Making your own Dreamcast MIDI Interface Cable

Saturday, 1st May 2021

I love an unusual accessory for a video game console or computer, and one such accessory is the Dreamcast MIDI Interface Cable, HKT-9200, which allows you to connect MIDI devices to the Dreamcast console's serial port. Only released in Japan and with only one piece of software released for it — the O・to・i・re (お・と・い・れ) sequencer – these are a somewhat hard to find accessory nowadays and prices for second-hand units are far beyond what I could hope to afford (at the time of writing there are two on eBay, both for over £300).

Fortunately, the user darcagn on the Obscure Gamers forum took some photos of the insides of the interface box and from that I could make a pretty good guess as to how the cable works.

MIDI uses a serial protocol running at 31.25Kbps (a speed that can be easily derived by dividing a 1MHz clock by 32). Rather than signal "0" or "1" bits with different voltage levels (as with a PC's RS-232 serial ports, for example, which commonly uses +12V for a "0" and -12V for a "1") it uses a current loop, with 5mA current on for a "0" and current switched off for a "1". To avoid ground loops, which are a big concern when working with audio as they can introduce intereference (e.g. a mains hum) on recordings, the two connected devices are electrically isolated with an optoisolator in the receiver.

At the very least I therefore expected to see some sort of optoisolator circuit on the adaptor's MIDI IN port and some sort of output buffer circuit on the adaptor's MIDI OUT port to convert between MIDI's current loop signalling and the Dreamcast's 3.3V logic on its serial I/O pins, and that is indeed what you can see from darcagn's photos. My worry was that there might be some additional Dreamcast-specific hardware inside the box, but fortunately there isn't – it's all off-the-shelf parts. My main concern was how everything was connected, as this can't be completely seen from the photos: would sending MIDI IN data to the console's serial port RX and relaying data from the console's serial port TX to MIDI OUT be enough? Some experimentation would be necessary!

Building a serial port connector

As mentioned above, the MIDI interface cable's box doesn't contain anything Dreamcast-specific, however this box is connected to the Dreamcast's serial port using a proprietary connector. To try anything out I'd need to find a way to connect a circuit to this port:

The serial port on the back of the Dreamcast

Fortunately, the port's contact pitch is the same as a PCI Express slot, and I was able to find a PCI Express slot for £5 which could be used to make multiple connectors! It will need to be cut down to size (and in half, as the Dreamcast's serial port only has contacts on one side rather than both sides of a PCI Express card) but with a bit of work will do the job.

Cutting the PCI Express connector to make the serial port connector

The above photos show the process of cutting the PCI Express slot down to size. The Dreamcast serial port has 10 connectors in it, so a block is cut that is 12 connectors long using a cutting disc – as this is quite a rough process an extra sacrificial connector is left on each end as it doesn't matter if this gets mangled by the cutter. The block is then cut in half, leaving more of the support structure from the bottom of the slot on the side of the slot we're going to be using. The outer two connectors are then removed if they haven't already been damaged, leaving the central 10 connectors, and the outer plastic is brought to the final width and tidied up with some hand files. The fit of the connector should be tested against the Dreamcast's serial port:

Testing the fit of the connector inside the Dreamcast serial port

There shouldn't be too much side-to-side movement but the connector will be very loose without something to hold it down against the contacts. In my case I found some 2mm thick ABS plastic sheet was the perfect material to make the backing piece for the connector, though you may find your choice of material depends on the thickness of your PCI Express slot. It will need to be 13.5mm wide (about ½") and a decent enough length to fit inside the enclosure you're going to use for the plug – in my case 4cms was about right. The plastic can be cut by scoring it with a knife and then snapping it over the edge of a table.

Backing support piece to hold the connector against the serial port contacts

You should also drill some shallow holes in the plastic, with the centres of the holes being 3mm from the end and 3mm from the sides. The serial port has a couple of bumps stamped into the metal surround of the serial port and these matching holes in the plastic piece allow it to snap into place. A stripboard track-cutting drill is perfect for this task!

Groove filed into the end of the piece for the raised ridge on the PCI Express slot

In my case the PCI Express slot also has a slight lip that prevents it from sitting flush against the backing support. I could have filed this flat but the connector is quite fragile so I didn't want to risk damaging it so I ended up filing a corresponding channel into the bottom of the backing support piece.

Test fit of the connector and support piece

At this point make sure that everything fits. If it does you can start wiring up! The photos below show the process – heat-shrink tubing and strain reliefs are very useful, so don't forget to install them before soldering!

Soldering the wires to the connector

Putting a scrap piece of circuit board material between the two rows of pins also makes soldering much easier. In my case I only soldered the six pins required for this MIDI interface cable:

  • 1: +5V
  • 3: GND
  • 4: RX
  • 5: TX
  • 8: GND
  • 10: +3.3V

Pins are numbered from left to right when looking at the serial port at the rear of the console (if in doubt, you can check the voltages of the end pins against the console's metal chassis ground).

At this point you may wish to double check that your solder connections are made correctly and that nothing is shorted out – try with the cable plugged into the Dreamcast as well, and check that adjacent pins are not shorted together (the only two that should be shorted are pins 3 and 8, the two GNDs). If you're happy with that you can glue the connector onto the backing:

The connector is now glued to its supporting backing

You may notice that this is actually a different connector in the photo to the previous one, and that's because I accidentally got glue into the connector's springy contacts and jammed them so had to start again – definitely not a fun mistake to make, so be careful!

Fortunately, the second one went more successfully. The glued connector snaps in and out of the console with a nice reassuring click thanks to the two holes drilled into the top surface. Before switching on the Dreamcast I tried wiggling the cable around to ensure that even when treated roughly it woudn't short out adjacent pins. When I was happy this was the case I switched on the Dreamcast and ensured that I was getting a consistent +5V from pin 1 and +3.3V from pin 10 – as these pins are at the far end of the connector these are the ones that are more likely to have problems with crooked connectors. In my case I found there was an intermittent fault with pin 1's +5V. This was because I hadn't glued the connector on particularly straight and so pin 1's connector was slightly back from the edge of the backing piece. I very carefully filed the backing piece's edge so that it was flush with the slightly wonky PCI Express connector, after which pin 1 made reliable contact.

Potting the connector inside a small enclosure with hot glue

When you're completely happy with the connector, you can make it more robust by putting it inside an enclosure. I have some very small project boxes that are perfect for this sort of thing, it's a bit bulky when compared to the official Sega product but it does its job well here and doesn't bump into the power connector or AV port connector.
I cut a slot in one end of the box for the connector to stick out of and a notch in the other for the cable strain relief to clip into. I surrounded the console's serial port with a few of layers of masking tape to ensure that when the connector was inserted there was still a small gap between the case and the plug to make sure that it could always be fully inserted and not held back by interference from the case (it also protected the console shell from accidental strings of hot glue!) I then plugged the connector into the Dreamcast, made sure that everthing was neatly lined up, and secured the parts in place with copious amounts of hot glue. Once this had set I added more hot glue to the rear of the connector to make sure it was all held as securely as possible, and then screwed the enclosure shut.

The finished serial cable

With all the effort spent on the Dreamcast end of the cable, don't forget about the MIDI interface box end! I'm fond of JST-XH connectors so crimped one onto the end of the cable, ready to plug into the circuit board. The finished cable is seen above!

Building a prototype MIDI interface cable

With a Dreamcast serial port cable to hand I was able to experiment and see what happened when using the O・to・i・re (お・と・い・れ) sequencer. Here's the circuit I ended up building on a breadboard:

The prototype MIDI interface cable on a breadboard

On the very left is the first serial cable connector I tried constructing. It looks more yellow than the final version as I used a piece of pad board for the backing piece instead of the 2mm ABS plastic – I thought I was going to solder the PCI Express connector piece to a circuit board rather than directly to the cable's wires, this turned out to be a mistake. One advantage is that it does have eight wires soldered to it – I thought I'd see what RTS and CTS were doing, but I ended up not using them. The MIDI port to the left is for MIDI OUT and has a large black chip to act as a buffer, the port to the right is for MIDI IN and has a small white chip as an optoisolator. Here's the corresponding circuit diagram:

The schematic for the MIDI interface cable

The position of the ports is flipped in this diagram when compared to photo of the breadboard, but it otherwise matches up!

On the left is the MIDI IN port. This uses an H11L1 optoisolator as they are still reasonably easy to get hold of today, are fast enough for use with MIDI and can be run directly from +3.3V. Its output on pin 4 is open collector so it needs the 270Ω pull-up resistor to the +3.3V rail. A +5V-demanding optoisolator could also have been used if it had an open collector output and the pull-up resistor on its output was still tied to +3.3V (we don't want to run +5V into the Dreamcast's +3.3V logic!) but this makes the wiring a little more complicated so sticking to a +3.3V-compatible part makes life easier.

The 36Ω resistor between the output of the MIDI IN circuit and the serial port's RX pin is there because it is in the official cable. The official cable also places ferrite beads on every I/O pin (and has a ferrite bead clipped onto the cable itself) which I have not replicated in my own cable, but they can't hurt and I suppose it could protect the console from certain direct shorts!

The MIDI OUT port uses a 74HC365N to convert from the serial port's +3.3V logic to +5V to drive the MIDI output. MIDI signals can be run from +3.3V (it's the flow of current that's more important than the selected voltage) but +5V seems more typical so I thought I'd stick with that, and as we need to buffer the signal anyway (I'm not sure how much current the Dreamcast's serial ports are designed to sink or source) using the +5V supply we have available to us made sense. The voltage threshold for a "high" input signal is probably a bit too high with the 74HC365N – +3.3V is pretty close to the recommended values in the datasheet, so the 74HCT365N version of the chip would give you more margin for error and would be a drop-in replacement. In my testing the 74HC365N does work well, though, and it's what I had available.

The cable used in this prototype has eight connections rather than the six in the final. This is because I did experiment with the RTS (pin 6) and CTS (pin 7) signals from the console, however as far as I can see these are just pulled low and high respectively and do not change from the moment the console boots, even when sending and receiving MIDI data within the O・to・i・re (お・と・い・れ) software. If they were planned to be used for some purpose then I'm not sure what, and with only one piece of software released to test with I'm not sure I'll find out.

In any case, with this circuit data from the MIDI IN port is translated to +3.3V logic levels suitable for the Dreamcast's serial port input (RX) and translated back from the serial port output (TX) to a current loop suitable to drive a device connected to the MIDI OUT port. In practice it seems this MIDI OUT port acts more like a MIDI THRU with the O・to・i・re (お・と・い・れ) sequencer – any data sent to the MIDI IN port comes straight back out of the MIDI OUT port, however if you record some notes in the sequencer and play them back afterwards they aren't played back out of MIDI OUT. At first I wondered if I'd made a wiring error (accidentally connecting MIDI IN straight to MIDI OUT) but the MIDI OUT is indeed under control of the software as it will stop relaying messages on certain screens.

Putting it all together in a nice box

I ended up using my usual pad board construction technique to build the final device:

The assembled circuit board for the adaptor

I followed the circuit diagram I'd drawn earlier (rather than copy the breadboard circuit) to ensure the diagram was correct. This was all assembled to fit in a ready-made ABS enclosure, into which I cut some holes. Historically I've had a hard time cutting neat round holes in plastic enclosures – I have a few hole saws and these are great for cutting through wood or acrylic but when trying to cut ABS they tend to bind and either rip the box out of the vice or just shatter it. I normally resort to drilling lots of very small holes around the perimeter of the circle and then try to file it to size, which is OK for buttons or sockets with overhanging parts to hide the inaccuracies but wouldn't do here! For this project I bought a very cheap step drill set on eBay (three bits for a fiver) with zero expectations but it did an excellent job, it kept the centre hole I'd started from and didn't need very much cleaning up. I wish I'd bought one sooner!

Using a step drill to make the MIDI port holes

The other enclosure challenge to deal with was labelling the two ports. I normally get away without labelling my projects because the function of each port can be guessed quite easily (e.g. inputs and outputs normally only plug in one way, or player 1 is on the left and player 2 is on the right) but in this case there's not much convention for where MIDI OUT and MIDI IN go (though the original Dreamcast MIDI interface cable puts MIDI OUT on the left, as does my M-Audio Midisport). For this I thought I'd try using some dry transfer lettering designed for model-making, and it seems to do a great job!

Dry transfer decal lettering marks the OUT and IN ports

I'm pretty happy with the final outcome – it's a bit of a pain making the serial cable for connection to the Dreamcast, but the end result works well and it saves spending an absolute fortune on the original Sega accessory.

The completed MIDI interface cable/box

Now I can get on with making music with my Dreamcast!

Vista and MIDI

Tuesday, 14th August 2007


I have a Creative Audigy SE sound card, which provides hardware MIDI synthesis. However, under Vista, there was no way (that I could see) to change the default MIDI output device to this card, meaning that all apps were using the software synthesiser instead.

Vista MIDI Fix is a 10-minute application I wrote to let me easily change the default MIDI output device. Applications which use MIDI device 0 still end up with the software synthesiser, unfortunately.

To get the hardware MIDI output device available I needed to install Creative's old XP drivers, and not the new Vista ones from their site. This results in missing CMSS, but other features - such as bass redirection, bass boost, 24-bit/96kHz output and the graphical equaliser - now work.

The Creative mixer either crashes or only displays two volume sliders (master and CD audio), which means that (as far as I can tell) there's no easy way to enable MIDI Reverb and MIDI Chorus.

A productive weekend

Wednesday, 3rd May 2006

What with the weekend having an extra Monday tacked on for good measure (Labour Day), I felt the need to be productive.

I also felt the need to listen to VGM files converted to MIDI, so rustled up a VGM to MIDI converter. There already is one (available on the SMS Power! site), but I could never get it to work.
Having never really puzzled out the YM2413 ('OPLL', FM chip) I limited it to the square-wave generating PSG.

First of all, you need to be able to convert a tone register value (from 0 to 1023), the period of the output square wave, to a MIDI key value (0 to 127, where every 12 keys represent an octave).
This is easiest if you have a real frequency (in Hertz) to work with, so I have the formulae:

Frequency = ClockSpeed ÷ (32 × ToneReg)
Key = 12 × Log2|Frequency × Constant|
ClockSpeed is the clock speed of the PSG in Hertz. Constant is a precalculated constant used to scale the frequency to a range so that 440Hz ends up being played as key A, octave 5.

As it is unlikely you'll get a round number with this, I rely on adjusting the MIDI pitch wheel. Now it's a case of detecting attacks (when the volume of a channel increases) and releases (when the volume of a channel is set to 0) to create MIDI key press and release messages.
Percussion (white noise from the PSG) is handled the same way, except that instead of using mapping frequencies to keys and pitch wheels it plays one of 3 different drums corresponding to the 3 different pitches of noise.

It works fairly well, and you can download the software and source from the site here.
You can also listen to some samples:

The problem with sound-related apps is that they don't provide very interesting screenshots, so I took on a little side project that I hoped would. (After all, journals are a bit dull if they're plain text).

tween1.png tween2.png

tween3.png tween4.png

Yep, it's that 3MHz, 32-colour, 8KB powerhouse the Sega Game Gear again.
Looking on, there is only one Game Gear demo on there. If I haven't got it in me to complete a full-blown game, I can at least try and contribute something. rolleyes.gif After all, it's a relatively simple Z80 system to write programs for.

Have the typically poorly-shot video to see it in action:

(The odd flickering horizontal band that sometimes appears is a case of me using up the 8 sprite per scanline ration).

Subscribe to an RSS feed that only contains items with the MIDI tag.

FirstLast RSSSearchBrowse by dateIndexTags