Vectrex Processor Upgrade?

[adamchevy]

In doing some reading about the 68A09 processor that runs at 1.5mhz I came across a 68B09(in the restricted section) and noticed that it runs at 2mhz. Seeing as its not the E version of the 6800 it doesnt need an external clock. I was just wondering if anyone has tried this upgrade? And if so how did it go?

[+save2600]

Be kind of cool if there was a simple upgrade like this that worked. Assuming the CPU is socketed, which I think it is. There's a few games that could stand a little bit of a speedup. Specially some home-brews that have a little bit of slow-down.

[adamchevy]

It looks like they go for $5.00 on ebay, and they are socketed. Thats quite an upgrade for $5.00.

[goldenegg]

I'd be surprised if no one has tried this yet. Either way, I'd love to know the results. I'm not willing to use my system as a Guinea pig though

[GroovyBee]

Looking at this data sheet its pin compatible with the 68A09 but you still need to change its crystal to increase its operating frequency. Without looking at the Vectrex schematic its impossible to say what impact it might have on other external components that derive their clocks from the 6809's E clock signal (output).

[raiderxx]

Looking at this data sheet its pin compatible with the 68A09 but you still need to change its crystal to increase its operating frequency. Without looking at the Vectrex schematic its impossible to say what impact it might have on other external components that derive their clocks from the 6809's E clock signal (output).

 

This site appears to have the Vectrex's schematic. Is that enough to go off of? If so I might be willing to guinea pig my Vectrex. Maybe..

[adamchevy]

I ordered a 68B09P ,so Ill let you know what happens when I put it in. The service manual says the crystal is at 6.0000 Mhz . So if my calculations are correct I need to replace it with an 8mhz crystal to get 2mhz out of my cpu. Would this make the games actually run faster than they are supposed to?

Edited November 25, 2015 by adamchevy

[goldenegg]

Would this make the games actually run faster than they are supposed to?

It can differ from game to game, depending how timing is handled in the code. Similar to how some old computer games run crazy fast on more powerful systems, while others are fine.

[vtk]

hey adamchevy thank you for trying the experiment; if not already on your test list please could you try the game 'Rockaroids' if possible..

 

Edited November 25, 2015 by vtek

[vtk]

ps. i don't know if it's useful at all but here's some interesting reading about the hitachi 6309 cpu (has 6809 'emulation mode')

 

http://fms.komkon.org/comp/CPUs/6309.txt

 

 

[GroovyBee]

This site appears to have the Vectrex's schematic. Is that enough to go off of? If so I might be willing to guinea pig my Vectrex. Maybe..

Its not my project but I spent 5 minutes looking at the schematic and you'll have to change the CPU's crystal to 8MHz, drop C210 and C222 to 18pF, replace the 6522A PIA with a 2MHz part (uses CPU E clock) and the AY3-8912 PSG will play music/FX at higher frequencies (it uses CPU E clock too).

[adamchevy]

Its not my project but I spent 5 minutes looking at the schematic and you'll have to change the CPU's crystal to 8MHz, drop C210 and C222 to 18pF, replace the 6522A PIA with a 2MHz part (uses CPU E clock) and the AY3-8912 PSG will play music/FX at higher frequencies (it uses CPU E clock too).

This is why I asked the question on atariage. Ive never upgraded hardware this old, and I knew there was a lot more to it . I really appreciate any and all help. It looks like Ill be adding to my parts list. Also, if theres a faster cpu that would work that has the same socket I might as well use that one. It looks like there is a hitachi that runs at 40mhz, but that might be pushing it. So far I found R65C22P2 that may solve the 6522A issue. Edited November 25, 2015 by adamchevy

[adamchevy]

Also wouldnt replacing the 6522A with the R65C22P2 solve the sound problem because its at 2mhz?

[GroovyBee]

Also wouldnt replacing the 6522A with the R65C22P2 solve the sound problem because its at 2mhz?

They are two different devices. If you increase the E clock frequency to the AY3-8912 PSG to 2MHz the games will still assume its 1.5MHz, thus making all sounds play a higher pitch. To make music/FX work you'd need hacked ROMs or you need to solve it in h/w by an investigation into clocking the AY3-8912 separately at its original 1.5MHz clock rate. You might need some kind of FIFO arrangement if it uses the E clock signal as an internal address bus latch signal of some kind.

[adamchevy]

They are two different devices. If you increase the E clock frequency to the AY3-8912 PSG to 2MHz the games will still assume its 1.5MHz, thus making all sounds play a higher pitch. To make music/FX work you'd need hacked ROMs or you need to solve it in h/w by an investigation into clocking the AY3-8912 separately at its original 1.5MHz clock rate. You might need some kind of FIFO arrangement if it uses the E clock signal as an internal address bus latch signal of some kind.

Looks like I need to do some serious homework. Should be a good holiday project.

[Richard H.]

Please do it, I REALLY want to see it working

[GroovyBee]

The AY3-8912 is good up to 2.5MHz so it won't mind being clocked at 2MHz. According to the 8912's datasheet, the E signal that supplies its clock will only be used for the noise, tone and envelope generators. The datasheet doesn't show the clock being involved in address/data bus transactions but does show it in the block diagram feeding tone, noise and envelope blocks. From that its probably safe to say that the PSG's clock can be supplied asynchronously at 1.5MHz. A simple 1.5MHz Pierce gate oscillator circuit could be used for that,. Personally I'd make this hardware change last.

The hidden problems will be any software timing loops in the routines that directly control the beams for vector generation.

Some BITD game's mask ROMs might not like the faster bus accesses either. If the mod doesn't work on one game, try others or even more modern homebrews that are likely to use faster flash chips.

[adamchevy]

The AY3-8912 is good up to 2.5MHz so it won't mind being clocked at 2MHz. According to the 8912's datasheet, the E signal that supplies its clock will only be used for the noise, tone and envelope generators. The datasheet doesn't show the clock being involved in address/data bus transactions but does show it in the block diagram feeding tone, noise and envelope blocks. From that its probably safe to say that the PSG's clock can be supplied asynchronously at 1.5MHz. A simple 1.5MHz Pierce gate oscillator circuit could be used for that,. Personally I'd make this hardware change last.

 

The hidden problems will be any software timing loops in the routines that directly control the beams for vector generation.

 

Some BITD game's mask ROMs might not like the faster bus accesses either. If the mod doesn't work on one game, try others or even more modern homebrews that are likely to use faster flash chips.

Thanks! I found the same thing when I was looking at the 8912's datasheet. Ive been looking at some oscillator options and I was thinking I would like to make this a permanent fix. It will be interesting to see how the games react for sure. I found a guy who did this in reverse and made a slower vectrex to make the games easier.

[Rastamafugg]

I know this is a long-dormant topic, but did anyone ever try to replace the 6809 in the Vectrex with a 6309?  I'm going to be opening mine up to apply the buzz kit when it arrives, so I was thinking of doing the swap at the same time, but I was hoping someone has tried it out before and posted their results.

[Malban]

see: https://vectorgaming.proboards.com/thread/2609/tried-6309

and the followup:
https://itwww.hs-pforzheim.de/daten/mitarbeiter/johannsen/vectrex_lab/projects/mc_6309/mc_6309.htm