SC-8820 Gripes

After a long while, I dusted off my Roland SC-8820 sound module which I bought used about 10 years ago. The goal was to drive it from an older PC via MIDI. This turned out to be a lot harder than expected, all because of a power supply.


The last time I used the SC-8820 was at least two moves ago and either I never had the matching power supply or I had lost it. But the SC-8820 can operate on USB power, so a missing power supply is not a problem, right? Wrong…

Yes, the SC-8820 can easily operate on USB power. But there’s a catch. On the back of the unit, there’s a little switch that controls whether the module takes input from USB, PC or Mac serial ports, or MIDI connectors. So I switched it to MIDI, fed it some data, and got absolute silence.

After brief head banging, this turned out to be a documented feature (not a bug!). The manual is fairly clear that when operating on USB power, the input selector switch does nothing and USB input is always used. This was probably intended to make things easy for USB users and make the module work regardless of the switch setting. It means that to use the MIDI inputs, the unit must not run on USB power. OK, off to find a power supply.

The SC-8820 (and many other Roland/Boss devices) has a standard 9V DC jack. I plugged in a no longer used 9V power supply from an external USB disk. The SC-8820 was unimpressed. OK, maybe the adapter was bad…

I also happened to have an old Yamaha keyboard around and could easily borrow its (likewise 9V) power supply, so I tried that next. The SC-8820 still didn’t power up. Since the SC-8820 worked fine on USB power (apart from the input source restriction), it couldn’t have been completely broken, yet with the external power supply didn’t even blink.

At that point, I started paying more attention to the diagrams on the back of the SC-8820…


SC-8820 Power

…and on the power supply used with the Yamaha keyboard:

9V Power Supply

Brilliant! Roland has the polarity reversed. Normal power supply has the center positive, but Roland needs center negative. That’s why a generic power supply wouldn’t work.

Why did they do that? Some think it’s just a way for Roland to extract more money. That is possible, but there’s some evidence that the reversed polarity is or at least was common in Japan, so perhaps the incompatibility was not intentional. If anyone knows the real reason, I’d like to hear about it.

Needless to say, there are adapters available to reverse the polarity… but you have to buy one first. Or re-solder a power supply. Oh well, another lesson learned.

Incidentally, getting the SC-8820 going with Windows 7 was another obstacle course. There are (both 32-bit and 64-bit) drivers for Windows Vista, but the driver installer refuses to run on Windows 7. However, it is possible to force Windows 7 to pretend it’s Vista and then the drivers do install and work. It’s just the installer being a bit paranoid.

With the driver in place, one is faced with the fact that Microsoft ripped out the MIDI related control panel bits in Windows Vista, and 3rd party software must be used to convince post-XP Windows to play on a device other than the built-in software synth. Mind you, the machinery is still all there, just the user interface is missing (how’s that for backward compatibility?). In the end, Windows was defeated and the SC-8820 works as it should:

Credit for the above MIDI conversion goes to Rosemary in Australia. The MIDI file was played by Windows 7 Media Player connected to the SC-8820 via USB.

This entry was posted in MIDI, Roland, Sound. Bookmark the permalink.

12 Responses to SC-8820 Gripes

  1. Krzyś says:

    IMO you were lucky that the SC-8820 just ‘didn’t even blink’ with reversed polarity of power supply. This might have easily ended with releasing ‘magic smoke’ 🙂

    I’m always double checking diagrams with polarity; if device does not have one, I’m checking polarity of capacitors or just trying to determine which lead is negative. My OCD makes me even checking AC/DC adaptor with multimeter for correct polarity.

  2. DOS says:

    Yes, I’ve worked at companies that have sometimes produced hardware that doesn’t have much protection in the power supply part of the PCB and wouldn’t deal so well with your “plug it in and pray” approach 🙂 Admittedly that could be related to the fact that they’re no longer around! If there’s no diagram I do tend to use your approach though, if I don’t care too much about the hardware.

    It’s probably worth checking that the power supply can deliver at least as much current as the device says it will draw, too. I’ve learned from problems with a Linksys WRT-160N that even when you use the power supply that came with the product, which is rated for the amount of power the device says it needs, that isn’t always enough – sometimes the device will draw a bit more power at startup or the power supply will deliver a little less power when it’s first switched on, so plugging the supply into the power point first and into your device second can help in that case, even though it doesn’t seem right to me. A better option is to use a power supply that can provide more current than the device requires.

    Incidentally, you can get nice universal power supplies that let you change the voltage to one of a few pre-defined settings, come with various connectors, and let you change the polarity. These might be useful if you have lots of old hardware with missing power supplies.

  3. Arbee says:

    As far as I’m aware there is no standard for polarity on wall warts and so like Kryzs I’m always super OCD about trying to substitute them. Back in the 80s you could buy a “universal” PSU with switches to go from 2-12 volts and flip the polarity, with a variety of connectors (multiples sizes each of barrel and ring-sleeve and the like). So I wouldn’t necessarily suspect Roland of doing anything wrong to sell more PSUs.

    Interesting that it can’t be both USB-MIDI and USB powered at the same time though – most of today’s gear can minimally do both, and a lot of modern synthesizers can additionally act as USB audio interfaces, for 3 or more functions through the one connector.

  4. Joshua Rodd says:

    Yep, Roland loves to have the polarity on their adapters “wrong”. When I had some Roland power adapter ends fail, I retrofitted a commodity power adapter by cutting the code and swapping the leads. (Equipment it was hooked up to was a Roland TD-8, if that matters.)

    I’ve always assumed it had something to do with their habit of using IEC C17/C18 connectors (ungrounded), which is something audio equipment manufacturers are wont to do.

  5. Julien Oster says:

    To add some anecdotical data:

    I remember that the very cheap universal wall plug power adapters I had had polarity switches, and that I always checked the polarity diagram on the device to set the switch plugging it in.

    The fact that all the universal adapters that I remember had this switch, and that the diagrams were prevalent, may give a hint that “reversed” polarity wasn’t so uncommon.

  6. Grey Panther says:

    Heh, I would have guessed that the power supply was of too low of an amperage 🙂

  7. jack says:

    for april fools joke microsoft release ms-dos mobile nokia lumia

  8. Michal Necasek says:

    I’m pretty sure the universal PSUs are still available, I just didn’t have one at hand.

    The SC-8820 is very similar to its contemporary SC-8850 and the earlier SC-88 and SC-55. It’s just a MIDI sound module with additional USB connection, nothing more and nothing less. Why the limitation when USB powered I’m not sure… but I wonder if using MIDI cabling might require additional power not available over USB. The power jack on the SC-8820 is rated for 9V/400mA (3.6W) which is more power than what USB 1.x/2.x can deliver (5V/500mA, or 2.5W). Just a guess.

  9. Joshua Rodd says:

    Given MIDI’s funky electrical interface (designed to run long distances without creating ground loops), I wouldn’t be surprised if the engineer who bolted the USB interface onto to the SC-8850 (thus creating the SC-8820) decided not to bother maintaining MIDI’s ground isolation from the USB ground. (USB is not good at all about avoiding ground loops.)

    Roland never (to my knowledge) released devices that violated the MIDI spec by creating ground loops, in sharp contrast to almost every PC sound card manufacturer whose joystick MIDI port and associated adapter would tie the MIDI in/out grounds together and also tie them to the PC’s chassis, creating all kinds of buzzing, hum, and high frequency noise.

  10. Michal Necasek says:

    That’s a plausible explanation for the restriction when using USB power.

    Reading product reviews recently, I saw that some Made-in-China USB to MIDI adapters likewise skip the optocouplers in order to keep the costs down… which causes all sorts of problems with some MIDI devices. And yeah, the cheap joystick to MIDI interfaces were infamous for that… the thing is that they worked some of the time 🙂

  11. Mark Boyle says:

    Hi Mike, I hope you are still enjoying your SC-8820. I recently bought an SC-8820 on ebay. Luckily it came with the power supply, so I can use it with a traditional midi cable. This is an amazing little module. The sounds are great and it makes my Standard Midi File arrangements really sing. However, much to my disappointment, finding working drivers so as to use the USB connection has been a real struggle. For some reason there does not seem to be an official repository maintained by Roland or Edirol for old chestnuts like this. This is disappointing because only USB unleashes the full potential of this module. I have a number of old PC’s, as well as Macs. So far, I have only been able to get a really old HP Windows 98 machine running a Windows 98 driver downloaded from kind of a random internet site to recognize the unit and play through USB. I’ve tried Windows XP drivers I have found on the net on two XP laptops and neither recognizes the unit. I don’t have any Windows 7 machines at present, but would consider getting one if I knew it would run this module. I also have a nice Toshiba Laptop (circa 2007) that came with Vista (but which I downgraded to XP for software compatibility reasons). Could you share how you got it to work with XP/Vista/Windows 7? Do you have a working XP driver? I’d also love to find a working Mac OS X driver, if any of your readers might have one!

  12. Michal Necasek says:

    Hum. Looks like Roland ditched the driver downloads for older devices. Last time I looked (a year ago?) the SC-8820 driver was still there. They did have the downloads available until at least 2015, but they probably decided that 15+ years is enough. That’s a bummer.

    If you search for you should find the 64-bit Vista driver which also works with Windows 7. You could try SC8820_MacOS_v130.tar.gz as a search term for the OS X version.

    I don’t think I have a working XP driver. I recall the software being a bit funny and having things plugged in/powered/started it just the right order. I’ve mostly been using the module with a MIDI cable which does limit the possibilities but always works.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.