Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Soviet PDP-11-40 (SM-4) computer repair Here’s a repair story from the past. After completing my electronics engineering degree in Poland in 1978, I started working at the Institute of Mechanical Engineering at a technical university (called a Polytechnic in Poland). The institute acquired a minicomputer. My duties included maintaining it and assisting scientists with programming it. To learn the internals of this minicomputer, I was sent on a course for a few weeks. The computer was an SM-4, a Soviet copy of the Digital Equipment Corporation (DEC) PDP-11/40. The story was that during the Vietnam War, Russians captured an American frigate with this minicomputer onboard and copied it. It’s hard to tell if those rumours were true, but during the course, the lecturers from the Soviet Union were using the original English documentation of the DEC PDP-11. The shape and position of the switches and lights on the front panel of the SM-4 look exactly like those of the DEC machine. One of my tasks was to regularly run test programs to check the health of the minicomputer. We had a collection of programs to test various components: the main processor board, memory, peripherals etc. Each test program was on punched tape. To run the program, it had to be loaded into the computer’s memory first. The computer did not have the equivalent of ROM for a startup program, so after switching it on, I had to manually enter the loader. Using switches on the front panel, I would set the starting address and the first instruction to Items Covered This Month • An old Soviet computer • A smart bin turned dumb • Repairing a Marantz PM630 amplifier • A broken standing fan • Compaq CQ56 laptop repair Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz Cartoonist – Louis Decrevel Website: loueee.com 84 Silicon Chip be stored at this address. After pressing the save switch, the instruction was stored and the address was incremented for the next instruction. For the rest of the loader, I only needed to enter the instruction codes and press the save switch for each one. There were twenty-something instructions to enter. The instruction code was an octal number rather than the hexadecimal we use today. For example, the instruction to clear a memory address (CLR) had code 050 (000101000). After a while, I just remembered those codes, so I could enter the instructions quite quickly. After entering the loader, I would put the tape with the program in the punched tape reader, set the address of the first instruction of the loader and press the run switch. The program was read from the tape into the memory and could then be executed. If there were no problems, the test would run to the end. I remember the first time there was an error: the test of the floating-point module stopped halfway. The address of the instruction that had an error was displayed on the front panel. It was time to apply what I had learned during the course. I switched off the machine, removed the front cover and unplugged the floating-point board. There was a so-called engineering panel, an extension that plugged into the rack in the place of a board and you would then plug the board into it. This way, the board was exposed outside the rack, with both sides fully accessible. It also had switches to set the address to stop the program and others to perform some operations. I plugged the engineering panel into the slot where the floating-point board originally was and connected the floating-­point board to it. The board was now outside the rack. After switching on the computer and loading the floating-­point test, I set the engineering panel break address to stop the program at the instruction where the test failed. I started the test program, and it stopped at the address set on the engineering panel before executing the instruction. At this point, the computer would normally execute this instruction, but using the engineering panel, I could step through the microinstructions that were processed to execute the program instruction. The computer was built with TTL 7400-series ICs (Soviet equivalents, to be specific). Australia's electronics magazine siliconchip.com.au Following the path of the signal at each microinstruction, I checked the inputs and outputs of various ICs. At one of the micro-steps, I found a NAND gate with all of its inputs high and the outputs were also high – all were 5V in the 7400 universe. There were no desoldering tools, so to replace a faulty IC in a DIP through-hole package with 14 pins, I cut all the legs off and removed the chip body. Next, I desoldered and removed from the board all 14 pins one-by-one. After cleaning out the old solder using a wire (we did not have desoldering braid), I soldered in the replacement chip. I cleared the breakpoint address on the engineering panel and executed the floating-point test again. This time, the test ran through without errors. I unplugged the board, took out the engineering panel and plugged the floating-point board back into the rack. I ran the test again just to confirm there were no more errors. Today, we would typically replace the board rather than go to the component level, especially since a single IC performs the work of thousands of individual chips. Only occasionally could we identify the faulty part. Cas Filar, Duncraig, WA. Sensor-triggered rubbish bin repair My youngest son asked if I would look at a rubbish bin with a sensor-actuated lid that had stopped working. The bin uses photo sensors to activate a motor that raises and lowers the lid when a hand is waved over them. My initial thought was that it would not be an economical repair. However, I enjoy a challenge, have spare time, and if it could be fixed cheaply, it would defer the cost of a new bin. That would be particularly helpful with the current cost of living and six mouths to feed in the household. The first step was to check the power supply, which consists of four 1.5V dry cells. All tested OK, so I proceeded with the disassembly. After removing ten self-tapping screws, I separated the lid assembly and quickly identified a likely source of the fault. The motor and gearbox assembly that drives the lid was coated with some kind of black, sticky, corrosive substance. I desoldered the wires from it, removed the motor/gearbox assembly and tested the motor using my bench power supply. There was no movement. I then began dismantling the gearbox, drawing a diagram to record the assembly sequence of the many nylon gears and shafts. All the gears were covered in the same sticky residue. After sliding the final drive gear off the motor shaft and removing two additional screws, the motor came free. Further testing confirmed the motor was lifeless. I carefully bent back the three retaining tabs securing the motor siliconchip.com.au endplate and slipped the rotor out. The brushes are thin strips of flexible metal, and they were bent and split. I tried re-shaping them and re-assembled it, but I found that the rotor would not turn. I suspected that bending the retaining tabs had disturbed the shaft alignment slightly. After some online research, I located a suitable replacement motor: the RF320 rated at 6V, 6000 RPM, 24mm in diameter, with threaded mounting holes in the right position. Fortunately, once received, the motor was a direct drop-in replacement. After thoroughly cleaning the gearbox and gears to remove the residue, I re-assembled the unit. The lid operated perfectly once installed. Another successful repair and one less item to be discarded. Phillip Webb, Hope Valley, SA. Marantz PM630 amplifier repair About eighteen months ago, my wife and I moved from the city to a rural town in the south-western part of the Wheatbelt in Western Australia. Being a collector of classic hifi equipment, I was recently asked if I was interested in acquiring some older Marantz hifi gear for free. I like older Marantz equipment, so my friend and I went over to an old abandoned farmhouse to grab the gear. The whole ‘box and dice’ was there: a genuine Marantz cabinet replete with a PM630 amplifier, TT530 turntable, ST530 tuner, SD530 auto-reverse cassette deck, CD54 CD player and some massive speakers. These had all been sitting in the farmhouse ever since it was abandoned many years ago for new digs. Australia's electronics magazine September 2026  85 Anyway, all the gear was quite dirty, and there was plenty of evidence of a mouse infestation from days gone by, judging from the deposits left behind on some of the gear. The speakers and the hifi cabinet were in too bad shape to bother with, but the rest of it, being Marantz ‘Champagne’ series, was loaded into the boot of the car with great enthusiasm. All of the equipment needed a good cleanup and a good looking at. I set about testing each item after a thorough cleanup, and as is typical of hifi gear of this ilk, the CD player, cassette deck and turntable all needed new belts. The cassette deck was the worst of them, with the main flat drive belt having turned to ‘goo’ over the years, leading to a very messy clean-up job. The turntable was a direct-drive type with linear tracking, utilising two small belts linked to motors and gears, one for the pickup arm tracking ‘sled’ and one for raising and lowering the arm. As for the CD54 CD player, the belt was a little loose, so it could not quite drive the loading tray mechanism in and out properly. New belts were duly ordered from my favourite supplier and fitted to each unit, restoring each of them to working condition, except for the cassette player, which would not auto-reverse reliably, so that was left for another day. Mostly easy fixes so far. Just for interest, the CD54 is essentially a first-generation CD player, employing the famous Philips CDM-1 swinging-­ arm laser mechanism and TDA1540 14-bit DACs. The unit is built like a tank, employing a heavy diecast chassis, and is quite sought after by collectors, me being one! It is CD54 number two for me. The tuner was relegated to the shelf, as I rarely listen to AM and FM broadcasts these days. The next repair candidate was the PM630 amplifier. It proved to be dead on power-up; the display and control panel LEDs did not light up as expected, so off came the cover to have a poke around. I managed to find service manuals for all of the gear on the interweb. Overall, the amplifier was in great physical condition, with no rust, no scratches, and no verdigris on the knobs, the latter being common in poorly stored units. Basic troubleshooting started with checking the main fuse, which was intact. Next, all the internal and onboard fuses were checked and also found to be intact. A check of the power supply voltages turned up no +5V rail (actually nominally around 5.7V), and no ±15V rails; the former revealing why there was no life from the microprocessor-­ controlled front panel. The lack of ±15V was not the best find either. Surprisingly or not, the main power amplifier and auxiliary supplies were all okay, and the speaker outputs had only a few millivolts of DC offset. So at least there seemed to be no faults in the main power amplifier circuitry, which was a welcome discovery. A subsequent visual check of resistors R807 and R808, both 27W fusible resistors for safety reasons, revealed that they looked pretty stressed. In fact, a deeper dive with a multimeter on the ohms range revealed they were both open circuit. Subsequently, I found that the ±V rails were both shorted to ground, explaining the unfortunate demise of the resistors. This did not inspire confidence, though, as there was a great deal of circuitry hanging off these supplies. I decided to turn my attention to the missing +5V rail first, noting that it was not shorted to ground, which was kind of good news. In fact, on closer inspection, the +5V supply output measured near zero, with the ‘input’ side sitting at around the expected voltage. Each of these supplies employs a zener diode ‘boosted’ by a series pass emitter-follower transistor, regulating the voltage to about 0.6V below the zener voltage. Anyway, since F803 and R810, a 10W fusible resistor, were intact, I measured a few voltages around the +5V supply, starting with the zener diode, a 6.2V type. I only measured a few tens of millivolts here, so I thought This section of the Marantz PM630’s circuit shows the power supply. Several of the electrolytic capacitors had failed, and not in the usual ways: with a high ESR or low capacitance. 86 Silicon Chip Australia's electronics magazine siliconchip.com.au that the zener had failed short-circuit, thus not providing any voltage at the base of the booster transistor. To be sure, I desoldered it carefully and measured it out of circuit using a power supply and series resistor, since I don’t have a fancy semiconductor tester. It measured 6.2V! OK, I thought, maybe there’s another culprit pulling down the voltage at the base of the transistor. Sure enough, a resistance measurement between the base of the transistor and ground indicated a near-short-circuit. It turned out that capacitor C818 was almost a dead short. This would be one of the few times I have ever come across a shorted electrolytic capacitor in my lifetime, especially since it seemed to be working well within its ratings. This capacitor was a 47μF 16V unit. With 5.6V across it in normal operation, I had no clue why it had failed. In any case, I was confident that refitting the zener diode and changing the capacitor should restore the +5V supply. This saw the +5V rail come to life along with the pretty lights on the front panel. Progress was being made! Having previously found the ±15V supplies shorted to ground, I now suspected C815 and C816. As it turns out, these were both shorted, which explains the missing voltages and fried resistors. Not having fusible resistors on hand, I had to order the same, along with new capacitors, and wait. Once they arrived, I fitted and changed all fusible resistors and small capacitors around these power supplies, including C813 and C814, for good measure. The failed capacitors were all small types from one particular Japanese manufacturer which, as it turns out, being of this vintage, are notorious for going bad. The good news is that all of this effort was rewarded with a now fully working PM630 to add to my collection of Marantz hifi gear! For reference, the circuit snippet at lower left shows the power supplies and components in question. Richard Kabzinski, Ellenbrook, WA. Standing fan repair It was a hot day, so I needed a standing fan for my work area. I got a spare fan out of the shed but as I was carrying it to my work area, the fan broke into two pieces, with the stand and shaft breaking off the main body and falling on the ground. As I had hold of the main body, the fan itself was not damaged. siliconchip.com.au The plastic had become brittle over time, and it just broke under the weight of the fan while I was carrying it. I went back to the shed and retrieved two dead fan motors with their control panel assemblies still attached. I would use one of these to repair the fan that was still working. I usually keep fan parts for repairs; I recently rebuilt a standing fan that my son picked up from the Op Shop that was missing the blade and guard; it also had no stand. The motor still worked, so I was able to rebuild it using parts from previously failed department store fans. These department store fans have a high mortality rate, as the motor has a non-resetting thermal fuse buried deep in the windings. When it blows, the motor is useless and the fan can’t be repaired, short of replacing the motor. The old fans from well-known brands do not have this thermal fuse, so it’s common for fans over 50 years old to still work. I went to undo the screws in the control panels, and of course they were “tamper-proof” types. These had an indented triangular hole that required a triangular bit. I’ve found these particular screws previously when I repaired a power board some time ago. At the time, I looked through about eight bit sets, and I did not have a bit for this type of screw, so it might not be common yet. Not having a suitable bit for this type of screw, I made a screwdriver to fit them from a piece of thick fencing wire. I used that to open the three control panels, ready for the repair. I checked the two spare control panels and found that one would not fit the fan I was repairing, but the other one fitted nicely. So I would swap out the broken parts with the good parts and get the fan operational again. Both control panels were wired the same way, so I just had to swap the wires from the broken one to the replacement one. The wires are just pushed into the terminals on the switch, and they can be easily removed by inserting a thin tool into the terminal next to the wire. I used the point of a compass for this job. The Neutral wire in the mains cable and the Neutral wire going to the motor were joined in a crimped connector; they were not even twisted together or soldered. That seemed dangerous to me. The blue wire coming from the motor was too short to reach the terminal in the new control panel, so I extended it by soldering on and heat-shrinking a short length of blue wire salvaged from the wiring in the fan that I took the control panel from. I also soldered the Neutral wire before refitting the original crimp connector to it. Australia's electronics magazine September 2026  87 Servicing Stories Wanted Do you have any good servicing stories that you would like to share in The Serviceman column in SILICON CHIP? If so, why not send those stories in to us? It doesn’t matter what the story is about as long as it’s in some way related to the electronics or electrical industries, to computers or even to cars and similar. We pay for all contributions published but please note that your material must be original. Send your contribution by email to: editor<at>siliconchip.com.au Please be sure to include your full name and address details. The disassembled Compaq CQ56 laptop. The very dusty heatsink is shown below, with the reassembled laptop shown next to it, waiting for a new screen & keyboard. While I don’t like these crimp connectors, they serve the purpose of covering the joint and are fine if the joint is soldered first. Reusing one saves using more heat-shrink tubing. With the replacement control panel fitted, I reassembled the fan, and it was ready to use again. It’s very handy having spare parts to be able to effect repairs, so I hang onto these dead department store fans. That was another successful repair that saved this fan from the junk pile. Bruce Pierson, Dundathu, Qld. Compaq CQ56 Laptop repair I was looking through some junk laptops that a friend gave me a while ago, and amongst them was a Compaq CQ56 laptop. It was missing the screen, keyboard, battery, right hinge cover and RAM. Could it work? I got a charger, plugged it in, pressed the power button and the laptop lit up. That was a good sign, but with no screen, I couldn’t be sure it was working. I had wrecked a non-working HP dv6 laptop recently. The Compaq’s screen connector was still there in the empty lid, so I had a look at it and it just happened to match the one in the HP. Its screen had some scratches, but it would do as a test screen. I didn’t have the correct keyboard for this CQ56, but I had a smaller one with the same connector, so that was worth trying. 88 Silicon Chip I disconnected the charger, installed some RAM, connected the screen and keyboard, reconnected the charger and pressed the power button. Holding down the F10 key, the screen lit up, and I was at the settings screen. Amazingly, this stripped-out laptop actually worked. Of course, the CMOS battery (cell) was flat, as expected with a 16-year-old laptop. It was time to dismantle it and start the rebuild. Fortunately, the optical drive was still in place; it’s always difficult to find a correctly fitting front and a rear retainer to suit a particular laptop. I dismantled what was left of the computer and decided I would upgrade the CPU while I had it apart to change the CMOS’s CR2032 cell. I looked online regarding what CPU I could use; because this laptop has a GL40 chipset, it only supports CPUs with a front-side bus (FSB) speed up to 800MHz. This ruled out using an Intel P8600 at 2.4GHz, which has a front-side bus speed of 1066MHz. The CQ56 came with an Intel Celeron T3500 CPU at 2.1GHz. Other supported CPUs include the Intel T9300 at 2.5GHz and the Intel T9500 at 2.6GHz. Unfortunately, I did not have either of these, so I had to settle for an Intel T4500 CPU at 2.3GHz. I had removed this CPU from another laptop where I replaced it with an Intel P8600. The T4500 was still a worthwhile upgrade. Australia's electronics magazine siliconchip.com.au After removing the heatsink and fan from the motherboard, I separated the fan from the heatsink and found the worst blocked-up heatsink that I have ever encountered. I’ve seen some badly blocked-up heatsinks, but this one is top! With the replacement CPU fitted and the heatsink and fan cleaned, I refitted the fan to the heatsink, cleaned off the old heatsink compound, applied a new smear and refitted the heatsink to the motherboard. When I went to replace the CR2032 cell, I noticed that the cell holder was broken, so I would need to replace it. I looked through my dead motherboards that I kept for spare parts when I wrecked non-working laptops and found a similar cell holder. I used my 20W soldering iron to remove it from the motherboard, then removed the cell holder from the CQ56 motherboard and fitted the replacement cell holder. Fortunately, the cell holders had the tabs at the ends, so this was an easy replacement. Next, I detached the lid from the main laptop body so I could rebuild the lid with a replacement screen. I unclipped the front panel from the lid, ready to replace the screen. I don’t know who stripped this laptop, but they didn’t remove the two screws from the bottom of the screen front cover and they just ripped the front cover off. It’s amazing that they didn’t break anything else when they did that. I had a good screen that came from a non-working Toshiba C850 laptop that I’d wrecked a while back. Even though this screen came from a different brand, it was compatible with this CQ56 laptop. All the screws to fit the screen were missing, so I grabbed a loose lid from a laptop I’d wrecked and retrieved all the screws I needed. I reassembled the lid with the replacement screen, reinstalled the motherboard and put the laptop back together again with a brand new keyboard that I’d ordered from eBay. I had several salvaged HP batteries that had come from dead laptops, so I fitted one. A quick test showed that the laptop was now working. I set the time and date and checked other settings in the BIOS, then saved the settings. The first battery I tried didn’t charge, but the third one did. Now it was time to install Linux, which is a good choice for a laptop this old with only 4GB of RAM. It would be a useless snail trying to run Windows 10 or 11 on it. I had downloaded the Pearl OS 8 ISO and burned it to a DVD earlier, so I used that. When I tried to install updates, all I got was an error message that the repositories did not have a release file, so no updates could be installed. Looking online, I saw that version 8 had been discontinued, but versions 12 and 13 had just been released recently. I read some reviews and they were all bad for version 13, but there was a good review for version 12, so I thought I would try it. I downloaded version 12, then burned it to a double-layer DVD, as it was 5.26GB, so it wouldn’t fit on a single-layer DVD. I installed that on the CQ56 laptop. The installation went smoothly, apart from a bug with the scroll direction for the touchpad, but after updating, that was fixed. This once piece of scrap now has a new life. A good result. Whenever I get a chance to buy non-working laptops for peanuts (or free), I pick them up because you never know when the good parts from them will come in handy. SC Bruce Pierson, Dundathu, Qld. siliconchip.com.au Australia's electronics magazine September 2026  89