Author Topic: Agilent 34461A corrupted flash  (Read 23005 times)

0 Members and 2 Guests are viewing this topic.

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #175 on: January 11, 2025, 03:55:59 pm »
I tried connecting through a network switch and also through my wireless router LAN connections and cannot make a connection to the 34461a. Front panel USB does not work either. Rear USB I don't know as boot log is showing I have 1 USB connection but it seems like it's not working. Not recoginized if I plug it into my laptop USB. So, this looks like if I need to use Ymodem if I want to attempt installing a new kernal.
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #176 on: January 11, 2025, 07:08:07 pm »
mmm seems the problems levels have increased

some talked about spear mcu damages, flash corruption,    some about some voltage regulators who failed in the mcu section ...  dang

it would need a working one and  probe into many places to get voltages, data points.  memory dumps etc ...      well done keysight well done   :--
« Last Edit: January 11, 2025, 07:10:36 pm by coromonadalix »
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #177 on: January 13, 2025, 12:30:43 am »
Tried to convert Nk.bin to nk.nb0 using viewbin.exe to get start and length which did not work giving error  "Missing B000FF) Not a BIN file".
Then was going to use cvrtbin.exe afterwards but assuming that won't work either. What's the reason for this? I'm trying these utilities because I don't want to have to install another programming environment to do the conversion or extraction of nk.nb0.
analogRF, you mentioned you have a good nk.nb0 file you can share? It would be greatly appreciated.

update... So, i decompresses nk.bin, used a small Russian made utility to create the nk.nb0 file. Then I uploaded it via Tera Term Y modem using loady waited the long 40 minutes. After it was done I ran "go 0x00320000" and nothing happened.  :(. Could be something wrong I did maybe like not have a good nk.nb0 file. Tried bootm also after another 40 minutes got wrong image format error.
« Last Edit: January 13, 2025, 05:31:50 am by Eng_tech »
 

Offline analogRF

  • Super Contributor
  • ***
  • Posts: 1024
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #178 on: January 13, 2025, 12:39:13 pm »
Tried to convert Nk.bin to nk.nb0 using viewbin.exe to get start and length which did not work giving error  "Missing B000FF) Not a BIN file".
Then was going to use cvrtbin.exe afterwards but assuming that won't work either. What's the reason for this? I'm trying these utilities because I don't want to have to install another programming environment to do the conversion or extraction of nk.nb0.
analogRF, you mentioned you have a good nk.nb0 file you can share? It would be greatly appreciated.

update... So, i decompresses nk.bin, used a small Russian made utility to create the nk.nb0 file. Then I uploaded it via Tera Term Y modem using loady waited the long 40 minutes. After it was done I ran "go 0x00320000" and nothing happened.  :(. Could be something wrong I did maybe like not have a good nk.nb0 file. Tried bootm also after another 40 minutes got wrong image format error.

I sent you the nk.nb0 file for 2.17. The FW in your device right now is 2.17. if you use another bin file it wont work
but where did you load the file in the memory? I am curious how did you come up with "go 0x00320000"?

you have to load it into 0x0361000 ("loady 0x0361000 115200" if you want to use ymodem) and then after finished loading with no errors
use "go 0x00362000" to boot

but like I said, you can use tftp if you set up a tftp server and put the nk.nb0 file in the shared directoy and then use
the dhcp command at p510> prompt.
believe me it will connect to ethernet and will get an IP instantly and download the file. Use the same address 0x0361000 for downloading the bin file
into memory.
still I am puzzled why your device thinks there is an ethernet cable connected when there is none but who knows...
let's see if it boots



 
The following users thanked this post: voltsandjolts, coromonadalix, Eng_tech

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #179 on: January 13, 2025, 01:16:43 pm »
Thanks, very much appreciated, I was looking at past posts and looked like it was mentioned to load the nk.nb0 file at 0x00320000. Anyway, thanks for pointing out my mistake because it looks like I was doing this all wrong. I think I tried DHCP command and didn't do anything but I'll try it again and see if it gets working. Really, thanks for the help. At work so, won't be able to try anything till tonight EST.
 

Offline analogRF

  • Super Contributor
  • ***
  • Posts: 1024
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #180 on: January 13, 2025, 06:06:53 pm »
Thanks, very much appreciated, I was looking at past posts and looked like it was mentioned to load the nk.nb0 file at 0x00320000. Anyway, thanks for pointing out my mistake because it looks like I was doing this all wrong. I think I tried DHCP command and didn't do anything but I'll try it again and see if it gets working. Really, thanks for the help. At work so, won't be able to try anything till tonight EST.

technically you can upload it anywhere in the ram as long as it is not used at the time but then you have to skip 0x1000 for the start or launch address.
however if you look at your boot record the FW always copies the bin file at 0x361000 and then starts at 0x362000 exactly the same as  in Keysight 3000/2000 scopes.

with DHCP command you need to give the correct addresses too just like above. or u can use tftp to upload it to memory Instead of LOADY from terminal and then use the 'go' command. your tftp server must be up and running.
if the unit still cannot connect to tftp server then I'd say there is still some hardware problem to be sorted out specially since she thinks the ethernet is connected while it isn't

I once had a case when the LAN controller chip LAN8710A needed to be replaced as well but never saw such a case again
« Last Edit: January 13, 2025, 06:11:03 pm by analogRF »
 
The following users thanked this post: Eng_tech

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #181 on: January 14, 2025, 02:54:13 am »
Must still be a hardware problem. Maybe I will go back a try reheating the spear process again. Or I have RAM issues. I get the same result as I have with booting from the NAND. It boots but gets a partial screen that is white with about 1/5 of the top of the display has a blue background. Here is the boot log. I did this with Ymodem. :( >:( :palm: :wtf:

Just for fun :-DD |O I'm going to do DHCP knowing full well it ain't gonna work.  I'll heat that Spear up more tomorrow at work. Just tried DHCP, gives no ethernet adapter found, same with tftp command

Also, any idea about the 2 leds next to the FPGA on the main board. The one closest to the FPGA is always flashing and the one farthest away is always on. Is this normal behavior?

Maybe this thing is too far gone. It partially boots but stops that was giving me some hope.


Code: [Select]
## Starting application at 0x00362000 ...
Windows CE Kernel for ARM (Thumb Enabled) Built on Mar  8 2013 at 17:05:33
Setting up for a Cold Reboot
Done Setting up for a Cold Reboot
Windows CE Firmware Init
BSP 1.0.0 for the SPEARHEAD600AB board (built Oct 19 2017)
Adaptation performed by ADENEO (c) 2005
+OALIntrInit
-OALIntrInit(rc = 1)
Initialize driver globals Zeros area...
pDrvGlobalArea 0xa0060000  size 0x800 (0xa0060800 -0xa0060000)
Initialize driver globals Zeros area...done
 OALKitlStart
Firmware Init Done.
OALIoctlHalEnterI2cCriticalSection init i2c cs
ERROR: C:\WINCE600\PLATFORM\COMMON\SRC\SOC\STM\SPEARHEAD600\DRIVERS\NandFlash\.\sh600_NandFlash.c line 57: ConfigTimming - Unable to open device registry entry
ERROR: C:\WINCE600\PLATFORM\COMMON\SRC\SOC\STM\COMMON\DRIVERS\NandFlash\.\stm_NandFlash.c line 1043: LLD_GetInfo - Unable to open device registry entry
++SER_Init: context Drivers\Active\10
SER_Init, dwIndex:2
GPIO_Select0 Register 0xB300_0024: 0x80000000
Control Register 0xB300_0010 : 0x00000040
RAS Select Register 0xB300_000C: 0xffffacf4
CORE_CLK_CFG 0xB300_0024: 0x80000000
SER2 got sysintr:0x00000013
SER2 Serial Port, new baud rate:0x1c200  (UARTCLK:83250000 IBRD:0x2d FBRD:0xa)
++SER_Init: context Drivers\Active\11
SER_Init, dwIndex:3
GPIO_Select0 Register 0xB300_0024: 0x80000000
Control Register 0xB300_0010 : 0x00000040
RAS Select Register 0xB300_000C: 0xffffacf4
CORE_CLK_CFG 0xB300_0024: 0x80000000
+OALIntrRequestSysIntr IRQ (1) already used by SYSINTR (19)
SER3 got sysintr:0x00000014
SER3 Serial Port, new baud rate:0x1c200  (UARTCLK:83250000 IBRD:0x2d FBRD:0xa)
OHCI\system.c, GCFG_USBH1_SW_RST
OHCI\system.c, GCFG_USBH2_SW_RST
-EDeviceLoadEeprom 80:09:02:01:BF:42
Phy not found
WaitForLink Start (ticks=3751)
Link Detected (ticks=3753)

 GMAC Init : 100 Mbit/s FULL DUPLEX (MII)
Flushed Transmit Buffer
phyCfg->dwSpeed 0x64
<--EDeviceInitialize

GMAC DMA status register = 0x600004
GMAC Device enable interrupt
DriverStart
GMAC Device enable interrupt
Resetting the USB-device silicon
sh600_pdd, IOCTL_BUS_POSTINIT
ERROR: C:\WINCE600\3RDPARTY\Agilent\HPP\Common\Drivers\stm320_UsbFnBusDriver\.\ufnbus.cpp line 1137: failed opening \Agilent Flash\SPD\usbOverride
                                                           ERROR: OALIoCtlHalGetDeviceInfo: Device doesn't support IOCTL_HAL_GET_DEVICE_INFO::SPI_GETUUID
Warning: you are calling IOCTL_HAL_GET_UUID, which has been deprecated.  Use IOCTL_HAL_GET_DEVICE_INFO instead.
ERROR: OALIoCtlHalGetUUID: Device doesn't support UUID
stm320_UsbFnBusDriver, set IST priority 96
« Last Edit: January 14, 2025, 03:15:13 am by Eng_tech »
 

Offline analogRF

  • Super Contributor
  • ***
  • Posts: 1024
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #182 on: January 14, 2025, 03:25:16 am »
I would insert a USB stick with FW file on it into the device before uploading the bin file and issuing the 'go' command. Still I think the correct kernel should look for update file if it finds a usb flash

for me this is still strange and does not match what I have seen on these meters. Maybe I am wrong but I have many logs of different meters and none
shows up  like this. I mean first it says phy not found and then it shows link detected and 100Mb/sec even though no cable is connected
this is not some internal ethernet communication. this is exactly for the LAN connection.
Code: [Select]
Phy not found
WaitForLink Start (ticks=3751)
Link Detected (ticks=3753)

 GMAC Init : 100 Mbit/s FULL DUPLEX (MII)
Flushed Transmit Buffer
phyCfg->dwSpeed 0x64
<--EDeviceInitialize

it could be that dll files of the WInCE are corrupted as well. The next message you will see after the last line of your log is about loading some dll files.
of course there could be some other stuff going on in between which we are not aware of but as far as the UART messages are concerned the next message is about some dll files. If that's the case we need to force the unit to boot the wince from usb (just like in the keysight 3000A scopes) and thus take the dlls from there
i dont know how it can be done but it must be similar to keysight infiniivision scopes.

still I think a hardware issue is the more likely problem. if you got the cpu from aliexpress, be ready to exchange it again as I had to a couple of times.

also maybe probing around that LAN chip could be useful too
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #183 on: January 14, 2025, 04:35:50 am »
I probably won't give up on it yet. I will probably just get another cpu again. I get them from a seller on ebay. They look original. The seller is top-chips and the chips are SPEAR320S-2-AA. Others being sold on ebay are Spear320-2. The seller is good, packages it in anti-static bags and the chips are cut from a reel tape and not loose. He has them cheap at $14 a piece last time I looked. Others sell them at much higher costs. Think I will pick up 2 more. Replacing them is no easy task but I figured out a way to perfectly align them on their pads. I will show you how I manage that if I can get an image of my little fixture/guide on here. It works really good.
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #184 on: January 14, 2025, 01:21:33 pm »
When I first did the YModem upload and booted I did have my USB flash drive inserted into the front panel with the firmware on it. The flash drive just flashed it's led like it was being read but the meter's screen stayed white. I left it alone giving it some time and watch the serial terminal for any indication that something was happening. After a while, I just shut off the meter and tried the same process but without the flash drive attached. It booted just like it boots off the NAND.
So, you speak of DLL files. They are not part of nk.nb0 that I uploaded via Ymodem? I'm confused. You think they are loaded in at another location in NAND? You could be correct that the boot process halts from a corrupted DLL. It looks like the process just stops as it is trying to display the blue graphics background. Gets a small part of the way and stops.
 

Offline analogRF

  • Super Contributor
  • ***
  • Posts: 1024
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #185 on: January 14, 2025, 01:31:44 pm »
When I first did the YModem upload and booted I did have my USB flash drive inserted into the front panel with the firmware on it. The flash drive just flashed it's led like it was being read but the meter's screen stayed white. I left it alone giving it some time and watch the serial terminal for any indication that something was happening. After a while, I just shut off the meter and tried the same process but without the flash drive attached. It booted just like it boots off the NAND.
So, you speak of DLL files. They are not part of nk.nb0 that I uploaded via Ymodem? I'm confused. You think they are loaded in at another location in NAND? You could be correct that the boot process halts from a corrupted DLL. It looks like the process just stops as it is trying to display the blue graphics background. Gets a small part of the way and stops.

yes if i understood correctly (others may correct me here) the main wince files are still being read from the nand. the nk.nb0 is just a minimal kernel to boot the device initially. i may be wrong about that but since everything is quite similar to keysight scopes, I believe the main WinCE files are loaded after your last line of boot record. if we can extract them from the FW file and instruct the meter to read them from usb then maybe....

are you sure there is no hardware issue with the display itself? or the LAN chip? those lan messages are still bothering me....
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #186 on: January 14, 2025, 02:32:33 pm »
Windows Ce  by default  load   nk. ***   in ram and decompress it  to run ...   normally the bin vs the b0 contain more data and infos .. but both should contain configs / drivers and the Keysight   dmm executable software  ....

at some point  Windows Ce could be connected to Windows if it contains what it need to do that communication, that way once connected you see WinCe drive and files and can copy them out, export the registry etc ...    not sure in this case, i don't have theses meters to play with

https://supportcommunity.zebra.com/articles/en_US/Knowledge/Windows-CE-device-does-not-connect-with-Windows-10-with-Creators-upgrade-via-WMDC


unless you do have somekind of ssd emmc  or elses, it start in nand ram locations as mentioned,  some  on other threads managed to shift the allocation / address to be able to boot the dmm  ??

the Spear 320 failures ? , no one has clearly stated why it is failing,  voltages ?,  heat ? degradation ?

for the other peripherals as we can guess  usb and lan  hardware should come 1st   to be able to help if it crash ?  uboot or others ...  the spear 320 should contain what we need ? unless  ...

old 2023 thread
https://www.eevblog.com/forum/repair/agilent-keysight-34461a-doesnt-boot/
https://www.eevblog.com/forum/repair/keysight-34461a-179388/

34465
https://www.eevblog.com/forum/testgear/keysight-34465a-reliability/
« Last Edit: January 14, 2025, 02:57:17 pm by coromonadalix »
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #187 on: January 14, 2025, 03:43:00 pm »
Pretty sure display is good. As when it first boots you get a white background with the words Keysight Technologies in black and at the bottom of the display is a black bar that increases in length as the booting process runs. After that, the screen would display a blue graphics background image that says on it Keysight Technologies with a load bar at the bottom that would say "initializing and verifying hardware". Only if the meter is working correctly.  But I only get the very top part of that blue graphics background and the meter stops booting at that point or loading the OS. So, maybe it's something with "initializing and verifying hardware". DLL, driver maybe? You would think it would at least display the full graphics background but it stops loading it or sending it to the display. Or maybe where the image is stored is corrupted. Don't know yet.
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #188 on: January 14, 2025, 05:28:31 pm »
previous fw  as 3.01 3.02  where about memory leaks, the latest v 3.03  ??


and a small note from keysight ?


have you tried the fw updater from keysight ?  +- 72 Mb
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #189 on: January 14, 2025, 05:32:48 pm »
This is whats contained in nb.nb0. I used a utility called dumprom.exe to extract all the files to a folder. It's a large log of what was dumped and it won't all fit in this message so I'm attaching it as a text file.

 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #190 on: January 14, 2025, 05:51:39 pm »
windows ce tools
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #191 on: January 14, 2025, 06:00:22 pm »
I've used it to create a USB drive containing the firmware that the meter orginally had. Of course the USB on the front panel does not work so it can't be used for updating firmware. YModem is only choice that works at this point. The log generated from dumprom is nice as it shows where each file is located and its length. Of what help that will at this point I don't know. I just wanted to see the files thats contained in nb.nb0.
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #192 on: January 14, 2025, 07:12:42 pm »
the worst  would be some both  spear problems and nand corruption ??       dang



but the keysight fw  talk about creating the build.xml file and the Nk.bin   and put them on the usb key ???

attached
 

Offline Neganur

  • Supporter
  • ****
  • Posts: 1174
  • Country: fi
Re: Agilent 34461A corrupted flash
« Reply #193 on: January 15, 2025, 04:36:56 pm »
managed to fix a white screen / no boot error on an 34465A by swapping the Spear (U201).

As received, unit did power up, but no fan/USB/LAN, white display.
Upon initial optical inspection of the front panel module, I saw that the Spear had shorted solder balls, and the power management IC (TPS650701) got very hot.
Replaced the Spear, short circuits were gone. Unit powered on, fan spun up. B/W Keysight logo on the display with a progress bar running, but halting.
Two reboots and the progress bar made it to 100%, device seems to power up normally and functions.

Last instrument error displayed is "+822 Controller and measurement board serial numbers do not match", which went away after a factory reset (Edit: seems to persist, error queue was just emptied).

The Spear is quite warm to the touch, I think I am going to place one of those sticky heat sinks on it.

Is it worthwhile dumping the EEPROM?
« Last Edit: January 15, 2025, 05:23:21 pm by Neganur »
 

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #194 on: January 15, 2025, 05:07:31 pm »
@neganur,  if you can do a dump and post it,  sure it would maybe help others ??

the best would be having nand dumps and others things like a "good booting"  log file

the FW is good for the 34461 34465 and 34470 meters ?
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #195 on: January 15, 2025, 05:56:31 pm »
Well just to isolate the problem with this meter I put my front panel on a known good unit to see if that unit would act the same during the boot process and it did. This eliminates the main pcb from having anything to do with the problem. So, the problem still resides with the front panel. I'm going to check out the best I can the Ethernet controller on the front panel because during boot the unit thinks it is wired to a network which it isn't. This maybe a clue of some kind. Odd that they put the ethernet controller on the front panel. I would think they would have that near the network connector on the main pcb.
 
The following users thanked this post: coromonadalix

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #196 on: January 17, 2025, 02:54:51 am »
Success! I think. I replace the LAN/Ethernet IC LAN8710A and the unit completely boots up now. Only one other condition exists and that is a little red box in the upper right corner that says "error". It passes self testing also. I am thinking that maybe the meter lost it's calibration information but could be wrong. Tried to do a self calibration but the meter is locked. Firmware updates are also locked. An odd thing that when this meter finally booted up it did a firmware update by itself. Sad installing firmware update in 2 parts. Same version. Kinda strange but maybe has something to do with the new spear microprocessor.
 

Offline analogRF

  • Super Contributor
  • ***
  • Posts: 1024
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #197 on: January 17, 2025, 03:06:07 am »
Success! I think. I replace the LAN/Ethernet IC LAN8710A and the unit completely boots up now. Only one other condition exists and that is a little red box in the upper right corner that says "error". It passes self testing also. I am thinking that maybe the meter lost it's calibration information but could be wrong. Tried to do a self calibration but the meter is locked. Firmware updates are also locked. An odd thing that when this meter finally booted up it did a firmware update by itself. Sad installing firmware update in 2 parts. Same version. Kinda strange but maybe has something to do with the new spear microprocessor.
great !
I told you I was very suspicious of that LAN chip because the log was strange.
you should be able to see the error in the error list. also the boot record might show what error is happening
can you connect to the ethernet? does the meter get IP?

no calibration is not lost. it is in a eeprom on the main board but I think there is a procedure to read that calibration and copy to nand because you changed the front panel. that could be the cause of error
you can update the FW either with lan or usb stick.
 

Offline Eng_tech

  • Contributor
  • Posts: 26
  • Country: us
Re: Agilent 34461A corrupted flash
« Reply #198 on: January 17, 2025, 03:31:05 am »
Jumped the gun I guess. It is randomly doing the same thing upon booting up. Sometimes the meter is working normally and sometimes I just get what I was getting before. It is a big step in the right direction I think. I have to find out why it is randomly working. Yeap, the error had something to do with calibration an serial numbers not matching. But for some reason after looking at the log the errors went away. Still gotta find out what is going on. Maybe I have a bad solder connection on that LAN chip. It was a bit of a pain getting it mounted and soldered. I used hot air on it both on top and below with and Chip Quik paste.
 
The following users thanked this post: voltsandjolts

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 7116
  • Country: ca
Re: Agilent 34461A corrupted flash
« Reply #199 on: January 17, 2025, 10:09:11 am »
Lan  chip  that's a 1st report i've seen  @analogRF was right

you should be able to realign it, but theses have a thermal pad under them ....    :-+   or  you may burn it  ?   

i would have tried the meter without it  just to see if it get working and stay stable in reboots ?  and do a factory reset ...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf