Ever since I completed work on the tiny retro system, it has displayed a peculiar habit of not booting after being disconnected from the mains for any length of time. At first I attributed this to the CMOS battery, but after swapping the battery three times. I had to conclude something else was a foot.
After consulting with various friends who hold a better grasp of electronics. I came up with a theory, which some agree with and others are less convinced about.
The LED PCB which I salvaged from a cycle light, is what drives the LED array on the front panel. Designed to run from 3xAAA batteries. I couldn’t help think the LED’s where some how continuing to draw power after the system was shut down. The only remaining source being the 3v CMOS battery.
Because the the regulator I built to drop the 5v feed to 2.88v is very simple. I can’t help but speculate that nothing is preventing the LED’s from continuing to draw power. Imagine a torch if you will, when switched on, the bulb will continue to illuminate until the source of power (batteries) has been depleted.
When the Nomad is told to shut down, the system does so. However the LED PCB is independent of the motherboard, in all but drawing power from a molex power connector. So the circuit is essentially open, just the power has been cut, so the lights go out.
As I said before, this is a working theory, which could be wrong and the motherboard itself could be to blame. So in an effort to fix the problem, I have rigged a CMOS jumper reset switch on the rear of the case.
Now when ever the system fails to boot, I no longer have to remove the lid to access the CMOS Reset jumper. I simply press the button on the rear. Now why dont they fit this to modern computers! 😛
Taking on the task of building a machine in a customised case is not without it’s challenges. As I soon discovered when I came to designing the front of the machine. With six LED’s, two switches and one illuminated power button. I had plenty to fill up the flat blank looking facia. But how was I going to go about it? I knew I wanted lettering under the various lights and switches to indicate function. But how exactly was I going to do it, without it look awful. Someone with money to burn, might have paid for chemical etching or some method that paints the letters on. On a budget I faced using a label maker and sticking labels on to the front facia. This was definitely not the look I was aiming for. The Altair had a wonder front to it and I wanted something similar.
It was then when I had an idea.
Placing the front panel under the flatbed scanner, I made a reasonably high resolution copy. I then loaded the image up and began designing where all the LED’s and switches would go. Having already sketched a rough design out on paper, with measurements of the panel and the components. I had a pretty good idea where things would be installed, which meant building it on the computer was fairly straight forward.
Printing a template out on my laser printer, I cut it out and taped it to the plastic panel. Then using a dremmel and a drill piece, I bore out holes for all the components. The LED’s where the trickiest part, make the holes to large and the LED’s would fallout, the fitting needed to be snug. With the holes made, I used a file to clean the panel up and removed the paper template. Offering the switches and light up, I found more filing/drilling was needing. So back to the dremmel I went.
It took a good hour of fiddling before all the components slipped in to their designated openings, but soon it beginning to look Altair-ish. I wont deny a sense of pride filled me, when I saw the front panel with all the components sitting in the case for the first time. Up until that point, the project had felt more a less like a pile of components. So seeing it come together for the first time was pretty satisfying.
With switches and light installed, I used a hot glue gun to secure the LED’s. With the wiring I know the chances of them falling out would be inevitable.
The next step was the actual facia or cover. Now this was the clever bit which I mentioned earlier. Having made a template on the computer, I had also incorporated in to the template all the various labels. Printing out another copy, I cut the out the holes for the LED’s and switches and offer it up to the front panel. Almost all the components pushed through the holes where they should. But several of the LED’s did not line up. As I had been cutting out the holes by hand with the dremmel, I guess I should be pleased at most of them aligned at all.
Taking measurements I went back to the computer and altered the layout, spacing the LED’s farther apart. This took me several attempts and many templates before I finally had one that worked. Finally with all the parts pushing through where they should. I was ready for the next stage. Making a final panel cut out, using thicker, high resolution card. When the printer spewed forth the finished version I saw a problem. Unlike normal paper the high resolution card had given the black ink a gloss look. With no way around this, I took a leaf out of my prop making skills and made a gamble. Using some Matt varnish spray I use for for water sealing decal paper. I sprayed the facia with three coats, with 4 minutes between coats. I then used a hair dryer to gentle dry the varnish. After 20 minutes it was dry to touch and had to my joy given the printed panel a Matt finish!
Using some 3M adhesive spray, I mounted the facia to the front of the case. Definitely heart in your throat work, as fowling up at this stage would have meant a big mess. As it where, it went on without a hitch.
One thing I was aware of when making the cover, was the need for it to be 3-4mm smaller then then actual plastic panel. As the panel it self, slid inside channels cut out along the front of the bottom and top sections of the case.
With the card attached the panel would have been to thick to fit. As the facia only needed to cover the visible surface of the front panel, it wasn’t much of a problem. Thou when the lid finally did go on, some trimming was need before everything went on properly. Even the best laid plans can run fowl!
With a melted PCB the front panel was as dead as an Aqua concert. So now I was faced with the prospect of having to rewire everything with a fresh LED PCB, providing I was able to get one.
Luckily as it was and I ended up buying two, as I knew what my luck was like.
Rewiring the front panel for a second time actually turned out to be very useful. As I was not only able to shorten the wires going from the regulator to the PCB, but also those from the PCB to the 4 LED’s on the front panel. Allowing me this time, the ability to mount the PCB on to the rear of the front panel. Once more using the handy retro 70’s solder/ray gun. I was able to get the whole job done in a surprisingly short space of time, not to mention ease.
Having done my research, I installed a 20 ohm resistor between the voltage out of the regulator and the PCB, preventing the circuit from drawing to much current. I had infact bought a 15 and 20, in anticipation of a 20 ohm resistor making the LED’s too dim. Luckily this wasn’t the case. After a quick test on the low voltage PSU proved the new circuit was working properly, I went about testing the larger Uniross. It was some time after all this, that I was to discover the Uniross was not actually powerful enough to boot the motherboard, not even as far as posting. But now I’m wondering off, so back to hooking things up.
With everything wired up on the kitchen side, I held my breath and press the power button, at first nothing, but then pressing a little harder the front panel came to life. five minutes passed and the blinkers where still working, I flicked through the patterns for a good 10 minutes, waiting for signs of over heating. Not a hint, the 20 ohm resistor had done it!! Hurrah
Powering the machine down, I went about placing the case back together. While the biggest hurdle was now over with, the next challenge, the one I had been dreading was finally upon me. Finishing the front panel. While the LED’s blinked, the Power LED worked, the Reset and Pattern change toggle switches worked. But the one thing which it needed was writing under each LED and switch, indicating what they where. How the heck was I going to do it? Decal’s? No, as I didn’t own a printer capable of printing white ink and yes they do make them. Etching? no as that would cost me a fortune and this was a budget, home brew project. It had to be something I could do myself. So after a lot of thinking, I stole and idea from my prop making skills.
In my next post I shall cover how I designed the front panel, it was a pretty simple job, time consuming, but a project that could be undertaken by anyone.
Above is the circuit I built for the purpose of powering the LED PCB. At the heart of the circuit is a LM317 adjustable, three terminal regulator.
Initially setup without a 20 ohm resistor in line with the voltage out. This amendment to the circuit was needed to control the current feeding in to the PCB. Which caused the first setup to overheat and fail. In plain English, the microchip on the PCB melted, well near enough!
For the correct voltage to output from the circuit, the values of resistor’s R1 and R2 had to be determined. As my maths is terrible, I cheated by using an online calculator to figure out the necassery values. Which turn out to be 120 x 150 = 2.81 Volts. This was under the 3v given to the PCB by the 2 AA batteries that originally powered it. But given batteries are a constantly depleating power source, I did not think feeding a constant 3 volts in to the PCB would be healthy in the long term. Which is why I choose 2.8v.
I have to say that when it came to making this circuit, I found it quite a challange. I’ve always had a love for taking things apart, rebuilding them. But that has always been easy to me. It’s like working on a three dimensional jigsaw, where the pieces are components, which all slot together in a specific way. But making a circuit from nothing, this was something else, especially as I did not fully understand circuit diagrams and symbols. Luckily consulting several people online via IRC, proved the most useful thing I could have ever done. It is thanks to those individuals i was able to understand what i was doing and actually turn a rough diagram in to something that functioned.
With it completed, I connected a universal PSU outting 3v 1.5mah and tested the regulator. Nothing, the thing had zero output. except from when i hooked the voltmeter to the V in and ground. So something was wrong. I had nothing coming from the V out and Adjuster. While to this day I’m not sure how it happened. R2 somehow became perminently open, perhaps it was a faulty component, who knows. But for the cost of a few pence, i wasn’t really bothered. In addition to this, i had the feeling that my brief fight with a stubborn soldering iron, which refused to get properly warm. Had resulted in the regulator being damaged. In the end, I bought full set of replacement components. Using my dad’s older 70’s soldering gun, yes a gun with a trigger. I was able to fit the replacement parts in minutes. Honestly i dont know how i’ve managed so long without having this tool, compared the standard wand iron i have. The trigger iron, applies heat when you need it.
With the new components in place, i tested the circuit and discovered huzaar 2.8volts!! When i saw the reading pop up on the multimeter I was relieved. As i was sure the reason for the fault had been something i had wired wrong. But no, the circuit layout as it turned out was good.
With the circuit tested, I set about putting all the parts together within the case. The end result was a rather full looking box of wires and parts and myself looking ever more worried something would surely come loose.
Switching from the universal PSU to the internal mini ITX power supply did not strike me as anything that would set off a red flag. Having taken in to account the 12v and 5v feeds, I had already worked out that i would steal power from a redundant HDD molex connector. When the machine powered up, it was on the kitchen side. Everything worked! Leaving the system on while I went for my camera up stairs. I thought five minutes turned on would tell if everything was working. I couldn’t have be more correct. On my return i noticed one of the LED’s was nolonger working. Flicking the switch, it came back in to life, while two others had died. Placing my hand over the microchip on the LED PCB to check the heat, the hot glue covering the chip stuck to my hands. It wasn’t warm, it was cooking! Powering down for 10 minutes and then back on, revealed the PCB was indeed cooked beyond repair.
It was at this point i realised something i had overlooked, the current! the universal PSU had a maximum output of 9v 1.5mah. But what rating was the Uniross adaptor i was now using to power the entire system through the ATX power connector? A lot more then 1.5mah.
Success!! After the not so successful first attempt at constructing a voltage dropping circuit last week, i purchased a full set of replacement components, hoping a second attempt would prove more fruitful.
Yesterday I replaced the regulator and 120 ohm resistor which proved to have blown, most likely because I cooked them when my soldering iron decided to have a moment and wouldn’t reach proper temperature. Armed now with a soldering iron from the 70’s and looking not so dissimilar to a ray gun (Thanks dad!). I quickly swappws out the faulty parts with new and began testing the circuit. 4.64volts, not the 2.88volts I had been hoping for, something was a miss.
Quickly hopping on IRC I sought the some advice from Thrashbarg, hoping that perhaps he would (A.) be awake, as he lives in the states and (B.) Be able to give some input.
In the end we used Tinychat to video conference, so that Thrashbarg was able to observe me testing the circuit and direct me to area’s he wanted to examine himself. As HAL might say from Space Odyssey, “it can only be a attributed to human error.” I soon realised I had not been connecting the circuit up correctly, like a clutz i had been neglecting to wire up the ground. As soon as this was done, the multimeter flashed up with a satisfying reading of 2.88 volts, reducing the 5.04volts that was being put in to it, SUCCESS!!
With the circuit working, I took it downstairs and installed it within the case, wiring it to the front panel the next step was supplying power. With some trepidation, I powered the ITX board and checked the voltage coming down the power lines and then in to the circuit. 5v was being dropped to 2.88v as intended, splendid! However the front panel wasn’t actually doing anything, it was dead. Flicking the Aux switch on the front, i realised why. The PCB that drives the LED’s is from a bicycle light and like any light circuit, there is an on and off.
So i hope you enjoy the video, the next phase will be to tidy up the case and get Lubuntu installed.
With ever increasing electricity bills and the measure of your carbon foot print, being something for serious thought. The thought came to me while doing my routine roaming of the internet for vintage tech. What about building a new computer, eco friendly, good for surfing and encompassing my favourite past time, vintage computers. A quick look will reveal a wealth of information on old computer mods. Mini ITX boards squeezed inside Atari 2600’s, Commadore 64’s and much more. However that wasn’t what I was looking for. I wanted to build an old looking computer, around a modern mini ITX motherboard as the platform.
With a fondness in my heart for the blinky LED’s of the MITS Altair 8800, I had a rough idea what I wanted. Grabbing a few sheets of paper, I sketched out a general design.
After a good search around my local electronic stores website, I discovered a perfect ABS plastic project box, coincidentally closely resembling the case of the new Altair mini. If great minds think a like, I was certainly pleased to discover I was using a case similair to very machine I was using as inspiration.
Next I was faced with a hurdle, how to reproduce the front panel. Over the years as the micro computer or as its commonly known today, the personal computer (PC). Coined by the brains at IBM the term PC eventually became the common word for and computer used in the office or at home. But as I was saying, the front panel! If you glance at a modern PC, you will commonly find two LED’s indicating power and drive activity as well as a power and reset switch. Compared to the computers of old, its become a bit bare. Now how the devil was I going to populate the front of my computer with only two active LED’s. Well a wise man once say “If you have it use it and if you don’t, fake it!” So that’s exactly what I planned. Utilizing some form of LED driver circuit I would make the lights blink, giving the impression of activity.