¿ªÔÆÌåÓý

  1. HP-Agilent-Keysight-Equipment
  2. Topics

HP54603B Software Gremlins

 

Hi HP Experts,

the HP54603B is a mid-90s digital storage scope (compact and nice besides the
missing ?-metal shield for the CRT). Mine has the HP54657A Measurement/Storage
Module installed. Most of the time it works flawlessly BUT there are
reproduceable problems with trace memory functions:
- Storing screen data to memories 4 and up (nonvolatile) seem to work fine
? until the device is switched off
- On the next power up, the recall (or preview) of traces 4++ leads either to
? a completely messed up screen display or a hard system reset (with all
? memory data wiped out)
- Trace memories 1 to 3 are ok

So the questions are:
- Is this a known problem ?
- Is there a firmware update available for the scope and/or the storage module ?
- Does at least a workaround exist to avoid loss of memory data ?

The software version of the scope is V.1.2 and "Yes", I've done the power-on
reset (module initialization) several times with no effect. The selftest functions
pass without error messages.

Many thanks for advice, help or solutions

Regards (& 73 :-)??? Andreas DG7RBN
 

I don't know specifically for that model, but your symptoms sure sound like a bad SRAM chip to me. Do you know if the self-test does a full test of the memory?

-Matthew
 

Hi Matthew,
the self test menue has an explicit function that tests RAM and it runs thru without error messages.
The problems seem to occur in the nonvolatile memory of the HP54657A module. The clobbered screen display vaguely shows the original image tilted to one side. It looks more like a classical programmer's bug (wrong increments of an adress pointer while copying data).
Regards????????????? Andreas
 

Hi Andreas,

There is two different Non-volatile/zero-power ram chips in the module.
Both have their own internal backup battery. So the battery in one of them is probably bad.

I would replace both, as the other one probably don't have much time left.

From the manual:
Saving Traces Hint
"For traces 1 to 3, the trace is saved in the non-compressed state. For traces
greater than 3, the trace is saved in 64 Kbytes of nonvolatile trace memory with
data compression. After a compressed trace is saved, the available area (in
percent) is displayed."

In my 54657A, there is one Dallas DS1244Y-120 (32kbytes) and a ST M48Z128Y-120PM1 (128kbytes). see picture.

So I would guess that its the ST part in yours that the battery is bad.

BR,
Askild



On Thu, Apr 9, 2020 at 12:00 PM Andreas Zilker via <dg7rbn=[email protected]> wrote:
Hi Matthew,
the self test menue has an explicit function that tests RAM and it runs thru without error messages.
The problems seem to occur in the nonvolatile memory of the HP54657A module. The clobbered screen display vaguely shows the original image tilted to one side. It looks more like a classical programmer's bug (wrong increments of an adress pointer while copying data).
Regards????????????? Andreas

20180525_173242_r.jpg

 

Hi Askild,

thanks a lot for the clear and detailed pointer.
With "nonvolatile trace memory" mentioned in the manual, I thought of something
more durable (EEPROM, FlashROM) than a battery back-up SRAM :-) Would have
been nice, if the programmers had added some CRC or at least a checksum to the
data compression throwing an error message if something fails. At least they had
time for the famous "Tetris game" :-P
As the price tag for the M48Z128Y replacement from Maxim is in the $30 range,
(and then you dont't know how old it is already) I'm thinking of setting up a
"discrete" solution with a $3 128kSRAM, a coincell and 2 Schottky-Diodes
(or a DS1210).

Cheers??? Andreas
 

Hi Andreas,

Yes, flash would be nice, but back when these where designed, that was not a good option for reliable long operation.

You could try to do what I did with the M48Z18 inside my HP 54645A scope.
I removed it from the PCB and put in a socket.
Looked at the IC, and it looked like an IC epoxied into a bigger housing.
So I heated it up and used a small screwdriver to pry the IC out of the epoxy.
The battery/batteries, it can be one or two on top of the IC. They where connected to the ends of the IC.
I just connected a new lithium battery to one side, see picture attached.

When dismantling, pay attention to where the + and - of the battery connects..

BR,
Askild


On Fri, Apr 10, 2020 at 12:45 PM Andreas Zilker via <dg7rbn=[email protected]> wrote:
Hi Askild,

thanks a lot for the clear and detailed pointer.
With "nonvolatile trace memory" mentioned in the manual, I thought of something
more durable (EEPROM, FlashROM) than a battery back-up SRAM :-) Would have
been nice, if the programmers had added some CRC or at least a checksum to the
data compression throwing an error message if something fails. At least they had
time for the famous "Tetris game" :-P
As the price tag for the M48Z128Y replacement from Maxim is in the $30 range,
(and then you dont't know how old it is already) I'm thinking of setting up a
"discrete" solution with a $3 128kSRAM, a coincell and 2 Schottky-Diodes
(or a DS1210).

Cheers??? Andreas

20190324_170726_r.jpg

1 - 6 of 6