74HC02 TTL CMOS logic IC obtainable anywhere.

Cheers,

pentiumspeed wrote on 2021-09-13, 23:21:

74HC02 TTL CMOS logic IC obtainable anywhere.

Cheers,

Ah good to know its easily obtainable ,
Still I don't know how involved adding the extra CM32L PCM ROM chip would be ,
I think you can have two roms linked together right?
Some bios are split into 2 chips and whatnot?

So like has anyone tried doing this to their RA-50?

I may have had too much beer, but it might be possible to piggy back the ROMs on top of each other, but lift pin 22 (!CE) on the top one, and connect it via an inverter (something like this: https://www.farnell.com/datasheets/158131.pdf ) with an input from pin 22 on the lower one. Pretty sure that's what the circuit diagram you posted at the start is doing (inverting inputs of an AND gate tied together makes it work like a NOT gate), making the RA19 output from the LA32 (pin 59) act as a select switch between the two ROMs.

But I've never tried doing this, so it might be a bad idea.

Sorry to bump an old thread, but has anyone attempted this? Won an RA-50 and this would be a nice mod to do (alongside the MIDI bypass).

You also need the CPU program ROM of CM32L to actually have the extra sounds usable, the ROM already there will not know anything about the extra sounds.

You can also use a 27C801 or some other 1MByte EPROM and a wire from A19 to the EPROM A19 to get the full sample set without needing to split stuff up between existing and newly added ROM.

Ah 0kay- I assumed it was the same CM32L ROM already in the RA-50 as it is labelled as such.

I'm still a bit lost as to what needs to be done (electronics isn't my strong suit!)

Replace CM-32L ROM on board with 'actual' CM-32L ROM
Flash CM-32L PCM ROM onto 27C801 chip.
Piggyback burned PCM ROM chip on top of TC534000P-7476 chip, lifting pin A19 and connect this to pin 59 of the LA32 chip?

I attach a picture of the RA-50 board as I can update it as I can annotate it for others once we figure it out 😀

EDIT: I'm an idiot and just saw snufkin's comment above *facepalm*. So you'd also need an 74HC02 to invert A19 so it's not selected at the same time as the original ROM?

You need the inverter chip only if you are going to use two 512KB EPROMs. In that case original A19 goes into !CE line of original (as it does already) and inverted A19 goes into !CE of the new ROM, all other pins are parallel. All other unused sections of the inverter chip have to have all the inputs signals connected to something, they cannot remain floating except maybe if you use a TTL and not CMOS type part but even then, it is always polite to connect any unused inputs to something (nearby GND, VCC, output of another section etc.).

In case of using single 1MByte EPROM you remove the original sample ROM entirely, put in the new one and connect A19 straight to A19 of the EPROM, while !CE line gets separated from A19 line and connected to ground. I personally would go with this option, since it can be done cleaner.

I have actually missed there was a CM32L program ROM in the machine already, in that case you don't need to do anything more about the program ROM.

Ah, that makes more sense! Thank you for the explanation.

I agree that it'd be much nicer to remove the ROM and install a socket + UVEPROM + A19 mod, although I'd need to invest in a good vacuum desoldering gun to do so.

So, I gave this a go using the piggybacked chip method, but sadly to no success.

Got a M27C4001 PROM and burned the upper half of the CM32L PCM ROM to it.

Got a SN74HC02N and wired it as follows:

Pin 2 of the SN74HC02N was soldered to the CE line on the original ROM, and pin 1 to the CE pin on the new ROM.

Powered on and no magic smoke. Tested standard LA voices and they were OK.

Tested channel 10 and played the upper notes, which I believe is where the extra CM32L sound effects notes are located. But no sound.

Any ideas? Have I made any schoolboy errors? I assume that pulling pin 3 to low was correct, as otherwise I'd probably get voice corruption if pulling it high and it being active as the same time as the original chip.

Admittedly the chip was piggybacked without soldering, but all pins had a connection.

Maybe there's another connection needed somewhere to 'tell' the operating ROM that both PCM ROMs are present?

Actually I might've been a bit hasty in my analysis.tried Beneath a Steel Sky and its definitely doing something. I need something to reliably test it.

So I committed to it and soldered the chip in place, just in case there were bad connections.

Unfortunately it seems that it still isn’t working. Comparing to Phil’s computer lab video on the MT32 and CM32, I’m still getting the MT32 sound effects for beneath a steel sky.

Not sure what’s next for this project really! I may leave it as is.

I’ve been poring over the data sheets and I think I’ve spotted my mistake. I’ve not connected VPP, as I assumed that it was only required for programming and not when in use… unfortunately I removed the leg, thinking it wouldn’t be needed!

I’ve just ordered a dremel to grind away some of the chip and I’ll connect it to VCC.

The logic as you connected it should work, and the VPP pin shouldn't really matter although the datasheets do tell it to connect to a specific state for normal operation and I suppose some EPROMs do need it not to float.

Game doesn't know what it is connected to so it think it needs to be specifically told that. I would think it is strange if a game uses features that aren't guaranteed to be present...
(unless there are identification commands that the MIDI device returns via its MIDI OUT that the game can listen to ? In such a case you need to mod the MIDI OUT signal path on the device also not just MIDI IN)

Sorry for the late reply but I’ve been ill.

Finally soldered VPP to VCC and I think it fixed it!

I’ve tried with Leisure Suit Larry and believe that it does in fact appear to be working! As per another thread here, the song ‘Salsa This’ uses the new laughing sample and I can hear it. Other recommended tests have been inconsistent however. Very odd.

But the laughing sample is definitely new to the additional PCM ROM.

Something isn’t quite right however. Having read the CM-32L manual I should be able to play the extra sound effects from key e5 onwards on the rhythm channel (10). Maybe this is being blocked somewhere?

I should really dump the MIDI stream for LSL to see how it’s doing it.

Well, I'm still awake due to the effects of this illness causing me to struggle to sleep... yay. Anyhow, I have at least worked out *why* the sounds aren't appearing.

They aren't being mapped!

Not sure why this is, as I'm pretty sure all RA-50s came with CM-32L ROMs - earlier ones had socketed labelled chips (as per the board pic above) and later ones unlabeled mask ROMs which also correspond to CM-32L units.

By using sfryers excellent MT-32 editor app, I was able to manually re-map the sounds, and after that all of them played as expected. I was also able to use the app to create a SysEx file, so this can be done easily.

Very odd behavior however - I wonder if there is a pin that has to be pulled hi/low for the firmware to 'detect' if the device is a CM-32L or a later model MT-32, as they are very similar (at least that's my understanding)

So VPP is important for this particular EPROM.

The rest is strange though, I would have expected things to just work. LAPC-I program ROM is worth to try too, it has same general hardware.

My board has the R15209214 masked ROM on it. Just had a quick google and people have verified seeing this on both the LAPC-I and CM-32L, so I'm pretty sure that's OK.

Had another look over the schematics for the CM64 and the later rev MT32 to see if there was anything obvious, but it seems not.

How odd!

I tested the SysEx file I created and Elite Plus now as the takeoff sound effects, so that works.

Honestly, it's really not the end of the world, I'm just curious as to why it's happening.

EDIT: I just realised that all of these devices have a 256K SRAM, perhaps this is where the tone data is stored.

That 256Kbit / 32KByte SRAM is the working memory of the CPU, nothing out the ordinary here since all LA32 hardware except the first generation MT32 have this much in this same way.
EDIT: But there is room for device specific conf, I see some wire links next to the gate array chip : R68A, R69A and R70A, they may select conf...
EDIT2: This is likely it, on other LA32 hardware the pins these connect to nothing at all (pins 6, 7 and 8 ). So it is worth doing the same thing here and perhaps then full compatibility is restored. It should be matter of removing everything from R68A/B, R69A/B and R70A/B positions.