Nord Lead 3 problem, does not boot [SOLVED]

[hozone]

I can set triggers but event if i capture a few output it became a little bit long to analyze the CMOS output bits.
It could be the CMOS. I can ask the official Italian support the repair cost. Cause if it's the ROM.. i don't think it will be available. What's the SMF (.mid file) here? https://www.nordkeyboards.com/downloads ... ord-lead-3

[pterm]

I agree it will be difficult to analyze the EPROM contents.
Please check with the Italian Support if it is possible to acquire the ROM. It must be pre-programmed (factory programmed) as it cannot be programmed in circuit. That is, a blank (empty) ROM is useless.
I don't know what the SMF (.mid file) is. - I guess it is the data file for reprogramming the OS. Please do not attempt this. It cannot recover the ROM (assuming ROM is the problem).

I reviewed the trouble shooting guide and found something item to check:
Please check that D0, D6, and D9 are pulled to GND with 10KOhm resistors --measure the resistance, but do not probe the resistor directly - the solder or part might be cracked and probing the part directly might close the crack.
These resistors configure the MCU's boot behaviour.

[23skidoo]

I agree that support would be the next step. I am curious, however, what exact EPROM chip is in there - can you get the part number or a good clear photograph of the top of it? Mine has a sticker over it and I don't have my board open to remove it - my last detailed photograph just shows the sticker. It's mounted in a removable socket, so that would indicate it could be popped out and read with a microcontroller, if one were daring enough.

[hozone]

No luck, D0, D8 and D9 are pulled down with 10k. I think I'll ask to Italian support, If they are going to replace my Board (if they have one old stock or repaired)or if they ask me too much (considering this synth worth 800/900EUR), I can try replace the MC68331, found on aliexpress at <10EUR, even if It could be the ROM.

@23skidoo
Find attached the EPROM, it's an SST 27SF010, mine had the N3 0.24C sticker on top. Ine is soldered on the board, no removable sockets on my board.
By the way my board is NORD LEAD 3 MAIN v.1.1
Beside this chip there is the AM29F0808 which i think is the OS flash.

[23skidoo]

That's a 128Kx8 flash chip, so that's almost certainly the OS, yes. Interesting that yours isn't socketed, my board has the same revision (1.1) and has that chip in a socket. It's very nearly pin compatible with the SST39 series, but those are 5V programmed with an unlock sequence whereas this one is erased with a 12V high voltage program line instead.
The other chip (which is an AM29F080B, not -08) is a 1MB flash chip with uniform sectors (and a very respectable 1M program/erase cycle minimum). This one is almost certainly patch storage.

[pterm]

@23skidoo - Thanks for identifying that ROM. I assumed it was the same EPROM used on NE2 (same package, same pinout). That EPROM only had 20 year data retention which made it seem a likely suspect for the problem.
This flash device's datasheet reports a 100 year data retention making it far less likely (but not impossible) a suspect.

@hozone
If the Flash device is reprogrammable, it requires Vpp (pin 1). Please check this pin with the oscilloscope during start up. This might have a FET to switch Vpp ON/OFF, so I expect it to be either 12V or 0V. --Without a NL3 schematic I must guess.

Also please check the +12V rail with your oscilloscope. If +12V is very noisy, it might be coupling noise into the boot Flash causing data bus errors.

Please post a large photo of the entire main board. I think the likelihood of a corrupt flash is much lower than I suspected previously. This makes me think the problem exists elsewhere, but I am running out of ideas, so I hope the photo of the main board will give me (or others) ideas.

[23skidoo]

Err, you don't want to put 12V anywhere near that pin. This EPROM doesn't have an erase safeguard and will happily blast it's memory to shreds with simple pinstate combos. That should only ever happen when updating the OS.

[23skidoo]

I have to say I am fairly impressed with how well specified the components in these synths are. They did not build them to last only 20 years.

[pterm]

Err, you don't want to put 12V anywhere near that pin. This EPROM doesn't have an erase safeguard and will happily blast it's memory to shreds with simple pinstate combos. That should only ever happen when updating the OS.

I totally agree. Users must not apply 12V to this pin. Sorry if this was unclear.
I expect to measure 0V or 12V per the table below: The 12V must only be applied if the chip is being erased / programmed, so I hope hozone measures 0V.

I asked for an oscilloscope measurement to see if any noise is coupled onto this pin from the 12V rail. If the OS resides on this device, the NL3 likely has a transistor to switch 12V onto this Vpp pin to allow updating the OS.
I noticed hozone did not list 12V among the rails he checked. If the OS resides on this chip, then the NL3 provides a path from the 12V rail to this pin. This means the goodness of 12V power rail needs to be checked.
If the 12V rail malfunctioned, the control of the Vpp transistor might be compromised allowing resistive coupling to the Vpp pin. If 12V is noisy, capacitive coupling onto Vpp remains a possibility.

Really, I'm hoping we can find a cause other than a corrupted Flash. Thanks for the second set of eyes on my advice. With possible language barriers, I think it essential to maximize clarity.

[hozone]

Thank you.
I've checked the pin1 of 27SF010. It's 0.30ohm (so directly connected) to 5V line, that is 4.985 read on my UT61E.
Italian suppor tells me they have to check the board, and if they are not able to solve the issue they will ship to Sweden.
I'm starting thinking this is the only way. I can replay the MCU, but the more i touch, the more the support will not take this board in consideration, this is just a supposition.
Attached high resolution board picture.