19.2.1998
Timo Raita <vic@iki.fi> http://www.iki.fi/vic/
Pasi 'Albert' Ojala <a1bert@iki.fi> http://www.iki.fi/a1bert/

PAL VIC20 Goes NTSC

Introduction

Recently Marko Mäkelä organized an order from Jameco's C= chip closeout sale, one of the items being a heap of 6560R2-101 chips. When we had the chips, we of course wanted to get some of our PAL machines running with these NTSC VIC-I chips. A couple of weeks earlier Richard Atkinson wrote on the cbm-hackers mailing list that he had got a 6560 chip running on a PAL board. He used an oscillator module, because he couldn't get the 14.31818 MHz crystal to oscillate in the PAL VIC's clock generator circuit.

We checked the PAL and NTSC VIC20 schematic diagrams but couldn't notice a significant difference in the clock generator circuits. There seemed to be no reason why a PAL machine could not be converted into NTSC machine fairly easily by changing the crystal and making a couple of small changes. Adding a clock oscillator felt a somewhat desperate quick'n'dirty solution.

Note that some old television sets require you to adjust their vertical hold knob for them to show 60 Hz (NTSC) frame rate. Some recent pseudo-intelligent televisions only display one frame rate (50 Hz in PAL-land). Multistandard television sets and most video monitors do display 60 Hz picture correctly. There is a very small chance that your display does not like the 60 Hz frame rate. Still, be careful if you haven't tried 60 Hz before.

You should also note that PAL and NTSC machines use different KERNAL ROM versions, 901486-06 is for NTSC and 901486-07 is for PAL. However, the differences are small. In an NTSC machine with a PAL ROM the screen is not centered, the 60 Hz timer is not accurate and the RS-232 timing values are wrong.

The Story

Timo:

At first I took a VIC20CR board (FAB NO. 251040-01) and just replaced the videochip and the crystal, and as you might suspect, it didn't work. I noticed that the two resistors in the clock generator circuit (R5 and R6) had a value of 470 ohms, while the schematics (both NTSC and PAL!) stated that they should be 330 ohms. I replaced those resistors, and also noticed the single 56 pF capacitor on the bottom side of the board. This capacitor was connected to the ends of the R5 resistor and was not shown in the schematics. As you might guess, the capacitor prevents fast voltage changes, and thus makes it impossible to increase the frequency.

Is this capacitor present also on NTSC-board? Someone with such a board could check this out. I removed the capacitor, and now it works. I didn't test the board between these two modifications, but Pasi confirmed that you only need to remove the capacitor.

Pasi:

I first tried to convert my VIC20CR machine, because the clock circuit in it seemed identical to the one a newer NTSC machine uses, except of course the crystal frequency. Some of the pull-up resistors were different, but I didn't think it made any difference. Pull-up resistors vary a lot without any reason anyway. I replaced the video chip and the crystal, but I could not get it to oscillate. I first thought that the 7402 chip in the clock circuit just couldn't keep up and replaced it with a 74LS02 chip. There was no difference. The PAL crystal with a PAL VIC-I (6561) still worked.

I turned my eyes to my third VIC20, which is an older model with all that heat-generating regulator stuff inside. It has almost the same clock circuit as the old NTSC schematic shows. There are three differences:

  1. The crystal is 14.31818 MHz for NTSC, 8.867236 MHz for PAL.
  2. Two NOR gates are used as NOT gates to drop one 74S04 from the design.
  3. In PAL the crystal frequency is divided by two by a 7474 D-flipflop.
I could either put in a 28.63636 MHz crystal or I could use the 14.31818 MHz crystal and bypass the 7474 clock divider. I didn't have a 28 MHz crystal, so I soldered in the 14.31818 MHz crystal and bent the pin 39 (phi1 in) up from the 6560 video chip so that it would not be connected to the divided clock. I then soldered a wire connecting this pin (6560 pin 39) and the 14.31818 MHz clock coming from the 7402 (UB9 pin 10). The machine started working.

I just hadn't any colors. My monitor (Philips CM8833) does not show NTSC colors anyway, but a multistandard TV (all new Philips models at least) shows colors as long as the VIC-I clock is close enough to the NTSC color clock. The oscillator frequency can be fine-adjusted with the trimmer capacitor C35. Just remember that using a metallic unshielded screwdriver is a bad idea because it changes the capacitance of the clock circuit (unless the trimmer is an insulated model). Warming has also its effect on the circuit capacitances so let the machine be on for a while before being satisfied with the adjustment. With a small adjustment I had colors and was done with that machine.

Then I heard from Timo that the CR model has a hidden capacitor on the solder side of the motherboard, probably to filter out upper harmonic frequencies (multiples of 4.433618 MHz). I decided to give the VIC20CR modification another try. I removed the 56 pF capacitor, which was connected in parallel with R5, and the machine still worked fine.

I then replaced the crystal with the 14.31818 MHz crystal and inserted the 6560 video chip. The machine didn't work. I finally found out that it was because I had replaced the original 7402 with 74LS02. When I replaced it with a 74S02, the machine started working. I just could not get the frequency right and thus no colors until I added a 22 pF capacitor in parallel with the capacitor C50 and the trimmer capacitor C48 to increase the adjustment range from the original 5..25 pF to 27..47 pF. The trimmer orientation didn't have any visible effect anymore. I had colors and was satisfied.

To check all possibilities, I also replaced the 74S02 with 7402 that was originally used in the circuit (not the same physical chip because I had butchered it while soldering it out). I didn't need the parallel capacitor anymore, although the trimmer adjustment now needed to be correct or I lost colors.

As I really don't need two NTSC machines, I then converted this machine back to PAL. I made the modifications backwards. I replaced the crystal and video chip and then was stumped because the machine didn't work. I scratched my head for a while but then remembered the extra capacitor I had removed. And surely, the machine started working when I put it back. Obviously, the machine had only worked without the capacitor because it had 74LS02 at the time. 74S02 and 7402 won't work without it. So, if you are doing an NTSC to PAL conversion, you need to add this capacitor.

Summary

PAL VIC20 To NTSC

This is the older, palm-heating VIC20 model with the two-prong power connector and the almost cube power supply.
  1. Replace the 8.867236 MHz crystal with a 14.31818 MHz crystal
  2. If UB9 is 74LS02, replace it with a 7402 (or 74S02 if you only use NTSC)
  3. Bend pin 39 from the 6560 video chip so that it does not go to the socket.
  4. Add a jumper wire from UB9 pin 10 to the video chip pin 39.
  5. Adjust the trimmer capacitor C35 so that your multistandard television shows colors.

PAL VIC20CR To NTSC

  1. Replace the 4.433618 MHz crystal with a 14.31818 MHz crystal
  2. If your machine has a capacitor in parallel with R5, remove it (the parallel capacitor). The values in our machines were 56 pF.
  3. If UB9 is 74LS02, replace it with a 7402 (or 74S02 if you only use NTSC)
  4. If necessary, increase the clock adjustment range by adding a capacitor of 11..22 pF in parallel to C50 (15 pF in the schematics, 5 pF in my machine) and C48 (the trimmer capacitor 0..20 pF). With 7402 you can do with a smaller capacitor or with no additional capacitor at all.
  5. Adjust C48 so that your multistandard television shows colors.

Trouble-shooting