Not too much to report this time. Jim has two of the new boards to test. One has a VNC1 and one has a VNC2. The VNC2 has problems loading larger files; I’m hoping it’s a firmware issue. Yesterday FTDI released new firmware, so when I get a chance, I’ll upload it to one of my boards and test it.
I’ve started looking into having the boards assembled for me, but so far the cheapest I’ve found is about $30 per board. So unless I get an order for 500 kits, I’m still going to be building them myself.
Most of my time this month has been spent hanging out with my son, born on Jan 3rd. It’s really an amazing experience but it does tend to soak up any free time during the day 🙂
I did a quick layout of the buffer board and sent it off to Advanced Circuits in Colorado – they have a good ‘Bare-Bones’ service and quick delivery. I ordered the buffer chips as well. When everything arrived, I realized that one of the chips was the wrong package size (TSSOP instead of SSOP), so that delayed things as well. After getting the correct size chip, I finally built and installed the board. I was still having issues, and after disassembling and tracing the code, I had to make some changes to the design of the 8155/8255 CPLD replacement. Once I did that, it passed the WRAM test – sort of. It was still only reporting 1 WRAM page instead of 64. Better than nothing though.
Next task was getting the board to boot to the main menu. It wouldn’t get past the check for the VNC2 controller, so I had to program that as well. I purchased a FT232R USB-to-serial board off of eBay, and wired it up to the board. After three or four nights of futile attempts, I put the 232R board under the microscope and discovered that the silkscreen labels for the TXD and RXD (transmit and receive points) were backwards. After that, the VNC2 was successfully programmed and the board booted to the main screen.
At this point, everything works except for USB. The VNC2 is correctly identified but defaults to the wrong command format. I have to dig in to the source code and make some changes. The new floppy controller appears to work fine, and most importantly, the DSS-1 plays sounds. The multi-page WRAM issue turned out to be a bad cable.
And finally, it turns out the buffer board was unnecessary – I reconfigured the interface ports using a method in a Xilinx app note that Jim pointed out to me – and if the VNC2 changes work, the board will be ready. I won’t even have to do a second revision 🙂
I wasn’t really happy with the result of cleaning the PC boards that I’ve built. Using flux remover always left residue on the board, and soaking in isopropyl alcohol overnight, although better, still wasn’t perfect. I did some investigating and decided to get an ultrasonic cleaner. The model I got, a Crest 1875D has a large enough tray that I can put a whole KLM-782 board in it. My friend Evan recommended Alconox products to clean the board, so I got some Detergent-8, and the combination of the two has been the bomb.
I also got a relatively inexpensive stereo microscope off of eBay, which has been a Godsend for soldering the fine pitch components especially on the memory board. Soldering with a loupe stuck in my eye wasn’t very good for my vision!
Four orders last week, a complete DSS-1, a board set, and two kits to ship out. This week, another complete DSS-1 (possibly two), and there are many others that are interested. I think I’m going to have to pull some boards from my already modified DSS-1’s….
I should start work on the new CPU board soon. I’m at the point where I’m just finessing the final design. I just found out that the Vinculum II is coming out, and am wondering if I should wait for that. The new board should run the existing code aside from the new floppy controller. For some reason they moved the control registers around in the memory map, but all that needs to change is the register definitions. Unfortunately this means maintaining two separate builds of code. The new board will also have a larger SRAM chip for future expansion.
The new ASR-10 SP3 board is completed as well. It’s all working without incident, apart from the 2N3904 transistors I had ALL being out of spec and not working right!! I’m sending one to Garth Hjelte at Rubber Chicken Software to do some beta testing.
Finally, after months of debugging, the USB storage part of the kit is functional. Jim has done fantastic work of getting it all up and going. We managed to squeeze all the needed functionality into a 72-macrocell CPLD. The USB transfer speed is very fast, utilizing DMA directly from USB to the wave RAM.
I’m posting the news on the Yahoo! groups DSS-1 user group, but first I’m going to offer the existing kits (at a reduced price) to those of you who have been patiently waiting.
Well, it turns out that the daughterboard had a LOT more problems than I would have liked. On the hardware side, solder bridges and shorts took a lot of time to track down. Also, similar to the main board, the main CPU’s ASTB (address strobe) signal needs to be buffered, as it’s not latching the address bits properly on its own.
Also a couple of dumb errors with the CPLD. I had several versions of the project, and the version I used had the pin placement wrong. Also had the decode logic wrong for the Flash ROM.
The good news is that Jim is working with the new hardware and I think we may see results in a couple of weeks.
Aside from this project, I’ve made some progress on the new revision of the main CPU board. All of the TTL logic is being replaced with two medium-sized CPLD’s and it will have a newer floppy controller and the VNC1L USB controller integrated.
So, I’ve finally gotten the USB daughterboard to boot up. Took a few weekends and some time on the logic analyzer. It doesn’t have the new code needed to run the USB module, but it does have the new expanded memory map and additional RAM. Note to self: next time a design has 10 mil track spacing and power planes, spend the extra money and get a solder mask. I spent most of the debugging time tracking down micrscopic solder bridges.
I sent a test board off to the programmer, Jim, and he sent me a test build of the software. It isn’t quite working yet, so I have to put it back on the logic analyzer. This time, I spent some money and got a proper socket probe for the flash memory, instead of soldering a bunch of test points on the PCB.
Also Jim wondered if it might be possible to upgrade the DAC to 16 bits. In theory it’s possible, since the RAM is now 16 bits wide, but the latches are only 12 bits wide and would have to be replaced. Also the software would have to be changed to allow the synthesis functions to work on full 16 bit words. And unlike the modifications so far, this would actually change (improve?) the sound of the instrument. I’m going to see how the USB mod pans out…then I’ll see if it’s worthwhile doing.
I just finished a grueling 8 week installation at my workplace. We gutted two of our studios and installed two brand new Euphonix System 5 Fusion consoles. Today is the first day they go online.
I got two more sets of boards back from manufacturing. The first adds support for flash memory, more RAM, and USB to the DSS-1 retrofit board. Can’t really test it out until the new software is written. Also got back the 16 Meg memory boards. I discovered much to my dismay that the DRAM chips I am using are not 5V tolerant, nor is the 3.3V side of the level translator IC I am using for the data bus. Back to the drawing board!
Hopefully I’ll get the add-on board for the DSS-1 built and sent to the programmer for testing this week.
I’ve started work on the SY-99 page.
Things will be on hold for a bit. We are installing a new Euphonix System 5 at my work, and the de-install of the old console is like hacking through the jungle with a machete. Instead of a machete I have my Greenlee cable shears.
If anyone wants a dual engine Neve DFC, I can point them in the right direction…
Well…better late than never I guess.
Response to the DSS-1 expansion project has been mixed, to say the least. Most people feel it costs too much money. Many people don’t realize that the project isn’t just a memory board that drops in place, it’s a whole new CPU board AND memory board. Not too much I can do about that, as Korg really under-powered the instrument when it was designed. Although in terms of 1986 tech, it probably didn’t make financial sense to make it more powerful. Korg tried to correct the shortcomings of the DSS-1 with the DSM-1, adding expanded memory, SCSI, and improved (?) operating system, but left out some of the good bits like the filter resonance control and DDL’s.
The DSS-MSRK addressed many of the DSS-1’s shortcomings but for various reasons never became widespread.
Another misconception is that I have ‘redesigned’ the original expansion kit. I haven’t. I guess you could say I’ve ‘reorganized and repackaged’ the original design, but I haven’t (and couldn’t) make changes to the software since I don’t have the source code nor the time to disassemble, analyze, and rebuild the software for a project that might interest 50 people. I did make an exact clone of the original DSS-MSRK PCB, but without the custom Korg chip to put on it, it’s not much use. The main ideo of this project was to simplify the original design and make it easier for an end user to install.
For those of you who haven’t been on the DSS-1 Yahoo user’s group, the biggest news has been the software updates to the project. I was contacted by Jim Babb, who was one of the original designers of the DSS-MSRK expansion kit. We had exchanged a few emails, and it turned out that he still had the original 1987 source code for the DSS-MSRK!!! Not only did he agree to send me the last version that was developed (which added support for 1.44Mb floppy drives), but agreed to further develop the software and add user-requested features (thanks Jim!!).
At this point the software (V3.3) supports HD floppy drives, but also improves the SCSI spec (now supporting up to 1GB drives), improves support for SCSI to CF drives, improves the diagnostic mode, and allows you to reassign the data entry and tuning sliders to various voice parameters. The sofware will now support a 16 MB memory board that will be going into production in a few weeks.
Of course this is only the beginning….stay tuned.