Zel's Fleet Blog - BX, Jag XJ-S, Sinclair C5, Mercedes 208D & 230TE, AC Model 70.

[myglaren]

What would life be like without zip ties now.
Fix just about anything.

[van ordinaire]

"cable ties" please - you're not in America now but, more importantly, I do wish people would be more responsible when disposing of/abandoning them.

With care, they can be used time & time again (often without getting shorter in the process), in fact I've given up buying them, any show ground will be ankle deep in them once the traders have broken down their stalls (although all too often they've cut the wrong end) & more substantial ones can be liberated from lamp-posts.

[Zelandeth]

Well performed live music never gets old, especially when it's by an artist you've followed for years.

Have followed these guys since well before they even were a group, the composer (on the piano above) was a very skilled soloist before picking up the rest of the group.

Have several albums now, including two signed on vinyl from backing their funding projects on Kickstarter, but this is the first time I've been lucky enough to see them live.

Moment of the week so far is getting the whole room full of people singing along with the chorus on a couple of the songs. That's one of those things which really don't get old to be part of and still makes my hair stand on end when it happens.

Hats off to the tech crew as well for getting the audio halfway listenable...the room is an absolute swine to mic up. It's really long, narrow and has a low ceiling...so to get anything resembling a useful level of audio at the back of the room basically means you've got to have the drivers at the back of the room...facing the mics. Not ideal. They tried having things rear facing set back a ways from the stage but that used to result in a horrendous resonance at one couple of square foot spot in the room which kept trying to turn the audience inside out, which while sometimes hilarious is generally frowned upon.

[Zelandeth]

Having decided that I'd had utterly enough of the outside world by about midday and having a complete failure to summon enthusiasm regarding doing anything actually useful...I spent a couple more hours fiddling around with the Sun.

Task number one was dealing with the vacuum hoses in the machine. I knew these had taken on the consistency of dried pasta and would disintegrate the moment I touched them. I wanted them changed before I started poking around as they were mostly still where they were meant to be...and I'd rather not play guessing games as to where they go.



By the time I was done, this is what was left. Plus a bunch of bits buried in the bottom of the case I'll need to get out with the vacuum cleaner.



Much better...



I discovered the hard way that that solenoid valve on the water purge line has mains on the (unsleeved) terminals even when the machine is off when it gave me a good old belt. Absolutely stupid not having the plug pulled before I did that.

Fiddly bit under the gas analyser done too. I'd missed the line heading off to the pressure transducer (far right out of frame) when taking this photo, it was sorted later.



With that sorted out I felt I could turn my attention to sorting the stuff which wasn't working. Two nuts removed allowed the whole front bezel to be removed...this vastly improved access.



This solves the problem of not being able to get to the computer.



This let me take a closer look at things. Not least figuring out what CPU the thing was running. I was expecting something 8-bit, maybe Z80 (half hoping actually as I've messed with them before and might have some hope of making heads or tails of the software) or 6000 series...Finding this however was rather a surprise.



That's a National INS8900D. For those not familiar with it, that's a 2MHz 16-bit processor... really wasn't expecting this to be a 16-bit machine... especially given it has its roots back in 1979...overkill?

The memory board. Yes, getting these EPROMs read and backed up is high on the priority list.



Not totally sure how, but I managed to forget to get photos of the MUX and I/O boards. I'll fix that later.

The board I was most interested in today however was this one.



The edge connectors actually looked pretty clean, though I gave it a scrub up anyhow. Based on prior experience though I figured the most sensible thing to do was to remove (carefully, using the right tool) each of the socketed chips and reseat them. Somewhat surprised to see a humble 555 timer in a socket. Have to wonder if they've had reliability issues with that IC.

With that done, slotted the card back in, powered on, and...



Rock steady display... I'll take that as a win. Having a display which would stay running for more than five seconds at a time I could let it complete the warm up process and run through the self test (having taken the opportunity to clean the CRT faces while it was in warm up as they were filthy).





Still surprised how sharp this display is... especially now there's not 1/8" of grime on it.

The faceplate was similarly filthy.



I wasn't surprised by the "service required" messages at the calibration screen. If I'd been abandoned in a cold damp shed since 2005 I'd be in need of a bit of TLC just the same.

I did a quick check on the gas analyser first to see if it would respond to a clean. First check though was of course to make sure that the IR source was "lit" correctly. The source here takes the form of a carbon rod, heated to the point where it just about glowing a really dim red (camera makes it look a little brighter than it is).



This is then focused via two parabolic mirrors onto a pair of sensors at the far end.







The spinning shutter at the source end allows the reading to be sampled from free air and the gas under test alternately.

It was given a really good clean (carefully...the mirrors are surface silvered) as the whole lot were filthy.

Sadly it didn't just miraculously come back to life...so further digging will be needed. Despite the errors,it was quite happy to continue into running mode.

First page you get is to enter the vehicle details - Number of cylinders, 2/4 stroke, and the timing offset. Once that is entered it presents you with the following screen.



The *** entries show where readings are missing due to the calibration issues. It will blank things out rather than potentially show erroneous data. This allows you to check things like the starter motor current, battery voltage during cranking. It also allows you to compare the current draw during each cylinder compression stroke to give a rough relative compression test to show if one pot is far lower than the rest.

Once that is completed, the engine running test page is presented.



I like the "bar graphs" drawn using the text...sneaky ways to do stuff like this without the overheads of bitmap graphics is half the fun of old kit from this sort of age.

Looks really the part with the front panel back in place.





So issues we need to look into:

[] Gas analyser inoperative.

[] Volt/ohm meter inoperative.

[] Temperature probe reading nonsense (see 160C reading above).

[] Suspicious of the vacuum reading.

[] Possible cap issues in the power supply as it shows a low line voltage warning at the calibration screen despite our line voltage being anything but low.

First port of call is going to be going over the machine end to end and reseating every socketed chip as I did on the CRT board, then see where we are. Will also make a point of checking for any signs of damage to the wiring where it enters the boom as it could be prone to chafing there.

Oh...and sort the dent in the power supply fan grill so it stops rattling before it drives me mad.



Interesting to see that the fans are actually 115V units...have to assume they're running from a step down transformer buried in that brick of a power supply.

Definitely making progress though.

Last random photo for the old tech enthusiasts...old computer equipment like this often displays interesting or odd artefacts on screen when rebooted. This is what the display shows on this for about a second when powered up.



Thought it might just be random nonsense in the RAM, but it seems to show the same every time.

[Zelandeth]

Today I picked up a couple of the correct F30WT12 tubes for the lit sign on the Sun 1215.

While T8 tubes physically fit and will work, the electrical specs are quite different. The T12 3' tube dates back to the original set of designs from the earliest days of fluorescent tech, whereas the T8 is a far later design which operates at a far higher voltage but lower current. As such an F30WT8 in this application would be overdriven, running at around 45W rather than the correct 30W, so wouldn't last long.

These however will be perfect. They're hard to find here as this size were never popular in the UK. These are new old stock circa 1993, so goodness knows where they've been hiding since then.



Replacement took all of about 30 seconds.



I remembered to check and clean both the reflector and inside of the diffuser this time. They were distinctly grubby.



Everything back together. You really can't see any difference externally to be honest.



As it's not the sort of equipment a lot of people will have used and it seems to have generated quite a bit of interest I thought I'd try to write up a bit of a guide to the basic startup and configuration of the machine. Figured some of you might find it interesting.

This isn't a full user guide...just a quick run through of the real basics from what I remember of using the one a mate had about 20 years ago and what I've seen messing around with this one.

Hopefully this will demystify all those buttons a bit.

When first powered on the machine shows a message informing you that it has started its warmup phase and gives a countdown (starting from fifteen minutes) to when it will be ready to use.



You can bypass this delay at any point by pressing #, but there's obviously a greater chance of the accuracy of the machine may drift as it warms up. The fifteen minute delay helps ensure that everything is stable before you put it into use.

The # button basically functions as the "next page" control throughout.



At the completion of the countdown, the machine waits for you to press # to continue before it will move on to the self test/calibration screen.



This takes a minute or so to run through...and obviously has a few errors reported in the case of my machine... hopefully these will disappear as time goes on and I work through things.



Next page is the program setup screen where things start to get a little more interesting.



The "set ignition selector" relates to two modes that the ignition side of things can run in. There's a table in the operator's manual telling you which cars which setting should be used with.

This is selected using this button.



Shown above in mode 1, and below in mode 2. Yes, this is a bit of an excuse to show off the flipdot indicators...even if they are a bit grubby at this stage.



"Set 2 or 4 cycle" is asking you to select whether the engine is a two or four stroke. There's a dedicated button for this, which like the ignition selector has an indicator built into the button itself.



Self explanatory really...4 for a four stroke and 2 for a two stroke.



The number of cylinders is then set using the left of the three buttons below. These have different functions in different program modes hence having several legends, it's less confusing than it initially looks.



The selection starts at 0 by default, and pressing the button cycles through 2, 3, 4, 5, 6, 8, 10, 12 then loops back to 2.

The last thing on this section is the timing offset. This is only needed I believe where the car uses a magnetic pickup to fire the ignition, this is usually offset a few degrees before or after TDC. You need to dial this in here.

Pressing "number select" initially changes the sign to + or - to reflect whether the offset is before or after TDC, then "cursor advance" moves to the first digit, with "number select" used to advance the number. This process is repeated until you've dialled in the correct offset.

The screen below is an example which would have been used on my old Saab from memory, you can see the cursor showing which digit is being entered.



Once you have advanced to the last digit it will tell you to press # to the next page. It won't let you proceed if there are things you've missed.



This brings you up to the "Cranking/Pinpoint Tests" page.



This gives you quite a bit of information on what's going on. However this is only half the information it can show you. An additional page is accessible using the "short" button on the control panel. This appears to indicate "short-cut" rather than that it shorts anything out to kill the ignition system or anything.

Here it's shown activated, hence the vivid orange indicator shown.



All of the buttons - even those which are purely momentary - have the indicators in them. Utterly unnecessary, but a nice touch.

This screen gives you some really quite clever diagnostic information.



This shows you the difference in engine speed, starter motor current draw as the engine spins over each cylinder. This can be helpful in showing if one or more cylinders has a significantly lower compression ratio than the others.

You might need to actually stop the engine from starting to get solid data from this test, and there's a control to disable the ignition system labelled as "engine kill."



Pressing this toggles the kill, and when it's enabled a flashing "engine kill" warning is shown on the display to warn the user. Having this feature on hand is obviously useful from a safety perspective as well.

The "Short" button latches, so pressing it again will drop you back to the Cranking/pinpoint Tests page.

Once you're done with that, pressing # will advance you to the main running test page.



This shows you pretty much everything you need to know. The displays look to update pretty rapidly, at least a few times a second. If wanted, you can pause the data on screen with the "Display Hold" button. This is directly below the # button.



When this is enabled a "hold" message flashes at the upper right of the screen to warn the user that live data isn't being shown.

A useful feature of this as well is that in the main screen above, when the display hold control is released, the original data is left in the screen with a new column being put up for the live data. This allows you two "old" readings to be shown along with the live data. This could be really useful if you're making small tweaks and wanting to double check what effect it's had.

The below display shows a "full" page with two columns of held data shown.



The live data is always shown on the leftmost column.

That's as far into the running of this unit as I've gone...but it gives you a basic rundown of how the computer works. Shows how it does a pretty good job of guiding you through everything...which in the mid 80s really wasn't a given!

Probably the most daunting looking controls relate to the scope...though I've never really done enough work with that to be able to talk through it from memory.



The Short button being grouped here rather than with the computer controls is just to keep you on your toes.

I'll try someday to write a bit of a how to for this too...I'll need to learn to drive it a bit better first though.

Likewise some of the advanced features of the computer...there's a whole additional layer in there I think.

For the sake of convenience there's a remote control which duplicates a number of buttons from the front panel to make life easier for the technician. #, Display Hold, Short and Engine Kill being those controls.



The "Print" button would trigger the optional printer. No timestamps or anything, it would just literally print a copy of what's shown on the screen. Sadly I don't have the printer, though I do have the interface card (found in the base cabinet) should I ever come across one.

This has been probably horribly tedious to most of you, but hopefully it's vaguely interesting to you if you're into old tech and want to see what would be happening when you're blindly mashing buttons to see what makes it tick.

If you have any questions please feel free to ask.

[myglaren]

That is a far more complex machine than I had thought.
The garage my dad sometimes worked in had one that he was quite impressed with and I can see why now.

[van ordinaire]

What amazes me is that your average mechanic was even expected to operate it, never mind be able to put it to good use. (says he, who still has not the slightest idea of what it's for)

[Zelandeth]

A couple of additional photos of the ridiculously overcomplicated status indicators built into the buttons.

It surprises me that they went to the lengths of building them into even just the momentary controls. You never actually *see* this one as the control is actually latched on software...the orange announciator in the button is normally hidden by your finger.



The actual indicator that this is active is a message that flashes at the lower right hand corner of the screen.



There is apparently a volts/ohms measurement capacity in there somewhere (which I've yet to figure out how to access), which is switched between the two modes using this control. Here it is in volts mode...



...and ohms mode.



Just surprises me this wasn't just done either on screen or using a status LED next to the control. I remember looking at this sort of control when I was building a power supply a few years back, and these things were *expensive* - even basic ones without legends like this. "We spared no expense" seems to be a running theme with this machine.

I've got a slightly more sane (though still offset) temperature reading now I've actually got the thermocouple plugged into it rather than a hall effect sensor!

I wanted to have a look at all the analogue cards to check for any additional socketed components, dirty contacts or dry joints.

Started out with the volts/ohms board as I know there are issues with that subsystem.



Had a brief "Ah ha!" moment when I spotted that toasty looking resistor...but checking it shows it to measure precisely the value that's stamped on the side of it. Nothing else obviously amiss.

The next board is the trigger control board.



This is quite important in that what it does is essentially listen for the ignition pulse firing on cylinder number 1, as that is the timing reference to which everything on the machine is slaved. Nothing amiss here that I can see.

Unsurprisingly, the next board along from that is the main timing board - this basically keeps everything in sync with the signal tracked by the trigger control board.



One of the more densely packed boards.

I initially thought that the next board (labelled "AMP" on the card cage) was going to handle a lot of signal amplification...but it actually appears to be the signal processing for the amps and temperature sensors.



This board does have a couple of dry joints that I'll give a tickle with the soldering iron tomorrow.

"CAL" is the next one, which I'm assuming given the precense of several relays, physically connects loads of known values across the inputs to undertake the calibration self test.



Again some of the soldering isn't great looking, but I couldn't actually see any dry joints...I may well reflow some of the heavier connections though as they could be better.

The fact that the only systems which are consistently failing the self calibration are the HC and CO meters is one of the reasons I want to have a really good look at the wiring from the I/O backplane to the sockets at the end of the boom.

Speaking of I/O, that's what the next board is tasked with dealing with.



Nothing wrong with this one that I could see.

Next up is the vertical pre-amplifier board, I believe this relates to the scope.



Quite obvious from the lovely old school hand routed traces that this one and the trigger control board are quite a bit older in design to the majority, though the date codes show they were actually made at the same time as the others.

"Logic Board No. 2" is next...though I don't recall seeing a number 1 anywhere!



I'm guessing a bit, but based on the hardware present that this is just handling some low level buffering or such like they didn't have room for in the main computer cage.

No signs of trouble anyway.

Finally we have an identical pair of cards containing quite beefy deflection amplifiers to drive the CRT (one card handles horizontal, the other vertical).



All of these boards could do with a good clean, as do the digital boards (which I'll grab proper photos of soon too) due to the close proximity of the cooling fan. Shame Sun didn't feel it worthwhile to fit a filter to it.

So nothing nothing obviously amiss there, aside from one resistor which has got a bit warm at some point and one capacitor on the amp/temp board that's got dry joints. I really did want to get them all out for a proper check over though. No socketed ICs or anything like that (which was responsible for the original display issue) in need of attention, but it was worth checking.

Looks like I might actually get a few hours free tomorrow afternoon, if so I'll hopefully get a bit more stuck into trying to work out what's going on with the faults.

The biggest irritation there really is that there don't seem to be many labelled test points, which would make checking to see if all the power supply rails are present and correct (there are a load of them!) a lot easier. Obviously given the card based construction it's a bit tricky to probe a lot of the machine when it's running as the cards are quite tightly packed together.

The other thing I wanted to check these cards for was whether they were hiding any more tantalum capacitors like I've seen on at least one or two of the digital boards, as once I've ascertained how many of them there are they will be getting replaced on mass. Wouldn't surprise me if that alone sorted a lot of the issues based on prior experience.

[Gibbo2286]

What amazes me is that your average mechanic was even expected to operate it, never mind be able to put it to good use. (says he, who still has not the slightest idea of what it's for)

As I said earlier we had Crypton Dynascope, a tool similar to the one in this thread, all the mechanics were trained to use it and we used it on every service.

[xantia_v6]

Where are the calibration settings stored? Non -volatile storage was not so readily available then.

[Zelandeth]

Where are the calibration settings stored? Non -volatile storage was not so readily available then.

I don't even know if they are as such.

The gas tester always takes alternate samples from free air and the gas under test...so it's always got free air there as the reference for zero. The gain controls to ensure that it's correctly tracking increasing volumes of HC/CO are simply multiturn potentiometers which would be altered with a bottle of calibration gas hooked up.

The ohms side of things zeros itself during the calibration, that's why it instructs you to short the leads together - I'd guess that also allows the machine to also check that that input is at zero.

I suspect everything else is just shorted to signal ground and possibly to regulated reference voltage lines during the calibration phase by those relays on the CAL card.

I don't think any of that is stored after power off, suspect that's why it goes through the calibration process on every boot (accompanied by much clicking of relays and solenoids).

[Hell Razor5543]

I like it when things are properly and solidly built. One of the best keyboards I have ever used was the Windows 95 keyboard (with a 5 pin DIN plug) from Fujitsu. If you were the sort of person who typed as though you hated the keyboard all that would happen was you got sore fingers (oh, and the keyboard worked properly nonetheless). I have taken it apart to clean it. From top to bottom it went;

Top casing,
Matrix casing (with LOTS of screws holding it to the steel backing plate),
Key cap,
Key head,
Spring,
Spring cup,
Rubber membrane,
Upper contact membrane,
Separator membrane,
Lower contact membrane,
Steel backing plate,
Bottom casing.



If you have ever seen the film "Wanted" you will know that James MacAvoys' character hits Chris Pratts' character with a keyboard (breaking the keyboard and, I think, knocking out a tooth). If he had used a Fujitsu keyboard there would be a lot more in the way of injuries, and the keyboard probably would still be useable!

[van ordinaire]

I bet they're not like that now.

[van ordinaire]

What amazes me is that your average mechanic was even expected to operate it, never mind be able to put it to good use. (says he, who still has not the slightest idea of what it's for)

As I said earlier we had Crypton Dynascope, a tool similar to the one in this thread, all the mechanics were trained to use it and we used it on every service.

Don't doubt it for one moment - that's why I'm amazed.

Until plug-in diagnostics I wasn't used to seeing anything that wasn't required for an MoT.

[Gibbo2286]

This is a version of the one we had, a bit later than ours and not as pretty ours was in a red cabinet.