Replacing OLED screen on an Agilent U1253A Multimeter

[AVGresponding]

Set your mind at ease; I have the display brightness set to low. At £70 + shipping, I want it to last as long as possible!

[gamalot]

Set your mind at ease; I have the display brightness set to low. At £70 + shipping, I want it to last as long as possible!

Thanks, your reply saved my day.

I've always wanted higher power conversion efficiencies in battery-operated devices, but the chip shortage is indeed causing a lot of trouble.

[--Oz--]

I will try and measure the current my meter drinks tomorrow and report back.
I would think adding three simple logic gate ic's would drink less than a mpu, but lots of variables

The stock nimh 9v energizer battery I got with my meter was 225mah. We know this 225mah is valid from energizer (they dont lie).
The li-ion 9v I use is rated at 600mah but tested to 520mah (its used and may have lost some capacity), but thats 2x the stock battery life.

With some good li-ion 14500 (AA size), they are 900mah (tested) and would give 4x the stock battery, that is low hanging fruit for little work.

If we could find some Li-ion pouch cells to fill the battery compartment (roughly 53mm wide X 17mm thick X 26mm high) more efficiently than round 14500 cells, should be at least 1200mah, or 5x the stock battery run time.
Maybe even use a wireless charger to charge the 2S Li-Ion pack.

[--Oz--]

I found a stock meter (no OLED display replacement).

This U1253B battery icon has these reaction.
Keysight Stock.                  Agilent Modded.
9.40v full battery bars,      8.1v full battery bars,
9.3 to 9.0v 3 bars,.             8.0 to 7.7v 3 bars,
8.9 to 8.3v 2 bars,.             7.6 to 7.2v 2 bars,
8.2 to 7.3v 1 bar,.               6.65v shows 1 battery bar then turn off.
7.26v "X" in battery,
Works ok down till 5.7v,
At 5.6v the numbers jump around,

Current consumption (mA):
Keysight Stock.             Agilent Modded
Vin     Lo  Med  Hi.      Lo. Med. Hi
10v.   23.  30.  34.      26.  32.  36
9.4v.  24.  31.  35.      27.  33.  37
8.4v.  25.  32.  37.      28.  35.  39
7.4v.  26.  34.  39       30.  37.  41
6.0v   29.  40.  44.      Turned off at 6.65v

Clearly the two meters have different firmware and possibly hardware changes.

For the fun of it I powered a populated adapter board with 3.3v, current was 0.81mA.

Edited: 12-13-2022

[--Oz--]

Will try and test my modded meter tomorrow.

The three IC's have very low idle current, total for the three is around 5uA.
All three work down to 3v, so lowering the 3.3v supply would work with the adapter board.

Two ideas.

1. Is there a "dark mode" for the OLED display? Could save some power there.

2. Is there a way to pwm the OLED display brightness (like you could could do with lcd)? Yea, long shot on that one, haha.

[--Oz--]

Edited my 2nd post above with the modded meter currents.

[tooki]

I found a stock meter (no OLED display replacement).

This U1253B battery icon has these reaction.
Keysight Stock.                  Agilent Modded.
9.40v full battery bars,      8.1v full battery bars,
9.3 to 9.0v 3 bars,.             8.0 to 7.7v 3 bars,
8.9 to 8.3v 2 bars,.             7.6 to 7.2v 2 bars,
8.2 to 7.3v 1 bar,.               6.65v shows 1 battery bar then turn off.
7.26v "X" in battery,
Works ok down till 5.7v,
At 5.6v the numbers jump around,
…

Clearly the two meters have different firmware and possibly hardware changes.

Maybe, but I kinda doubt it.

The big changes (which were still little) were between the U1253A to U1253B. AFAIK there are no differences between an Agilent ’53B and a Keysight ’53B, other than the housing colors and boot logo. (And included accessories.)

The ’53A is designed for 7.2V NiMH, while the ’53B is designed for both 7.2V and 8.4V NiMH batteries. The ’53A and Agilent-branded ’53B’s shipped with 7.2V batteries, while Keysight-branded ’53B’s shipped with 8.4V ones.

In the ’53B, you can select the battery type in the settings, so that the battery indicator will be more accurate.

It might be interesting to set your Keysight ’53B to 7.2V battery and then repeat your testing.

[gamalot]

Will try and test my modded meter tomorrow.

The three IC's have very low idle current, total for the three is around 5uA.
All three work down to 3v, so lowering the 3.3v supply would work with the adapter board.

Two ideas.

1. Is there a "dark mode" for the OLED display? Could save some power there.

2. Is there a way to pwm the OLED display brightness (like you could could do with lcd)? Yea, long shot on that one, haha.

When the brightness adjustment button is pressed, the processor will send a 'contrast' instruction to the OLED display. By intercepting this instruction and modifying its parameters, the brightness of the OLED can be controlled at the desired brightness.

I was thinking of using a simple DIY light meter to find out the relationship between current and brightness (most likely it is non-linear), so maybe there is a chance to find a balance point with slightly lower current but enough brightness.

[--Oz--]

OT: Battery options, revisited.

The stock battery was 225mAh (1.62wh), I have two of these used and they test to 160mAh, complete rubbish 

I mentioned "the 9v I use is 9v 600mAh (tested to 520mAh) or 4.4Wh and does not last very long imo."

I am thinking two Li-Ion 102550 pouch cells in series with 1400mAh capacity (10wh)  might just fit an unmodded case. But I did not find Li-Ion, it was LiPo which has lower capacity, so some possible capacity/runtime gain there.
This would be 6 times the runtime of a brand new 225mAh cell.


Could make a simple 3 conductor connector to balance charge the 2S pack.
Of course with a battery BMS, you could not use the standard charger. I think I would go with simple $1 USB 2S charge bms board, with would charge at 1A and charge in less than 2 hours.
Or maybe even a wireless charge system for convince like cell phones use.

Thoughts?

[Darkover]

This is what I am using now for 3-4years,

> Of course with a battery BMS, you could not use the standard charger.

Perhaps you should not, but you can. Unfortunalty you have to charge the battery 2-3times in a row,
because there is a time limit in the chargealgorithm of the 1253.

But my battery has some electronic to prevent from overcharge, just in case. 

[zoiasso]

Built PCB using satanistik design with abit change here and there. Using 1309 display, there is column offset. I have bunch of spare pcb if anyone need it, just throw some beer my way. Gerber file attached, so anyone can just send to manufacturer like seeedstudio or other cheap pcb builder. It's panelized so go grab your friend to share the mfg cost. You can attach the PCB directly to 1253B pcb.
U1 is 4071, im using Nexperia HEF4071BT, mouser pn 771-HEF4071BTD-T
U2 is 4078, im using TO CD4078BM, mouser pn 595-CD4078BM
U3 is 74AHC1G00W5-7, im using Diodes inc, mouser pn 621-74AHC1G00W5-7
Decoupling/bypass capacitor use 0805 smd cap. I'm using 100nF 16v just because it wander near my soldering station.... 
Aligning the PCB to go to the display daughterboard is trivial, I forgot to print alignment mark on the PCB, just scrape off a bit on the edge of the PCB to mark the edge location of pad 1 and pad 31, align it on the daughterboard and solder it like you would soldering any smd component.

File for the PCB is uploaded to Seeedstudio fusion galery as well, not making any money out of this link, it just make it easier for any of you that need the PCB in the future.

does this work with ssd1305z or ssd1306 ?

[onesystem]

This is in regards to U1273AX.
Obviously same screen problem.
On Keysight website there are two branches of firmware. Latest versions is 3.03 and 4.02
I compared the two differences, and major difference is in a file called "u1273ax_v303.ag" vs "u1273ax_v402.ag"
I am not a programmer so I do not know the .ag file extension. Surface search states that .ag file is a "applix graphic" file extension, which may or may not be this particular file. If anyone knows, please fill in the missing info.
Here is why I am posting this here. Since it is a definite that SSD1303 has been discontinued, it follows that current replacements by Keysight would be SSD1306 or SSD1309, which means automatically that firmware would be adjusted. So I figure that is why there are two branches of firmware - due to display differences... Let me know if my reasoning makes sense.
Maybe one of you does have firmware version 4.02 on your meter and can confirm if your OLED has a different driver chip on it - SSD1306 or SSD1309. Keysight used SOLOMON Systech company for their 1303 OLEDs.

Maybe problem these days is quite simple - buy the same screen that Keysight currently swapping their meters with, update with 4.02 firmware and there you go. I just got quoted for 402USD to swap my dim screen to a new one...

Can anyone confirm or deny my reasoning? I find it quite improbable that Keysight is still swapping their screens with SSD1303 drivers due to it's discontinued state.

[onesystem]

Here are both versions of U1273AX firmware for investigation:

V3:
https://drive.proton.me/urls/N55QYAYX80#j24y3g6IIvZT

V4:
https://drive.proton.me/urls/5J0FE0ZPMC#invyx6Y26hik

[onesystem]

Me again.

Main chip on meter is RENESAS D78F0547.
Here are links to programming manual:
https://drive.proton.me/urls/G0F5REFDBC#tp38ti26zHNf

and the datasheet:
https://drive.proton.me/urls/SF51ZB1RF4#R0M7AdtgC0SV

Is there a way to decompile the firmware update using the programming manual data? Or is it a 100% lost cause?

[onesystem]

Agilent U1273AX / Keysight U1273AX

Original Screen is SSD1303 driver, SOLOMON SYSTECH

Digital Multimeter microprocessor is D78F0547, datasheet and programming docs are in earlier post.

Keysight has firmware updates, links to them added in earlier post.
I erroneously thought firmware version 3 and version 4 are due to different screens. That was an unreasonable assumption. I do not know the difference but I can't verify if it is due to a different screen SSD version. In fact I have not seen on the net anyone claiming that their meter is with anything newer then 1303 from the factory.

My focus on now is how to disassemble the firmware, amend it's code to match a newer OLED driver, compile it back and push it over to the multimeter.

Searching the net I found a guy who wrote a disassembler for this type of chip. Here is the GitHub page:

https://github.com/mnaberez/k0dasm

His code is specific to a different chip though, so I had to update two of his files, command.py and symbols.py, with hardware vectors and special function registers for D78F0547.

Here are the links to these two updated files:

command.py - https://drive.proton.me/urls/SERRTAHHZM#XDrd7wwvzOkc

symbols.py - https://drive.proton.me/urls/NXZ45SNNMW#aJBO1dpW3PZv

Inside firmware you will find 18 files called Binary.NewBinary1 through Binary.NewBinary18. Those are the firmware for our chip (at least this is what I can see so far, please correct me if I am wrong).

Running each of these files through k0dasm, we get output which is to be processed through ASxxxx 5.50 Cross Assembler (https://shop-pdp.net/ashtml/asxget.php). Obviously before that it would have to be inspected and edited to amend the communication with OLED screen driver.

In terms of loading our edited firmware onto multimeter, Keysight allows to roll back the update to a previous version which in turn will allow to update it back to current version. Just swapping the Binary.NewBinary1 through Binary.NewBinary18 inside the update file with our amended ones then will allow to upload the firmware onto multimeter.

Where I have a steep learning curve ahead is to understand the output code from k0dasm. It is assembly and I do not know it. Maybe someone here can help, or maybe all of this is somehow a dead ending without me realizing it - let me know!

I hope to learn a lot as a result of this project. And obviously simplify replacement of the OLED screen.

[onesystem]

Cool leow4738! I am glad there is a spot where you can buy the SSD1303T10 OLED screens!

Except they charge an arm and a leg for it....

There is a firmware patch for U1273AX that anyone here can do to get the meter to work with OLED screen with driver SSD1309 which you can buy for 5th of the price - around $10 US dollars. And this patch assures that you could replace a fresh OLED screen on that meter any time now!
https://www.eevblog.com/forum/repair/agilent-u1273ax-dead-display/msg5282710/#msg5282710


I do not know if the firmware is the similar and if the main chip on U1253A is the same as on U1273AX. Maybe someone is interested to take a look and see if similar patch can be done for it too.

[DavidAlfa]

Please anyone give the modded firmwares a try!
https://www.eevblog.com/forum/microcontrollers/8-bit-78k0-d78f0547-help-with-firmware-modification/msg5144169/#msg5144169

[msquared]

The firmware upgrade at the link above works. 

I changed the SSD1303 display in my U1253A with a SSD1309 and upgraded the firmware with from the link above.

With just a few minutes of work and $10 for the display my meter is working again!

I'd suggest watching Stuart Rogers video for a few tips on successfully getting the old display removed from the PCB.


I've included a couple of pics of before the firmware upgrade and after.