Jump to content

MT-640 Questions


JEFFV145

Recommended Posts

I'm new to keyboards- the only one I've ever owned is an MT-640 which I really like. I'm curious if there is a similar sounding Casio keyboard with full size keys. I've had the MT-640 since I was very young and I really like the sounds, it's very dear to me but is in rough shape (missing keys and all). Right now I'm looking for an upgrade and curious what you think is best. I've used the MT-640 on a couple recordings, most recently my Christmas album, which was done on a Tascam 414 (link) https://fountaindew.bandcamp.com/album/christmas-in-the-fridge See attached photo. Thanks!

s-l1600.jpg

  • Like 2
Link to comment
Share on other sites

I have an MT640 too. It's an interesting keyboard, a little gimicky in some respects. I have not found a use for it in my recordings yet, certainly not compared with my other Casios.

 

As for an upgrade, it's a different (later) sound engine, but look for a CTK1000. They're rare, however, they have full size keys with velocity sensitivity. Lots of interesting sounds (with limited editable parameters) and effects, the usual drums and rhythms as well.

 

Otherwise, look for any of the full size later CT, and 90's CTK models to give you a similar experience to the 640.IMG_20180208_130131.thumb.jpg.4257154da17ac9d97dfcc5139a1c01a1.jpgIMG_20180208_130145.thumb.jpg.71e4aed42fc4668fb38b761fe03e1930.jpg

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

  • 3 months later...

Any of the 1989 or so full keys Casio keyboards will give you the very same sounds (you can easily spot them because they all look similar). The most common 1989 model seems to be the MT-240, but any of the "40" keyboards from that time (or the MT-260) will have the same timbres with one or two differentials. They were very popular back in the day. The "oddity" among them is the DM-100, which is a two keyboard unit with equivalents of a MT-240 and a SK-8 on top (sans rhythms). It's interesting to note that when the 1989 line was announced, it was supposed to be Casio's "MIDI line" - including the DM-100, which eventually was released *without* MIDI. Even promotional material from back then had it as containing MIDI I/O. Did they remove it to cut costs?

Link to comment
Share on other sites

Maybe they couldn't figure a way of getting MIDI to both keyboards? From what I understand, the DM100 was pretty much two separate models cobbled together in one unit. Perhaps it was either too complicated and/ or too costly to incorporate a midi controller that could operate both keyboards?

Link to comment
Share on other sites

The Casio DM-100 contains only one half of an SK-8, i.e. it lacks the (rom sample) percussion IC. While the main CPU (MT-240) of the lower manual can handle midi, the SK-8 never was designed to support it, so it would have needed a complicated additional microcontroller to handle both keyboards, which would have significantly increased cost.

Yamaha VSS-200 is another such mashup cucumber; it contains the FM keyboard hardware of PSS-270 combined with the sample keyboard VSS-30 (some buttons omitted), which resamples the FM of the PSS-270 to treat it with effects. But the wiring only assignes both CPUs to different keyboard split sections, not dual manual.


ROM hacking?

Casio MT-240, MT-540, MT-640 and their many fullsize variants (MT-640 has none) are all built around the CPU "NEC D938GD" and only have different external roms.

The MT-540 contains 2 16bit ROMs with all important pins wired parallel except A17. At the HN62404P also the BYTE pin 31 (the other ROM lacks it) is wired to +5V to prevent 8bit mode. Despite both pinouts are like an eprom 27C400, the HN62404P contains here a little trick that makes it appear empty (all #FF) when read in an eprommer, namely the OE/ pin is inverted and thus needs +5V instead of 0V to output data. (In an adapter I bent it away and connected it with Vdd to do this.) Likely this was done to simplify multiplexing by toggling between both roms through the OE/ line (but it also might have been a stupid copy protection attempt). With HN62404P removed, the MT-540 still boots properly, but lacks the piano and effect sounds (there are also some popping glitch noises), which shows that the operating system and most samples are in the smaller 2nd ROM D23C2000AC-1 (without that it can not run at all). I later verified this in MT-540 service manual. I compared a Casio CSM-1 rom dump with my MT-540 roms and they are identical.

In opposite to this, the Casio MT-240 contains a "HN62404P D27" as its only ROM (can not run when removed), and the OE/ pin of this one behaves normal (not inverted). The same ROM was also used for the lower keyboard of Casio DM-100 and in Hohner PSK50 (seen in service manual). Hohner PSK55 (seen in schematics) has a "HN62404P H84" instead. (According to parts list in service manuals, the software number seems to be the lower 3 digits.)

Although I haven't done this, I think it may be possible to upgrade the MT-240 (and related keyboards) with MT-540 sounds. For this you would need to copy both MT-540 ROMs on each a 27C400 eprom. Desolder the MT-240 ROM and install a 40 pin DIL IC sockets in its place and another one in the 2 unlabelled hole rows next to it. Install 2 short wire bridges in the empty hole pairs between both sockets and above the empty one. A little complicated is that you need to cut the OE/ line to pin 12 of the HN62404P socket and invert it (using a CMOS inverter IC or transistor). Connect the inverted OE/ to pin 12 of the HN62404P socket and the uninverted OE/ to pin 12 of the unlabelled socket. The eprom with the contents of the MT-540 HN62404P (software number G93) is inserted instead of the old HN62404P, and the eprom with D23C2000AC-1 content into the unlabelled socket (both facing to the same direction, i.e. notch facing at PCB rim).

And of course you need to add the missing effect pads and switches to the keyboard matrix.

If you want to keep also the old behaviour, you may additionally solder the original HN62404P piggyback onto the new eprom in its place (wire its /OE line uninverted). Disconnect the /CE input (pin 10) and install a switch that connects either the upper (old) or the 2 lower (new) ROMs with the /CE output of the CPU. Add a pullup resistor (about 10 kOhm?) against +5V at both banks to disable the unused bank. (Flip the switch with the keyboard powered off, else it may crash, which may be suited for shitshot.)

A mechanically simpler variant to avoid the piggyback soldering would be to use a larger capacity eprom (27C800?) with 42 pins (lower pins are the same pinout) and burn additionally the content of the old ROM into the upper half. Connect the switch output at the A18 address line to switch between both banks. It may even be possible to use only a single eprom with enough capacity (27C160) by burning the D23C2000AC-1 contents into an additional bank and connecting /OE only(?) to A19 to simulate the multiplexing. But this is all theory; I haven't done any of these and can not promise that it works.

Many of these mods may also function with Casio MT-640. According to its manual it e.g. has 10 additional preset sounds through MIDI, those likely also exist as matrix eastereggs. Its 2 ROMs are "NEC D23C2000AC-1 503" and "HN62404P Z29", those might be also addable as a bank switch mod fot MT-240.

I haven't tested this, so I don't know if e.g. the different software numbers of the D938GD CPUs may prevent it.


  • Like 1
Link to comment
Share on other sites

  • 1 year later...

Yes, I dumped these roms and a couple of others. But I haven't analyzed the file format.

The MT-540 ROM contains no complete sounds, but only many small fragments of those an algorithm puts the sounds together (wavetable synthesis). With the program "Wavosaur" I could import the samples and play them. The proper import settings are 16 bit, little endian, unsigned. Pitch depends on the played note (16000Hz sounds recognizeable). But these are not only plain samples; there are also quaterwaves, envelope curves (lookup tables) and program code sections, although fortunately at least the samples are not compressed.

As far I remember, there may be some clipping artifacts in the waveforms. The Casio algorithm treat a direct change from the highest to the lowest value (which doesn't happen in properly sampled audio) as a signal to add or subtract a highest bit, so the genuine bit resolution is higher than the 16 bits you see.

Link to comment
Share on other sites

On 1/30/2020 at 3:16 AM, CYBERYOGI =CO=Windler said:

Yes, I dumped these roms and a couple of others. But I haven't analyzed the file format.

The MT-540 ROM contains no complete sounds, but only many small fragments of those an algorithm puts the sounds together (wavetable synthesis). With the program "Wavosaur" I could import the samples and play them. The proper import settings are 16 bit, little endian, unsigned. Pitch depends on the played note (16000Hz sounds recognizeable). But these are not only plain samples; there are also quaterwaves, envelope curves (lookup tables) and program code sections, although fortunately at least the samples are not compressed.

As far I remember, there may be some clipping artifacts in the waveforms. The Casio algorithm treat a direct change from the highest to the lowest value (which doesn't happen in properly sampled audio) as a signal to add or subtract a highest bit, so the genuine bit resolution is higher than the 16 bits you see.

Sounds like a proprietary companding scheme. One would expect other data besides audio in there too. I have a friend who has figured those schemes out in the past (Alesis, Yamaha and Casio drum machines), with other machines and decoded the data to regular PCM. If you could share the images (if you still have them) that would be great. It would be a very useful resource for anyone trying to repair a machine with "ROM rot" as well I think?

Link to comment
Share on other sites

The Casio patents may help to decode it. It is not really data reduced (unlike MP3 or such) but only sections of recognizable samples alternating with short data sections. The operating system with all those softsynth algorithms is inside the main CPU, not these roms. Some Casio roms (e.g. MT-750) can not be listened that easily, but only because data and address lines were shuffled to simplify PCB layouts. Long ago Dtech helped me to decrypt it (I don't find the output anymore) - apparently he used a correlation algorithm. I also dumped the CTK-1000 rom, which was the largest stuff my Willem eprommer could read (using plenty of flimsy homemade adapter sockets).

Link to comment
Share on other sites

  • 2 weeks later...
  • 2 weeks later...
  • 3 years later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.