Fluke 5500a repair attempt

[dcmxx]

Some months ago i received an almost complete Fluke 5500a MFC that is missing the A9 Main CPU Board. It even haves the SC300 option.
Since then i have waited to come across with one of that boards without luck

Instead i have come across a A9 board from a 5820 osciloscope calibrator wich is the exact same A9 board for the Fluke 5520a MFC but with different firmware.

Altough the board is not exactly the same as the A9 of the 5500a i'm almost sure that if I program the firmware of the 5500a's board in the 5520a's board it will work

One of the main differences are that the uses bigger flash memories and instead of one for the lower word and one for the upper word in ROM Bank 0, the 5520a's  board uses a single bigger flash memory.

I removed all the other function boards form the calibrator and connected the 5520a/5820a A9 board to the main backplane board, and turned on the instrument. It energizes both displays and tries to start up. But without the boards it stay stuck there as expected.

The problem I have right now is that I have no source for the firmware dump of the flash memories, so if anyone has it or is able to dump it, it will be greatly appretiated. If there's some cost assosiated, as long as it is rasonable, i don't have any problem in paying it. The needed dump is from the U1, U26 and U27 memory chips of the A9 board for the Fluke 5500a MFC

I attach pictures of the 5820a/5520a board

[Ugur]

Hi,
Are you sure that the A9 card is the reason why the device does not turn on and all the lights are on?

[dcmxx]

The 5500a is missing the A9, so if i turn it on, it doesn't do anything. So i'm pretty sure the problem is the missing A9 board. thats why i'm trying to replace it with the a9 board from a 5820a/5520a

[ddrl46]

Besides the firmware being different between the 5500A and 5520A A9 boards there is also a difference in the pinout of the cable which goes to the A3 motherboard.
You might want to compare the schematics of the Fluke 5500A with the CPU board listed in the Fluke 57LFC service manual (which is the same one used in newer Fluke 5520A units).

I'm actually currently working on putting together are 5500A unit from parts which I've gathered, and for this project I've had to recreate a few of boards. One of these being the A9 board.
There are currently still a few issues I'm ironing out, but I may have some boards left over when I'm done. I also have access to the firmware, but I'll have to do do some work on that before I can share it.

[dcmxx]

that's a really nice proyect, if you consider selling the boards i'm definetly interested in one, even if I make the 5520a's board work, due to the fact the A9 board from the 5520a is not the same size as the original A9 of the 5500a.

I also thought about the posibility of recreating the board from the available schematics, but i haven't had much free time lately so that's not an option right now.

hope you could share the firmware dump when it's possible. and thanks for the hint about the pinout, i will check it out

[dcmxx]

I attach here the dump from the U5 and U6 memory chips with the 5820a's a9 firmware, in case some one needs them one day

[ddrl46]

Attached is the Fluke 5500A firmware.

Some notes:
U1 (the 28F400) occuppies the following address range: 0x00000 to 0x7FFFF
U26 and U27 combined occupy 0x80000 to 0xBFFFF

When powering up the calibrator it is normal for it to hang and not throw an error when the A6 + A8 assemblies are not fitted.
When experimenting with the calibrator it may be useful to know that you can get into the bootloader menu via the "HOST" RS232 port by holding the RESET key when powering up the calibrator (9600/8/N/1).

Just wondering, but someone wouldn't happen to have a spare empty Fluke 55xxA chassis laying about gathering dust?

[dcmxx]

Hi. thanks a lot!

i going to try it soon

i'm cheking the differences beetween the 57lfc a9 schematic and the 5500a a9 schematic. it will take some modding to make it work

[dcmxx]

The calibrator starts to show signs of life with the firmware provided by ddrl46, but it just hangs

What I have done so far was to remove the rs232 transceivers to make room for an easy modding of the pinout of the connector that goes to the motherboard.

Although the 57LFC A9 board schematics has been very useful, is not exactly the same board as the 5520a/5820a, there are some minor differences.

With the firmware files provided i splitted the U1 bin file, in low byte and high byte files, as the 5520a A9 board uses 2 E28F004B5B memories, each one holding one byte of a 16bits word. Opposed to the U1 MT28F400B5 from the 5500a which holds a full 16bits word in each address

After that I merged the low byte file with the U26 dump and high byte file and U27 dump, as  U26 is addressed as low byte's memory and the U27 as high byte's memory. I programmed the low byte file to U5, and high byte file to U6

What’s next is that I have to address the part of U26 and U27 as they are normally selected using CS1, and in the 5520a A9 CS1 is not used. And the fact they start at the end of the splitted U1 dump 

[ddrl46]

Nice to see it starting to show some signs of life.

By comparing the two schematics it looks like the FIG / TIG data lines need to be modified quite a bit.
In the 5500A these are connected pretty much directly to the processor its TXDA (via an inverter) and RXDA lines, in the 57LFC schematics these appear to be handled by the 68681 DUART.

Unfortunately the 5500A does not show an error when it is unable to communicate with the in-guard circuitry and just hangs.

By the way, did you press the reset button while powering up the unit in the second picture thus causing it to enter bootloader mode? Or did it enter this mode on its own?

[dcmxx]

Yes I got there pressing the reset button. But it's intermittent sometimes it shows, sometimes don't

But what I’m seeing is that those messages are independent, and are stored in the display board

I modified the pinout of the motherboard connector, that's why I removed the rs232 transceivers

In fact the cpu board was stalled, one of the reset lines was low. I corrected that and the processor tries to read the fw from the memories, I see activity in the address lines, and data lines, but i don’t see any activity in the tig, fig, tds or fds lines, also I don't see any activity in the CS1 line that should select the u26 and u27 memories. So I’m stuck. i don't really know what’s happening

This board is requiring much modding, and right now I don't want to cut any traces.

Any chance to sell me one of your A9 PCBs? I could populate it with many of the components of this board

[ddrl46]

I've made a few mistakes on my A9 board related to the RS232 interface which I would like to correct first.
Most importantly however I have verified that my board is able to boot correctly and talk to the in-guard circuitry.

I'll let you know when I have these new boards available.

[dcmxx]

I'm looking forward for the new boards

I will keep trying to make this board work and post if I make some progress

[dcmxx]

I made a mistake when I splitted and merged the bin files for the memory chips. The processor was crashing and asserting the HALT line.

So, now I really made some progress.  I corrected the bin files. And reprogrammed.

I also found that the 68C681 should be populated or else the firmware also doesn't run correctly. The IRQ line has to be disconnected from the processor as the 5500's A9 uses IRQ4 and the 5520's uses IRQ6.

Now when I turn it on I get consistent behavior, the Starting up screen shows each time, CS0 and CS1 seems to be working as expected. If I press the reset key when turning it on, I get the RESET key hold down screen correctly.

I start to hear some relay switching after some seconds of power on, but the calibrator doesn't seem to fully start up.

[Srikanth]

Hi,

         I have same problem Fluke 5500a calibrator, Please guide me how to flash the firmware in to main CPU board, I tried to communicate with RS485 Host port through Hyper terminal no response from the CPU card. All supply voltages are showing correct. Due to power fluctuation may be firmware got corrupted. Thanks for your support.     

[ddrl46]

Srikanth:

Try connecting a null-modem type RS232 cable (or USB -> RS232 adapter) to the SERIAL 1 port (the male one).
Open a terminal program (i use termite rs232), configure it for 9600 baud, 8 data bits, 1 stop bit, no parity, no flow control, linefeed enabled.
Press and hold the reset key on the calibrator front panel and power it up, check to see if the calibrator displays "RESET KEY HELD" on the right hand display.

From the terminal window see if you get a message indicating that the RESET key has been held and press return to enter the command line, send a "?" to the calibrator to get a menu listing of all options.
You can check to see if the firmware is deemed to be corrupted by the calibrator by running the "checksum" command which is displayed in this list.

If your firmware is indeed corrupted, reprogramming should be possible via the RS232 interface using the "serial outguard download" command. I've not yet figured out how this command is supposed to work.

dcmxx:

I've ordered the revised A9 CPU boards and they should be here in a few weeks. When I've completed testing them I'll contact you.

[dcmxx]

I programmed the memory chips with a cheap TL866II programmer, but obviously you have to remove them from the board and put them in an adapter and then in the programmer.

ddrl46 has already posted the firmware dump so if that's the cause, you should be able to bring it back to life

[Srikanth]

Hi ddrl46,

              Thanks, firmware is corrupted on serial port "Flash Checksum fail". firmware can be upload via serial or do we need to use Jtag, please advice.

Thanks. 

[Srikanth]

Thanks, I got firmware, please help on how to download.

[ddrl46]

Thanks, I got firmware, please help on how to download.

Like dcmxx mentioned, your best bet is to de-solder the three flash chips from the A9 cpu board and flash them using an EEPROM programmer. A TL866II with the proper ZIF sockets should do the trick.
Keep in mind that all calibration data in your unit will be lost, and that full re-adjustment will be required after flashing.

If someone feels up to the challenge it would be possible to update the flash memory via the serial port with the "SRECV" or "GRECV" commands in the bootloader menu.
Unfortunately I was unable to figure out how these commands work, them seem to throw back an "Aborted" message regardless of what I do. Figuring this out would likely require disassembly of the bootloader.

[Srikanth]

Dear ddrl46,

                Thanks for the help, I reprogrammed the all the ICs via TL48 programmer calibrator is up now.

[dcmxx]

After some time after opening up this thread, and more time since i got this calibrator, it finally has come back to life

The idea of modding the Fluke 5820's A9 to use it as a Fluke 5500a's A9 board, although possible requires a lot of work and modding that involves cutting traces. So i dropped that idea

But thanks to ddrl46's amazing work the Fluke 5500a has now a working A9 board. He took the time and effort to have develop a working A9 board

The board came and was installed without much trouble.

But after powering up the calibrator the first time with the new board I realized that calibrator had other issues, and that probably was the reason that it got his a9 board removed.

I was getting the Error 104, and if I executed the diagnostics I also got Error 399

Also thanks to ddrl46's help, i was pointed in the right direction, and started to look for the problem

The problem was located in the A8 board, the voltage generation board. Initially found a missing component, Q19.

After some troubleshooting, realized that the Error 104 is always triggered by the _Fault line that it's shared between boards. Leaving connected just the A8 and the A6 board narrowed the problem

And after some time i noticed that under some circumstances the 65v supplies were faulty. +65 were quite high (72v) and -65v supply was down (-15v)

The transformer it's good, and also the capacitors, diodes and transistors. But i found 2 of 4 1ohm resistors that connect the 65COM to SCOM open, so the 65v secondaries from the transformer were floating in relation to the SCOM 

I replaced the four resistors in the A8 board (R55, R56, R63 and R65). The supplies were good again and I tested the calibrator. That showed no more errors.

[dcmxx]

The calibrator it's being tested right now, but it doesn't seem to have any other error.

The zero function it's executed without problems. I could also adjust and store calibration coeffients.

So now this Fluke 5500A seems to be fully working, and it also have the SC300 scope option.

There are two details that are missing to me. The encoder knob and the fan filter are missing.

If any one knows where i could purchase them please let me know.

In the mean time i'm going to order a chinese replacement button membrane for the calibrator