Adventures with the Nixdorf 8870 Mini-Computer
  • Welcome
  • Contact Me

It works... perhaps not...

27/2/2011

1 Comment

 
The last couple of days have been rather up and down for me; perhaps you can guess this from the title of this entry.
Yesterday morning Heike and I set off for Birmingham. Sometimes things come to those in need, and just as I discovered that I needed a keyboard to get the DAP4 from Johannes working, what should pop up on EBay but a Quattro 25 machine complete with three DAP4 terminals complete with keyboards.

I should also point out that at the same time, I’d had an Email from Lawrence; an ex-Nixdorf 8870 engineer, who told me of a chap in some far off land who happened to have several DAP4 keyboards for sale. The postage was going to be expensive as they are rather heavy, so I decided to keep him in reserve until I saw what would be in my treasure from Birmingham.
Picture
Environmental Computer business card.
We arrived at the pre-appointed time and Steve from “Environmental Computer” was there to meet us. We had a brief chat about the 8870 and screens. I showed him what happens when the battery backup plate is left in a discharged state for too long and then he helped us load all the stuff into the car. I happened to notice that he had many UPS units lying around all over the place, in fact, they have a LOT of stuff lying around all over the place, and if you’re looking for vintage or newer technology then you may want to give them a call or check them out on EBay; they’ve got an old British Telecom exchange sat in the corner that caught my eye; but was told to stop looking at it as there wasn’t room in the car.

Anyway, Steve said that he would happily throw in a small UPS for me if I had room in the car, and I was VERY grateful. I’ve been having problems with the breaker tripping the power to the workshop and whilst all my PC equipment is happily running off a large UPS, I didn’t think it had spare capacity to run an 8870 as well. The UPS after it had been on charge for a couple of hours seems to be working but the batteries do need replacing. Anyway, thank you Steve.

Needless to say, we raced home as I was eager to plug in one of the keyboards, fire up my terminal and watch my 8870 spring into life, then came the first of many ups and downs of the day. Plugging in the keyboard and switching on the VDU resulted on the VDU coming on for a second... and then going off and after repeated attempts the fault wouldn’t clear.

It was either a keyboard fault (seemed unlikely), VDU power supply or VDU CPU card. My gut feeling was it was the VDU PSU. No problem I thought as I had four new (well 25 year old) terminals to choose from.

The horrors of time.

Picture
Damage after battery leaked. (Click for larger images)
Taking the cover off the first of the Birmingham VDUs made my heart sink. The lithium battery inside the VDU CPU card has leaked and eaten everything it could find. This included part of the CPU card, the aluminium protection plates and more of an annoyance, the V24-remote switch-on card. This “was” the system master port. However, the PSU seemed intact and it's quite a simple job to strip these VDUs down and interchange their main modules (more on this in a seperate article). The PSU was transferred over to my original VDU (as this one had been cleaned and now looked almost like new), power applied and switched on. Nothing… not a peep. So, another PSU was harvested and installed… this one hissed like escaping gas, and then finally just died. I was now starting to get worried.

However, 3rd time lucky and eventually the VDU came on and just sat there with the green keyboard light on. 
After pressing the parameter load key, <CR>, 0-1-8-5 and <CR> it accepted the parameters. Things were looking up, so next I connected the VDU to the system and switched on; a few seconds later the 8870 chassis fans started and the 8870 began its boot phase.

Incorrect CORESIZE.

Picture
Incorrect CORESIZE.
Low and behold, “CR” appeared on the VDU; this means that it wants you to specify the number of the workstation program to load or you can let it load the default. The main difference between workstation programs is what devices you want to connect to the terminal; I didn't care at this point. Since this was the VDU that came with the system and anyway, by the time the CRT has warmed up it’s too late to override the default setting, I decided to just sit back and let it run.

I could see the L1, L2, L3 & L4 lights flashing on the keyboard so I knew the VDU was loading it’s workstation program that was being supplied by the 8870, and I started to get a nice warm fuzzy feeling that the system would boot ok.

Then I saw the “INIT” message. This is your chance to change the default boot device from something other than “0”. However, I expected to see just the word “INIT”, but I actually got the message “CPU# 0 INIT”, and that really surprised me. This meant that the operation system was a version designed to support a multi-processor environment, so it had to be at least NIROS 7.0 loaded on the disk; but we were always told that you had to have a 1559 CPU as a minimum; not the 1537 CPU that’s currently in the system. Looks like the sales boys and girls had been up to their old tricks again.

So, I left the system to IPL and see what would happen…and I waited… and waited… and waited…

Eventually I got a rather worrying messages that the core size was incorrectly configured. Assuming that this system hadn’t been tampered with, and why would somebody just exchange the CPU or remove memory, then there was obviously something wrong but nothing I could do at this stage.

Picture
File allocation probems.
What should happen next is that NIROS boots it’s emergency minimum configuration, and that’s exactly what it did, after deciding that there were some index / file entry problems. This of course is also a problem, especially if it decides to delete an operating system file.

Picture
NIROS EX 1.0 Log-on Screen.
After a couple of minute’s, joy of joys… the log-on screen appeared and by the look of it; User Administration hadn’t been installed (so I had a pretty good idea what all the system passwords would be), and what's more, the system had NIROS EX 1.0 installed. Since I've not tampered with anything and this is the operating system loaded, I can only assume that everything is as it's supposed to be... so why the CORESIZE problem??

Picture
Welcome Screen.
I entered the EX 1.0 system password (NOOFFICE) and a couple of seconds later, I was in !!

Picture
LIBR file listing.
Entering a “LIBR” command to see what would happen, I could hear the disk heads moving and a Library listing started to appear.

To start TAMOS from SCOPE (# prompt), you just enter "START" <CR> and this looked like it was going to work, then gave me a nasty error #3 and threw me back to SCOPE. Then I remembered, the system had booted in its emergency configuration mode and wasn’t going to allow normal operations until the primary configuration had been sorted.

I went through CONFPROG (System configuration tool) and removed all the PLC and Floppy Disk drivers (as these boards had been removed from the chassis), and reduced the number or partitions; anything to save memory.

Again during the IPL more file allocation problems were reported but once again, it didn’t pick on anything important. This time it let me log-in and start TAMOS.

First thing to do was attempt to get an operating system backup; the only working operating system I have is on this disk and it seems to be failing.

I’ve been buying up old SMC tapes off EBay whenever I see them; the older the better as there’s no information available as to what size tape is required but I know the LU size would be 66MB maximum. PUs (Physical Units) are usually devided into multiple LUs (Logical Units). A typical customer installation would always have a minimum of two LUs and on the M25, these two LUs are contained on one physical drive unit. The 8870 dosn't allow a complete drive (PU) backup; you can only backup on an LU basis.

I did have a lucky find on EBay a few weeks ago. A chap in Germany was selling a single Nixdorf SMC tape that had S1 written on it. This means that it probably came from an 8870 but as to which 8870 I had no idea. Anyway, it was worth the couple of Euros. I’d also found some old SMC’s in the USA and had obtained those.

Picture
TAMOS LU > SMC Backup.
I inserted a tape and attempted to format it. It failed. I tried a second tape and this worked fine, so I started a backup running.

Unfortunately it failed. So I tried a different tape… and another… and another. All the tapes would either fail during formatting or during the backup process.

I was getting desperate so decided to service the SMC drive that had come from the Birmingham machine and try that. I shut down the machine and swapped over the SMC drives.The system re-booted this time without error, but the backups were still a problem, then the VDU locked up; it just stopped responding. The ALME was polling the VDU and I could see the red poll LED in the small plastic window below the bottom right-hand corner of the CRT flashing. Normally, you would have no problems switching off a terminal and re-loading it, but this time the terminal wouldn’t re-load. The keyboard was on, the ALME and the DAP4 poll LEDs were all flashing, but the terminal seemed dead.

After re-entering the VDU parameters several times I decided to swap the terminal CPU card with a spare and that seemed to do the trick; the terminal sprang back into life. So at this point, I've now only got one working terminal out of four.

I ran a LIBR listing to the printer and that seemed ok, and decided that I needed to change the printer assignment to specify 66 lines per page and started TAMOS; the system hung.
During the re-boot it starting picking up file allocation errors again and would often hang at this point. After several re-starts it came back to the log-on screen.

Picture
TRAP #255 and system HALT.
Attempting to log-on brought back a trap #255 and then the system crashed.

Another few re-boots and more file allocation problems and the system came back-up again.

I logged in and attempted to dump the memory contents to the printer but it started complaining that there was no printer driver assigned, then the system crashed again.

Picture
TRAP #81 and system HALT.
Another couple of re-boots it started complaining about the memory configuration. It seems obvious that the disk was getting more and more corrupt as I use it so at this point, I gave up.

I'm hoping that the drive is just corrupted and if I could format it and reload a clean working operating system I would probably be back in business, but since I don’t have a clean operating system to load, for now, I’m stuffed. If the drive is faulty, then I'm going to need a replacement. There are some on EBay but they are EXPENSIVE !!

It’s been a rough 24 hours really. At one point I was even praying to Heinz Nixdorf himself to allow his creation to boot correctly; it even worked once, but I suspect that too many years in storage have taken their toll on the hardware.

This morning I thought I would try and boot from the drive that came from the Birmingham system.

Sometimes the system would just sit there, other times it would start the boot process and either hang, or report file allocation problems. All I know about this drive is that it contains an older Niros release; probably 5.1 judging by the copyright date.

So for now, I can't really take this any further.

Anybody got an SMC with Niros EX 1.0 on it ? Or any NIROS operating system for that matter.

1 Comment

At last... my M25 has arrived.

23/2/2011

1 Comment

 
Picture
8870/M25 running after being serviced and repaired.
Well, it’s been an exciting couple of days.
I arrived home from work yesterday to find a huge pallet sat on the driveway; I guessed it was the M25 from Johannes in Germany.

It was starting to get dark, raining, and the package was heavy. I had to quickly break apart the packaging and removed the printer and VDU to a safer place. Next we carefully moved the M25 off its pallet and because the casters had been removed, we “walked” the M25 it into the garage and out of the rain.
Now safe and dry in the garage, the first task was to get the front and back covers off the machine. I thought I was going to have to drill off the locks as there were no keys, but as it happens only one of the four locks had been secured and a quick “yank” pulled the front cover lose.

It was immediately obvious that the machine had been used in an exposed workshop type environment as it was very dirty inside. This was actually a selling point of these machines as they didn't really mind the dust or room temperatures. Removing the plate retaining screws, I managed to remove all the plates, disk drive, SMC and FD drives and Power Supply which made the chassis a lot lighter and easier to relocate to my workshop.
There’s a fuller description on how I serviced the machine elsewhere on this site, but after cleaning and some basic testing, the machine was reassembled.

Picture
CPU status showing 26 - VDU does not reply.
After some fiddling around I managed to get the system to boot and joy of joys; status 26 was output on the CPU display which means “VDU does not reply”. Since there was no terminal connected this is exactly what I’d been hoping for, then, I hit a small problem.
I opened up the package that contained the VDU and discovered there was no keyboard. This is a problem. Without a keyboard, I can’t start the VDU and so can’t boot the computer. So, for now, I’m stuffed. For those that are wondering, it has to be a Nixdorf SAS keyboard for a DAP4. I’ve left the VDU in the garage for now until I can sort something out.

In the mean time I moved the ND11 printer in the workshop and promptly dismantled it, vacuumed it out, cleaned it, and put it back together. Loading some continuous stationary and pressing the test button produced the expected test page but it’s obvious the ribbon has dried out and I’ve not been able to fine suitable replacements yet. So, I squirted some WD40 onto the ribbon; just a quick spray every 3 inches or so on the ribbon tape. Once I’d been right round the ribbon I re-installed it in the printer and re-ran the test print; it was much better. The ribbon needs time now for the WD40 to migrate evenly over the ribbon material which will happen on it’s own over the next few days.

It's been an exciting 24 hours but it looks like I'll have to wait a bit longer before I can attempt to boot the machine properly.

1 Comment

An 8870 at last...

17/2/2011

6 Comments

 
One of the things I’ve been after is an actual 8870 machine. I regularly search EBay and other auction sites and often spend several hours surfing the net trying to track down old equipment and information, and you’d be surprised what turns up if you kick enough rocks over.
Picture
8870/M25 with SMC and 8' Floppy Disk
A couple of months ago whilst looking for 8870 pictures, I stumbled across a picture that somebody had posted to an automotive related bulletin board and asked, in German of course, what sort of computer was shown in the picture. Well, it was an 8870/M25 with a DAP4 screen and what looked line an ND11 printer. Unfortunately the post was rather old; May 2010. However, nothing ventured, nothing gained so I sent Johannes an Email, asking if he still had the computer, and he replied a couple of days later that he had, but didn’t know anything about it.

So, I replied telling the basic history of Nixdorf, and a little something about the computer he had, and, if he would be willing to sell it and what I intended to do with it. To cut a long and very happy story short, yes, he would be willing to sell it.

Now Germany isn’t really that far, but it’s amazing how difficult it is to arrange to get a box the size of a small refrigerator from German to the UK. The shipping company wanted to know the dimensions, weight, required that it be securely mounted on a pallet and then boxed up.
Johannes of course is a busy chap, and it’s not fair to expect him to bundle the machine up like that, but, he offered. I will be eternally grateful for him doing this as the only other solution was for me to go there and get it. It would be cool to have pictures of the machine as it was found and packaged and then shipped and Johannes has taken lots of great pictures. It also shows how much work he’s had to do to get a pallet, cut it down to size and then pack everything in as tight a space as possible.

Today I received an Email from Johannes telling me that the transport company will be collecting the computer tomorrow and it should take between two and five working days to get to me.
From what I’ve seen from the all the pictures, I’m not overly confidant that the computer will be actually working, and getting spare parts is almost impossible. However, there are only really three things that can go wrong; some more serious than others.

The 8870 uses a completely propriety VDU and communication protocol, so if the single supplied terminal is faulty, I’m not going to be able to boot the system. There’s a good chance that any large electrolytic or X2 capacitors in the main PSU have dried out, or that the hard drive has either become corrupted of gets damaged in transit.

The simplest of these to rectify are dried out capacitors. It’s just a simple matter of finding replacements of the same specification. A faulty VDU may be repairable depending on what the fault is, but a dead hard disk is a real problem. Remember that all hardware is proprietary so it’s not a matter of finding a disk unit that’s compatible with the controller, its finding a drive that capable of behaving exactly like the original drive. Then there’s the problem of locating an operating system. Johannes has checked and there are no backup tapes stored with the machine. Still, no point in worrying about it just yet.

The plan is once the machine arrives to photograph it, and then strip it down. This will be a bit of a challenge as we don’t have the cabinet keys for it so I’m going to have to drill the locks out. I’ll then remove and open up all the plates and clean out all the dust that always accumulates. Next I’ll remove all the fan modules and clean them. I’ll also open up and clean the PSU. Once that’s all been done, I’ll replace the PSU and fans and switch the PSU into service mode. Hopefully with the load from the fans, the PSU will start up and then I can check the voltages to make sure it’s running within expected parameters. I’ll leave it running like this for an hour or so to make sure there are no small explosions. All being well, I can then reassemble everything and attempt an IPL.

However, working or not it's a great find and will give me a chance to document and photograph the hardware.

Watch this space.

6 Comments

Never… NEVER… throw anything out...

17/2/2011

0 Comments

 
I’ve been accused of being a horder, and I am to a degree.  No, I don’t fill the house with old newspapers going back to the 1960’s but I do tend to acquire a fair amount of technology, and a lot of it usually isn’t in prime condition.  Just for the record, I don’t store my acquisitions in the house. I have a large garage and a new custom built workshop for that. But I have unfortunately, in the past, thrown things away that were obviously never going to be needed again… until now that is.

When I stopped working with the 8870 in the early 90’s, I kept some of the technical documentation that I’d religiously carried around for many years. During the heavy rains In 1998, we were unfortunately flooded and since the insurance company placed a skip on the drive I took the opportunity to fill it with things that I would never need. Whilst I’ve been searching around for 8870 documentation, many people have told me that they’ve only recently junked the stuff they had.

If you’ve got some old computer technology at home (and I don’t mean some old PC graphics card from the 1990’s), but perhaps for some reason you kept that old deck of punch cards (new or used), or some manual from an old mainframe, don’t throw it. At the very least, if you must get rid then put it on Ebay; and it probably won’t even cost you anything. Please don’t deprive some geek working all hours some valuable piece of information that they will pay for. You also have the pleasure of knowing that your junk is being recycled and going to a good home.

Oh, if it’s 8870 / Nixdorf related items, please contact me first :-)
0 Comments

I can’t translate German… and neither can Google...

16/2/2011

1 Comment

 
Actually, that’s not fair. Google can translate German, it’s just not very good at translating technical German.

I have two 8870 manuals that are key to understanding exactly how the 8870 works under the hood as it were. The NIROS manual (version 3.3 for now), and the Field Engineering Manual. The problem is they are in German, and my German is rather limited.

My partner Heike, is German by birth, so I printed out the NIROS 3.3 manual, sat her and the manual down and said, “you translate it, and I’ll key the English into my laptop… should only take a couple of hours”. It took around thirty minutes to get a half decent translation of a single paragraph. The problem is Heike doesn’t know technical German and her translations were too literal. This is exactly what Google’s language tools did. I decided I was going to need help.

I went to the Internet huning around for an independent or freelancer who would take pity on my, and translate my 160ish pages for the price of a bag of peanuts, but I didn’t have much success.

What I did stumble across was a website Freelancer.co.uk and the concept is rather simple. You create a project that defines what you want doing, your specify a budget range and let people ask additional questions or bid to do the work.

So, I created a project, uploaded a low-ish quality scan of the manual for bidders to see, set the project bidding stage to run for 10 days, and sat back really expecting for nobody to reply. Within 10 minutes I’d had several offers ranging from the peanuts I wanted to pay, up to £250 (which was my maximum). I also had one chap ask if I was serious as he charges £30 per page; odd really, I didn't force him to bid. The next day I had a look and had around 15 bids for the job. The site on-line help advises you not to just decide on price, which makes sense but what DO you use to help you decide.

I remembered than when Heike and I tried to translate a very simple passage we got into a terrible mess. So, I sent the four lines of German to my friend VAXMAN and asked for his translation. I then picked four bidders based purely on gut feeling, and if the English they used in messages they had sent me, appeared half decent, and asked if they would translate the four lines for me.

After 48 hours, only one person replied, and she replied almost instantly and her translation was pretty accurate; well,good enough for me anyway.

So, Ioana in Romania is currently in the middle of translating my NIROS 3.3 manual. We’ve agreed a price, and the work will be divided into five equal parts. I get to see the quality of her work, and she gets paid after each part. Stage payments are all handled by the Freelancer website so it really is quite simple. If she does a good job, I’ve got another couple of thousand pages of manuals for her to translate as well. Lucky girl.
1 Comment

Recovering old operating system tapes...

16/2/2011

0 Comments

 
One thing my emulator is going to need to work, is an original operating system. If we were talking about Windows, DOS or even CP/M, then it wouldn’t be a problem. But were talking about a propriety 15+ year old operating system from a company that no longer exists and used on a single range of computers which are all but extinct.

So it’s lucky that when the company I worked for after working for Nixdorf went out of business, I spent my final minutes there on-site (we really didn’t get much notice to vacate the building),backing up my 8870 development system, and I kept the tape.

The problem is that these 9 track half-inch tapes don’t age well if they are stored in less than ideal conditions. The tape is actually made from a plastic material with a magnetic coating bonded to it’s surface using a hygroscopic glue (absorbs moisture from the atmosphere). What all this means is that if you start running an old tape through a tape drive there’s a good chance that the tape is going to fall apart and be ruined.

Lucky for me during one of my scouring the Internet for information sorties, I found VAXMAN. Go check out his site. We started Emailing as you do, and I happened to mention my problem with the tape, though I was more worried that I didn’t have an actual tape drive to read it on. VAXMAN offered to recondition the tape (by placing it a nice warm oven for a while which apparently doesn’t damage the magnetic coating or the data), and the reading it’s contents onto one of his VAX machines. He could then spool the data into a tape image that I could then read.

I got an Email from VAXMAN the other day saying the tape had arrived, but it may not be good news. Whilst the tape is physically intact there appeared to be a lot of parity errors when the tape was read, but hopefully, theres a good reason for this.
Picture
Cipher 880 1/2' 9 track tape unit.
The 8870/M45 I used to create the tape had a Cipher 880 tape unit installed. These can read/write at either 1600 or 3200 bpi. Unfortunatly the tape unit that VAXMAN uses works at 6250 or 1600bpi. The Cipher default is 3200bpi so this could explain why it can’t be read. It’s not all bad news as he does have other tape drives that will do 3200bpi, but it’s not a simple job for him to reconfigure all the hardware.
Another problem may be the “mode” that was used to write the tape. The 8870 supports two modes of operation; streaming and start-stop mode.

Streaming mode is used for fast system backups. The problem is that tapes written in this mode weren’t designed to be read on other manufacturers systems. I’m not saying hat you can’t read them, it’s just that usually, if we were writing tapes for other manufacturers computers, they had to be written in start-stop mode. My tape is a standard system backup tape, so was written in streaming mode, so it’s possible that the tape is perfectly fine, it’s just that I need an 8870 to read it on. It could also be that the tape is completely knackered. I’ll wait and see what VAXMAN suggests as to what to try next since he's got the hardware... and all the work to do.
0 Comments

It's alive....

16/2/2011

0 Comments

 
It's been a bit of a battle getting this far, but at last, this site is up and running. Whats a bit of a shame is that I can't seem to post blog entries in the past. A lot has happened up to this point and I really need to find some way of documenting everything that's happened. As it happens I've already got some fasntasic news to write about but I'll drop that in a new posting. So, welcome to my blog, I hope you enjoy, and please, feel free to comment.
0 Comments

    Author

    I'm a software developer, an engineer and I love vintage computers, but the 8870 has a very special place in my heart.

    Archives

    August 2017
    January 2016
    May 2015
    June 2014
    November 2012
    June 2012
    February 2012
    July 2011
    May 2011
    March 2011
    February 2011

    Categories

    All
    1/2 Inch
    2516
    2708
    8870
    9 Track
    ALME
    BASIC
    Boot
    Bootmenu
    Chassis
    Comet
    Confprog
    Connection
    Dap
    DAP4
    Dap Emulation
    Dc600a
    Diablo
    Dicolldata
    Disc Pack
    Disk
    Documentation
    Downloads
    Drive
    Ebay
    Entrex
    Eprom
    Ex
    First Post
    Format
    Formatter
    Freelancer
    German
    Germany
    Hawk
    History
    Ihss
    Init
    Ipl
    Iris
    Junk
    Kienzle 6100
    LIBR
    M25
    Manual
    Micro 7
    Nd11
    Niros
    Nixdorf
    Nros
    Pcb
    Perish
    Programmer
    Programming
    Ram
    Ribbon
    Rom
    Smc
    Start-stop
    Streaming
    Tamos
    Tandberg
    Tape
    TDC3309
    TDC 3309
    Translation
    Ultratec
    Vaxman
    Vdu
    Wd40
    Wireing
    Xlp300

    RSS Feed

(C) Joe Farr, 2011-2017