Looking for Mettler Scale ROM backup

[jamarjan]

I swapped the crystal oscillator - the waveforms didn't change
the oscillator waveform itself doesn't look very bad, but the clock waveform from uC is not rectangular
on the other hand, both ADC oscillator waveforms are perfect

My integrated circuit doesn't program at all - it can be read (it's filled with blanks), but it can't be programmed
gmac34 I also ordered RAM d446c - they'll be here tomorrow

[dustooff]

Hi,
with regards to the MCU oscillator, I usually scope the ALE and PSEN pins to see if the MCU is trying to access any of the external memory devices, ALE should be 1/8 of the XTAL frequency, see timing waveform on page 25.
If these external memory timing pins are active then the main XTAL clock should be fine.
Then probe each data pin and address pin to check for less than optimal logic levels, sometimes a addr/data line can be stuck.
regards
Andrew

Edit: correction ALE is 1/6 of CLK frequency.

[jamarjan]

fzabkar analyzing what you wrote I made a list of my parameters and dumped rom
I'm trying to recreate the configuration of my deleted rom 
I have problems with positions 10 and 11
can you tell me how to write it correctly
it's a bit strange that the first digit is omitted
I'll stare at this ROM dump for a while 

	jamarjan				gamc34
Model name	PM4600	hex	mod_hex		PM4600	hex
ID	735485				874543
00	306874	4AEBA	AEBA		306874	4AEBA	AEBA
01:	460672	70780	0780		460672	70780	0780
02:	360252	57F3C	7F3C		341052	5343C	343C
03:	197833	304C9	04C9		263354	404BA	04BA
04:	531476	81C14	1C14		659849	A1189	1189
05:	687621	A7E05	7E05		910597	DE505	E505
06:	525688	80578	0578		197998	3056E	056E
07:	1048247	FFEB7	FEB7		326074	4F9BA	F9BA
08:	268031	416FF	16FF		486143	76AFF	6AFF
09:	131068	1FFFC	FFFC		720879	AFFEF	FFEF
10:	59	3B			197208	30258	0258
11:	27136	6A00			621824	97D00	7D00
12:	393011	5FF33	FF44		1048348	FFF1C	FF1C
13:	230473	38449	8449		682179	A68C3	68C3
14:	680701	A62FD	62FD		739325	B47FD	47FD
15:	327697	50011	0011		196620	3000C	000C
16:	489575	77867	7867		527291	80BBB	0BBB
17:	480512	75500	5500		232192	38B00	8B00
18:	917504	E0000	0000		655365	A0005	0005
19:	531245	81B2D	1B2D		581909	8E115	E115
20:	220672	35E00	5E00		886257	D85F1	85F1

[fzabkar]

Code: [Select]

10:   59          0003B       003B
11:   27136       06A00       6A00

There is a typo here:

886257   D85F1   85F1

886527   D86FF   86FF  <- this is correct

[jamarjan]

Hi Andrew and others.
Flashed and reinstalled new EPROM - no changes  :'( Still waiting for RAM chip.
On ALE i have a nice square wave of 1,23 MHz
On PSEN similar with a little bit of noise on a high state.
XTAL is about 7,3 MHz
In most cases I have no activity on pins of channel 1 (P1.X) where communication with ROM should be.
Once every few restarts there is activity with waveforms not resembling those I found in the ROM datasheet (in / out).
screenshots attached
These waveforms appear once in a while (every 5th time the scale is used), usually there is no activity or it occurs very shortly after the scale is turned on - I am not able to record it on the oscilloscope.

[gmac34]

Did you check continuity between the uvprom and the board? Check that all data line are going trough the connector, also check for corrosion or bent pins.

[dustooff]

In most cases I have no activity on pins of channel 1 (P1.X) where communication with ROM should be.

Is port 1 actually connected to the ROM/RAM devices?
My understanding of the data sheet says...
Port 0 is also the multiplexed low-order address and data bus during accesses to external program and
data memory.
and
Port 2 emits the high-order address byte during fetches from external program memory and during
accesses to external data memory that use 16-bit addresses.

In the past I have repaired two PM4600's one for VFD problems, the other was failing to startup, like your current situation, unfortunately I can't recall the specifics of how the repair went. And I currently own it.

Edit: found app note AN457 (but not on NXP's site)  look at Figure 20, the 74HC373 has to hold the low address while the MCU presents the data byte.
Check the 373 output for an address, may need a LA to do that.

[jamarjan]

I've had a brief break in fixing the scale due to a business trip.
I replaced the UPD446C CMOS SRAM, 74HC373 chip, uC P8031, and the resistor ladder (it looked like it had a crack).
Still no change.
The clock waveform is still not square.
On startup there seems to be some activity on the uC communication with the ROM.
However, it seems to me that it does not load the program.
I guess I actually need to equip myself with a logic analyzer and take a look at the uC - ROM communication.

[gmac34]

Sorry to hear that the part’s cannon approach didn’t lead to anything.
In a few days I will be back home, if you need I can probe around on mine and compare things.

[jamarjan]

I managed to play around with a logic analyzer
I am posting screens from channel P0, P1 and from the EPROM
it seems to me that the microcontroller is not reading data from the eprom
just why it is not displaying any error on the VFD
gmac34 - can you check how your scale behaves when you unplug the connector ( blue) from the weighing module? and how about when you unplug the interconnect board with the rear connectors?

[gmac34]

nothing happens if I remove the communication board, I think it was an optional anyway, and I get error 4 ( Error in temperature measurement) with the cell disconnected.

[jamarjan]

Thanks for checking and the information provided
Unfortunately my efforts are still without progress
I also checked the proper operation of the optotransistor controlled by the integrated switch
Everything seems to work ok
When switched on the scale consumes about 9W of power
Under thermal imaging nothing heats up above the expected level
I'm trying to analyze what could have gone wrong due to badly filtered / unstable power supply voltage but without a schematic my skills are already exhausted
I guess I have to hunt for a twin used scale

[dustooff]

maybe there are some mainboards available somewhere,
https://www.ebay.com/itm/134203907609

[gmac34]

Is there any activity on pin 9 of the microcontroller (reset)?

[gmac34]

Non having a schematics is bummer, but one last thing that comes to mind is to check the IC1, the PC74HC373P. but I don't fully understand what is it doing there, by the position it would seem like it would interface between the ROM and the micro.

[PA0PBZ]

Non having a schematics is bummer, but one last thing that comes to mind is to check the IC1, the PC74HC373P. but I don't fully understand what is it doing there, by the position it would seem like it would interface between the ROM and the micro.

The lower 8 address lines and the data lines are on the same processor pins, so they need to be 'demultiplexed' and latched. See reply #31 in this thread.

[jamarjan]

I connected the logic analyzer to the 373 latch
below is a screenshot with the channel descriptions
unfortunately I don't really know how to interpret it

[gmac34]

I cannot help you with that, but I have one last thing I think you should check: I would check the data input on the vfd controllers, should be pin 39, to see if they are receiving any data. A bad controller might be another explanation for a blank screen

https://www.digchip.com/datasheets/parts/datasheet/477/SN75518N-pdf.php

[jamarjan]

Thanks for the suggestions!
Actually, the diagnosis after repairing the power supply and restoring stable voltages at the control points I just started with the VFD controllers
Unfortunately there is no data reaching the display board (voltages are ok)
The only life I get on the display board is if I run the scale without the cassette. Then all segments light up. I can turn the display off and on by pressing the spacebar button.

It seems to me that driving the display is secondary to loading data from the EEPROM. And here, unfortunately, I have no information exchange between the uC and the memory.

[gmac34]

I see, well, for what is worth the display lighting up with no cassette is normal behaviour

[gmac34]

there is a "for now" cheap scale on eBay that could be a good parts donor
given the state of the vfd I would imagine the cell to also be mechanically damaged but the main board may be ok

https://www.ebay.it/itm/335651376850

[jamarjan]

Thanks. The current bidder is me ;-)

[gmac34]

Good luck then.
In the meantime I was trying to figure out which values in the EEPROM might be those of the linearity adjustment as my PM3000 is a bit out of specifications, it's a challenge to figure out what is what.
The most confusing part is what happens to the parameters when they get written in memory, for example the first parameter for all 3 of my scales is written as AEBA even tough it is three different numbers written on the various parameter sheets. 3 different numbers that wen converted to hexadecimal end in AEBA but I cannot understand what the first part might be there for

[gmac34]

Well, it seems you got it. the main board has the same part number between the pm6000 and 4800 hopefully you will just have to move over the reference resistors R41 and R42, and the memory

[jamarjan]

Yes, I bought the scale and it's on its way to me now
I can tell from the pictures that the scale has been mistreated so I don't have high expectations
I will let you know when I install the circuit board, and ill also try to dump the memory to have another bin for analysis ;-)