Welcome to my Blog

To content | To menu | To search

Sunday, January 12 2025

2 years of silence : Digging into the PSS-A50 firmware...

By lionel nollet on Sunday, January 12 2025, 16:37

It's a few years already I started to dig in the QSPI flash of the PSS-A50 (and a bit in the PSS-F30 one). I desoldered the flash chip in order to extract the data with a QSPI reader.

Then I started at first to identify the data blocks in order to understand how the samples are built and accessed depending on the selected voice and key pressed. Visually by watching some repetitive patterns in an hex editor (HxD).

Thanks to Ghidra I could go deeper in the firmware, and confirm the YMW830 SWLL was based on Armv7 big endian. Then try to identify the GPIO ports and try to locate the registers of the audio core in the SOC. Not all the firmware is run XIP from the Flash (Executed In Place, means runs directly from the QSPI Flash -with a small buffering-). I could notice that part of the SW was copied from Flash to RAM, and ran from RAM. But I had to admit that some subroutines were called in the (supposed) internal ROM of the YMW830 SWLL.

I could get a better understanding using a (very) high speed logic analyzer connected directly on the QSPI in order to get a "real time" dump of the µC access, while the PSS-A50 is running (needed to be decoded later). Which could help to understand the boot process. It boots at first in SPI mode, copy some code to RAM, and then switches to QSPI mode start the execution of the rest of the firmware.

My assumption was that "they" probably blown the jtag fuses in order to lock the access to the internal ROM... so I did not spend any time trying to connect a JTAG probe (that I don't have btw).

But a few days ago, I discovered here that the JTAG access was enabled : https://mastodon.social/@whitequark/113694931960849696

Catherine/@whitequark could connect to the JTAG using her own "glasgow" HW (kind of "Bus Pirate") : https://glasgow-embedded.org/latest/intro.html

Catherine could extract the QSPI firmware, which is not the main outstanding thing from my current situation, but she could also extract the 48k Internal "ROM" from the YMW830 SWLL.

There's a mega.nz/file link in the discussion flow.

PSS-A50, PSS-F30, PSS-E30, SHS-300, PSR-F50, PSR-F51, PSR-F52, PSR-E253/YPT-255, PSR-E263/YPT-260, PSR-E273/YPT-270, PSR-E283/YPT-280 are all based on this YMW830 SWLL.

Now I'm wonderning if the 48k ROM is the same in all these products, especially for the PSS-A50, PSS-F30, PSS-E30 I own ? => answer is below in this post

The thing is that only the PSS-A50 has a dynamic touch response, while PSS-E30 and PSS-F30 are not managing dynamic touch in the SW, despite the fact the keyboards are the same from a HW point of view (dual switch per key - except some diodes shall be added on the PSS-E30/F30 PCB - PCB is the same for both pss-A50, PSS-F30, PSS-E30). PSS-A50 had an additional chip to manage the MIDI USB connection. Dynamic touch response is possible on PSS-E30/F30 ? That is a point I would like to investigate for a hack .

Then, also, a few days ago, someone made a great discovery : It is possible to access to an internal console of some Yamaha keyboards, using the USB connection and a Python script : https://github.com/portasynthinca3/swl01u

How awesome is that ?!

Porta/portasynthinca3 was also able to extract the firmwares using a FT232R-based JTAG dongle / OpenOCD (internal ROM of the SOC and external Flash), and performed a partial retro-engineering of her PSR-E433 firmware!

__Porta discovered that Yamaha made a shell that runs on top of MIDI SysEx messages, on the USB. She wrote a Python script in order to be able to access to the console.__

And that was not enough : she even was able to add/write her own arm code inside the ram, just by using the console. And the result is here : https://psi3.ru/blog/swl01u/

A GIT "issue" was open as a discussion flow, and it looks that so many Yamaha keyboards are compatible with this script !

Let's go back to the PSS-A50/PSS-F30/PSS-E30.

Earlier, when I went deeper inside the firmware/binary of the PSS-A50, I already suspected an old arm generation because of the typical arm 32 bits codeop starting with E1 xx xx xx / E5 xx xx xx / E9 xx xx xx / EA xx xx xx; which was confirmed by Ghidra and the best matching was Armv7; which confirmed the YMW830-V was a SOC based on a tone generator plus an Armv7 core. I suspected earlier a Renesas SH 'something' because I could see a string "Renesas", which was confusing when considering aother string "This code can only run on a Thumb compatible processor".

Of course, I was also intrigued by all these keywords I could also see this curious "login #0000" which made me thinking about an embedded console.

Recently, I try to make my JTAG/SWD probe with a blue-pill running BlackMagicProbe, but I did not success to have it talking with the YWM830. Don't know what is wrong on my side. I suspected the missing pull-up resistor on TDO, but no change when adding one.

Anyway, the shell.py from Porta "Works on PSS-A50".

My suggestion is to add this in the script, it helps to identify the right naming of the ports and in case, change them to fit your instrument (because PSS-A50 is : 'Digital Keyboard 0' / 'Digital Keyboard 1') :

print("MidO-IN =", mido.get_input_names(),"\n") print("MidO-OUT =", mido.get_output_names(),"\n")

login
passwd? #0000
login OK
> help
logout
help
?
info
ver
stack
perf-on
perf-off
perf-disp
d
dp
d   xxxxx
d/s xxxxx
m   ADDRESS DATA
m/b ADDRESS DATA
m/w ADDRESS DATA
m/l ADDRESS DATA
regist
voiceset
ots
func
mdb
backup
chkprm
freearea
jobconid
dispinfo
swmask
ioport
getsw
pushsw
sendsw
> info
DevelopName        PSS-A50
DevelopNumber      #4007
Main DevelopNumber #4007
Make data & time   MAY 07 2019 14:07:51
J/E Select         English
> ver
Main  Version      1.01
Boot  Version      1.00
Style Version      No Data
Song  Version      1.00
Param Version      1.00
Wave  Version      No Data
Pattern Version    No Data
Unshi Version      No Data
Adss  Version      No Data
Manufacturer                   : 0105 : YAMAHA.
Product version                : 0200 : V.2.00.00

Also, it's working with the PSS-E30 and with the PSS-F30 (not tested with the F30, but both same family as the PSS-A50, but without USB data interface). There's no data USB on the PSS-E30 / PSS-F30 products, but on the PCB, the Rx/Tx of the YMW830-V is easily reachable.

I could connect a cheap USB<->midi DIN (I had to modify it in order to extract the Rx/Tx at 3.3V levels - using a cheap level converter 5V<->3.3V). This USB/midi is seen as the following ports : 'USB MIDI Interface 0' / 'USB MIDI Interface 1'

And then, once connected to the PSS-E30 UART, same magic :

(and FYI, the internal 48k ROM content is the same for both PSS-A50, PSS-E30, PSS-F30)

login
passwd? #0000
login OK
> help
logout
help
?
info
ver
stack
perf-on
perf-off
perf-disp
d
dp
d   xxxxx
d/s xxxxx
m   ADDRESS DATA
m/b ADDRESS DATA
m/w ADDRESS DATA
m/l ADDRESS DATA
regist
voiceset
ots
func
mdb
backup
chkprm
freearea
jobconid
dispinfo
swmask
ioport
getsw
pushsw
sendsw
> ver
Main  Version      1.00
Boot  Version      1.00
Style Version      1.00
Song  Version      1.00
Param Version      1.00
Wave  Version      No Data
Pattern Version    No Data
Unshi Version      No Data
Adss  Version      No Data
Manufacturer                   : 0105 : YAMAHA.
Product version                : 0200 : V.2.00.00
> info
DevelopName        PSS-E30
DevelopNumber      #4006
Main DevelopNumber #4006
Make data & time   MAR 22 2019 10:59:11
J/E Select         English

For curious people, here are the other main commands results for PSS-A50 :

Except these 3 ones, I'm a bit affraid to try : getsw pushsw sendsw

> perf-on
Performance Start
> perf-display
Command Error
> perf-disp
                            -9Sec  -8Sec  -7Sec  -6Sec  -5Sec  -4Sec  -3Sec  -2Sec  -1Sec   0Sec (MAX=100%)
ID 00 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 01 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 02 Task                    1.8    1.5    1.5    1.5    1.5    1.7    1.1    1.6    1.6    1.8
ID 03 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.1    0.2
ID 04 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 05 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.2    0.1    0.5
ID 11 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 12 Task                    0.2    0.2    0.1    0.2    0.2    0.1    0.1    0.1    0.1    0.5
ID 13 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 14 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 15 Task                   13.3   13.7   13.3   13.2   13.4   13.1   13.7   13.4   13.2   13.2
ID 16 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 17 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 18 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 21 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 22 Task                    0.9    1.5    0.6    1.2    1.2    1.1    1.8    0.6    1.5    0.8
ID 23 Task                    1.7    2.0    2.2    1.8    2.2    2.2    1.8    2.2    1.8    2.3
ID 27 Task                    2.5    4.3    3.6    3.8    4.7    2.5    4.4    3.9    3.3    5.8
ID 28 Task                    1.9    1.5    1.4    1.4    1.4    1.7    1.3    1.6    1.6    1.6
ID 30 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 31 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 33 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 37 Task                    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0    0.0
ID 43 Task                   77.2   74.9   76.9   76.4   75.0   77.1   75.4   75.7   76.2   72.8
>
> regist
========= Regist Parameter =========
Part param      Main    Dual    Left
 On/Off     :    127     000     000
 Voice No.  :    000     000     014
 Bank MSB   :    000     000     000
 Bank LSB   :    000     000     000
 Prg Change :    000     000     033
 Volume     :    100     100     090
 Octave     :    000     000    -001
 Panpot     :    064     064     064
 Rev Depth  :    008     008     000
 Cho Depth  :    000     000     000
 Fil Cutoff :    000     000     000
 Fil Reso   :    000     000     000

Common Param
 Rev Type   :    002
 Cho Type   :    001
 Panel Sus  :    064
 Style No.  :    000
 Acmp On/Off:    000
 Style Sect :    000
 Acmp Volume:    255
 Transpose  :    000
 Acmp Split :    054
 Split Point:    054
 Harm On/Off:    000
 Harm Type  :    001
 Harm Volume:    255
 Tempo      :    120
 Apg On/Off :    000
 Apg Type   :    000
====================================
>
> voiceset                                              => Preset 1 Piano when PSS-A50 is turned on
======== Voiceset Parameter ========
Part param      Main    Dual    Left
 On/Off     :    127     000     000
 Voice No.  :    000     000     014
 Bank MSB   :    000     000     000
 Bank LSB   :    000     000     000
 Prg Change :    000     000     033
 Volume     :    100     100     090
 Octave     :    000     000    -001
 Panpot     :    064     064     064
 Rev Depth  :    008     008     000
 Cho Depth  :    000     000     000
 Trns On/Off:    001     001     001

Common Param
 Cho Type   :    001
 Harm Type  :    001
 Harm Volume:    255
====================================
>
> voiceset                                              => Preset 5 Organ
======== Voiceset Parameter ========
Part param      Main    Dual    Left
 On/Off     :    127     000     000
 Voice No.  :    004     004     014
 Bank MSB   :    000     000     000
 Bank LSB   :    000     000     000
 Prg Change :    016     016     033
 Volume     :    111     111     090
 Octave     :    000     000    -001
 Panpot     :    064     064     064
 Rev Depth  :    010     010     000
 Cho Depth  :    036     036     000
 Trns On/Off:    001     001     001

Common Param
 Cho Type   :    001
 Harm Type  :    001
 Harm Volume:    255
====================================
>
> voiceset                                              => Preset 12 Guitar
======== Voiceset Parameter ========
Part param      Main    Dual    Left
 On/Off     :    127     000     000
 Voice No.  :    011     011     014
 Bank MSB   :    000     000     000
 Bank LSB   :    000     000     000
 Prg Change :    027     027     033
 Volume     :    091     091     090
 Octave     :   -001    -001    -001
 Panpot     :    064     064     064
 Rev Depth  :    011     011     000
 Cho Depth  :    064     064     000
 Trns On/Off:    001     001     001

Common Param
 Cho Type   :    001
 Harm Type  :    001
 Harm Volume:    255
====================================
>
> PSS-A50 turned off and on again to have 1 Piano again
> ots
=========== OTS Parameter ==========
Part param      Main    Dual    Left
 On/Off     :    127     000     000
 Voice No.  :    000     000     014
 Bank MSB   :    000     000     000
 Bank LSB   :    000     000     000
 Prg Change :    000     000     033
 Volume     :    100     100     090
 Octave     :    000     000    -001
 Panpot     :    064     064     064
 Rev Depth  :    008     008     000
 Cho Depth  :    000     000     000
 Fil Cutoff :    000     000     000
 Fil Reso   :    000     000     000

Common Param
 Harm On/Off:    000
 Harm Type  :    001
 Harm Volume:    255
 Apg On/Off :    000
 Apg Type   :    000
====================================
>
> voiceset                                              => Preset 14 Bass
> ots
=========== OTS Parameter ==========
Part param      Main    Dual    Left
 On/Off     :    127     000     000
 Voice No.  :    013     013     014
 Bank MSB   :    000     000     000
 Bank LSB   :    000     000     000
 Prg Change :    032     032     033
 Volume     :    080     080     090
 Octave     :   -002    -002    -001
 Panpot     :    064     064     064
 Rev Depth  :    005     005     000
 Cho Depth  :    000     000     000
 Fil Cutoff :    000     000     000
 Fil Reso   :    000     000     000

Common Param
 Harm On/Off:    000
 Harm Type  :    001
 Harm Volume:    255
 Apg On/Off :    000
 Apg Type   :    057
====================================
>
> func
======== Function Parameter ========
Part param      Main    Dual    Left
 Volume     :    080     080     090
 Octave     :   -002    -002    -001
 Cho Depth  :    000     000     000

Common Param
 Rev Type   :    002
 Cho Type   :    001
 Panel Sus  :    064
 M EQ Type  :    000
 Acmp Volume:    255
 Song Volume:    100
 Transpose  :    000
 Master Tune:    000
 Acmp Split :    054
 Split Point:    054
 Wide Type  :    000
 Rev T Level:    064
 LocalOn/Off:    127
 Ext Clock  :    000
 Metro Sig  :    004
 Metro Vol  :    100
 Demo Group :    001
 Demo Play  :    001
 Demo Cancel:    000
 A Pow Off  :    030
 Battery Sel:    000
====================================
> PSS-A50 turned off and on again to have 1 Piano again
> func
======== Function Parameter ========
Part param      Main    Dual    Left
 Volume     :    100     100     090
 Octave     :    000     000    -001
 Cho Depth  :    000     000     000

Common Param
 Rev Type   :    002
 Cho Type   :    001
 Panel Sus  :    064
 M EQ Type  :    000
 Acmp Volume:    255
 Song Volume:    100
 Transpose  :    000
 Master Tune:    000
 Acmp Split :    054
 Split Point:    054
 Wide Type  :    000
 Rev T Level:    064
 LocalOn/Off:    127
 Ext Clock  :    000
 Metro Sig  :    004
 Metro Vol  :    100
 Demo Group :    001
 Demo Play  :    001
 Demo Cancel:    000
 A Pow Off  :    030
 Battery Sel:    000
====================================
>
> mdb
Not Supported
>
> backup
======= Backup Ram Parameter =======
Part param      Main    Dual    Left

Common Param
 Panel Sus  :    064
 M EQ Type  :    000
 Master Tune:    000
 Split Point:    054
 Demo Cancel:    000
 Master Vol :    008
====================================
>
> chkprm
========== Check Parameter =========
Part param      Main    Dual    Left

Common Param
====================================
>
> freearea					=> "0" is strange, because the track recorder was far to be full
Disk Free Area
 0
>
> jobconid
Jobcon ID
 [Alert  000]
 [Popup  000]
 [Hold3  000]
 [Hold2  000]
 [Hold1  020]
 [Base   004]

 [Active 004]
 [Char Active Jobcon ID   004]
>
> dispinfo
Display ID
 002 (Jobcon ID:004)
>
> swmask
Mask Pattern
 None
>
> ioport
I/O Port
 Port0 0x000A6AEC
>
> getsw
SW History
new
 052 (0x34)
 052 (0x34)
 053 (0x35)
 052 (0x34)
 052 (0x34)
 052 (0x34)
 052 (0x34)
 141 (0x8D)
 141 (0x8D)
 141 (0x8D)
old
>

no comment

Tuesday, November 15 2022

PSS- stereo mod : mod PCB integrated in a PSS-F30 & Flash DUMP

By lionel nollet on Tuesday, November 15 2022, 10:14


I recently bought a second-hand PSS-F30 in order to integrate my stereo mod PCB and confirm it is also valid for other keyboards of the PSS- series.

This time, I made the integration a bit easier because I did not use 1.27mm headers. I directly soldered some wires. I also just wired the missing DACR+/DACR- channel

Only two caps (C601 and C602) are removed from the PSS-F30 PCB.
Thin 5cm long wires are used for the audio signals, and thicker 8cm wires are used for power supply. Wires are twisted, once glued.

PSS-F30_top_PCB_wires_half_glued_small
PSS-F30_top_PCB_wires_glued_small
PSS-F30_top_PCB_wires_small
PSS-F30_under_PCB_small
My mod PCB is put in place on the main PCB using (two pieces, one on the other) thick double-sided tape

PSS-F30_top_PCB_full_wired_small

I'll post the pics later a video with the stereo audio

Now, the PSS-F30 also sounds stereo ! For some sounds (as for the PSS-A50) it's not so remarquable, but for others, it's really pleasant
I expected that the (many) rythms/styles integrated in this PSS- variant would sound much more beautiful, but in fact it's not so outstanding (but sounds better anyway)

Since I also bought this PSS-F30 for experimentations, I desoldered the SPI QUAD Flash memory in order to get a DUMP of its content and have a look on "what's inside", if anything remarkable
I also plan to do the same with the PSS-A50 Flash memory in order to compare them and maybe -replace some voices- or use voices from PSS-F30 on PSS-A50 ?

I'm not sure that someone ever performed a DUMP of the flash for this/these instrument(s), so maybe "I'm the first" !
For copyright reasons, I won't share the content, but you'll get the information about how I proceed

In order to make easier the removal/plug/modifications on the Flash, I mounted it on a 1.27mm male header (modified), and plugged on a 1.27mm female header soldered on the PSS-F30 PCB.

PSS-F30_1.27_headers_small
PSS-F30_Flash_On_PCB_small
Currently, it's not so clear for me about the architecture, especially if there is part of the SW in the YMW830-V Processor (it has a JTAG, so most probably a code memory). Most propbably, there could be a bootloader in the YMW830 (maybe the same for all PSS-series). Nothing happens if we turn the PSS- ON (while the Flash is not connected).
It looks strange because it's not possible to enter in a "test mode" (explained in the Service Manual), and this test mode is supposed to help to identify (error code) if the Flash is not connected. Anyway.

I'll also identify where the non-volatile data is saved (volume setting for example) inside YMW830 or in the external Flash Memory. Just needed to perform a DUMP with the default settings, and another DUMP with the value changed. If nothing changes in the external Flash, it means that the data is stored in the processor.

I'll do the same with the PSS-A50 in order to try to locate the value (if any for the PSS-F30) that could be used to define the "Touch sensitivity" : Currently, the PSS-F30 HW doesn't implement dynamic touch. The chance is that the keyboard PCB and the rubber keys/contact is nearly the same as the one of the PSS-A50 : dual keys (press and release) are present. Just, a diode is missing for each key. Also, just one wire is also missing between the Keyboard PCB and the MainBoard PCB. So maybe, it might be possible to active a dynamic touch in the PSS-F30, if the SW implements the function and if it can be activated dynamically by changing a value in the memory.

I also noticed that the midi messages are sent out on the Tx pin of the YMW830, and that could help to implement a wire(less BT) midi to record what you play !
Talking about midi, I could easily extract (from the flash DUMP) the midi files for the 30 songs stored in the PSS-F30. I did it for the first song. In the DUMP, the songs are stored exactly like any .mid file on your PC.

To be continued...

one comment

Saturday, October 8 2022

PSS A50 stereo mod : mod PCB integrated PSS PCB : details

By lionel nollet on Saturday, October 8 2022, 20:42


I'm happy with my integration. I now enjoy a lot playing with my PSS-A50.
It's a real pleasure, especially when playing with headphones.

To begin, have a look on my latest video : I hope you'll enjoy the all world around known "Get Lucky".
Maybe it's the first time this song (kind of intro in fact) is played (at least on youtube) on a Yamaha PSS-A50.
And for sure, it's the first time on a stereo PSS-A50.
https://youtu.be/QGN_a2lWWMY

Please also have a look there (he was inspired by my mod) : https://darekasan-net.hatenablog.com/entry/2022/09/19/045703
(Chrome can translate in case)
DarekaSan implemented an additional jack output using a PCB shield for Raspberry Pico (it seems the PCB is only available in Japan), and cut the PCB shield to isolate the lowpass filter section.
In this implementation, no stereo sound will be connected to built-in PSS-A50 headphones jack.


About my own PCB integration, here are some details :

At first, I soldered two 4 pins male headers (1.27mm) on my PCB (H3 and H4).
Then, I inserted two 4 pins female headers (1.27mm) and soldered two wires : black on female H3 for AGND and red for female H4 +5V

Mod_PCB_w_wires_small
The pins of the female headers have been bended in order to lower the header profile

Also, don't forget to solder on the mod PCB a 6 pins 2.54mm female header (or, in case, only 2 pins may be enough just for the R+ / R- RIGHT channel) in order to plug the wires coming from of the Yamaha SWLL YMW830-V IC.

I also put some "Koptan" (the Chinese low quality Kaptan) in order to have an insulation with the PSS-A50 PCB; because the idea is :

Mod_PCB_wo_wires_w_kaptan_small
1. Plug the female headers on the Mod PCB male header
2. Put some hot Glue on the PSS-A50 PCB (at the places where the two 4 pins headers will be, once the mod PCB will be pluged on the 2x4pins male header I soldered the last time on the PSS-A50 PCB
3. While the glue is still hot (don't waste time), plug the PCB mod at the right place on the PSS-A50 PCB
4. Solder the red and black wires on the Capacitor C035 pins (on the bottom of the PSS-A50 PCB)


The assembly will look like this :

Mod_PCB_Integrated_PSS_A50_Small
If you unplug the mod PCB, here's how it looks like :

PSSA50_PCB_predisposition_for Mod_PCB_small
That way, the mod PCB is mecanically firmly inserted in three locations, which makes it stable from a mecanical point of view.

Anyway, I'm aware the fragile location (or weak point) is the 2x4pins male header soldered on the PSS-A50 PCB, because the original PCB pads are not suited to hold this kind of component. In order to make the meca more robust, it would be relevant to put some cyanoacrylate glue around the PSS-A50 PCB and the bottom of the 2x4pins male header (once soldered of course).

one comment

Wednesday, August 10 2022

PSS A50 stereo mod : mod PCB integrated on PSS PCB

By lionel nollet on Wednesday, August 10 2022, 13:45


I finished my integration work.

Mod_PCB_Integrated_PSS_A50_Small

Mod_PCB_Integrated_PSS_A50_Vertical_Small

The question we would like to know the answer : Will it work ?
The audio/video answer will be provided in a new Youtube video I'll prepare later.

The written answer is : "Yes, it works fine, as expected", with no reworking needed. Also, no additional noise (from ear, subjective, with headphones) compared to the genuine PSS-A50.
I'm really happy with the result and the nice integration.

1. I'll publish later an article with the details of this integration.
2. I'll also publish later another possible integration of this PCB, without (the complexity of) using 1.27mm headers, but (easier) just using wires instead.
3. I'll also publish later an article about audio noise when recording from the usual audio input of the PC, if the PSS-A50 is USB connected to your PC (sending midi to the PSS). Solution : use audio 1:1 transformers (usually 600 Ohms) in order to de-couple the ground

Please also have a look there (about my mod) : https://darekasan-net.hatenablog.co...
Or there (about a midi Bluetooth mod) : https://darekasan-net.hatenablog.co...
(Chrome can translate in case)

3 comments

Monday, August 8 2022

PSS A50 stereo mod : PCB delivered !

By lionel nollet on Monday, August 8 2022, 08:38


Got the PCB samples ! Last step is now coming !

PCB_JLCPCB_small

one comment

Friday, July 29 2022

PSS A50 stereo mod : latest video : Back to stereo

By lionel nollet on Friday, July 29 2022, 16:45


Here's my latest video.

I let you enjoy the difference, especially in the 2 last minutes of the video where you can hear the difference between the mono and the stereo sound !
Headphones are recommended to enjoy the flavour !

https://www.youtube.com/watch?v=bAUvMFMGzyE

Latest_Video_Back_to_Stereo_small

one comment

Thursday, July 21 2022

PSS A50 stereo mod : mod PCB Assembly with PSS PCB

By lionel nollet on Thursday, July 21 2022, 13:32


The PCB design is close to be frozen
On the top left corner, the 2x4 pad for the female header (bottom of the PCB)
In case your iron ability is not proven to solder a 1.27mm header on the PSS PCB, it's still possible to solder just 2 wires from the PSS PCB (near the headphones amp; C601 and C602 to be removed) and the 2.54mm pads on the left side of my PCB (OUT : R- and R+)

Two wires are also needed from the DACRPP/DACRMM pins of the Yamaha SWLL YMW830-V IC; to be soldered on the 2.54mm pads on the bottom side of my PCB (IN : R- and R+)
And last but not least, two wires for the power supply (to get from the capacitor C035 (under the PSS PCB) on the 2.54mm pads on the left side of my PCB (+5D, AGND)

PCB_Top_with_OpAmpPCB_small

I used two 1x4 pins 1.27mm headers (bottom left and top right) that are used as a spacer (same height as the 4x2 male + female header) in order the PCB keeps well in place (two 4 pins male headers soldered under my PCB, and two 4 pins female headers glued on the PSS PCB)

The PCB has a breakable part in order to be plugged on the male 2x4 header, soldered on the PSS PCB (in case we decide to remov the mod and find the original mono sound configuration.

PCB_Top_Small
PCB_Under_Small

one comment

Monday, July 18 2022

PSS A50 stereo mod : The mod PCB

By lionel nollet on Monday, July 18 2022, 18:55


This is my PCB project, but still brainstorming if I can improve in order to ease the connection, especially with the Yamaha SWLL YMW830-V. It would be nicer to have a 0.65mm pitch kind of header or connector, instead of wires.

Here's a first (and old) proposition for the PCB : https://www.youtube.com/watch?v=z9ano9jffsM

The newer implements the pads for the 1.27 2x4 pins header
The PCB has pad jumpers in order to be able (when using two PCBs, stacked one on the other) to recreate both LEFT and RIGHT channels. Normal use case is to feed the existing Head Phones amplifier (RIGHT Channel input) on the PSS-A50 PCB with the OUT R +/- of my PCB.

Target is to connect it between DACRPP/DACRMM and R+/R- inputs of the HeadPhones amplifier IC601.
Power Supply will come from +5V and GND under the normal size cap C035 (+5D / AGND) located on the right of the 3.5mm HeadPhones Jack; near the screw.

one comment

PSS A50 stereo mod : The 1.27mm 2x4 pins mod

By lionel nollet on Monday, July 18 2022, 18:03


First step is to remove C601/602/603/604 and adjust and solder a 8 pins 1.27mm male smd header (2x4) in order to have something clean :
C601_C604_Removed_Small
The easy way for me to desolder the caps is to use two irons, in order to prevent heating too much. My brand new TS100 with the TS-I tip on the left, and my old good JBC30S with R05D round tip on the right (the smallest available).

This pic is for illustration only (the header is already soldered; The iron tips are on another cap, just for the picture).

Desoldering_Cap_Small
Then, the 1.27mm header has to be adjusted before beeing soldered :
The two top-right pins stay at their nominal length
The two pins below are cut at the half length
The four pins on the left are cut that way only one third of the length remains

Header1.27_Full_Small
Header1.27_Cut_Small
The picture below is for illustration only, since the header is not located on C601-604. As you can see, bottom parts of the smd header is located exactly on the pads.

Align_1.27Header_Cap_Small
Alignment of the header is a crutial operation, because the pads on the PSS-A50 PCB are not exactly designed for this kind of header. It is very risky if you are not experienced or if you don't have the right iron / tip. Don't do it if you are not used or fell uncomfortable.

The result after the soldering operation :
(it looks dirty on the picture because of the flux, but the header is in a well condition and not melted).
During the soldering, it's mandatory to have the female conterpart connected on the header in order to avoid having the pins bent.
Header_Male_Female_Small
Header1.27_Soldered_1_Small
I used Kaptan tape in order to protect the headphone amplifier from the heat and avoid any solder projection on the 4 very tiny pins of the chip. This is a chinese "Koptan". The quality is not very good and might melt if heated too much. Once the header was soldered, I removed the tape. I also used a flux gel in order to ease the soldering and ease the capillarity.

Header1.27_Soldered_1_Small
Once soldered, check for every pins, and also cross-check that there is no short circuit and no open connection
1 x x 8 ("L OUT-" written on the PCB, from existing OpAmp)
2 x x 7 ("L OUT+"written on the PCB, from existing OpAmp)
3 x x 6
4 x x 5

pins 6 and 7 are shorted, because of the PCB trace, that's normal.
pins 5 and 8 are shorted, because of the PCB trace, that's normal.
pins 1 to 4 are going respectively to the close pins of the headphone amplifier IC
You can use a needle in order to have a thinner test point with your multi-meter in order to check the continuity and absence of short circuit

one comment

PSS A50 stereo mod : Breadboard implementration for missing mono right audio path

By lionel nollet on Monday, July 18 2022, 17:55


The main step I went thru was to remove C601/602/603/604 and adjust and solder a 8 pins 1.27 male smd header (2x4) in order to have something clean, easier to prototype in, and, in case I want to go back to the initial setting : plugging a 1.27 female header on it, with 4 caps soldered on the header.

I implemented on a breadboard the missing Symmetric Low Pass filter for the right channel. Surprisingly, even with a basic LM324, the result is very satisfying, no major difference with the existing left audio from a noise point of view.

Breadboard_Mono_Right_Small

one comment

Saturday, July 16 2022

PSS A50 stereo mod : more intrusive, bring the right audio signal to the jack

By lionel nollet on Saturday, July 16 2022, 22:41


There's a nice article here from PJD about my experimentation : http://sandsoftwaresound.net/yamaha-pss-a50-stereo-mod/
It is very pedagogic with nice illustrations. I also think the differential approach is relevant in order to lower the noise.

I retro-engineered the Op Amp block, and... I get the same schematics.
Not so a suprise because the global picture is nearly the same as the one we can see in the PSR-F50 service manual.
I confirm also the resistors values and the 1µF capacitors, but I could not get a solid value for other small caps; that's why the schematics provided by PJD is interesting, because even if I try to reconstruct using the circuit simulation https://www.circuitlab.com/; it's better to have the good picture.

Retro_Schematics_PSS_A50_Small
The only major difference I have on my side is R501 and R502 for which I have 1,5k (instead of 2,2k, which is not dramatic since the divide bridge is still 1/2 in order to add the half direct voltage).

My initial idea with this mod was a kind of proof of concept, in order to identify if it's worth going deeper (i.e check if the stereo samples were present and played back). Which is the case.
Now, my target is to create a small PCB with the missing op amp, and replicate the mono circuit. Because of the components shortage, and since I plan to make my PCB via www.jlcpcb.com, I identified three possible alternatives : TL972, TS462 or MAX4477. For availability and cost reasons, I'll use the TL972.

Now the question is : only replicate a "Right" channel or replicate both in order to have a common design on the same PCB ? Well, I'll replicate only one channel. In case performances is worse (ground loops, noise... it might be easier to fix), and in case performances is better, I will just have to use two PCB (probably stacked, so I have to think about the design) : one for the new right channel and on for the left channel that will go into the headphones amplifier.
The left channel path for the loudspeaker will remain, using the original NJM2740M Op Amp

Most probably, I will unsolder the 4 caps C601 to C604 and wire the outputs of my PCB. TBC.

Now the Block Diagram with the integration of the Right channel looks like this :
BlockDiagram_PSS_Mono_Right

Thursday, July 14 2022

PSS A50 stereo mod : audio line-out

By lionel nollet on Thursday, July 14 2022, 22:09


I focused on this PSS-A50 product because I read many interesting articles from "pjd", on his site : http://sandsoftwaresound.net/yamaha-pss-a50-look-inside/
When focusing on the PCB, I noticed that only pins 1 and 2 (Left channel : DAC L+ and DAC L-) of the Yamaha YMW830-V Processor and tone generator (IC101) are wired on the PCB, not pins 5 and 6 (DAC R- and DAC R+). It is known that this product is "mono" (one speaker, and despite the 3.5mm headphones jack is stereo, the sound outputs "mono").
This product is declined in three variants : PSS-E30, PSS-F30 and PSS-A50). If we consider the two first variants are entries and design to cost, it can be acceptable to have a mono product.

I was wondering if the "high" variant PSS-A50 (it has a MIDI feature that the two others don't have) would have the same limitations ("mono" sound), and especially if the tone samples in the 16Mbits flash memory are mono or stereo.
My first proof of concept will consist in recreating a stereo audio path from the YMW830-V pins 1,3 (left) and 5,6 (right). For that, I would re-create a stereo asymmetrical low pass filter in order to feed the audio signal in any audio "Line-in". The audio input of my PC in my case.
At this step, I did not retro-engineered the PCB to recreate the schematics. Just wanted to check if the sound outputs stereo or not.

The Block Diagram of the audio section looks like this :
BlockDiagram_PSS
In order to get the stereo signals again, I wired the DAC L+/DAC L- and DAC R+/DAC R- outputs of the SWLL to an external active Low Pass filter with a basic LM324 and a few resistors/capacitors (same schematics as for PSR-F50... visible in the service manual of the PSR-F50)
BlockDiagram_PSS_Stereo_LineOut
I wired the DAC L+/DAC L- and DAC R+/DAC R- outputs of the SWLL to a 2.54mm header : PSS-A50 PCB small
and connected this header to my Low Pass filter implemented on a breadboard : Stereo Breadboard small
The result is visible and audible in this video : https://www.youtube.com/watch?v=cwDBfVab3vc

one comment

A few words about me

By lionel nollet on Thursday, July 14 2022, 21:37


I'm about 50 years old, I live in France, in Paris area.

I work in the automotive field, especially in the infotainment systems.

https://www.linkedin.com/in/lionel-nollet-b780a55b/

My interests are in electronics, hardware, embedded systems programming.
I'm also learning again the piano, on a Yamaha P-125.

I bought a few weeks ago a Yamaha PSS-A50, especially "for fun" but also in order to analyze the HW, and potentially to "mod" it.