member
Rank
Member

Here are the cache mods:

The 486SXL and to a lesser extent the 486DLC require some modifications to get the full benefit of the cache although they will still be substantially better than the original 386 even without them. There are two signals that need to be sent to the processor:
1. A20 Gate. This is used to signal when the A20 gate is active causing wraparound of the 1st 1MB of memory to exactly simulate the behaviour of the original 8088
2. Cache flush on DMA. This is used to flush the cache whenever there is a DMA operation by other hardware. Unless you install some expansion card like a SCSI controller, the only time this happens is when the floppy drive is read.

The A20 Gate mod is a single wire but the cache flush requires a 74HC00 logic gate IC to generate the signal.

This diagram shows the pinouts of the CPU and the 82385 Cache Controller (both viewed from pin side) together with the flush circuit:

Flushmod.jpg
Filename
Flushmod.jpg
File size
448.78 KiB
Views
2080 views
File license
Fair use/fair dealing exception

And this is an overview of the mods on the underside of the motherboard (yellow wires are the flush mod and blue wire is the A20 mod):

overview.jpg
Filename
overview.jpg
File size
1002.81 KiB
Views
2080 views
File license
Fair use/fair dealing exception

The A20 mod is a wire from Pin 8 of IC 43 to pin F13 (A20M#) of the 486 CPU (blue wire)

a20-cpucc.jpg
Filename
a20-cpucc.jpg
File size
771 KiB
Views
2080 views
File license
Fair use/fair dealing exception

The cache mod is made as follows:
Wire from The ISA MEMW# signal to pins 1 and 2 of the 74HC00. (The MEMW# signal can be found on the rear end of R517)
Connect pins 3 and 4 of the 74HC00 together
Wire from pin M11 (BHLDA) of the 82385 cache controller to pin 5 of the 74HC00
Wire from pin E13 (FLUSH#) of the 486 CPU to pin 6 of the 74HC00

Mounting the 74HC00
I used a surface mount 74HC00 and mounted it face down on the PCB adjacent to C29 which is a decoupling capacitor with +5v and 0V available.
I connected pin 7 to the front side of C29 (0v) and used a bare wire to connect the rear side of C29 (+5v) to pins 9,10,12,13 & 14.
Pins 8 and 11 which are outputs have been cut off close to the body of the chip and do not make contact with the wire.

Warning: The position and orientation of R517 and C29 are based on the PCB in my systems, later PCB issues may have moved them.

I have added pin numbers on the 74HC00 in the photo below: (numbering is reversed from normal as the IC is face down)

chip-memw.jpg
Filename
chip-memw.jpg
File size
1011.06 KiB
Views
2080 views
File license
Fair use/fair dealing exception

These mods require my patched BIOS to work (see front page).
The A20 mod is auto detected by the patch code but the cache mod can't be detected unless there is a floppy in the drive so the patched BIOS assumes it has been made if it detects the A20 mod and configures the CPU accordingly.

Testing the mods:
If you get the A20 mod wrong, the BIOS will not configure the CPU for either mod.
If you get the A20 mod right but the cache flush mod wrong, the T5200 will probably hang when trying to boot a floppy.

I'll update this post with some more testing details including some test software soon.

Reply 103 of 534, by Vipersan

User metadata
Rank Oldbie
Rank
Oldbie

Very detailed Ian ..and now added to my to-do list ..
I assume the modded firmware Im using (Hamby bios) is suitable ?
It will however have to wait as I am currently testing the 510CDT with CFs ...plus my eyes need a break.
😉
Fleabay / China comes up trumps again..
got 10x SOP14 mini pcbs for a £1 ...and 10x 74HC00D sm nand chips for £1.50 ..
I didnt really need so many ..but they are cheaper in lots of 10 from China than singles are in the UK.
go figure..

Attachments

  • sop14.jpg
    Filename
    sop14.jpg
    File size
    45.08 KiB
    Views
    2066 views
    File license
    Fair use/fair dealing exception

Reply 104 of 534, by IanB

User metadata
Rank Member
Rank
Member
Vipersan wrote:

I assume the modded firmware Im using (Hamby bios) is suitable ?

The modded Hamby BIOS is suitable but it must be the 1.2 update, not the earlier 1.1 update.
This update number is displayed on bootup.

Reply 105 of 534, by kultakala

User metadata
Rank Newbie
Rank
Newbie

Think i wont modify anything else for now but great work all this!

My japanese T5200 now has a new floppy, a flash IDE disk and since today... a 486 SXL-40 Processor.
Thanks to IanB for the hint! 😎

I ran Norton Sysinfo to measure the improvements.
Values on the left are from the original one, values on the right from today with new disk and new cpu.

T5200 Sysinfo Overview.jpg
Filename
T5200 Sysinfo Overview.jpg
File size
1.45 MiB
Views
2057 views
File license
Fair use/fair dealing exception

And some other numbers...

Booting to DOS took 20 seconds with the original one.
With new disk 11 seconds, plus new cpu it takes 9 seconds.

Starting Windows 3.11 from DOS prompt took originally 46 seconds, new disk 33 seconds, now 26 seconds.

And starting Opera 3.62 with loading the Google start page took originally 104 seconds, with new disk only 58 and now only 30 seconds!

Great 😀

Reply 106 of 534, by IanB

User metadata
Rank Member
Rank
Member
kultakala wrote:

My japanese T5200 now has a new floppy, a flash IDE disk and since today... a 486 SXL-40 Processor.
Great 😀

Another satisfied customer 😀

If you want to try cooling the 486SXL, Vipersan posted in his thread about some experiments using a thin copper sheet to transfer the heat to the floppy/HD cage:
A couple of retros I'm working on.

Did you manage to get the other T5200 booting again after disconnecting the battery?
If you still have a problem, the flash of the 3 leds on power up seems to indicate that the CPU is running so maybe one of these would help to diagnose the problem:
https://www.ebay.co.uk/itm/Diagnostic-PCI-4-D … B9/322654320657
There are loads of them on Ebay, this one is from Hong Kong. You can get them from European suppliers but they will cost more.
It plugs into an ISA or PCI slot and displays a progression of diagnostic codes as the BIOS starts up so if it freezes on a code that can be used to determine what the problem is.
I used one recently to diagnose a faulty 486 motherboard.

Reply 107 of 534, by kultakala

User metadata
Rank Newbie
Rank
Newbie
IanB wrote:

If you want to try cooling the 486SXL, Vipersan posted in his thread about some experiments using a thin copper sheet to transfer the heat to the floppy/HD cage:
A couple of retros I'm working on.

Did you manage to get the other T5200 booting again after disconnecting the battery?

Thanks, i will make some measurements regarding the temperature and maybe try something out.

I had no time yet to give the second T5200 a try but will do this week probably.

Reply 108 of 534, by OldCat

User metadata
Rank Member
Rank
Member
kultakala wrote:
My japanese T5200 now has a new floppy, a flash IDE disk and since today... a 486 SXL-40 Processor. […]
Show full quote

My japanese T5200 now has a new floppy, a flash IDE disk and since today... a 486 SXL-40 Processor.

And some other numbers...

Booting to DOS took 20 seconds with the original one.
With new disk 11 seconds, plus new cpu it takes 9 seconds.

Starting Windows 3.11 from DOS prompt took originally 46 seconds, new disk 33 seconds, now 26 seconds.

And starting Opera 3.62 with loading the Google start page took originally 104 seconds, with new disk only 58 and now only 30 seconds!

Attachments

Reply 110 of 534, by Vipersan

User metadata
Rank Oldbie
Rank
Oldbie
IanB wrote:
Another satisfied customer :happy: […]
Show full quote
kultakala wrote:

My japanese T5200 now has a new floppy, a flash IDE disk and since today... a 486 SXL-40 Processor.
Great 😀

Another satisfied customer 😀

If you want to try cooling the 486SXL, Vipersan posted in his thread about some experiments using a thin copper sheet to transfer the heat to the floppy/HD cage:
A couple of retros I'm working on.

Did you manage to get the other T5200 booting again after disconnecting the battery?
If you still have a problem, the flash of the 3 leds on power up seems to indicate that the CPU is running so maybe one of these would help to diagnose the problem:
https://www.ebay.co.uk/itm/Diagnostic-PCI-4-D … B9/322654320657
There are loads of them on Ebay, this one is from Hong Kong. You can get them from European suppliers but they will cost more.
It plugs into an ISA or PCI slot and displays a progression of diagnostic codes as the BIOS starts up so if it freezes on a code that can be used to determine what the problem is.
I used one recently to diagnose a faulty 486 motherboard.

Hi Ian ..
the disgnostic ISA/PCI arrived this morning ...
It came with jumper leads ?
no idea what they are for ...and a small booklet which at first glance is meaningless ..though error codes are listed.
there s also a small pushbutton onthe board labelled S1 ...and again ...no idea what it is for.
I guess I have a learning curve to climb.
The booklet is sort of chinglish or at least assumes you have some knowledge to start with...
perhaps you can point me at site with better info for a novice such as myself ?
rgds
VS

Reply 111 of 534, by kultakala

User metadata
Rank Newbie
Rank
Newbie
IanB wrote:

If you want to try cooling the 486SXL, Vipersan posted in his thread about some experiments using a thin copper sheet to transfer the heat to the floppy/HD cage:

Did you manage to get the other T5200 booting again after disconnecting the battery?

Hi,
when i disconnect the battery and power up my spare T5200, it starts, displays the memory check, then some error messages and i can go to setup by pressing F1.
Everything looks fine there and when saving the settings nothing else happens... i rechecked the cables and so on but all looks good.
Hmmm... i will take it apart again completely... somewhen 😀

My "main" T5200 is working fine...
Regarding the cooling, i checked the temperature with my thermal camera.
While idling the 486 SXL-40 cpu reaches about 66°C (151°F).

After playing Prince of Persia for 30 minutes, i loved this game back then... max. temperature was almost 75°C (167°F).
I played with the keyboard and cover beneath mounted and took it off just before measuring, looks like this:

Toshiba Thermal 1.jpg
Filename
Toshiba Thermal 1.jpg
File size
123.82 KiB
Views
1977 views
File comment
CPU max
File license
Fair use/fair dealing exception

The Display on the front has a hot spot at 41°C (106°F)

Toshiba Thermal 2.jpg
Filename
Toshiba Thermal 2.jpg
File size
115.97 KiB
Views
1977 views
File comment
Display
File license
Fair use/fair dealing exception

And on the back hottest outside part is over the power supply with 34,5°C (94°F):

Toshiba Thermal 3.jpg
Filename
Toshiba Thermal 3.jpg
File size
116.36 KiB
Views
1977 views
File comment
Back
File license
Fair use/fair dealing exception

Maybe i will try out some copper heat sink... 😀

Reply 112 of 534, by IanB

User metadata
Rank Member
Rank
Member
Vipersan wrote:
the disgnostic ISA/PCI arrived this morning ... It came with jumper leads ? no idea what they are for ...and a small booklet whi […]
Show full quote

the disgnostic ISA/PCI arrived this morning ...
It came with jumper leads ?
no idea what they are for ...and a small booklet which at first glance is meaningless ..though error codes are listed.
there s also a small pushbutton onthe board labelled S1 ...and again ...no idea what it is for.

I think the jumper lead is for connecting the speaker output to the beeper on the diagnostic card so you can hear BIOS error beeps and the button is a reset button.
There are two digit pairs, the left one displays the current status code and the right one displays the previous status code and it can be useful to see both.

You can look up the codes in the booklet which has tables for Award, Phoenix and AMI BIOSes but the table isn't that comprehensive as some of the codes tended to change with different BIOS generations but it should give you a rough idea of what's going on.

Even without knowing the exact meaning of a code, it's still very useful when there is no video output on startup, if you get no code changes at all then it likely means that the CPU isn't running, if you get a couple of code changes then a stall, it's could be a memory issue as it's failing initial memory tests, if you get lots of code changes but no VDU output then it's likely a video card problem etc.

Reply 113 of 534, by IanB

User metadata
Rank Member
Rank
Member
kultakala wrote:

when i disconnect the battery and power up my spare T5200, it starts, displays the memory check, then some error messages and i can go to setup by pressing F1.

Glad to hear it's not dead, I found that sometimes when the CMOS is corrupt that you can get no output on startup which is why I suggested disconnecting the battery.

kultakala wrote:

While idling the 486 SXL-40 cpu reaches about 66°C (151°F).
After playing Prince of Persia for 30 minutes, i loved this game back then... max. temperature was almost 75°C (167°F).

I found very similar temperatures when testing my system (I used an infrared laser thermometer)
The CPU is rated for 85°C so still OK but not much safety margin.
A cpu heatsink on its own reduced the temperature by 10°C and a fan/heatsink combination by about 30°C but I think Vipersan's solution with a Copper strap to the steel drive chassis also looks promising.

Reply 114 of 534, by REFCODES.ORG

User metadata
Rank Newbie
Rank
Newbie
IanB wrote:
Vipersan wrote:
Here is a zip with all the main BIOS and VGA BIOS versions found so far including patched versions: […]
Show full quote

Here is a zip with all the main BIOS and VGA BIOS versions found so far including patched versions:

The attachment T5200 BIOS pack 1-2a.zip is no longer available

This contains BIOS versions 1.20, 1.30, 3.00, 3.10 & 5.30 and the 24K and 32K VGA BIOS
All versions except for 1.20 have been patched and there are two different versions of each patch, one with XTIDE that uses workspace in low memory and one that uses workspace in high memory. See the readme.txt file in the zip for more details.
I'm currently using "T5200 BIOS V5.30 (Award) patch 1.2 with LOMEM XTIDE 640K.BIN" but the others are included in case of any compatibility problems.
As the T3200SX is very similar, I've patched some of the BIOS versions for that as well:

The attachment T3200SX BIOS pack 1-0.zip is no longer available

Hello everybody,

I am new to this thread and with much interest I was reading this thread! Very impressive!! I got a T5200 as well (with BIOS issues), put everything apart, cleaned the plastic parts, first put some "MIPA 1K-Kunststoffprimer-Spray" layer on them parts and then the actual color "Mipa Mix-Acryl Farblack Spraydose 400 ml halbglänzend RAL 7024" - looks great and the color layer is tough.

Now I am reassembling the T5200, with GOTEK Floppy drive and a CF-Card reader. This setup worked perfektly fine with my T5100, but the T5200 causes trouble. So I came accross this thread and the BIOS problems + solutions (great work!).

In my T5200 I have two DIP40 Chips below the keyboard: One labeld 006D, being a TMS 27C210-20JL which is a 1048576-Bit EEPROM (1024K), the other labeld 012B, being an INTEL micro-controller D8749H with 2K x8 EPROM on board (keyboard?).

The TMS 27C210 I managed to read, the D8749H is not supported by my miniprom programmer. The content of the TMS 27C210 looks like it starts off with a Video Bios? "Phoenix Advanced Video BIOS Version Version "00 0D 0A" Copyright (C) 1984-1989 Phoenix Technologies Ltd". I wonder because the T5200 has this extra VGA-BIOS chip near the ISA-Slots. Other strings in the BIOS indicate that it might be a "Normal" BIOS in it as well? ("Set default values", "Save and reboot" and the like). I also dumped the Video-BIOS, located on a 27C256 chip labeld 198B.

Well, my 5200 BIOS does not look like the Phoenix 3.00 nor like the 3.10 posted in this forum. I wonder if I still can use a patched BIOS from this forum. Does anybody know about the BIOS of mine T5200 (attached) and whether I can patch my T5200 with one of the provided ones? Are all the other BIOS dumps collected somewhere on Github?

Attached find the BIOS from the 27C210 chip labeld 006D and the Video-BIOS, located on a 27C256 chip labeld 198B from my T5200, maybe it helps someone when I share 😀

Filename
T5200 - BIOS + VGA-BIOS DUMP.zip
File size
50.49 KiB
Downloads
84 downloads
File comment
My T5200 BIOS + VGA-BIOS
File license
Fair use/fair dealing exception

Thanks everybody! 😀 Siegfried

Reply 115 of 534, by IanB

User metadata
Rank Member
Rank
Member
REFCODES.ORG wrote:

The TMS 27C210 I managed to read, the D8749H is not supported by my miniprom programmer. The content of the TMS 27C210 looks like it starts off with a Video Bios?
Attached find the BIOS from the 27C210 chip labeld 006D and the Video-BIOS, located on a 27C256 chip labeld 198B from my T5200, maybe it helps someone when I share 😀

Thanks for posting your BIOS and VGA dumps, your BIOS is V5.10 Award. I don't know why there is another VGA BIOS in the first 64K of the EPROM but I have seen this as well with the V1.30 Award BIOS. It can be removed and replaced with 'FF's and it doesn't make any difference. It wasn't present in the V1.20 Award or the V3.00 Phoenix BIOS versions but I'm not sure about the V3.10 Phoenix or the V5.30 Award as those versions were read using a BIOS dump program rather than an EPROM programmer and the first 64K doesn't appear in the memory map.
The T3200SX does use the first 64K for it's VGA BIOS and the one in your 27C210 exactly matches that so perhaps the various Toshiba systems in production at that time shared a common BIOS build system and it put a VGA BIOS in the first 64K even when it wasn't needed. This extra VGA BIOS won't actually run on the T5200 as it has data at offset $6000 - $6800 and that space is occupied by some VGA control registers in the T5200 memory map which is why it requires a special VGA BIOS in the 27C256 with a gap in that area.
Your 27C256 VGA BIOS is the same as the one provided by Hamby, (i.e. the 1988 1.04 01 24K version with the 1991 date at the end) and as you suspected, the INTEL D8749H is used for scanning the keyboard.

REFCODES.ORG wrote:

I wonder if I still can use a patched BIOS from this forum.

I recommend using the V5.30 Award patched version as that is the one everyone else is using at the moment and it is a slightly later version than the one in your machine so may have some bugfixes.

REFCODES.ORG wrote:

Are all the other BIOS dumps collected somewhere on Github?

No, but they are all in the zip file at the bottom of the first page of this thread. I occasionally update the zip in that post with any new versions.

If you are looking for blank BIOS EPROMS, the 27C1024 is pin compatible with the 27C210.

Reply 116 of 534, by REFCODES.ORG

User metadata
Rank Newbie
Rank
Newbie

Hello Ian,

IanB wrote:

your BIOS is V5.10 Award.

thanks so much for the valuable information !

IanB wrote:

I recommend using the V5.30 Award patched version as that is the one everyone else is using at the moment and it is a slightly later version than the one in your machine so may have some bugfixes.

I will do so!

IanB wrote:

No, but they are all in the zip file at the bottom of the first page of this thread. I occasionally update the zip in that post with any new versions.

Great, I will update from time to time 😀

IanB wrote:

If you are looking for blank BIOS EPROMS, the 27C1024 is pin compatible with the 27C210.

Good info! My luck, because I ordered a 27C1024 without taking a look at the chips in my T5200 😀

I will post my doings as soon as the chip arrives 😀

Reply 117 of 534, by Vipersan

User metadata
Rank Oldbie
Rank
Oldbie

If it's not too much trouble Ian..
I could do with some pointers on how and where to route 40 wire ribbon cable from inside the metal cage where the conner was to a pcmcia interface card inside the ISA cavity to the rear of the T5200..
A description would suffice..
pictures would be even more helpful..
rgds
VS

Reply 118 of 534, by IanB

User metadata
Rank Member
Rank
Member
Vipersan wrote:
If it's not too much trouble Ian.. I could do with some pointers on how and where to route 40 wire ribbon cable from inside the […]
Show full quote

If it's not too much trouble Ian..
I could do with some pointers on how and where to route 40 wire ribbon cable from inside the metal cage where the conner was to a pcmcia interface card inside the ISA cavity to the rear of the T5200..
A description would suffice..
pictures would be even more helpful..

Sorry, that's one thing I haven't actually done, although OldCat posted a pic of his mod for a T3200SX on the first page of this thread which routed the IDE cable to a CF card on an ISA slot so it should also be possible on the T5200.

Reply 119 of 534, by OldCat

User metadata
Rank Member
Rank
Member
IanB wrote:
Vipersan wrote:
If it's not too much trouble Ian.. I could do with some pointers on how and where to route 40 wire ribbon cable from inside the […]
Show full quote

If it's not too much trouble Ian..
I could do with some pointers on how and where to route 40 wire ribbon cable from inside the metal cage where the conner was to a pcmcia interface card inside the ISA cavity to the rear of the T5200..
A description would suffice..
pictures would be even more helpful..

Sorry, that's one thing I haven't actually done, although OldCat posted a pic of his mod for a T3200SX on the first page of this thread which routed the IDE cable to a CF card on an ISA slot so it should also be possible on the T5200.

Present! To be completely honest, I didn't do it all by myself - another guy in Poland who has a knack for fixing old computers helped me with it (cheers, Greg!). Please find below the photo of IDE ribbon placement in T3200SX:

P1080186.JPG
Filename
P1080186.JPG
File size
352.73 KiB
Views
1853 views
File license
Fair use/fair dealing exception