[Tutorial] Making your own PS1 Modchips

Mayumi 2.1 for pu7
Stealth 2.8a for pu8

 

Ok I'm about to give up on this stupid K150. What is the most reliable way to flash 12C508As? I have a bunch of them so I don't want to get something that is incompatible. Has anyone tried a PICkit?

 

Ok got it to work using a Mayumi v4 hex. I'm guessing the issue was with the Oscal value, since MM3 uses it and Mayumi v4 doesn't. I don't know what I was doing wrong, but I also have no idea which driver and version of Microbrn I'm supposed to be using with this k150, and the seller won't message me back.

Is there any reason to use MM3 over Mayumi v4? I think I'm just going to stick with Mayumi unless there is a compelling reason to use something else and revisit this if I end up owning an early model PS1 and need stealth 2.8 or MM2.

 

Mayumi v4 is fine, I supply that and not mm3.

 

so whats stated in the first post is fine but use Mayumi 4 instead of MM3?

Anyone have links to all the files?

 

I'm no expert on this, but it's my understanding that Mayumi v4 is more reliable since it uses the timing from the board rather than the internal oscillation of the chip, which makes it more in sync with the PS1. I have no idea if I explained that correctly, but the general consensus seems to be that Mayumi v4 is just better than MM3.

Here is the pack I downloaded, but I can only vouch for the Mayumi hex because that's the only one I got to work:

https://mega.nz/#!0sU10AgL!vv9U4mhn4_zjzZBvml0d8d90XFpBg_P4byPAonW2BEk


Here's another resource:

http://www.eurasia.nu/modules.php?name=Downloads&d_op=viewdownload&cid=43

 

Honestly, it shouldn't make any difference at all - the code inside the HC05 on the PS1 mainboard resets the timing for each character, so the data rate would have to be off by at least 11% to corrupt the SCEx strings. This is a clock error that's outside the worst-case spec for the chip even at supply voltage and temperature extremes (I.E. 2.5V VDD and 125 degrees C). I personally suspect that the idea there was a problem using the RC oscillator came from people that were using cheap third-party programmers that didn't preserve the factory oscillator calibration - and in this case, using the external oscillator connection will fix things because the corrupted calibration value is no longer being used.

Personally, I don't like using the external oscillator connection for a couple of reasons:
1) The external clock is derived from the audio master oscillator - it's 96 times the CD sampling frequency, so it comes out at 4.2236Mhz. Since the PIC has a specified clock speed of 4MHz, this represents a 5% overclock. Sure, these chips have plenty of margin, and should have no problem working, but it's still bad practice.
2) Connecting up the clock requires you to run a wire with ~4MHz on it around an area of the board that has some very low level signals on it, and at that sort of frequency there is significant coupling of the signal even just using parasitic capacitance. An interesting experiment is to solder a wire to that clock and then route it near the RF amp chip on the board - the other end doesn't even have to be connected to anything - then monitor the eye waveform test point while playing a disc. You can see the noise pickup on the scope. For best (or worst) results, do this with a PU-18, since the long ribbon cable gives you reduced signal anyway.

I'm not suggesting that the Mayumi code is bad, because it isn't - but the idea that it's "just better" is not as clear cut as you seem to think.

 

Wow, awesome info. I searched the web for a few days and the best explanation of the difference I could find was what I said earlier. I've also read claims of the MM3 chip failing to boot at times while playing imports while the Mayumi v4 chip didn't, but I don't know exactly how true those are. Whatever the case, I'm going to have to try them both out now.

 

I would assume that as long at the clock is correct the boot behavior of the MM3 should be identical to the Mayumi 4.0 - the stealth mode code between the two chips is extremely similar. So similar in fact that I'm pretty sure the MM3 was copied from the Mayumi code (that way around because the Mayumi V4.0 shows considerable similarities to the V2 code, which strongly suggests the Mayumi V4 was the original).