Highly Liquid Forum facebook twitter google plus rss feed
  #1  
Old 07-28-2013, 08:37 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default Murom Aelita

I am wondering if the key matrix on the Murom Aelita will be compatible with the UMR2, or if it might be doable with some modification. It doesnt behave the way the matrix article describes..

I have observed the matrix behavior with a scope, and I can detail it here but I am more of a hobbyist than an engineer so I might be missing something obvious.

44 keys, 8x11 matrix?

8 pins that respond like this:
8: F3-C4. normally low at 0v, key press F3 is ~7v, F#3 a little lower maybe ~6v, G3 a little lower, finally down to 5v at the lowest for C4.
7: F3-C4. normally around 9v, key press F3 comes down to ~7v, F#3 a little lower, G3 a little lower, down to 5v at the lowest for C4
6: F2-E3 same as #8
5: F2-E3 same as #7
4. F1-E2 same as #8
3. F1-E2 same as #7
2. F0-E2 same as #8
1. F0-E2 same as #7

So paired high and low pins for each octave, with a descending voltages as keys ascend between ~7v and 5v.
11 pins that respond like this:
1. normally 0v, raised to 5v for each D# on the keyboard
2. normally 0v, raised to 5v for each E on the keyboard
3. same for all Ds
4. same for all C#s
5. same for all Cs
6. Bs
7. A#s
8. As
9. G#s
10. Gs
11. F#s
Any solutions out there?
Here is a russian synth site that has more details about the murom aelita including schematics but they are not in english. http://www.ruskeys.net/eng/base/aelita.php
Reply With Quote
  #2  
Old 07-29-2013, 03:33 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Quote:
Originally Posted by mbroers View Post
I am wondering if the key matrix on the Murom Aelita will be compatible with the UMR2, or if it might be doable with some modification. It doesnt behave the way the matrix article describes..

I have observed the matrix behavior with a scope, and I can detail it here but I am more of a hobbyist than an engineer so I might be missing something obvious.

44 keys, 8x11 matrix?

8 pins that respond like this:
8: F3-C4. normally low at 0v, key press F3 is ~7v, F#3 a little lower maybe ~6v, G3 a little lower, finally down to 5v at the lowest for C4.
7: F3-C4. normally around 9v, key press F3 comes down to ~7v, F#3 a little lower, G3 a little lower, down to 5v at the lowest for C4
6: F2-E3 same as #8
5: F2-E3 same as #7
4. F1-E2 same as #8
3. F1-E2 same as #7
2. F0-E2 same as #8
1. F0-E2 same as #7

So paired high and low pins for each octave, with a descending voltages as keys ascend between ~7v and 5v.
11 pins that respond like this:
1. normally 0v, raised to 5v for each D# on the keyboard
2. normally 0v, raised to 5v for each E on the keyboard
3. same for all Ds
4. same for all C#s
5. same for all Cs
6. Bs
7. A#s
8. As
9. G#s
10. Gs
11. F#s
Any solutions out there?
Here is a russian synth site that has more details about the murom aelita including schematics but they are not in english. http://www.ruskeys.net/eng/base/aelita.php
Hi Mbroers, welcome to the forum.

Can you post a schematic of the keyboard switch matrix? That is the best first step.
Reply With Quote
  #3  
Old 07-29-2013, 03:56 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

Its hard for me to tell which schematic page on the ruskeys.net link will be most helpful, hopefully one of these will do the trick.

Heres one that shows the key switches I think:
http://www.ruskeys.net/pasp/aelita/c/aelsxem1.jpg

This is a diagram of the actual board the pins from the keyboard feed into, you can see on the right side and right bottom the references to the keys and the octaves:
http://www.ruskeys.net/pasp/aelita/c/aelsxem5.jpg

I think this is the schematic of board 5 which might be the one you are looking for:
http://www.ruskeys.net/pasp/aelita/c/aelsxem17.jpg

These can all be found here, with additional schematic pages:
http://www.ruskeys.net/eng/pasp/aelita/pasp.php#сs

Thanks for your review, I successfully applied the UMR2 to a Yamaha PSS380 and am very hopeful I can get the same response from this oddball russian synth.
Reply With Quote
  #4  
Old 07-29-2013, 04:38 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Quote:
Originally Posted by mbroers View Post
Its hard for me to tell which schematic page on the ruskeys.net link will be most helpful, hopefully one of these will do the trick.

Heres one that shows the key switches I think:
http://www.ruskeys.net/pasp/aelita/c/aelsxem1.jpg

This is a diagram of the actual board the pins from the keyboard feed into, you can see on the right side and right bottom the references to the keys and the octaves:
http://www.ruskeys.net/pasp/aelita/c/aelsxem5.jpg

I think this is the schematic of board 5 which might be the one you are looking for:
http://www.ruskeys.net/pasp/aelita/c/aelsxem17.jpg

These can all be found here, with additional schematic pages:
http://www.ruskeys.net/eng/pasp/aelita/pasp.php#сs

Thanks for your review, I successfully applied the UMR2 to a Yamaha PSS380 and am very hopeful I can get the same response from this oddball russian synth.
A few more questions:

- Is this a monosynth? Do the keys "alias" when you press many at once?

- When you were using the scope, were you able to see actual logic waveforms? Your text only mentions variable DC voltages.
Reply With Quote
  #5  
Old 07-29-2013, 08:10 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

This is a 3 osc mono synth, im not sure I know what you mean by alias. I think it only plays the highest note held but I can verify.

Ah great point - I know in the past I have seen the logic wave forms (i think) but this time around I had the scope set to dc to measure what was happening at those pins and only saw the dc voltage.

I will see if I can find the logic wave forms again - does the logic wave form look like several square waves that move at different speeds based on the keypress? IIRC I saw this after the keyboard decoder feeding into the DCO circuit.
Reply With Quote
  #6  
Old 07-30-2013, 04:54 AM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

I have posted an album of scope screens here: http://imgur.com/a/80Xc2

Here you should see the main decoder board that accepts the signals from the keyboard on the left and has about 41 output pins on the right.


Here is an example of the pattern I see on the output pins that I am guessing is the logical waveform. It looks like a complex saw/square wave and as I press keys from low to high on the keyboard the cycle frequency increases. Some pin outs do nothing, some look more like square waves than the complex waves. If a key creates a response on one of the pins, the cycle frequency remains whatever it was last pressed at until a new key is pressed, then it switches to that frequency and stays there even when the key is released.

example slow cycle low keypress complex


example mid keypress complex


example low keypress square


example mid keypress square


example high keypress square
Reply With Quote
  #7  
Old 07-30-2013, 01:43 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Looks like it uses active-high select pulses.

Have you found a +5V supply rail with which to power the UMR2?

Given that the UMR2 is powered by 5V, the next thing to make sure of is that any signal voltage connected to the UMR2 does not go below UMR2 ground (0V) or higher than 5V. Is that the case with the signals that you are observing?

If not, it will be wise to add some voltage limiting circuitry between the UMR2 and host. This can be a simple arrangement of diodes & resistors. I will post more if you like.

Once that is taken care of, the best approach is probably to try to retrofit "1 key" by just connecting 1 select & 1 data line to the UMR2. If that works right, you can add more lines to cover more of the keyboard.
Reply With Quote
  #8  
Old 07-30-2013, 02:40 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

Im still a little confused - Im thinking the data lines are the pins from the keyboard, but some are always high at ~9v and some are always at 0v for the octave ranges, and some only react to individual keys, (all d's for instance go to one pin, all c's to another). Im not sure which pins I would use for the data lines if this is the case.

Same for the select lines, which Im assuming are the pins that go out from the decoder board to the synthesizer. Im not sure which ones I would use, some look like square waves, some like the complex waveform. Any guidance on which pins to target for the data and select lines?

I think the synth runs at 15v or maybe 13v so I would have to bring the voltage down to 5v, if you have a recommended circuit that would be helpful. Thanks again!
Reply With Quote
  #9  
Old 08-05-2013, 03:48 AM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

I believe I found a 5v + rail, it seems to measure around 5.6 - 5.8V but it is described on the right side of this schematic as + 5 B. Seems legit? I ordered the PCB and will attempt it asap.

Do you have any idea from the description which select / data lines seem the most likely to work or is there another post that describes a similar 1 key at a time approach in more detail?
Reply With Quote
  #10  
Old 08-05-2013, 05:49 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Quote:
Originally Posted by mbroers View Post
I believe I found a 5v + rail, it seems to measure around 5.6 - 5.8V but it is described on the right side of this schematic as + 5 B. Seems legit? I ordered the PCB and will attempt it asap.

Do you have any idea from the description which select / data lines seem the most likely to work or is there another post that describes a similar 1 key at a time approach in more detail?
If it's less than 6.0V, then the UMR2 will be OK. You can also add a crude voltage regulator as described in the UMR2 hardware manual by using the included 220 ohm resistor and zener diode.
Reply With Quote
  #11  
Old 08-05-2013, 05:52 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Quote:
Originally Posted by mbroers View Post
Do you have any idea from the description which select / data lines seem the most likely to work or is there another post that describes a similar 1 key at a time approach in more detail?
If you can simultaneously view two signals in your scope, it will help a lot.

A select signal will be a repeating pulse that "passes thru" to a data line when the correct key is pressed. This will be pretty easy to see if you have the correct signals on your two scope channels.

I'd use visual inspection of the keyboard PCB to find a pair of lines (1 select and 1 data) that meet at a specific key. Monitor them both in the scope while you operate the key switch. It should be apparent which signal is the select and which is the data.
Reply With Quote
  #12  
Old 08-05-2013, 05:55 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Also, do you intend to use the synth as a sound module (MIDI in), MIDI controller (MIDI out), or both? From that, I will continue with some ideas for limiting the voltage on the UMR2 select & data pins.
Reply With Quote
  #13  
Old 08-05-2013, 09:01 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

Thanks, yes I can look at two signals at once on the scope. I am planning on using this just as a module, not as a midi controller.
Reply With Quote
  #14  
Old 08-06-2013, 03:00 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Here's a method for limiting the voltage on each UMR2 i/o pin so that it does not exceed 5V.



The positive voltage at the cathode of each zener diode is limited to 4.7V. The resistors limit the current. I'd start with a value of 150 ohms for R1, R2 and R3.

You should be able to observe the voltage limiting effect with the scope by comparing the raw signal from the host to the limited signal that appears at the corresponding UMR2 pin.

If the UMR2 setup procedure can be successfully completed with a basic 1 x 1 matrix, then try normal operation as a sound module with just that one key retrofitted.

If that succeeds, then you can repeat the voltage limiting circuitry on the other select & data pins and retry the setup procedure with a larger matrix.
Reply With Quote
  #15  
Old 08-12-2013, 04:02 AM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

Unless I am missing something obvious I don't think the aelita will work with the umr2.

I checked my successful umr2/yamaha pss380 mod to compare, and that behaves as described in the keyboard matrix article. The select lines in this case are a pretty constant square pulse down to 0V from 5V (active low) and when the right key is pressed it allows that same select voltage to flow through the data line.

On the aelita, with the same scope settings that verified the pss380 matrix, I cant find similar behavior on any of the pins from the keyboard to the decoder board, or on the pins from the decoder board to the synth board. There are some signals that look the same as the pss380, but as I press different keys the frequency changes as described in a previous post, on the pss380 the frequency of the select pulse is constant.

I'm guessing its a different kind of key determining circuit that isnt compatible with the umr2. Is there anything I should try to confirm or any further information I can provide?
Reply With Quote
  #16  
Old 08-12-2013, 11:20 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Quote:
Originally Posted by mbroers View Post
Unless I am missing something obvious I don't think the aelita will work with the umr2.

I checked my successful umr2/yamaha pss380 mod to compare, and that behaves as described in the keyboard matrix article. The select lines in this case are a pretty constant square pulse down to 0V from 5V (active low) and when the right key is pressed it allows that same select voltage to flow through the data line.

On the aelita, with the same scope settings that verified the pss380 matrix, I cant find similar behavior on any of the pins from the keyboard to the decoder board, or on the pins from the decoder board to the synth board. There are some signals that look the same as the pss380, but as I press different keys the frequency changes as described in a previous post, on the pss380 the frequency of the select pulse is constant.

I'm guessing its a different kind of key determining circuit that isnt compatible with the umr2. Is there anything I should try to confirm or any further information I can provide?
There is no rule that states that the select signals have to be uniform, or of a stable frequency or pattern. It could be that the microprocessor on the keyboard reacts to keystrokes in a way that disrupts the stability of the signal or alters its pattern. This can make it tough for the scope to keep track the signals (triggering etc). I think this is especially common with older keyboards that have less computing power available.

Given that you have determined that the keys are indeed wired in a matrix with a diode for each switch, it's very likely that the keyboard works in a way that is expected by the UMR2.

Have you tried getting the UMR2 setup procedure working for just "one key" by connecting only one select/data pair? If you use the voltage limiting circuit above, there is no risk to either device. The UMR2 is completely passive during the setup procedure.

If you can complete the setup procedure as expected, it is likely that things are working...

Last edited by John; 08-12-2013 at 11:22 PM.
Reply With Quote
  #17  
Old 08-13-2013, 04:48 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

Ok, fair enough. I read some things about different keyboard matrixes and better descriptions of what the aelita might use (bus matrix?) and knee-jerked when I saw the proper response on the pss380.

I got everything hooked up and ready to test (power, midi, pgm switch) and did closer observations on the pin ins/outs on the matrix decoder board and although I got pulled away. Before I could start the one key at a time approach I wanted to see if the pattern I am observing makes any sense and if you could guess what a select line and what a data line is in this configuration, Im still having trouble determining this - nothing looks at all like a data line to me and the select lines seem iffy at best.

The pins in from the keyboard are already described above in the thread (are they the data lines? there are 8 for the semitones but 11 for the octaves, nothing appears to 'pass through' select signals).

The pins out from the decoder board to the synth could use a better description - I got confused about how to measure them and was switching to AC and posting pictures that may have been misleading:

There are 3 groups of 10 pin outs that have the roughly the same behavior (i think there are three groups because its a 3 vco mono synth, but who knows).

The last 6 pins of each group all have a pulse chain pattern that stays between 0 and 5v and changes in frequency as the keys are pressed. The pulse chain frequency stays constant depending on the last key pressed. Im guessing these are select lines, but there are 18 of them in total.

Group 1 first four pins looks like this:
1. pulse chain 0-5v
2. 0v constant, no reaction to any keys
3. smaller pulse chain, looks more distinctly like a square wave, but from 0v to 2v instead of 5v like pins 6-10, mostly at 0v with positive pulses, constant frequency depending on last key press, like this picture previously posted:
http://i.imgur.com/5RIkGjq.jpg
4. 0v constant, no reaction to any keys

Group 2
1. 0v no reaction
2. 0v no reaction
3. smaller pulse chain like pin 3 of group 1.
4. 0v no reaction

Group 3
1. 0v no reaction
2. smaller pulse chain like 1.3 and 2.3, but inverted where its seems mostly at 2v with pulses down to 0v
3. 0v no reaction
4. same as pin 2, smaller pulse chain but inverted

Tonight I will attempt to wire up what I think might be considered a data line (semitone pins in from the keyboard) with what i think is a select line (one of the pins 6-10 of the 3 groups of 10 outs) but if this provides any additional clarification that would provide better direction, please advise.

Thanks!
Reply With Quote
  #18  
Old 08-14-2013, 01:57 PM
mbroers mbroers is offline
Junior Member
 
Join Date: May 2013
Posts: 10
Default

I didnt have much luck, the lines I thought were select (0-5V pulse chain on the outs) prevented the UMR2 from going into program mode, when I tried using one of the other outs from the board as a select (the smaller square wave pulse chains) it went into program mode but didnt respond to any key presses on three differently behaving attempted data lines.

Again, I might be missing something but based on my experience with the pss380 this is a different beast altogether and might require something like a custom solution to duplicate the voltages coming in from the keyboard. Thats my hunch at least.
Reply With Quote
  #19  
Old 08-19-2013, 04:17 PM
John's Avatar
John John is offline
Moderator
 
Join Date: Jan 2009
Posts: 3,007
Default

Quote:
Originally Posted by mbroers View Post
I didnt have much luck, the lines I thought were select (0-5V pulse chain on the outs) prevented the UMR2 from going into program mode, when I tried using one of the other outs from the board as a select (the smaller square wave pulse chains) it went into program mode but didnt respond to any key presses on three differently behaving attempted data lines.

Again, I might be missing something but based on my experience with the pss380 this is a different beast altogether and might require something like a custom solution to duplicate the voltages coming in from the keyboard. Thats my hunch at least.
Well, thank you for your trying. From this end I can't guess too much more, but if you are desperate and want a solution that will work for certain, there is always the approach of wiring relays in parallel to the key switch contacts. Then, a MIDI decoder like the MD24 can drive the relays.

Sorry for not having some better advice for you!
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT. The time now is 09:19 AM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2017, vBulletin Solutions, Inc.