Closer shots of the chips on the board.

If I'm reading the date code on the PCB correctly it says July 7th 1995. Not as old as I expected it to be. 😀

Also, unsurprisingly, the usual physics-defying Japanese battery technology means that that CR2032 that has the same red dot QC marking as the cap in the left corner is still good and reads over 3.1v after nearly 28 years of service. Seriously, how do they do this!?!? I have other Roland devices with batteries that are still good after 31+ years. And yes, I've checked it over and there are no signs of tarnish or leakage.

Also, here are some randomly placed pictures of the components on the PCB, since I don't believe there are any other pictures of these online anywhere. May be helpful for replacing caps, etc. some day. Not that it's likely to need any work any time soon with Roland's use of quality components.

And here's a shot of the ribbon cables that connect the top cover\controls to the main board.

I don't know if this tells us anything, but there is a ROM chip on the board marked as MB838000CPF 0L6 AA , and I found something similar mentioned in this document:
https://www.nightfallcrew.com/wp-content/uplo … VICE_MANUAL.pdf

It says MB838000CPF-G-0L6 and it says it is an "8M Mask ROM (Wave)". So what does that mean? 8MByte or 8mbit? I'm guessing its 8mbit\1Mbyte. More importantly, are there any other Roland modules from the mid 90s that have just one of these? It looks like the SC55MKII might, but I'm having trouble finding good pictures right now.

Sample ROM is 8Mbits / 1Mbyte which puts it near the SC7 / SCB7 category.

Tiido wrote on 2023-02-03, 12:55:

Sample ROM is 8Mbits / 1Mbyte which puts it near the SC7 / SCB7 category.

Nice, good to know!

I think someone needs to make a chart for Roland synths to show their polyphony, sample ROM size, etc.

Also, from what I have been able to discern, it seems like Roland can get much higher quality samples into 1MB ROM than what you'll find on a cheap 1MB wavetable card, right? Like they use some really good compression, plus the sounds are better from the start?

The main sound chip is same as SC55mk2 (TC6116) so you have 28 channel polyphony. That 1MB holds ADPCM type compressed samples from what I know, which trades space for extra noise but in a big mix with effects on top it won't be audible so it is a good tradeoff. Overall there simply is more care in those samples compared to cheap solutions which goes a very long way, you don't necessarily need big samples to have good sound *coughmultihundredmbytesoundfontsthatfailtoimpresscough* 🤣

Tiido wrote on 2023-02-03, 17:28:

The main sound chip is same as SC55mk2 (TC6116) so you have 28 channel polyphony. That 1MB holds ADPCM type compressed samples from what I know, which trades space for extra noise but in a big mix with effects on top it won't be audible so it is a good tradeoff. Overall there simply is more care in those samples compared to cheap solutions which goes a very long way, you don't necessarily need big samples to have good sound *coughmultihundredmbytesoundfontsthatfailtoimpresscough* 🤣

The manual says it has 24 note polyphony, which is more in line with the original SC55 rather than the MKII or the SC7... do you know why there'd be a difference?

And that's pretty much what I thought regarding the sample ROM size. It is amusing that there are these MASSIVE sound fonts out there, and yet one of these ancient Roland synths with relatively tiny sample ROMs can produce excellent sounding music when used properly. It's kind of like hooking up a $50,000 guitar to a basic amp setup and putting it in the hands of someone who can't play (like me)... no matter how good the guitar itself sounds, it will be limited by the rest of the gear and the musician playing it. 😁

The board you got definitely has the later TC6116 sound chip which has 28 note polyphony capability, earler R15239148 sound chip found in SCC1 and SC55mk1 indeed does have 24 note polyphony.

Tiido wrote on 2023-02-04, 01:18:

The board you got definitely has the later TC6116 sound chip which has 28 note polyphony capability, earler R15239148 sound chip found in SCC1 and SC55mk1 indeed does have 24 note polyphony.

Cool! So this should basically perform like an SC55MKII with a smaller sample ROM? Or, put another way, like an SC-7 with GS?

I am unsure if there's room for GS in than 1MB ROM but perhaps some subset. One will have to see what it reacts to etc. (I don't know any details for that)
It will definitely be adequate for most games, GS bits are irrelevant to majority of them.

Tiido wrote on 2023-02-04, 01:32:

I am unsure if there's room for GS in than 1MB ROM but perhaps some subset. One will have to see what it reacts to etc. (I don't know any details for that)
It will definitely be adequate for most games, GS bits are irrelevant to majority of them.

I just mentioned GS because it says it is GM and GS compatible. I guess it never occurred to me that they could label it as such without putting all of the GS samples in there too.

But, as mentioned earlier in the thread, I really don't understand the practical difference between GS and GM anyway outside of MIDI production.

While we're on that subject... for retro gaming, what would be the difference between starting this module in GS mode vs putting it in GM mode using a sysex message? Would it only matter for specific music tracks that were created for GS? Or does GS mode make a GM track not sound right?

GS adds variation instruments etc. that are accessed via banking mechanism that GM already has, and standard way to access and alter effects. GS is supposed to work as GM until those extra features are accessed, and if a GM device is being fed GS specific data it just gets ignored and you're subject to potentially very wrong sounds. There are some games that use those GS variations for SFX etc. but I know no further details.

Majority of games do GM and will work just fine on a GS device (as long as no special confs are entered prior, either from user interface or by the music data itself, so a reset may be necessary sometimes).

Tiido wrote on 2023-02-04, 04:17:

GS adds variation instruments etc. that are accessed via banking mechanism that GM already has, and standard way to access and alter effects. GS is supposed to work as GM until those extra features are accessed, and if a GM device is being fed GS specific data it just gets ignored and you're subject to potentially very wrong sounds. There are some games that use those GS variations for SFX etc. but I know no further details.

Majority of games do GM and will work just fine on a GS device (as long as no special confs are entered prior, either from user interface or by the music data itself, so a reset may be necessary sometimes).

I see. I guess that makes sense then because many SC55 modules are just labeled as GS, and they seem to work fine overall for games. I think there's an old thread around here regarding that actually.

Anyway, what that means is that the RA-30 should operate totally fine as a sound module for games and MIDI tracks and all the user has to do is hold "Other" while powering it on to put it into sound module mode. Hopefully I will get some time to compare it to some other modules eventually.

Thank you again for answering all these questions Tiido. 😀

It's hard to tell from a quick search exactly where the RA-30 lies, but according to the manual, it has the tone set of the SC-50 (basically, the SC-55mkII set minus the MT-32 tones, 354-128 = 226).

To my knowledge, Roland used delta compression (2:1), so 8Mbit = 1MB chip = 2MB of samples. The SC-55 and the SC-55mkII used 3 x 1MB chips = 6MB samples. The in between SC-50 had 2 x 1MB = 4MB of samples, with the same specs as the RA-30 in terms of patches.

Obviously, there is a discrepancy between 1 wave rom chip that was mentioned earlier and 2. I think there's in fact 2, as can be seen in this picture:
download/file.php?id=156456&mode=view

The chips are the HN6244... ones. I'm guessing these are the 2x8Mbit wave roms.

But, as mentioned earlier in the thread, I really don't understand the practical difference between GS and GM anyway outside of MIDI production.

Basically you already get it - GS isn't relevant for gaming, for *the most part*. For Sierra games as an example, none use GS instruments.. but many games drum parts are tailored for the GS spec, even when using the default drums. So presumably some drum notes won't sound correct on a non-GS GM device (if that makes sense).
Very occasionally a game uses GS drumkits themselves, but I'm not sure if the game themselves actually pass that data on.

Bottom line, you won't ever need to put the unit in a specific mode, either should be fine for gaming.

As for the whole 'someone should catalogue this stuff' point, you're absolutely right, and why a year or so ago I started the "Sound Canvas Comparison Project" here on Vogons. I haven't had the time to devote to it I'd like to, so it's stalled for now, but there's already some information uncovered in the thread, just not the recordings I wanted sorted for listening purposes. Maybe one day I'll get it going again.. 😀

Spikey wrote on 2023-02-06, 04:15:

It's hard to tell from a quick search exactly where the RA-30 lies, but according to the manual, it has the tone set of the SC-5 […]
Show full quote

It's hard to tell from a quick search exactly where the RA-30 lies, but according to the manual, it has the tone set of the SC-50 (basically, the SC-55mkII set minus the MT-32 tones, 354-128 = 226).

To my knowledge, Roland used delta compression (2:1), so 8Mbit = 1MB chip = 2MB of samples. The SC-55 and the SC-55mkII used 3 x 1MB chips = 6MB samples. The in between SC-50 had 2 x 1MB = 4MB of samples, with the same specs as the RA-30 in terms of patches.

Obviously, there is a discrepancy between 1 wave rom chip that was mentioned earlier and 2. I think there's in fact 2, as can be seen in this picture:
download/file.php?id=156456&mode=view

The chips are the HN6244... ones. I'm guessing these are the 2x8Mbit wave roms.

It would be neat if it actually had 2 x 1MB sample ROMs, but those are 256K SRAM chips:
https://pdf1.alldatasheet.com/datasheet-pdf/v … 256BLFP-7T.html

And the fact that they connect directly to a non-MIDI related, non-Roland microcontroller makes me think that they are just used for running the arranger part of the RA30.

There is also a similar 256K SRAM chip connected to the main Roland chip.
https://datasheetspdf.com/pdf-file/660189/Hit … 65256BLFP-12T/1

There appears to only be one ROM in there, aside from the large DIP chips on the right side, which are probably the firmware itself and maybe the data for the preset backing rhythms.

Unless Roland is known for doing double sided PCBs with ROMs on both sides, then I don't think this could have anything but a 1MB ROM. Just checked it, there's nothing on the back side of the PCB. Thankfully, it still sounds fantastic. 😁

EDIT: Also, if you want to add this data to your Google Doc at some point, it appears that the RA30 uses an NEC UPD6379 DAC and an NEC UPC4750 OpAmp.

the socketed ROMs are program code + data for the CPU, along with the two SRAMs. PSRAM next to sample ROM is effect memory.

Got my hands on the service manual for this unit. Better scans will be made tomorrow.

https://imgur.com/a/u2TVV4k

Thanks so much! Really appreciate it you scanning it, it's hard to find service notes for the unusual Roland MT/SC variants.