Announcement

Collapse
No announcement yet.

Acquired a Praeludium II that needs work - questions

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    So the up-stop pads on the Pedal board are going to be a problem. They won't stay in place, and they are really worn down. I think I have a solution. I ordered 1/2" neoprene, 1/5" felt, and a 1" arch punch on Amazon.

    Punching the felt required soaking and freezing it to the punch can cut through. They are glued together with wet suit contact cement, designed for neoprene. This is the result:

    Click image for larger version

Name:	Up%20stop%20pad.jpg
Views:	118
Size:	80.3 KB
ID:	650544

    Now I just need to make 31 more.

    Comment


    • #17
      I admire your determination to complete the Praeludium and wish you luck. It's not the best build quality. In my friend's case all the switches and the speakers worked! But without a working CPU there was of course no MIDI. Hence the complete redo. Since you would like to use this console with HW, have you verified that that the MIDI out works? If it does, that is a big deal.

      You decry the lack of swell MIDI data so I offer a suggestion. Since the swell and crescendo pedals are identical in operation, in my friend's case we assigned one to the swell and the second to the great/pedal. In your case, why don't you keep the swell pedal as is but convert the crescendo pedal only, for use with HW as an expression pedal. This requires converting it from 12V to 5V, wiring up the LDR in a voltage divider configuration and connecting it to the analog input on the Arduino. The Arduino can then merge this data with that coming from the Galanti and you'll have, at least partially, what you want.

      If you wish to go this route, let me know and I can put more details on my website. There's already a picture there of the modified swell pedal circuit board. I'll check this post periodically to see how you are doing.
      John

      Comment


      • #18
        Originally posted by Coenraads View Post
        I admire your determination to complete the Praeludium and wish you luck. It's not the best build quality. In my friend's case all the switches and the speakers worked! But without a working CPU there was of course no MIDI. Hence the complete redo. Since you would like to use this console with HW, have you verified that that the MIDI out works? If it does, that is a big deal.
        Yep, MIDI out is fully working on Great, Swell, and Pedal. I'll keep the LDR conversion in mind. That's a good idea. I would never use the crescendo pedal anyway. That, like the auto-pedal, is for piano players.

        As for the CPU, I'll leave this here, in case someone is searching for this info: Everything on that board is pretty much stock, cheap, and readily available. There is even a modern drop-in replacements for the Intel 8032: the 80C320. It is faster (fewer clock cycles per operation) and lower power, but still fully instruction compatible. The EPROM is the only dodgy part, but I'm betting the same place I'm getting my replacement EPROMs from, has the code for it, and could burn copies.
        Last edited by mwdiers; 02-22-2019, 03:56 PM.

        Comment


        • #19
          Your source for the EPROMS can rebuild damaged boards or supply replacement boards for your organ.

          Comment


          • #20
            Originally posted by don60 View Post
            Your source for the EPROMS can rebuild damaged boards or supply replacement boards for your organ.
            Hello, Don.

            Yeah, I assume that's the case. He really knows his stuff.

            My missing EPROMs arrived today. I'll try them out tonight.

            I think I am going to backup all 13 EPROMs on this instrument so that I am prepared for anything that might fail in the future.
            Last edited by mwdiers; 03-04-2019, 01:14 PM.

            Comment


            • #21
              I got the chips installed tonight. The result was not great at first.

              With the four missing chips populated, I was getting a high-pitched sound that seems to be a Chimes component playing on all Great keys on top of the stops that are actually on. Mixture III was also completely whacked, and playing erratically.

              I tried each pair of chips by themselves (the EPROM and its corresponding M114).

              With IC3 and IC8 populated, the same problem happened (the noise, and the wacked out Mixture III).

              With IC2 and IC7 populated, the problem went away, and I had working 2’ Fifteenth and 4’ Flute stops.

              I also swapped the two new M114s. Both of them worked fine in IC7.

              And so, on a whim I swapped the CPUs between Slave Board 1 and Slave Board 2, and installed all chips. To my astonishment, everything started working perfectly.

              I’m not sure what to make of this, but I’m a bit concerned that it won’t stay working.

              The CPUs on the slave boards are Intel 8052s. The 8052 had an internal masked ROM that was factory programmed, in this case, with GEM-specific code. I know how to dump the masked ROM. Can this just be flashed into a modern MCS-51 variant such as the Atmel AT89S52?

              Second, one of the replacement EPROMs I received was specced as a 256k chip in the schematics, and yet the chip I received was a 128K chip. Yet it works fine. Is this because the original ROM didn’t actually contain more than 128K? Seems kind of strange.

              Comment


              • #22
                The problem may have been fixed by the removal and reinsertion of the CPU chip. Perhaps nothing is wrong with it, just wasn't making perfect contact in the socket.
                John
                ----------
                Church: Allen MDS-45 with Allen MIDI-DIVISION-II expander
                Home: Allen Renaissance R-230 with expanded four-channel audio and MIDI-DIVISION-II
                Shop: Bunch of organs in varying conditions, some good, some not...
                Half of an incredible two-man organ service team -- servicing all the major digitals in Arkansas churches
                https://www.facebook.com/pages/Birds...97551893588434

                Comment


                • #23
                  Perhaps. I guess time will tell. One of the first things I did was remove all the socketed chips, and lubricate them with a very thin layer of vasoline, so it may indeed just have needed reseating.

                  When my chip reader / burner arrives, I am going to attempt to read the masked ROM on the 8052 CPU, the "GEM 10711".

                  Some versions of the MC-51 processors from other manufacturers have an internal EPROM with "security bits" that you could burn in. These prevent the contents of the program memory from being read by anything but the CPU itself. However, I've not found any information that this was possible with the mask ROMS on the original Intel 8052. I guess I'll find out.

                  If I can read the ROM, I am going to try flashing an AT89S52 and see if that is a viable option as a replacement for the 10711 CPUs, because with the demise of General Music, that's the one part that cannot be replaced except by scavenging them from old instruments. If this works, I'll provide the result to my source for the EPROMs so this becomes an option in the community for anyone needing to repair these instruments in the future.

                  Now for a really challenging project, I may try to disassemble the source code on the CPU Board EPROM and see if it is possible to expand the MIDI capabilities of this instrument with a software upgrade. I've never done anything like this before, and it's not likely anyone else would benefit from it, so it will mostly be for learning purposes.

                  Comment


                  • #24
                    I'm totally fascinated with what you are attempting, and obviously you have some expertise in this area. On the ROM, you're more likely to find machine language than source code. Even if you can read the ROM, at least make sure you have a disassembler for that CPU to do the heavy lifting of turning all those ones and zeros into assembler codes. Even then, the process of reverse engineering this code, involves a very steep learning curve.

                    I don't want to take the wind out of your sails, but do go into this with your eyes open before expending too much money and effort. On the other hand, if you are doing this for fun, go for it.

                    I also think you need to assess what can realistically be achieved. If your intent is to expand the Midi capabilities, make sure the CPU has access to the data you would like to encode as part of the Midi output stream. The crescendo signal does appear to be digitized by some comparators onto five of the eight data lines. Now I could be wrong, but it looks to me like this would only give you five levels which is pretty crude and wouldn't be that useful. The swell pedal on the other hand doesn't look like it gets digitized at any point and I doubt the CPU would have access to it.

                    As I mentioned before, I would leave the swell alone, convert the crescendo to 5V and read it with an Arduino. And then, use the Arduino to merge this with the Midi out from the Galanti.

                    The pistons, share a common data bus with the keyboards and maybe the CPU might have access to them. Simpler would be to use some of the pistons for HW and leave some for the Galanti. Or, install double pole switches behind the (toe) pistons. Or, mount a row of new pistons under the swell. This rail is heavy metal and may be a bit difficult to drill. Then scan these pistons with the Arduino.

                    You can even tap into the lighted stop rail. The lights are 12V, 1W incandescent bulbs. You can connect the LED (with suitable series resistor) from an opto-isolator across the bulb and then scan the phototransistor with the Arduino. This would be a neat way of operating the HW stops. This in no way would interfere with the operation of the Galanti. I must admit I have never tried this but it should work. And I doubt the small additional current involved will stress the driver transistors used with the bulbs.

                    I think I can be confident in saying that adding an Arduino is going to be much easier than hacking the software.
                    John

                    Comment


                    • #25
                      I don't have the schematic in front of me, but wouldn't five bits coming from the crescendo pedal imply 32 levels? I cannot imagine that the encoding is a one-hot or cumulative scheme that represents only five levels.

                      Of course the contents of the ROM would be machine language, not assembly language code. I am sure that MW understands that point and just said "source code" in a generic sense.

                      Comment


                      • #26
                        Admittedly, I was being pedantic. My Apologies to MW.

                        About the crescendo, I too had a tough time believing that the crescendo only has five levels, but I do have the schematic in front of me and it clearly shows the output of the LDR feeding five comparators each biased to a different voltage by a resistor chain. Each one outputs directly to one of five lines on the data bus. In fact, these outputs are labelled: pp, p, mf, f and ff. Could this be a memory issue? Perhaps someone with a Praeludium can verify this five step crescendo.
                        John

                        Comment


                        • #27
                          5 steps for the PII crescendo sounds correct. The Praeludium III, which came a couple years later and was more advanced in several ways only had 9, and that was a lot more than I remember the PII having.

                          Comment


                          • #28
                            Originally posted by don60 View Post
                            I don't have the schematic in front of me, but wouldn't five bits coming from the crescendo pedal imply 32 levels? I cannot imagine that the encoding is a one-hot or cumulative scheme that represents only five levels.

                            Of course the contents of the ROM would be machine language, not assembly language code. I am sure that MW understands that point and just said "source code" in a generic sense.
                            Correct. I would be disassembling to assembly and working in that. There's tons of info out here on the MC-51 architecture, and lots of free tools to do this. It may well be too much for me. It's been 20 years since I've touched assembly (on an x86), but at least the 8-bit processors are MUCH simpler.

                            Comment

                            Working...
                            X