Document SER-DATA-IN protocol

[probonopd]

Let's document the SER-DATA-IN protocol of the Acoustimass module so that we can use the Acoustimass module without the original Music Center.

Update 2023: See below for documentation provided by Bose.

[garci66]

I am have a slightly different accoustimass system (lifestyle 12) and my control center is half dead. My subwoofer has a slightly different connector (its a regular DIN instead of mini-DIN) and does not accept SPDIF but I would expect the serial protocol to be the same.

So far, the protocol is not evident.. but seems simple enought to be "clonable" from a working system.

The serial line is always pulled up. I am seeing what seems to be the encoded commands using a similar method to the NEC codes from the remote (as in short and long pulses) but the pulses seem to be 1msec or 2msec in duration. There is a 8 msec "start" bit of sorts and also a 2-msec "stop" bit. The comands seem to be 13 bit in length which is quite weird. Might be 12-bit + parity but I havent't checked yet.

I think its totally possible to control the accoustimass system without the console. I'll have to keep trying.

[probonopd]

Excellent. Let me know if I can help by looking at something, although I do not have access to the system right now (might take a couple of weeks).

[garci66]

Im still trying to figure out the encoding to be honest. But I think we can just dump the commands in a format similar to the "raw" format similar to irdb. Should be easy enough to reproduce. Basically the line has a 5V pull-up and there a transistor should pull it down. I have a model 5-V music center and an accoustimass 12 subwoofer/amp. I will see if I can hack something up.

…

On Thu, Jul 11, 2019 at 2:59 AM probonopd ***@***.***> wrote: Excellent. Let me know if I can help by looking at something, although I do not have access to the system right now (might take a couple of weeks). — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2?email_source=notifications&email_token=ACQV3WF6Q5KZTSP3F4XMVNLP63D2NA5CNFSM4ETMAE22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZVTA7A#issuecomment-510341244>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACQV3WEP45LQQEGRK32Q2BDP63D2NANCNFSM4ETMAE2Q> .

[garci66]

Here are some images of the traces I captured last night. The narrow pulses are 1msec in width and the wide ones are 2msec. The long pulse down at the start (some sort of "start bit") is 8msec.

From what I am seeing, it seems that the encoding is something similar to the IR encoding with short-short or long-short pulses. And then the "stop" bit seems to be the only "long" low bit (as well as the start).

[probonopd]

Interesting! Hope you can figure something out by comparing some codes.

Maybe @bengtmartensson recognizes this type of code from just looking at it?

[bengtmartensson]

Looks like biphase encoding to me. IrpTransmogrifier is a good tool for analyzing this.

[garci66]

I though it was bi-phase but if you look carefully, there are no "long" zeros (other than the start and stop bits). So far, it seems to be more a case of "pulse width modulation" where the low portion is always 1msec (except for start/stop bits) and the high part is either 1msec or 2 msec. I arbitrarily chose "1" to be 2msec pulses and 0 for 1 msec pulses.

Using this mechanism, I was able to decode all the codes from the remote. Code length is still weird (as in 13 bits or 20 bits in total).. but I guess BOSE was not expecting to be compatible or interoperable with anyone else.

I have attached captures in "IRDB RAW" format (as in indicating the length in micro-seconds for each pulse, whether low or high for all the keys. I have not attempted to make any sense out of each yet, but im pretty confident that just reproducing the same codes will trigger the accoustimass amp to switch inputs, change volume. mute, etc..

We should also note that there are 50 msecs between two codes (plenty of keys send two codes).

The circuit is super simple. A pull-up drives it yo 5V and we pull down the line with an arduino directly or through a transisitor.

bose remote codes.xlsx
bose remote codes.txt

If we just want to control the sub, we can just trigger those codes and I guess the accoustimass will react in the same way as if the control center would be connected.

[garci66]

For example, using the encoding I described above, and SSSS being the 8-msec long Start bit and TTTT the 2msec stop bit, these are the keys for volume and surround vol:

VOL+  SSSS|1001100010001|TTTT
VOL-  SSSS|1001100010010|TTTT
SVOL+ SSSS|1001100011001|TTTT
SVOL- SSSS|1001100011010|TTTT
MUTE  SSSS|1001100010101|TTTT

And for the function switch (surround mode)

SURR: SSSS|100110001110100000100|TTTT
S+CE: SSSS|100110001110100000011|TTTT
STER: SSSS|100110001110100000010|TTTT

Finally, some of the Function keys:

VIDEO: SSSS|1011000011010|TTTT
VIDEO: SSSS|1001100010011|TTTT

FM:    SSSS|1010101001010|TTTT
FM:    SSSS|1001100010011|TTTT
AM:    SSSS|1010100111010|TTTT
AM:    SSSS|1001100010011|TTTT

All the keys are logged in the excel and CSV file

It seems as if the second key press of the function key is the same for all functions. Which could be an attempt to turn on the subwoofer just in case.

[probonopd]

Wow, awesome work @garci66. I am trying to convert those codes into something IrScrutinizer understands, so that it can easily be converted from there into various formats.

Looking at VIDEO as an example, I wonder whether

--START|, -8028, +2000, -1008, +996, -1012, +1996, -1012, +1996, -1012, +992, -1012, +996, -1008, +996, -1012, +992, -1012, +2000, -1008, +2000, -1008, +996, -1012, +1996, -1012, +992, -2016,|END--

can be written as

+8000, -8028, +2000, -1008, +996, -1012, +1996, -1012, +1996, -1012, +992, -1012, +996, -1008, +996, -1012, +992, -1012, +2000, -1008, +2000, -1008, +996, -1012, +1996, -1012, +992, -2016

or even be cleaned up a bit to read

+8000, -8000, +2000, -1000, +1000, -1000, +2000, -1000, +2000, -1000, +1000, -1000, +1000, -1000, +1000, -1000, +1000, -1000, +2000, -1000, +2000, -1000, +1000, -1000, +2000, -1000, +1000, -2000

Which would translate, according to IrScrutinizer, to Pronto Hex

0000 006D 000E 0000 0130 0130 004C 0026 0026 0026 004C 0026 004C 0026 0026 0026 0026 0026 0026 0026 0026 0026 004C 0026 004C 0026 0026 0026 004C 0026 0026 004C

Does this look right?

[probonopd]

I am a bit confused as to why something else than the would get sent to the Acoustimass module regarding the input source such as VIDEO, AM, FM, CD, TAPE or actions like REWIND. Isn't that something that is entirely handled in the Music Center?

Or are you talking about controlling the Music Center, not the Acoustimass module?

Both are interesting, but my end goal is to be able to se the Acoustimass module without the Control Center.

[bengtmartensson]

Just to show you what you can do with IrpTransmogrifier:

First a slightly edited version of bose remote codes.txt above:
bose.remote.codes-edited4transmogrifier.txt

Transmogrifying:

$ irptransmogrifier analyze  --trailingg 2000 --bit-usage  --parametertable --named  bose.remote.codes-edited4transmogrifier.txt    
VIDEO              {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x61a}
VIDEO-alt          {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x313}
FM                 {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x54a}
FM-alt             {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x313}
AM                 {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x53a}
AM-alt             {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x313}
CD                 {1005,msb}<-1,1|1,-1>(1,A:3,-2){A=0x4}
CD-alt             {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x55a}
CD-alter           {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x313}
CD (already in CD) {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x313}
TAPE               {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x42a}
TAPE-alt           {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x313}
STOP               {1005,msb}<-1,1|1,-1>(8,-2,A:3,-2){A=0x6}
Play / pause       {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x611}
SKIP back          {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x613}
SKIP FWD           {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x612}
SEEK back press    {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x615}
SEEK back release  {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x61a}
Seek fwd press     {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x614}
seek fwd release   {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x61a}
SURR vol +         {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x319}
SURR vol -         {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x31a}
VOL +              {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x311}
VOL -              {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x312}
SURROUND           {1005,msb}<1,-1|1,-2>(8,-2,A:20,2,-2){A=0x31d04}
STEREO+CENTER      {1005,msb}<1,-1|1,-2>(8,-2,A:20,2,-2){A=0x31d03}
STEREO             {1005,msb}<1,-1|1,-2>(8,-2,A:20,2,-2){A=0x31d02}
MUTE               {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x315}
OFF                {1005,msb}<1,-1|1,-2>(8,-2,A:12,2,-2){A=0x315}

Bit usage analysis:
A       **000*****0*******
Duplicates analysis:
A       (a:1)(a:1)000(a:1)(a:1)***0*******      1,1,3,1,1,1,1,1

Parameter table:
VIDEO                   61a
VIDEO-alt               313
FM                      54a
FM-alt                  313
AM                      53a
AM-alt                  313
CD                      4
CD-alt                  55a
CD-alter                313
CD (already in CD)      313
TAPE                    42a
TAPE-alt                313
STOP                    6
Play / pause            611
SKIP back               613
SKIP FWD                612
SEEK back press         615
SEEK back release       61a
Seek fwd press          614
seek fwd release        61a
SURR vol +              319
SURR vol -              31a
VOL +                   311
VOL -                   312
SURROUND                31d04
STEREO+CENTER           31d03
STEREO                  31d02
MUTE                    315
OFF                     315

[probonopd]

This is truly amazing @bengtmartensson

[garci66]

I don't think you need your first "+8000" part on the signals.. basically the line is always idle at +5v (it's pulled up by both the accoustimass and the console it seems. (or on the accoustimass side when it's powered up). Other than that, the graph you show seems perfect.

If we are using an aruduino and its internall pullups then we need to leave the line "idle" at "1" or "high-z". And then we start by writing a 0 for 8msec (the start bit), then we write a the whole pattern and finally we write a 0 for 2msec (the stop bit).

If we're using an external transistor and external pullup we might need to invert the output first but not sure. The system seems to have a 10K pull-up according to the service manual.

Also, regarding why other keys (than volume, mute, surround) get passed along.. the service manual says the following:

When an external source is selected (AUX, VIDEO, or TAPE), the transport commands (FF,
FR, etc.) are passed through the serial data jack via Q401 and its associated circuitry.

I guess this means that you could have a BOSE (or bose compatible) tape deck and receive commands via the serial Bus (and why the control center has two serial ports in parallel so that you can connect other devices).

Finally, just wanted to mention that I used the remote and the music center configured with the "house code" of 0000 and also the speaker config for 00 (I believe its speakers A). didn't test what happens when using other settings.

I will try a very simple arduino sketch to send basic commands to the sub and see if it "reacts accordingly". Not sure how many "orphan" accoustimass amps are lying around without a console, but it would be good that we're able to "revive" them... ;-)