exxos's Atari, Amiga & retro forum

Well the replacement ATTinys arrived today and I had a quick go with one and it programs just fine. I can feed it my I2C clock program and it boots just fine and it also programs straight back with an LED blink test to prove it's all working.

The original one still doesn't work but I'll come back to that later. I was hoping I'd do more today but I came off my bike this morning and have been feeling battered and bruised, so not in the mood for doing much, but I really needed to know the new one worked as expected.

Fingers crossed I can finally get the next step done soon, which is wire the ATTiny to the clock generator so that the ATTiny injects the code into it on boot. That will hopefully do away with the RPi I'm currently configuring it with.

The two devices can then be used standalone with the 800 and powered from it and then I hope to fully replace the crystal.

Oh! I hope you're not too bruised!

Fingers crossed on getting the programmable clock replacement done. It may be a useful project to make into a generic clock generator in future.

Might be a nice little project to make this in to a single PCB affair?

stephen_usher wrote: 13 Jan 2026 21:51 Oh! I hope you're not too bruised!

Thanks. Still feeling sorry for myself, but more worried about what I'll be like tomorrow. I have some cuts on my right foor from the pedal and a huge graze on my left knee, and other than being stiff across the shoulders and rib cage I'm mostly good. I can certainly advocate strongly for the use of bike helmets, had I not been wearing one I think I'd have needed a trip to the hospital. Money well spent IMHO and I'm shocked by how many cyclists I see without helmets.

stephen_usher wrote: 13 Jan 2026 21:51 Fingers crossed on getting the programmable clock replacement done. It may be a useful project to make into a generic clock generator in future.

I really don't think it'll take long at all now, and with an ATTiny board plus the clockgen I see it as an easy way of replacing some crystals and the boards I have offer 3 clock signals too and it should all be programmable over USB, all you need is the correct register settings for the Si5351, and you can have all 3 clocks running as you want.

PhilC wrote: 13 Jan 2026 21:55 Might be a nice little project to make this in to a single PCB affair?

I was thinking that. Prototype first though :)

PhilC wrote: 13 Jan 2026 21:55 Might be a nice little project to make this in to a single PCB affair?

I was thinking that. There are quite a few times that specific, and unusual, clock frequencies are needed where the original crystal is basically no-longer available.

stephen_usher wrote: 13 Jan 2026 22:01

PhilC wrote: 13 Jan 2026 21:55 Might be a nice little project to make this in to a single PCB affair?

I was thinking that. There are quite a few times that specific, and unusual, clock frequencies are needed where the original crystal is basically no-longer available.

Maybe even a switch or two to change frequency?

I could certainly use something like that when I try and update my Atom to 2 MHz & 4 Mhz but still want it to be 1 Mhz compatible without too much hassle.

PhilC wrote: 13 Jan 2026 22:20

stephen_usher wrote: 13 Jan 2026 22:01

I was thinking that. There are quite a few times that specific, and unusual, clock frequencies are needed where the original crystal is basically no-longer available.

Maybe even a switch or two to change frequency?

I could certainly use something like that when I try and update my Atom to 2 MHz & 4 Mhz but still want it to be 1 Mhz compatible without too much hassle.

OK, so if you wanted to change frequency you'd have to use a switch to switch from the clock 0 to clock 1 output, it's not possible to use a switch to just get the SI5351 to change frequency, any frequency would require the registers to be changed.

So to do that you'd need to program both clocks at the same time and just use the switch to swap outputs. There are 3 clocks all together so it would be easy enough to figure out. @PhilC please shout if you want any help around this.

However I do have good news!!!

I hooked up the ATTiny to the Si5351 and powered the pair up via the USB socket on the ATTiny board, and it works just fine:


In other words, the ATTiny is sending the config to the Si5351 and the clock is starting as expected:


The only sort of issue is that having timed the startup of the ATTiny through to when the clock signal appears on the output, it's 6 seconds. Thats a long time in 8 bit startup world. I'll see if the ATTiny code can be made to go faster, but it may not.

So I hooked it up to the Naked 800 but still powered over USB:


And I powered on the moment the clock appeared on the scope. It booted perfectly :)


Next up I figured I need to try the whole thing powered not from USB, but from the Atari 800 itself. It seems that despite a 6 second delay, the machine boots just fine as soon as the clock comes online :)


Obviously I'm pretty happy with the outcome of this, and despite the delay the machine just starts up fine. Of course it may be possible to use some method to keep the SI5351 powered, a bit like a CMOS battery on a PC, so I may investigate that.

Next up some way of connecting the boards together on a PCB would be good to get rid of all the wires, but I have to say the results of this experiment have been far better than I envisaged when I started.

One other thing I'll probably try is to switch the frequency to 3.579575 Mhz and fit an NTSC CPU board I have spare (Very different to the PAL CPU board), just to potentially verify if the machine will run as expected as a US spec but with an Si5351 generated clock.

I just need to generate the register codes using Clockbuilder Pro and we're off to the races.

I guess that if it had been a problem, i.e. the CPU needing a certain number of clock cycles before it came out of reset as the 486 requires, then you could program the ATTiny85 to hold reset low.

The delay you’re seeing is from the USB bootloader code. Someone who tried to use those boards with my vDrive sound code had the same issue. The original bare chips start instantly.

stephen_usher wrote: 17 Jan 2026 16:42 I guess that if it had been a problem, i.e. the CPU needing a certain number of clock cycles before it came out of reset as the 486 requires, then you could program the ATTiny85 to hold reset low.

The delay you’re seeing is from the USB bootloader code. Someone who tried to use those boards with my vDrive sound code had the same issue. The original bare chips start instantly.

Yes, it's the Micronucleus bootloader this specific board has, which contains a 6 second delay. I've done some digging and there are versions with much lower or zero delays, so I'm going to look into that, probably tomorrow now.

Clearly the 800 doesn't care about the wait, it just boots when clock is available, and if I can get the bootloader replaced easily (Should be no issue TBH) then that problem will just go away.