Rigol DS1054Z Repair dead keypad help needed!

[vlad0623]

Hi,
I've got a Rigol DS1054Z that has a dead keyboard module (Keys and LEDs are not working). I've tried using it with Rigol UltraScope software, and everything else seems to be working fine. I disassembled it and checked it visually. Also, I've measured the power lines for the keyboard module. All seems to be fine.
So I assume that the problem is the FPGA itself. It has a built-in flash, and it could be a dead chip or the internal flash is corrupted somehow. I'm not sure what was done to this device to get it in such a state. I've found here on the forum a few threads about reverse engineering this scope. From that, I've learned that It seems like this keypad FPGA could not be reprogrammed by the firmware upgrade. Also, the keypad module should use the UART line of the microcontroller.
I found one of the lines and tried to measure the data transmission when I switched channels on and off from the PC. It looked like it was transmitting some data to control the corresponding LEDs. There wasn't any responses from the keypad itself.
The ideal solution would be to rewrite the code for the original FPGA, but I have no experience with that, so it might be challenging. As an alternative, I could probably use some MCU to emulate the original keypad behaviour and replace that FPGA.
I really need some help with that. I have no experience disassembling firmware. So it would be really helpful if somebody could provide a protocol for this keypad and UART baud rate either by disassembling the firmware or by capturing the traces with a logic analyzer.

[m k]

Is it possible there is some sort of front panel lock?
Don't know how it would affect the LEDs.

[adinsen]

I haven't read the threads on reverse-engineering, but it sounds like your best bet wold be to either remote control the scope from a PC or getting a new board. Reverse engineering the protocol to make a custom PCB would probably be more expensive than getting a new keypad board from Rigol.

[vlad0623]

Is it possible there is some sort of front panel lock?
Don't know how it would affect the LEDs.

I don't really think so. I'm unable to find such an option in the user's manual. During boot, all the LEDs stay off, which is wrong considering there should be a self-test. I also clearly see some communication happening on one of the traces that look exactly like some LEDs should switch on and off a few times before it's finally booted up.

I haven't read the threads on reverse-engineering, but it sounds like your best bet wold be to either remote control the scope from a PC or getting a new board. Reverse engineering the protocol to make a custom PCB would probably be more expensive than getting a new keypad board from Rigol.

I don't really see how it could be very expensive? I think of maybe studying an FPGA and trying to reprogram the onboard one. Or use something like Atmega or STM8/STM32 for that. I think it's quite possible to make it for under $10, a keypad module would be much more expensive. It will definitely take some time, but it's not quite an issue, because electronics is more like a hobby for me. So it might be a fun project and an opportunity to finally try learning some FPGA stuff.

But if you or somebody could suggest any options to buy such a keypad (or maybe a preprogrammed FPGA or a physically damaged module with a working FPGA) that is not very expensive, I'm open to offers. Please post it here or in a PM.

[adinsen]

But if you or somebody could suggest any options to buy such a keypad (or maybe a preprogrammed FPGA or a physically damaged module with a working FPGA) that is not very expensive, I'm open to offers. Please post it here or in a PM.

I checked ebay and ali and there are none. I think Rigol will supply them - at a price. Sounds like you have a fun project there, good luck!

[GGMM]

Hello,

Demonstrate and check the panel before.
I think I saw that sometimes an encoder can short-circuit or something? It shows and cleans.
Let’s do a search on the net.
Some examples here if it looks like it.
See also the flat cable if not cut.

cdt

[GGMM]

Look also here, it's Hantek panel Oscilloscope, but that will give some ideas for tests



cdt

[vlad0623]

I have double-checked everything and found a suspicious VCC that was around 0.7v. I've looked to a datasheet and found that it should normally be around 1.5v, so I've tried to power it from a bench supply (removed that SOT23 and soldered a wire), and IT WORKED!!!. It's obvious that the problem is caused by a failed voltage reg circuit. So it's now a much more trivial repair.


Now I need to figure out what that SOT23 part is and how that circuit is supposed to work. Or maybe just replace that with something like 1117-1.5v? Does somebody know what it could be? It looks like some BJT with a base connected to a divider (01C measures as 10k and 30C is around 20k).

[adinsen]

I have double-checked everything and found a suspicious VCC that was around 0.7v. I've looked to a datasheet and found that it should normally be around 1.5v, so I've tried to power it from a bench supply (removed that SOT23 and soldered a wire), and IT WORKED!!!. It's obvious that the problem is caused by a failed voltage reg circuit. So it's now a much more trivial repair.

Hooray! Great to hear  I hope someone else can identify the failed component as I don't have a clue 

[wasedadoc]

I have double-checked everything and found a suspicious VCC that was around 0.7v. I've looked to a datasheet and found that it should normally be around 1.5v, so I've tried to power it from a bench supply (removed that SOT23 and soldered a wire), and IT WORKED!!!. It's obvious that the problem is caused by a failed voltage reg circuit. So it's now a much more trivial repair.
(Attachment Link)
(Attachment Link)
Now I need to figure out what that SOT23 part is and how that circuit is supposed to work. Or maybe just replace that with something like 1117-1.5v? Does somebody know what it could be? It looks like some BJT with a base connected to a divider (01C measures as 10k and 30C is around 20k).

The 'Q' designator is conventionally used for transistors. Datasheet for G1 marking: https://www.s-manuals.com/pdf/datasheet/b/f/bfs20_zetex.pdf . Nothing special.

[vlad0623]

The 'Q' designator is conventionally used for transistors. Datasheet for G1 marking: https://www.s-manuals.com/pdf/datasheet/b/f/bfs20_zetex.pdf . Nothing special.

Yes, it's obvious that it's a BJT. I have the S8050, which I already soldered there, and it seems to work pretty well. The measured voltage is around 1.47v to 1.5v. I just can't understand how such a simple, low-power circuit could fail? Is it a factory defective transistor? Since I couldn't really find any other cases of such failures online, I think it's OK to just leave it with the original design. What do you think, guys?

[GGMM]

Hello,
 i thing it's a NPN transistor
See if some others on mother board and test with multimeter to compare values of jonctions with and equivalence transistor
do some rétroingénierie

some indication here ??


 Dave speak of smd code in the video , I had to lower the sound if you go all the way. lol!!!

here i found the frontboard schematic
https://fr.scribd.com/document/383431220/Rigol-DS1054Z-Schematic-FrontEnd

cdt

[GGMM]

Happy it's work

cdt

[vlad0623]

Thanks to all, especially to GGMM (you encouraged me to check everything once again).
By the way, I got this scope on US eBay for around 170 USD, including shipping (without probes). I think it was worth the risk.

[GGMM]

hello,
There are some goods things on Ebay with scope...
For me, 1052E for 60€..Lol
For your 1054Z, humm.... See youtube if you want to perfect it, you know the word to add in your search. LOL

cdt