Page 1 of 4
Guts of a Virgin (Electribe 2)
Posted: Tue Jan 06, 2015 9:47 am
by cntrlchng
Posted: Tue Jan 06, 2015 10:06 am
by dutchcow
Awesome! Have you discovered what chpis/cpu it has? Also wonder if the sampler version will have different hardware. Or newer e2 models with beefed up specs.
Posted: Tue Jan 06, 2015 10:08 am
by wasstof
Yeah, I'd say THAT'S done for your Warranty!!
Thanks for stripping it back, I'm impressed by how tidy it all looks in there, I used to have a job making mobile phones and some of Those boards looked like the Dogs Dinner!
Posted: Tue Jan 06, 2015 10:24 am
by cntrlchng
dutchcow wrote:Awesome! Have you discovered what chpis/cpu it has? Also wonder if the sampler version will have different hardware. Or newer e2 models with beefed up specs.
All of the IC's on the control board and main board are in the photos. The control board has an
ARM Cortex-M3 (MB9AF141LA) with 64KB with integrated flash memory. Midi-in goes to one of the serial data pins, and it's gotta be running in UART mode. there are de-populated jumper header spots on the main board, i'm guessing read/write protect and debug interface is taken out for production versions.
Posted: Tue Jan 06, 2015 1:36 pm
by Tom 62
I wonder for what IC7, J2 and J3 is good for
Posted: Tue Jan 06, 2015 3:11 pm
by Poumtschak
Thanks and sorry about your warranty.
Looking at the TI AMI1802B on the
CPU board, it seems consistent with the strings previously found in the firmware:
Poumtschak wrote:Tom 62 wrote:All in all it will be a question of the given CPU/DSP power of the unit.
Speaking of which, you can gran a few hints from the 1.03 firmware posted at korg's japan, when viewed in a hex editor :
././src/Dsp/DspIf.cpp...
././src/MCU/Component/AM180xMMSDCardController.cpp..
././src/MCU/Component/AM180xSpi.cpp.
././src/MCU/Component/AM180xSystemConfiguration.cpp.
././src/MCU/Component/AM180xUSB.cpp.
././src/sdcard/mmcsd/SDCommand.cpp..
././src/sdcard/SDCardDriver.cpp.
././src/sdcard/SDCardDriverTask.cpp.
././src/SerialFlash/SerialFlash.cpp.
././src/usb/cusbdc.cpp..
././src/usb/cusbdrv.cpp.
././src/usb/cusbdc.h....
././src/usb/cusbreq.cpp.
Looks like it is ARM9 (
TI Calixto AM180x) based, maybe with the associated DSP
I wonder if the Analog Device ADSP-BF523 BLACKFIN is the main DSP, or just here for the MFX/Kaoss Pad or other non audio purposes.
Posted: Tue Jan 06, 2015 7:17 pm
by DrHoo
Well done man, sonbody had too do it, looks very sturdy.
Posted: Tue Jan 06, 2015 11:32 pm
by cntrlchng
Poumtschak wrote:Thanks and sorry for your warranty.
Looking at the TI AMI1802B on the
CPU board, it seems consistent with the strings previously found in the firmware:
Poumtschak wrote:Tom 62 wrote:All in all it will be a question of the given CPU/DSP power of the unit.
Speaking of which, you can gran a few hints from the 1.03 firmware posted at korg's japan, when viewed in a hex editor :
././src/Dsp/DspIf.cpp...
././src/MCU/Component/AM180xMMSDCardController.cpp..
././src/MCU/Component/AM180xSpi.cpp.
././src/MCU/Component/AM180xSystemConfiguration.cpp.
././src/MCU/Component/AM180xUSB.cpp.
././src/sdcard/mmcsd/SDCommand.cpp..
././src/sdcard/SDCardDriver.cpp.
././src/sdcard/SDCardDriverTask.cpp.
././src/SerialFlash/SerialFlash.cpp.
././src/usb/cusbdc.cpp..
././src/usb/cusbdrv.cpp.
././src/usb/cusbdc.h....
././src/usb/cusbreq.cpp.
Looks like it is ARM9 (
TI Calixto AM180x) based, maybe with the associated DSP
I wonder if the Analog Device ADSP-BF523 BLACKFIN is the main DSP, or just here for the MFX/Kaoss Pad or other non audio purposes.
how did you get this info? where?
Posted: Wed Jan 07, 2015 4:26 am
by Tom 62
cntrlchng wrote:Poumtschak wrote:Thanks and sorry for your warranty.
Looking at the TI AMI1802B on the
CPU board, it seems consistent with the strings previously found in the firmware:
Poumtschak wrote:
Speaking of which, you can gran a few hints from the 1.03 firmware posted at korg's japan, when viewed in a hex editor :
Looks like it is ARM9 (
TI Calixto AM180x) based, maybe with the associated DSP
I wonder if the Analog Device ADSP-BF523 BLACKFIN is the main DSP, or just here for the MFX/Kaoss Pad or other non audio purposes.
how did you get this info? where?
If you open the os update file with any hex editor you'll find some more or less interesting ASCII strings inside the binary code.
Posted: Wed Jan 07, 2015 7:40 am
by Poumtschak
Tom 62 wrote:If you open the os update file with any hex editor you'll find some more or less interesting ASCII strings inside the binary code.
Exactly.
And a lot of free space for future updates too.
KLM-3315 - Main board
IC4 -
Texas Instruments AM1802B (ARM9)
IC6 -
Analog Devices ADSP-BF523 BLACKFIN (DSP)
The ADSP-BF523 is marketed as a multimedia co-processor.
I have somewhat the intuition that the Kaoss part is running apart from the rest, hence it's the only effect that survives a pattern change, and yet MFX is not exported to wav or ableton. Which could mean it's not part of the main voice generation + IFX architecture, that possibly runs on the ARM9 alone.
I have no idea what I'm talking about, but found these readings about DSP vs ARM core for real-time audio processing interesting :
http://a3lab.dii.univpm.it/webdav/docs/ ... E-Pisa.pdf
https://dsp.stackexchange.com/questions ... cessing-on
Posted: Thu Jan 08, 2015 6:43 pm
by d2ba
512 MB DDR ram --this is great news
Re: Guts of a Virgin (Electribe 2)
Posted: Sun Jan 11, 2015 11:09 am
by Poumtschak
Okay, I've been looking at your pictures again, and - correct me if I'm wrong - I think it should read like this :
Direct links to your pictures @imgur added when possible.
Any embedded systems dev to educate us, given what we know of the e2 idiosyncrasies from the early adopters point of view ?
Anyway, hope this helps.
Posted: Sun Jan 11, 2015 11:47 am
by dutchcow
Thanks for sorting out the ICs! Seems to me this device has more than enough muscle power to do pretty much anything we want. The crux lies with the coders at Korg. They should've taken enough time to properly optimise and bug test the firmware before releasing it. Getting a good optimised firmware that takes full advantage of all the HW takes time.
Posted: Tue Jan 13, 2015 1:30 am
by cntrlchng
I have a
DediProg SF600+ SPI "in system programmer", which can read and write to IC8 (MX25L12835F) on the KLM-3314 board, and I also have the adapter (
Pomona SOIC-Clip Model 5250) to clip onto the IC, without having to remove it from the board.
I initially tried using the adapter in the default pinout, but realized later that the RESET pin coming from the programmer was not connected to the RESET pin going to the IC, so I modified the adapter to account for this. All of the other pins are connected correctly.
However, I was never able to detect the EEPROM IC correctly, it was always acting as if the clip wasn't attached at all. I use this tool all the time at work, and this should have worked...
Before anyone asks, I did have the voltage set correctly, and I even went in and manually set it the config to the Macronix MX25L12835F IC, but it still was never able to detect. Any ideas?
Posted: Tue Jan 13, 2015 2:27 am
by Tarekith
Did you check the carburetor?
