Category Archives: Electronics

More generator issues

Checked the output of the generator with nothing connected.

running

With no load it was causing the capacitor to saturate at about 46 Volts. Not so good!

fractured

Also, the main voltage regulator chip had vibrated loose and fractured off. I made a bracket to secure it to the board better.

secured

Bolted down securely, and relocated the capacitor; however with only the rectifier in place I’m reading a fairly significant AC component on my multimeter. I’m going to hook my oscilloscope up to it and see what the output looks like. I have a feeling possibly one (or more) of the diodes have failed.

Processor and a monitoring panel (plus holiday oddments)

After the previous discovery that the -12V rail was collapsing under load, I read up on the datasheet for the 8080A processor. It states that the voltage rails must be brought up in a particular order (-5, +5, +12) and if the -5 is missing then there’s a very high chance of damaging the processor. The 8085 negated this and required a single, positive power source- the 8080 has strict requirements.
The 4116 RAM states that it should be brought up with the -5 first then the +5 and then the +12 but any of the rails missing should not cause damage to the chips. Bonus there.
So, I decided just to fling a processor at it. For $4 a pop, I decided that it was worth a shot, having fixed the power supply.

8080A

Same behavior upon power-up. At least I have a couple spares now.

repeating trash

Pulling the video RAM does show a different pattern on the screen so that at least is moderately positive- it initializes the memory in the same fashion each time it it switched on. Pulling the CPU and I/O cards and powering up the video card only liberates a solid beep, no crackling warbles.

voltmeter

Bought some little voltmeters from China- self-powered with internal self-regulated reference. They work from about 2.5 Volts all the way up to 30 which is impressive.

contrast

Also, bought some red transparent acrylic to serve as a mounting plate and to act as a contrast-improving device for the displays.

cut out

Marked up, cut with a hacksaw to size and sanded down smooth. Gently heated it up with my butane torch and bent it into a shape that stands up by itself.

measured

Measured up for the meter locations, drawing the positions on the front in Sharpie.

pva

Spread a thin layer of PVA glue on to each meter. It’s nice because it dries clear and doesn’t attack the plastics.

row of meters

Five meters stuck on and waiting for the glue to dry.

heat shrink

Trimmed the wires down to length and soldered on a length of multi-core wire, and using a spoon as a heat shield, added some heat shrink tubing to each join.

clips

Soldered some test leads onto the other end of the cable and marked them up as to which is which.

plugged in

Got everything hooked up onto the appropriate locations on the PSU.

meter on

With a few labels to show what’s what (and a couple others because I got carried away) the output of the PSU is monitored fully. Everything is in spec, all loaded down with all boards fitted. Now that I have this set up, I’m happier in the knowledge that everything is doing what it’s meant to in terms of power.

Next up, pulling it all out of the chassis to get the analyzer clipped on properly.

new light

Back to the house, I bought a new 4′ Lithonia light fixture to replace the junky Lights of America one that refused to light one tube and barely lit the other. Wired it up, fitted the old 25W tubes and flicked the switch. Nice bright light with instant start, too.

recapped tv

Recapped a TV for someone at work. The ventilation grilles at the back had become lodged with dust due to the location it was kept in (TV stand with no access to the sides) and as such, had cooked itself. Uprated the heat ratings of the replacements (all the originals were 85C, the new ones went in all 105C with a few in really warm locations at 135C) and ran it for an evening. Worked well.

Jeep

Traded the Challenger in against a new vehicle- Jeep Grand Cherokee. A bit more practical than ferrying the family around in a 2-door muscle car.

drip drip

Changed the ignition amplifier in the Renault and also a new distributor cap and arm, along with the crank position sensor. The CPS was at fault causing the non-running but the ignition amp appears to have been a bit marginal because it has a lot more torque now. Also, the power steering has stopped leaking and is holding fluid. The gearbox is leaking though, along with a weep of coolant onto the exhaust manifold that I need to find the source of because it smells when I start the car up. Other than that it is actually working moderately well.

Eighty-Eighty

I decided that I had wanted to give myself a little bit of mental exercise so I started work on the Lanier again.

Lanier on table

This began by dumping the thing on the table, pulling the CPU card out and trying to begin modeling its schematics in KiCAD. This is slow, problem-riddled (the board traces go back-front-back-front and are difficult to follow).

Made a decision to power it up after far too long sat up doing nothing. Pulled the PSU out and brought it up slowly on the Variac.

slowly, slowly

Same old nonsense, the thing didn’t do a whole lot of anything to begin with, a bit of a beep, floppy drive lights on. Hooked it up to my new logic analyzer on what looks to most likely be the bus on the backplane.

logic analyzer

There’s a bit of life, and the behavior begins to change. The floppy drives begin to clatter and rattle. I solder a bypass onto the video card to make the screen always show what’s on-screen (blanking bypass) and it’s got some nonsense changing here and there, occasionally filling the screen with alternating patterns.

logic traces

The logic traces look “dirty”, there’s a lot of glitches and the machine in general seems rather unhappy.

omooikkiakka

Some of the random content on the screen can be moderately amusing though.

Right around this point I was contacted by a guy in Ireland who used to service these things, still has one, schematics, discs and spares- everything I thought I’d never got for it. He says it’s a real early model with the screws holding the case on at the sides (probably circa 1975).
His suggestion was check the power supply, because random freaking-out type behavior is normally bad power.

power

I pulled the PSU and checked the voltages present. Everything was where it was supposed to be. Plugged the boards back in and same problem. Started to probe voltages on-load and found that the -12V rail had dropped down to +0.4V which is rather odd. Put a light bulb across the rail- it flashed then went out. Pulled the video card out and the rail stayed stable.

Zener

The RAM in the machine was all 4116, which requires +12, +5 and -5V supplies to work. I had replaced the CPU board with 4164, which needs only +5 so that cuts down on heat and current requirements significantly. The video board still uses 4116 so would definitely not be working without -5, which is derived from the -12 rail. The circuit is rather simple, a diode to the -12V rail, through a 330 Ohm resistor and the other side of that held to a stable voltage by a 5.1V Zener diode. I pulled one leg of the Zener off, removed all the RAM chips and put the video card back in and powered the machine up. The screen filled with [] symbols, which was expected as the RAM is not present and all lines would remain at zero. The -12V rail stayed stable also so the Zener was suspect. I’ve had them fail in the past and turn into almost a dead short. Found a temporary replacement in my box of stuff (half the Wattage rating but OK for a temp try).
Back into the chassis and still the -12V rail collapses. Start tweaking the crowbar adjustment and manage to get it to an unhappy -11.5V before it would just trip out over and over. Turn the PSU over.. and the -15 rail, which the -12 comes from.. is the only line I’d not replaced the main filter capacitor on.

capacitor

It’s all leaky up top, the vent is showing signs of having vented out. Not good. That’s most likely going short, drawing too much current and when the boards are added to that draw, the over-current crowbar is kicking in and shutting the rail down.
I had a dig and find a 15000uF, 50V capacitor with clip lugs. Soldered that in on short flying leads and powered the PSU up. -12 rail is at 12.04V, -15 at 13.3 unloaded so I adjusted them and then hooked up the computer boards. Made a slight tweak to all the voltages and got them within 0.01V stable.

There’s still something that’s not happy because switched on cold it beeps (with a somewhat crackling warble modulating the beep) and the drives click. Once it’s been on for a while it begins to not beep any more and it still goes a bit haywire but at least now I know that the main supply is good. I need to order the correct screw-in capacitor and a higher Wattage Zener for the -5 rail supply on the video card, but this is positive progress.

Mac back on track

Despite the little set of ceramic capacitors I put in having a combined value in an acceptable range for C1, it would appear they were either breaking down or otherwise incapable of sustaining the horizontal frequency of the screen scan.

caps being meaured

So, a different set of capacitors was used and that brought the screen back to life.

working again

That appears to overall conclude the work on this machine.

Broken Macintosh

Very short-lived, this one. I took everything apart and adjusted up the screen so it was central.

Nice and even

Set it on the side to do a bit of a burn-in test, and the decoupling capacitor for the horizontal yoke let out all its magic smoke, the screen went a bit funny and I switched it all off quickly.

c1

Decided to build the little capacitor network suggested in the manual, of several ceramic capacitors in parallel to build up the capacitance but reduce the ESR.

c1c1c1c1

Put it back together and again we have screen but.. the horizontal isn’t aligned any more. The two interlaced pieces of the screen, all stretched out – do not line up.

interlaced

I’m going through the schematic, I’ll check all values and make sure that the capacitor failing did not draw too much current through the flyback and fry one of the windings.

Rejuvenated Macintosh

I decided that the wobbly-screen Macintosh needed to have a bit of work done to make it a bit more trustworthy and reliable. The electrolytic capacitors had been installed in it since 1985 (8113th device produced in the 24th week of 1985 in Fremont, per the serial, so a real early 512k).

mac on table

Took it to pieces after ordering a bunch of spares (total about $15, so not bad), with the replacement caps rated at least one voltage rating over what was fitted and 105C rather than 85C as fitted originally to try and keep them a little more stable.

logic board

Not many capacitors on the logic board, though it is a 4-layer PCB so extra care had to be taken. I’ll give them this, it’s a well made PCB.

logic board closeup

Tried to keep the solder-work neat. Don’t think I did too bad a job, to be honest.

work in progress

Got stuck in doing the power and sweep board, which has the majority of the capacitors.

powered up

First test, and things were promising until the brightness control decided to stop working.

broken brightness pot

Turns out the center carbon track had completely fallen apart. Checked the schematic, fairly standard 16mm 1MOhm linear pot, so went on the hunt for something similar in the electronics catalogs online.

new pot

Closest modern variant I could find. The board is the right size and the connecting pins the correct length, important in this application because of the way the computer is assembled.

take apart

Carefully undid the new piece, as the thing was essentially backwards from how it needed to be.

sweep

This side is where the old one had the handle coming out. I removed the crimped section and split it all down, had to file the ears of the board a little to make it fit but managed to put it back together again, and have it test good and smooth operation across its range of travel.

reassembled pot

Put that back in the circuit and the brightness is now adjustable again. Checked the voltages and the 5V rail (the master) was a bit low at 4.72V so that was tweaked up to 5.00V. The 12V rail, slaved from it sat happily at 12.2V so that’s all well within limits.

5 volts

The screen became a bit intermittent- traced that eventually down to a couple broken solder joints on the connector to the end cap of the tube.

broken joints

Removed the old solder and re-added new.

working now

All up and working again. Need to adjust the centering of the screen a little, I think I must have knocked the centering rings when I put the wiring back in. Overall, a lot better than it was.

New Arduino Project – matrix LED display

I salvaged a couple of displays from some old equipment that was being thrown away. Pulled up the datasheet and started trying to make my Arduino Uno display things.

gibberish

After bit-bashing in some dots, I managed to get it to correctly display a letter, this case a lowercase “a”.

a

Now it actually works- I’ve made it scroll text input from the Serial port, full ASCII table. 3.8kb of code, not bad compiled from C.

nmmm

The above was from a leafed/paged test of the full ASCII table I made it loop.

Wipe and wait

My car has a 4-position windshield wiper stalk; off, intermittent, slow, fast.

While the motor needs to be pulled and serviced because there’s little difference between slow and fast, the intermittent wipe had always been somewhat pathetic.

Selecting intermittent would make the relay click, and as soon as the wipers returned to the parked position, click again and they would almost immediately wipe again.

Pulled the dash shroud down and located the clicking box at the top left. (Top right is the key-in/seatbelt buzzer, bottom left the cruise control stalk, above that with the finned heat sink the dashboard dimmer control).

relay delay box

Took the circuit board out and drew up the circuit diagram by following the traces, to get a better understanding of the circuit and its method of operation.

board

Culprits were most likely the electrolytic capacitors. One had a high ESD and the other was meant to be 22uF, it measured in at about 100nF. Replaced them both. Now the wipers wipe and wait about 5 seconds before wiping again on intermittent, and added bonus, they now wipe 3 times after spraying screenwash onto the screen. Not bad for 45 minutes’ work and spares I had in stock.

Co-processing.

I decided to buy a little something for myself, so eBay was opened up and I hit “Buy It Now” on an item.

Package

3 weeks later, it arrived from Kowloon, Hong Kong. The reason for the package? This empty socket.

socket

As if by magic, it was filled with a suitable device:

387

Upon fitting a 387 co-processor, a jumper needed to be shorted to inform the board to use it.

jumper

CheckIt determined that the thing was working (which is good, as there are surprisingly quite a lot of fake/broken chips circulating the market).

checkit

It was then tidied up into the desk. Makes a significant improvement where a math co-processor is utilized.

lotus 3

A bit of fun, reminder of days gone by.

Film photography and a 386

I went to an estate sale at the end of last week and picked up a few things, the computer below and a camera, of the notably interesting items.

386

A 1993 Magnavox (Philips) 386 SX/20, with optional CD-ROM sitting on top. It was grimy on the outside, filthy inside as to be expected from an older system.

A1

After a thorough cleaning, the Canon A-1 camera. 35mm film camera with fully automatic exposure (AE). View through the viewfinder (this shot before I had cleaned the lens and filter, so it’s a bit hazy) shows the aperture and exposure time in f/stops and 100ths of a second respectively.

viewfinder

It’ll calculate that and set itself if you leave it in fully programmed mode.. fancy stuff. Bought a new battery for it and a foll of Fuji 400.. have been shooting some pictures so once that’s finished I’ll see if they came out OK.

wolfenstein

Got the computer running after giving it a thorough clean and fixing a couple contacts. Found a cable to replace the one in the back to link to the CD. Added a network card so it has CD interface, network and sound. Multimedia on a 386, before “multimedia” was a coined phrase!

cd working

Added a larger hard drive in (540Mb, or rather, 30Gb but partitioned down to the max the BIOS can handle without an overlay) and got the CD to read. 1x so it’s not fast but it does work!