-
Tektronix AFG3021 died. Firmware update gone horribly wrong :-(
Posted by analogRF on 27 Jan, 2018 15:38 -
I got this AFG3021 with FW 1.0.4 Sept 2005 off ebay which had little self cal issue but not a biggy. However, yesterday I tried to update the firmware through the USB flash. I tried 3-4 usb flashes I had around (all >=8GB) and with FAT32 format and none of them worked. They were not recognized and the generator could not read or write to them. However, at some point when a USB flash was in there (with FW on its root) I turned on the generator. It booted up normally and got to the first generator screen (with sine wave showing etc...) but hanged! Would not respond to keys and there was no message on screen. After 30-40 seconds or so I decided to turn it off and on again :-(
Unfortunately it didn't boot anymore and has died completely now. As soon as the power is turned on, all LEDs on the panel light up and the screen is turned on but is blank white (well more like blank pinkish color) and sits there. I tried to connect it to computer from the USB on the rear side but there is nothing in there. The computer would not even make the connection sound. This model does not have GPIB or LAN options (AFG3021B has them)
Of course my first thought at this point is that the flash memory that holds the firmware and system status is corrupted.
There are two chips with tektronix part number stickers on them, I think they must be the flash memories as other chips are well known. see the attached pictures.
Please please somebody help me out here. What can I do to restore this unit? Is there any way I can re-flash the memory on board through something like ICSP or any other way? Is it possible to buy and replace the flash memory chip(s) pre-programmed? Or get them blank and program them and then solder them on board. anything that can be done?
Is it possible that some hardware has failed and not the firmware? I didnt check the power supply coming to the cpu board...
-
by the way, that edge PCB connector (next to usb) does not connect to anything (see the last picture) I guess that is used by the manufacturer to troubleshoot or program these units
also the two parallel pin headers (above the usb to the right) do not connect to anything.
-
U202 which is an Intel flash chip holds the firmware. If you scope it's pins do you see any activity?
Btw, do check the power supply before anything else. -
Sorry for the long delay. I finally had a chance to probe around the main board and see what is going on around the RAM and Flash memory and the CPU. Here are my findings but I cannot draw any conclusion as to whether the RAM chips are blown or the flash memory is blown or the firmware is corrupted.
Any help would be greatly appreciated. Specially if someone with AFG3021 or AFG3022 could probe around the pins and let me know how my findings are different than his/her working unit.
On BOTH of the two Micron CMOS RAMs (both are x16)
CKE, DQML, DQMH, CS, WE, CAS, RAS pins are all HIGH (constant 3.3V)
BA0,BA1 and address lines A7-A11 are LOW (constant)
Address A0-A6 are active (live signal)
On U208 (left IC) I/O data lines are ALL stuck at LOW (no activity)
On U204 (right IC) I/O data lines DQ0-DQ7 are live (active) but DQ8-Dq15 are all HIGH (seems more like Hi-Z)
As you can see all the controling signals on both chips are exactly the same but they are showing different data on the I/O data lines
Specially the CS is stuck at HIGH, too!
On the FLASH memory:
CE1, CE2, BYTE# pins are LOW
CE0 pin however is live and active, so is A0 pin
also A1-A8 pins are live (ctive) but A9-A23 are all stuck at low. But I don't think its unusual
VPEN pin #15 is stuck at HIGH and that worries me the most :-(
RP pin is HIGH
WE pin is HIGH
OE pin is live (active signal)
All I/O data lines DQ0-DQ7 are live and active but DQ8-DQ15 are HIGH
I will attach the datasheets for both chips
It is very hard to probe the CPU pins but I could see that it is active and for example address lines A0-A8 have activity
any idea on how to proceed is greatly appreciated
-
by the way the other chip with TeK part number sticker on it (U108) is a CPLD from LAttice (ispMATCH series)
the part number is LC4256V-75TN100-101.
one thing that seems strange to me is that ALL the control signals, enabling/disabling signals, etc... to the two RAM chips are exactly the same
even the address lines are the same BUT one of them shows signals on its I/O data lines while the other chips data lines are all stuck at LOW.
Maybe the CPU is trying to write to the RAMs but cannot do it somehow...all the command lines and WE are stuck at HIGH on both RAMs
any thought?
can someone with a working unit probe around the chips and kindly let me know his findings?
-
it seems I am the only person writing in this thread for myself
I am beginning to think that the failure perhaps had nothing to do with the USB plugged in when I turned it on
because: (1) that same USB flash was not readable by the instrument. When I first put it in, it could not read from it.
(2) I had plugged in 2 other USB flashes and turned the instrument on before and nothing happened. I have no reason to believe
that this USB flash was any different.
(3) there was no message when it was booted, just hanged (basically died) and the next time I turned it on, it never booted again
Still it might be a case of firmware corruption, but what are the other possibilities that I can check? RAM failure? CPU failure?
and if it is the firmware corrupted, how can I know? what would be my options?
thanks
-
It's kind of a crap shoot at this point what the problem might be. If the power supplies are all good and it didn't perform a firmware update then the highest failure point is likely the ram. I don't have one to compare to. If the ram is available and cheap it might be worth swapping it out. You could also remove the flash chip and dump it to see if it looks ok.
-
I have ordered the RAM chips ($0.83 each!) and I am waiting for them to get here.
I will have shot at replacing them first.
With regard to reading the flash, I don't exactly know how to do it
I wish it was a microwave/RF circuit issue or some analog electronics :-(