Saturday Haul R&R

[SpidersWeb]

In other news, Wang made their serial connector the wrong way around - so need to grab a gender changer before I test those to check for output. The keyboard is needed for diagnostics (the 5 LED's provide rather detailed information of failures on startup - similar to the beeps of a failed PC but much much more detailed) so I'll likely just delay that until those two parts arrive.

I'd previously forgotten about a machine I had still sitting in the Camry - a PC general with a missing case lid. I finally got around to testing it. Internally I found a 386DX-40 with VLBus. The motherboard is an OPTI-495 - traditionally a 486 VLB board, but this one has the 386 CPU soldered in, 386 BIOS, and an 80387 socket.

I spent much time fiddling about with 30 pin SIMMs - which can be quite frustrating trying to find four matching and working ones, while potentially having connection issues on a motherboard you're not even sure works. Eventually I did find two sets of working SIMMs, that after gentle encouragement are connecting reliably. I removed the very leaky CMOS battery and moved it in to a new case. I now have a working 386DX-40 with 8Mb RAM (8x1Mb 70ns 30pin SIMMs), VLB graphics, SB16, and a 170Mb Conner hard drive The case is fairly small in foot print - so will make a good DOS gamer box - and it came with 1.2 and 1.44 floppy drives as well.

One complication I did have but worked out fairly quick - was the Multi IO controller it came with had become a jumper donor at some point - eventually I found the right jumper and position to re-enable the floppy controller using good old fashioned guess-n-check. FDC is backwards - straight through = DS0, twisted = DS1 (with drives set to DS1). Was something I'd heard of but actually never had before (on a card anyway).

To finish I need to install the floppy drives properly, solder up an external battery, screw everything down and do a software install.

[SpidersWeb]

Wang Professional Computer
Removed video cards from the Wang, checked baud rate, but no serial output. It requires DSR to be set at power up, and no video cards present for serial to be used as console-output. I'll need to check the TechRef for pinouts to make sure I did it correctly and make sure DSR is actually being set. The fact it says "DSR" instead of "DTR" makes me wonder if the Wang is setup to be DTE rather than DCE - bah it all ends up quite confusing. I'll wait for the keyboard and pretty diagnostics traffic lights before I spend too much more time on it.

With the graphics cards removed - the red LED at the rear of the machine flashes instead of staying on. This is a good sign, to me it indicates that the machine has detected an issue and is moaning about it - rather than just going nutso. This would make sense - given that it was unable to set console output or input correctly - both errors are classed 'fatal' in the manual. I've also noticed this Wang PC has two 256KB memory boards and 128KB on board - giving it a total of 736KB RAM - not too shabby.

386-DX40
Can't find my spare 'external AA battery' enclosures. Genius.
I did freak out a little bit, I finished building it and realised there was no 4 pin external battery connector. However it turns out, there was a 4pin connector which I had ignored because it had a jumper in the middle. Looking up various verisons of this board on TULARC revealed that it was standard to:

Pin 1-2 CMOS clear
Pin 2-3 CMOS normal
Remove jumper - and use pin 1 and 4 for your battery

So now it looks like I wont have to completely disassemble the thing and solder something in. This solution makes perfect sense, but I hadn't come across it with the jumper in the middle before.

Concord 286XT
I have another EPROM programmer in the mail, but in all honesty, I'm really tempted to just order a generic 286 motherboard. I have nice 16bit RAM cards, sound cards, MFM controllers... all just sitting there with no 286 to go in to. (If you missed it, the Concord 286XT is an 8bit ISA bus XT with a 286 CPU). I never considered motherboard replacement before because my case said '286XT' on it - but the new case just says "Concord 286" - I guess someone in marketing thought it'd be better to drop the 'XT' part lol

Unrelated
My IBM XT's get regularly powered up - at least monthly. This weekend one of them failed - the 1987 Enhanced 5160. It appears to be another shorted cap (I've already removed 3 from it on previous occasions).
What's interesting about this is - 1983 5150 motherboard from the US - all original capacitors - not a single one has blown, 1987 5160 motherboard, lived in NZ since new, like firecrackers.

Also unrelated, running Mac OS X in VMWare, easier than I thought.

[tezza]

Concord 286XT
I have another EPROM programmer in the mail, but in all honesty, I'm really tempted to just order a generic 286 motherboard. I have nice 16bit RAM cards, sound cards, MFM controllers... all just sitting there with no 286 to go in to. (If you missed it, the Concord 286XT is an 8bit ISA bus XT with a 286 CPU). I never considered motherboard replacement before because my case said '286XT' on it - but the new case just says "Concord 286" - I guess someone in marketing thought it'd be better to drop the 'XT' part lol

Looks like you've been busy in the shed like I have today.

Before you go order a 286 motherboard, I might be able to help. I'll send you a PM

[SpidersWeb]

Thanks for the motherboard!

I've ended up installing it in my Concord 286XT case (instead of the 286 case) because it's in nicer condition - it's actually still white on the front.

Some details:
- NiCad battery - no leakage but 0.34V - however it appears to be recharging and holding it's charge. I'll remove it at some stage in favour of an external setup once I get a few more battery backs delivered.
- Made in 1990, MORSE 286 with Siemens 286-12
- Oldest version I've seen of the American Megatrends BIOS (AMI-BIOS) - called '286-BIOS 1.04' - setup and diagnostics in ROM (seperated), no purple and green - just monochrome, all settings in one screen, but the important stuff like custom drive types is there. The later coloured version was quite popular in the 386 era - much prefer it over the alternatives like Pheonix.
- 1Mb RAM on board, plus you can add two pairs of 30 pin SIMMs
- can allocate the 384KB leftover memory as XMS (counted after SIMMs) or as EMS
- did not shoot me with a capacitor

Memory mapping was interesting, on board 384KB of leftovers is counted after SIMMs - so if any SIMM is installed but fails - memory drops to 640 instead of just 1024. SIMMs are great in a 286, because you only need pairs of two - not a set of four.

Currently I've got it running with VGA and 3Mb RAM - having trouble deciding which hard drive to install - I was going to use an NEC D5126H (fast seeking 20Mb MFM) which suits the look of the machine, but this board was born in an era of IDE (even has solder points for IDE on board), so I'm tempted to go through my drives and find something IDE from 1990. I guess that'll leave my AT RAM cards and WDC controller free for the next AT.

[tezza]

Glad you're finding it useful. It was probably one of the last and fastest 286 boards made. I can't imagine there being much of a market for them beyond 1990, with the 386 then becoming popular.

[SpidersWeb]

Lower price point I'd suspect. With that in mind, and the fact this isn't the IDE model, I think I might run with the D5126H - I think the 40Mb version would be more appropriate but don't have one sadly.
Also realised the case has a rail setup, so I'm going to have to go through everything trying to find where I put my spare drive rails :S

[SpidersWeb]

D5126H installed. Very impressed by the ROM diagnostics, everything from an array of video card tests through to com/parallel/floppy tests. It's like having a mini version of CheckIt in ROM. Hard drive stuff was fairly extensive as well - Hard Drive Preformat, Auto-Interleave with speed ratings, Media Analaysis (which fills in the bad track table), Performance Test (I got 8ms track to track and 214KB/s at interleave 2), Seek Test, and a test for repetitively writing and verifying the maximum cylinder.

The D5126H recently arrived seized from my "Saturday Haul", and it's had a hard life by the looks, 98 bad sectors, but the rest of the drive is reliable. No weird noises, no bearing noise, seeks perfectly and data on other areas is reliable. Being the H model the seeks are faster but you lose that cute chirp of the normal D5126 - instead it's a much more deeper note - similar to an ST-251.

Just need to clear the SD card on the camera, install some drive rails and a few accessory cards and get some software installed.

Edit: was finishing the build, and test fit the case lid, I always forget to check some mundane detail - like "is the WD 1003 card miles too high to clear the case lid?"
So either, 8bit RLL, 8bit MFM, or 16 bit IDE are the options now. Hard choice, I either lose IRQ 5 + speed, or lose the "I fight bears with my spare time" hard drive noise.

Edit 2: Cancel that, found a WD-1006! (newer version (1988) of the same card, half the size), my evil plan of keeping a large collection of spare bits is working well

Edit 3: Wang PC Keyboard arrived, many immature jokes ensued, FedEx only managed to break off two keys and unclip the two halves. Thankfully Wang hardware was service friendly making it tolerant to abuse and easy to clip back together. The importance of this keyboard is it has 5 diagnostics lights and a speaker - which is the minimum the Wang PC needs to communicate errors - it's also uni-directional, so wiring up anything else was unlikely to work even if it supported the same protocol. I'll possibly start looking at going through my two Wang PCs (now armed with a keyboard) later in the week or next weekend. Wang machines are starting to become a side collection to the IBM gear.

[SpidersWeb]

Concord 286XT
Well Morse 286-12 now. Booting DOS 5.0 like a champion. Makes all the right noises and I had just enough drive rails to build it properly.
1.44 + 360K + 20M, SBPro, 256K VGA, 3Mb RAM, 2xSerial (16540's).

Update: Hard drive has gone poo - left Wolf3D demo running and after a few hours I got "Ready error reading drive C". I'll need to investigate further, but it's likely I'll end up having to swap it for another D5126 or ST225. ST-251 would've been good, but the Wang PC-240 stole that. ST-4038 would've been perfect as well, but it has a circuit board problem. Also have a 10Mb Tandon - but that's just as reliable as the drive that's already in it - and it's only 10Mb. This failure could explain the issues I had when using it on the AT&T - however it's own drives had the same fault - might be worth trying a confirmed 100% drive and if successful - seeing if the 30Mb CDC drives work OK in a PC.

When I connected the 360KB drive as B, the BIOS autodetected both it's presence and 40 track limitation and automatically changed the CMOS setting. My 386 AMI machine will detect a 40 track drive and change it's settings too - but only if it's set to anything other than Not Installed.

Wang PC
Spotted this on ebay: http://www.ebay.com/itm/Vintage-WANG-Co ... 2ecedbebae

I'd never get it shipped at that price, but maybe screen and printer?
Only alternative I've thought of might be a MultiSync CRT - but over the last few months I've had 0 luck tracking one down in NZ.
Or I could get one of those "converts anything to VGA" boxes - but did they have 800x300 mono in mind when they made that?
Or I could just run it over serial or parallel I guess. Decisions decisions.

Also, how did Wang emulate 720x348 on a 800x300 display? I wonder if you lose 50 pixels, or the monitor supports multiple frequencies.

[SpidersWeb]

Soooo been working on the Wang.

Keyboard is hooked up, cool beeps and the traffic lights (it has 6 LEDs) during startup were neat - the LEDs race from one side to the other, then start flashing like a binary panel on an Altair, if a light pattern stays lit - that's an error code - all my lights turned off! So it's passing it's pre-boot checks (CPU, base memory, etc etc etc).

Along with the flashing LED (on the back of the machine) the keyboard clicks. I think what's happening is the video card is missing, so it is polling through the options and each click is a restart.

Unlike the IBM PC - the Wang is very flexible in it's options. As per the tech ref, no CO (console output) is considered fatal - but it supports more than just video cards - it'll test for option ROMs, serial (baud rate set per motherboard switches) and parallel. It's also possible the ticks are just a very long memory test - but I did leave it running for quite a while, and the flashing LED only appeared after removing the video cards.

I'm going to try straight through and null modem cables now. Motherboard is set to 9600 baud (default). Fingers crossed I get something.
Parallel is another option - but I can't remember what the deal is with DOS and setting the pins as inputs instead of outputs - the Wang has a Centronics port just like a printer, so a standard printer cable is what I have to use, just data will be flowing in the opposite direction.

This machine makes the IBM PC look like a PCjr. Other than size and price - this was a much better machine.

I also have another Wang PC to try, which I got from Andrew many moons ago but never booted it up because I was told the magic smoke had escaped - might be worth a retry on that one as well.

Unrelated: picking up a ~90 year old typewriter tomorrow - 35 bucks and it looks complete and besides a little rust - undamaged. Think it's a Remington Standard 12.

[tezza]

Not sure if anyone has had this - Willem PCB3:

- start write,
- after the first byte it gives up and says "Buffer 0x75, chip 0x35"
- go "read chip" - and the first byte comes back as 0x75 (so the chip had actually recorded it correctly)

I can't see anyway to force it to verify afterwards instead of immediately after writing. Guess I'll try some different chips first.

I had lots of issues with my Willem. In the end I got a TOP853 as seen here:
http://www.classic-computers.org.nz/blo ... eproms.htm

[SpidersWeb]

Yep, read your blog I ended up getting another Willem clone, a PCB50 - just because it was so incredibly cheap and it wasn't terribly important - otherwise I was going to give the TOP a try - I do love the fact it's jumperless.
I'm hoping at least the new 5.0 version is more reliable.

In other news - I've got a Wang PC attempting to boot from floppy, pressing "R" it tries again.

Last night I did work with serial cables - no luck - but the Wang would occasionally spam my receiving PC with blank characters. I did a test on the serial port, and DSR should have been high > +3V - but instead it was 2.6V. Lines being held low were -8V. This explained why my PCs weren't picking up anything (DTR, DSR etc) because 2.6V is not high enough for RS-232. So I replaced the motherboard - it's enourmous - but it's a card so only requires one screw and slides in/out.

Results were the same, I wonder if 2.65V is what Wang used for RS232 :/ Kind of annoying if that's true, I'll need to dig more.

But anyway, the replacement motherboard - a 256KB 8086 (not 8088) with 16 bit expansion bus - made some other changes.
The self test pattern is a lot longer, the keyboard looks like a christmas tree for quite a while, then it beeps for boot. Having a 256KB model installed is quite neat, because with my two memory boards, I'll be running it at it's maximum 768KB.

Unrelated: typewriter - checked serial number - works out around 1926 - still types! Remington Standard 12.
Unrelated: picked up a "portable" 1979/1980 model National colour VHS video camera system - including the original "I have a VCR on my hip" satchel.

Edit: removed floppy from drive, and it booted from the hard drive - an ST225. Still can't see what's going on, but I think my Wang PC is working fine.

[SpidersWeb]

Quick Demo of the keyboard lights / startup seuquence: http://youtu.be/bmGo4-I7v6g

[SpidersWeb]

Got the Wang using the serial port instead of the text/graphics cards.

"WANG PROFESSIONAL COMPUTER REV 2.40"

sweeeeeet

HDD starts loading but locks up the machine, I'm thinking it's 90% likely that the hard drive needs a LLF and is starting to lose it's marbles.
Also possible the software installed needs the text/graphics cards install - but that's a bit weird - the OS etc was all text.
Third option is a memory issue the Wang isn't detecting on power up or some other hardware fault - fingers crossed it's not those.

So I guess I should try and recover it's moldy installation disks wish me luck.

PS one 256K memory board has a parity error, so only running 512K not the 768K max.

Edit 1: trying to image the disks, Wang thought it'd be just hilarious to make them single sided, but on the opposite side the IBM PC used for single sided. Disks are in better shape than I remembered, Wang wont start off them but it's drive probably needs a clean - or maybe I've been inserting them upside down (just thought of that - label down perhaps? bah)

[SpidersWeb]

Not having much luck.

Original floppy drive doesn't seek - even when used on a normal PC.
Replacement drives on a PC, can read the original Wang disks fairly well (occasional bad sector at end of disk), but Wang keeps reporting the same "Format error".

If it's of interest to anyone, the way the disks are recorded is rather 'special'.

Head 0 - tracks 34-39 contain data, 0-33 completely unformatted
Head 1 - tracks 0 to 39 contain data

I'm guessing that "head 0" 5 cylinder spot is for boot data? File allocation?
(Edit: tech ref says for boot/verify it just reads the first 4KB, can also boot via serial which is interesting)

Anyway, makes it a pain in the bottom to image - because imaging tools can't see anything on Head 0 Track 0 and fail. I had to adjust a lot of options in IMGDISK to get it to even read the things - but it's the usual 250Kbps 9 sectors 40 track stuff once you hit a formatted track.

Disks are:

Basic Support Services - x2 (128K and 256K)
System Software - x5
Installation Utility - x1
Advanced Options Diagnostics - x1
IBM PC Emulator x 1 (apparently you put this in, boot, then insert a MS DOS disk (not included in package), and get a fairly incompatible operating environment to play with)

The floppy labels are not in the usual spot, just a bit unusual, the disks go in to the envelopes sideways.

[SpidersWeb]

Another user has pointed out to me, it's quite possible somebody used these disks in their Apple II - 35 tracks, singled sided, would explain the missing chunks.
From reading the tech ref, there is no indiciation that the Wang would look on H1 for data, more likely Head 0 Track 0 Sectors 1-8, which is blank on these.

They also sent me DOS 2.1 for the Wang, so I'll write those images to disk tonight and try those. I don't think this version will like the 20Mb hard drive though, my disks are a later version.

I also switched the ROM chips between motherboards.
It's now a "WANG PROFESSIONAL COMPUTER REV 2.50".

The other motherboard fails it's self test. I have a code to lookup, but no plan to diagnose unless I need to.