The Commodore VIC=20 came in different versions with two very different power supplies. The earliest models had all the voltage regulation circuitry inside the computer. The external supply was simply a 9 volt AC transformer. These have a strange, hard to find, two pin connector. But since a transformer is a very reliable device, the power supply itself will quite likely last almost forever, barring any abuse. The computer, with all the heat producing stuff inside, may be a different story.
The other type of VIC=20 power supply has all the voltage regulating electronics in the "lump in the line" power box. These have a 7 pin DIN socket on the computer and a 7 pin DIN plug, with three pins missing, on the supply. These are notorious for failing in a manner that causes to voltage to go well above the 5V the computer wants. That high voltage will often take out some of the (hard to find) chips inside the computer.
Since my VIC is 35 years old I worry about its power supply failing and taking the innards with it. Plus, I plan on operating it on a 12V gel cell battery. So, I decided to build two things: a replacement switching regulator that regulates twelve volts down to five, and an overvoltage crowbar that can be added between any power supply and the VIC. The supply is built and described here. The overvoltage circuit should be coming soon.
There wasn't a lot of design work to put in. I used an LM2576-5 Simple Switcher from TI (formerly National Semiconductor.) I wanted a switching regulator for efficiency that could provide more than enough current for the VIC and whatever add-ons I need. Three amps seemed plenty, since the VIC itself takes around 1.5. I also wanted a simple circuit and easy availability. The 2576 is easily available and requires very few external components. Using the charts and graphs in the datasheet, and a bit of experienced guesstimation, I came up with the circuit below. Most of the parts are easily available from Tayda Electronics. The only unusual part is the inductor. There are plenty on the market that would work fine, but I wound my own. I needed a 100 uH inductor what could easily handle the peak currents involved. I had some T80-26 powdered iron toroids that are made just for such a thing. I wound 49 turns of 22 gauge wire on it for about 100 uH. Unless you really know what you are doing, I would recommend either doing the same or following the recommendations for inductors in the datasheet.
So far, I have built and tried the supply. I haven't done any real testing, but I will. I will update this site with test results when I do. Although the supply should nominally handle 3 amps, it is probably best to limit it to about 2.5. That should still be plenty. To get the full three amps it would need a diode rated for about 4 amps and perhaps a better inductor. The core should handle it fine, but the wire size is a bit small.
The power supply should work with any DC input from 8 volts to 15 volts, with a maximum output current of about 2.5 amps. It is intended for use mainly with a 12 volt gel cell battery, but any decently smooth supply in that range should work. I will probably build an AC supply in the future.
A crowbar circuit gets its name from the idea of placing an actual crowbar across power terminals. If you did that, it would short all the current through the crowbar, causing the output voltage to drop (almost) to zero and hopefully blowing any fuse or circuit breaker. The electronic crowbar is intended to have the same effect.
The general idea of the crowbar over-voltage protection circuit is that a voltage detector watches for over-voltage and if it sees the voltage go too high it turns on a latching switch that shorts out the supply. That causes the output voltage to go to near zero and causes the fuse to blow, shutting off the power.
In the circuit above I have added the crowbar circuit to the previous power supply. The voltage sensing components are R1, D3, and D4. D3 is a 5V1 zener diode and D4 is a small-signal diode with a voltage drop around 0.5 mA at low current. Between these two diodes, current is blocked until the output voltage reaches about 5.5V to 5.6V. If the voltage gets to that point, current will flow through the diodes and the resistor. The voltage across the resistor (E = I*R) will feed into the Silicon Controlled Rectifier (SCR) D2. The SCR is the latching component we need. It acts somewhat like a transistor: no current flows from the Anode to the Cathode until it is "triggered" by current into the Gate. The voltage drop across R1 creates the needed current flow into the Gate terminal. Once the SCR "fires" by being triggered with Gate current, it will conduct current through itself from the Anode to the Cathode until that current drops below a minimum value, even if the Gate current is removed. In essence, it latches on.
So far this is a typical, standard crowbar circuit. If you Google for crowbar circuits you will see many like this with lots of explanations. In most cases, the Anode of the SCR will be connected directly to the regulator output. In that case it shorts the output to ground causing the output voltage to drop and eventually blowing the fuse. Here, I've added a couple of components.
I added D5 and D6, 1N4001 rectifier diodes. The reason for that is that I want to short the output to ground, but the regulator has current limiting circuitry that may prevent the fuse from blowing, or at least delay it. So, I wanted to connect the SCR to the regulator output and also to the input side ensure the fuse would blow quickly. To do that, the higher input voltage must be isolated from the lower output voltage. The diodes take care of that. Now, the SCR can short both input and output to ground through D5 and D6 without allowing the 12V input getting to the 5V output. Pretty simple for such an important task.
Let's talk a little about component values. First, I really should have used a 5V6 zener diode (or even a 6V2) instead of the zener plus signal diode. But I didn't have one. This combination works well enough, but the 5V6 (e.g. 1N4734) is recommended. With it, the 1N914 isn't needed. You can choose a different value if you want to change the trigger point. But I don't recommend going above 6V2 (1N4735.)
Next is the rating of the SCR. I have specified an S306B1. It is an 8 amp SCR which is really overkill for this application. But they were cheap (from BG Micro. Any SCR with a high enough voltage rating and 3 amps or more current rating should be fine. Watch out for trigger current. A "sensitive gate" SCR is probably best. And since the SCR should only conduct current briefly on VERY rare occassions, a current rating somewhat lower would probably be fine. But SCRs are cheap; spend $1 and get a good one.
Finally, the isolating diodes D5 and D6 are 1N4001 rectifiers rated at only 1 amp. The same logic applies: they should only be used briefly and rarely. The data sheet for them says they can handle 30 or more amps for a single surge. That should blow the fuse pretty quick. It should never be a problem.
And there we have it. A complete power supply with plenty of current and over-voltage protection to power your VIC20 from a 12V system, for a grand total of about $10 or less.