Some UK suppliers of the Superboard II used to advertise that they were selling boards that had been modified for 50Hz video refresh. (Not to be confused with the UK101 which was already 50Hz)
Here are a couple of Computing Today adverts from different suppliers advertising 50Hz Superboards:
Does anyone have any details of that mod or even have one of those boards?
Superboard II 50Hz mod
-
IanB
- Posts: 13
- Joined: Sat Jan 11, 2020 4:52 pm
Re: Superboard II 50Hz mod
Well I've managed to obtain such a 50Hz superboard as one turned up on ebay recently (strange that it should appear so soon after asking the question, especially as I've never noticed one like it before):
Closeup of converter board:
Here's what the output looks like:
For comparison here's 60Hz mode:
It outputs 306 lines compared to the standard 256 lines as this source summary from the RGBtoHDMI converter shows:
The two trimpots adjust the vertical position of the screen and the amount of blanking at the top of the screen.
The blanking can be seen in this power up screen:
When the pots are adjusted incorrectly it becomes clearer how the board works:
It works by displaying the first six or so text lines then resetting the counters so that the display starts over again resulting in an extra 50 video lines.
The reset puts a glitch into the first line of the restarted screen but this line is never viewable on 60Hz machines anyway because it is generated during the actual field sync pulse and in any case the next 2-3 lines are usually off the top of the screen on 60Hz machines leaving the typical usable 24 lines.
The blanking pot can be adjusted to hide as many lines as required at the top of the screen and the other pot can be adjusted to centre up the remaining lines.
The chip numbers have been sanded off but just looking at the layout of the board and the relevant signals it is most likely that the two rightmost chips are 74LS163 counters and the one between the two pots is a 74LS123 monostable. The remaining leftmost chip is 14 pin so probably 74 series gates but I will know more once I have traced the circuit and measured some signals.
It requires fairly minimal modifications to the superboard and looks like it should be easy to make switchable.
I will eventually post a schematic once I've traced out the signals.
- 50Hz superboard.jpg (438.04 KiB) Viewed 2509 times
Closeup of converter board:
- 50hz_closeup.jpg (627.39 KiB) Viewed 2509 times
Here's what the output looks like:
- capture3.png (1.3 KiB) Viewed 2509 times
For comparison here's 60Hz mode:
- capture6.png (18.16 KiB) Viewed 2509 times
It outputs 306 lines compared to the standard 256 lines as this source summary from the RGBtoHDMI converter shows:
- capture2.png (3.44 KiB) Viewed 2509 times
The two trimpots adjust the vertical position of the screen and the amount of blanking at the top of the screen.
The blanking can be seen in this power up screen:
- capture1.png (6.03 KiB) Viewed 2509 times
When the pots are adjusted incorrectly it becomes clearer how the board works:
- capture4.png (1.53 KiB) Viewed 2509 times
It works by displaying the first six or so text lines then resetting the counters so that the display starts over again resulting in an extra 50 video lines.
The reset puts a glitch into the first line of the restarted screen but this line is never viewable on 60Hz machines anyway because it is generated during the actual field sync pulse and in any case the next 2-3 lines are usually off the top of the screen on 60Hz machines leaving the typical usable 24 lines.
The blanking pot can be adjusted to hide as many lines as required at the top of the screen and the other pot can be adjusted to centre up the remaining lines.
The chip numbers have been sanded off but just looking at the layout of the board and the relevant signals it is most likely that the two rightmost chips are 74LS163 counters and the one between the two pots is a 74LS123 monostable. The remaining leftmost chip is 14 pin so probably 74 series gates but I will know more once I have traced the circuit and measured some signals.
It requires fairly minimal modifications to the superboard and looks like it should be easy to make switchable.
I will eventually post a schematic once I've traced out the signals.
-
IanB
- Posts: 13
- Joined: Sat Jan 11, 2020 4:52 pm
Re: Superboard II 50Hz mod
I've reverse engineered the 50Hz converter and here is the schematic:
Also as a pdf:
A lot of inputs that should be pulled high have been left floating which is not the recommended way to do it but it does still appear to work. (The PCB had been laid out for some of those inputs to be tied high but the the resistor to do that was not fitted).
It functions as follows:
U4 and U3 are an 8 bit counter which is preloaded with 0b01100111 (103) and that counts up to 256 at line rate (15.7Khz) (the carry out is passed through U1B and used to preload the counter for the next count)
This results in a count of 153 lines which is half a ~50Hz frame of 306 lines so the Q3 output of U3 changes at 100Hz.
This is divided in two by U1A to give 50Hz and inverted 50Hz on it's outputs.
The 50Hz output triggers the U2A monostable and the output of that is connected to the Vsync monostable on the Superboard (Pin 1 of U65).
The U2A monostable also triggers a second monostable U2B to generate a blanking pulse to hide the extra 50 duplicate lines.
The inverted output of the U2B monostable is connected to pin 13 on U56 on the Superboard. There is already a TTL signal being fed into this input which is a delayed version of the 3.9Mhz clock and that is normally used to make the screen characters dotted. This shorts together two TTL outputs which is also not recommended but acts like an AND gate as the low level on either output 'wins' the conflict. This might be an installation error because the blanking would still work if pin 13 was lifted up to isolate it from the board although the characters would then be solid rather than dotted.
The 50Hz inverted output is passed through an RC network (C3, R3, R4) to produce a very short 50Hz negative going pulse that resets the counter chain on the Superboard by driving the preload inputs on pin 9 of U30, U59, U60 & U61.
- superboard_50hz.png (129.64 KiB) Viewed 2296 times
A lot of inputs that should be pulled high have been left floating which is not the recommended way to do it but it does still appear to work. (The PCB had been laid out for some of those inputs to be tied high but the the resistor to do that was not fitted).
It functions as follows:
U4 and U3 are an 8 bit counter which is preloaded with 0b01100111 (103) and that counts up to 256 at line rate (15.7Khz) (the carry out is passed through U1B and used to preload the counter for the next count)
This results in a count of 153 lines which is half a ~50Hz frame of 306 lines so the Q3 output of U3 changes at 100Hz.
This is divided in two by U1A to give 50Hz and inverted 50Hz on it's outputs.
The 50Hz output triggers the U2A monostable and the output of that is connected to the Vsync monostable on the Superboard (Pin 1 of U65).
The U2A monostable also triggers a second monostable U2B to generate a blanking pulse to hide the extra 50 duplicate lines.
The inverted output of the U2B monostable is connected to pin 13 on U56 on the Superboard. There is already a TTL signal being fed into this input which is a delayed version of the 3.9Mhz clock and that is normally used to make the screen characters dotted. This shorts together two TTL outputs which is also not recommended but acts like an AND gate as the low level on either output 'wins' the conflict. This might be an installation error because the blanking would still work if pin 13 was lifted up to isolate it from the board although the characters would then be solid rather than dotted.
The 50Hz inverted output is passed through an RC network (C3, R3, R4) to produce a very short 50Hz negative going pulse that resets the counter chain on the Superboard by driving the preload inputs on pin 9 of U30, U59, U60 & U61.
-
bxdanny
- Posts: 336
- Joined: Thu Apr 16, 2015 2:27 pm
- Location: Bronx, NY USA
Re: Superboard II 50Hz mod
Could the same general idea be used simply to make a 540 board, or a slightly-speeded-up 600, conform more closely to the NTSC specification? OSI video circuits normally have a horizontal-to-vertical frequency ratio of 256, while real NTSC video has 262.5. Most popular home computers used a ratio of 262 even, I think. So keeping the counters that access video memory reset for six scan lines after the start of vertical sync would increase that ratio to 262, and allow more of the top two rows of characters to be visible.
Keeping the counters reset until the vertical sync pulse was done would make the ratio too high rather than too low, but not by all that much, as I think vertical sync lasts about 12 scan lines, meaning there would then be a total of about 268 scan lines per field, allowing the first two character rows to be fully visible.
Does this make sense? Has anyone tried something along these lines?
Keeping the counters reset until the vertical sync pulse was done would make the ratio too high rather than too low, but not by all that much, as I think vertical sync lasts about 12 scan lines, meaning there would then be a total of about 268 scan lines per field, allowing the first two character rows to be fully visible.
Does this make sense? Has anyone tried something along these lines?
No current OSI hardware
Former programmer for Dwo Quong Fok Lok Sow and Orion Software Associates
Former owner of C1P MF (original version) and C2-8P DF (502-based)
Former programmer for Dwo Quong Fok Lok Sow and Orion Software Associates
Former owner of C1P MF (original version) and C2-8P DF (502-based)