Replacing OLED screen on an Agilent U1253A Multimeter

[--Oz--]

Off Topic (but pertaining to this U1253B meter):
The 9v I use is 9v 600mAh (tested to 520mAh) or ~4.4Wh and does not last very long imo.

What are my options for longer life? (links)

I was thinking two 14500 li-ion protected cells would fit and they are 800~900mAH, ~7.1Wh.

Or ghetto it on the back with:
2s1p 18650 3500mAH cells on the back for 29Wh, or 6x my current battery,
2s2p 18650 3500mAH cells on the back for 58Wh, or 13x my current battery,
2s3p 18650 3500mAH cells on the back for 87Wh, or 20x my current battery.
TIA

PS: If I ever go this route, I know not to use the meter in high voltage applications, mostly I am under 50V.

[--Oz--]

Ordered boards from china and got them in 14 days. Will shoot for next week to assemble and test.

[--Oz--]

Success! Thanks to ALL that helped!

If anyone is interested in a assembled board (no display) and is located in CONUS, I am willing to sell a few to recoup my costs.
wforacer(at)rocketmail(dot)com
and please use A1253B in the email subject line, this will help me not miss the email.

Again, thanks to everyone for the awesome help here!

[--Oz--]

More pics of the finished project.

[vishaldotgupta]

please confirm the resolution of the oled.

oled presently available has slightly less horizontal resolution resulting in loss of two columns on either side.

in the DMM pic it seems the resolution parts is okay

[--Oz--]

please confirm the resolution of the oled.

oled presently available has slightly less horizontal resolution resulting in loss of two columns on either side.

in the DMM pic it seems the resolution parts is okay

With all the reading I have done on this subject months ago, seems the original display was not the standard 128 columns wide and was 132 wide.
Seems this display is OK in not having the right most column display on the far left column.
I will look for a meter with original still working display and compare, when I find one, I will report back.

[wraper]

please confirm the resolution of the oled.

oled presently available has slightly less horizontal resolution resulting in loss of two columns on either side.

in the DMM pic it seems the resolution parts is okay

With all the reading I have done on this subject months ago, seems the original display was not the standard 128 columns wide and was 132 wide.
Seems this display is OK in not having the right most column display on the far left column.
I will look for a meter with original still working display and compare, when I find one, I will report back.

It's a standard display and with same resolution. 132 columns are supported by SSD1303 controller, not actual display. IIRC you get a few columns wrapped around and bottom row shifted. Proper solutions were given in this thread which do not cause these artifacts.

[gamalot]

Thanks for all the very useful information, I have seen solutions for combinational logic, EPROM, FPGA and CPLD.

I'm trying to solve the problem with an STM32 MCU and everything seems to be working fine so far.

[moeyekung]

Hello.This can do on ssd1309?
Thank you.

I friend had problems with an Agilent U153A and asked me to find a way to fix it.
(Sorry Jonh for the delay...)

The original yellow OLED display was very very dim.

I followed the idea from Alexej Krom (with a twist) and worked well enough. Thanks Alexej.
http://www.alexej-krom.de/replacing-the-oled-ssd1303-display-with-ssd1305/

I found a pin compatible green display with SSD1305 controller, from ebay.
https://www.ebay.com/itm/Green-2-4-inch-128x64-OLED-Display-Module-I2C-Serial-SPI-SSD1305-w-Tutorial-/291844721833?hash=item43f34d0ca9&tfrom=301226745140&tpos=unknow&ttype=price&talgo=origal

The unsoldering and soldering the new one was almost easy... Nothing broke. I had more trouble positioning the display to match the front window.

I tried first an old UV EPROM that I had in hand but it was too slow and could not follow up the MCU control signals.

I used a different FLASH memory at 5V, SST39SF020A-70-4C-NHE PLCC-32, from ebay.
https://www.ebay.com/itm/123259658853
It was not a big surprise, it arrived with some BIOS firmware as a bonus!

I programmed it with a mini programmer from ebay called TL866CS and the optional PLCC32 adapter.
Worked great.

I used the 5V level signals side to make the soldering easier. Thats the reason I used the 5V FLASH.
The FLASH datasheet says it has 10mA consumption. Will see how much faster it drains the battery.

The mirror image was fixed and only some artifacts in the lower side of the display are shown when the fast moving bar graph is displayed and only in some screens. The readings and the other graphics are not affected.
I don't see any vertical line offset like others are mentioning.

OLED and Memory total, about $20.00.

Here are some pics and the schematics.

[gamalot]

Try different splash screens and have so much fun. 

[exe]

Am I correct that the issue of mirrored screen could be easily fixed in firmware (possibly just in initalization code), it's just keysight is not interested in it?

[wraper]

Am I correct that the issue of mirrored screen could be easily fixed in firmware (possibly just in initalization code), it's just keysight is not interested in it?

Yes, it can be easily fixed in firmware. A few lines of could would do the job. U1253A was discontinued a long time ago, so don't expect any. Dunno about U1253B and U1273A made in recent years, they actually could use display with a different controller as SSD1303 seems to be either obsolete or very scarcely available.

[flyglas]

Ordered boards from china and got them in 14 days. Will shoot for next week to assemble and test.

I have also ordered some circuit boards. If someone comes from Germany, I can send him some. Just send me a PN.

[gamalot]

Hi everyone, I need some help from you guys.

I'm still working on my U1253 OLED replacement, the first version of the PCB has some minor flaws, but it seems like all the expected features have been implemented.

My goal is a "perfect" replacement that should be better than the original one. Not just display all information correctly, but it also has a customizable splash screen and lower power consumption.

Preliminary test results show that my design consumes about 7mA less current than the original (31mA vs 38mA, battery voltage 9V, brightness setting high, DCV 5V range, reading 0.0000).

I watched Shahriar's (The Signal Path @Hugoneus) video. His meter with the replacement display draws more than 40mA, but one test obviously isn't enough, I really want more.

I would be very grateful if you could measure the power consumption of your U1253 multimeter (either the original screen or any version of the replacements) under the same conditions and post your test results.

thank you very much!

[--Oz--]

@gamalot, you can double your battery run time by using a 650~750mAh li-ion 2S 9v battery.
And could do more capacity with pouch lii-on cells.

[gamalot]

@gamalot, you can double your battery run time by using a 650~750mAh li-ion 2S 9v battery.
And could do more capacity with pouch lii-on cells.

Yes, choosing a battery with a higher capacity is a solution, but the high power consumption design of the original display and some replacements is still unsatisfactory to me, and no matter what kind of battery being used, a display with low power consumption is always a better choice.

[--Oz--]

The larger capacity battery is the lowest hanging fruit. So the original display board had 12v, 9v and 3.3v. I think all of these could be looked at for improved efficiency. The 3.3v is a linear reg is a no brainier to switch to a skip mode buck regulator. Regarding the boost regulator(s), a larger lower esr inductor will help a little, playing with switching freq can also help. You might want to try lowering the 12v and 9v outputs to shave some mA, same with 3.3v. Maybe an option to dim the backlight even more for low light applications. Is there settings in the driver to reduce power, like lower hz update rate/etc?
Milling out the case for a larger li-ion battery would be many fold more run time than all these small incremental improvements, but some would look down on that. 🤣

[TheSteve]

Custom boot logo is pretty nice.

I admit in my mind a patch to the original firmware would be the best overall solution.

[--Oz--]

Yes of course, but what's the chance that is going to happen? About the same as your daughter cleaning her room.

[TheSteve]

Yes of course, but what's the chance that is going to happen? About the same as your daughter cleaning her room.

Well I certainly wasn't suggesting Keysight would do it. Firmware for many devices have been patched by creative owners. This seems pretty "do-able", just takes one inspired person(and sadly maybe a few bricked meters).

[--Oz--]

Its semi easy to copy hardware (relatively), the FW is Agilent's secrete sauce, and to debug FW that cant be read out of the mpu is pretty tough. I agree, sure would be the cleanest way.

[gamalot]

Hacking the firmware looks like a good idea, but when there is an MCU or FPGA on the replacement display, the customized splash screen is very simple, and even some tricks can be made.

One of the features in my current implementation is that it can show a different splash screen every time you power up or wake up the meter (sequentially or randomly).

[gamalot]

Aside from those icing on the cake features, I think the most important thing I discovered and solved was why the MCU reads from the OLED display's RAM.

I've noticed that some readings only happen when the battery's voltage drops below a certain threshold (on my U1253B it's around 8.3V), and if the read back data is not what it expects, it shuts down immediately.

I'm not sure if the same thing happens with the U1273A as I don't have one. If you happen to have one and replaced the display module with one which has an MCU or FPGA, you can verify if it works at low battery voltages (my implementation works down to 6.2V)

[AVGresponding]

Aside from those icing on the cake features, I think the most important thing I discovered and solved was why the MCU reads from the OLED display's RAM.

I've noticed that some readings only happen when the battery's voltage drops below a certain threshold (on my U1253B it's around 8.3V), and if the read back data is not what it expects, it shuts down immediately.

I'm not sure if the same thing happens with the U1273A as I don't have one. If you happen to have one and replaced the display module with one which has an MCU or FPGA, you can verify if it works at low battery voltages (my implementation works down to 6.2V)

I have such a 1273A. I shall investigate.



At 6.0V (this meter uses 4x AAA) the meter pulls 26mA, absolute madness... the first battery warning, an "x" in the battery icon, comes on at around 4.6V, and it's using 32mA. The "Change Battery" text comes on at 4.25V, and the meter shuts down after a few seconds. It will power back up by pressing a button, but shuts off again quite quickly. The current draw at 4.25V is 36mA.

[gamalot]

Aside from those icing on the cake features, I think the most important thing I discovered and solved was why the MCU reads from the OLED display's RAM.

I've noticed that some readings only happen when the battery's voltage drops below a certain threshold (on my U1253B it's around 8.3V), and if the read back data is not what it expects, it shuts down immediately.

I'm not sure if the same thing happens with the U1273A as I don't have one. If you happen to have one and replaced the display module with one which has an MCU or FPGA, you can verify if it works at low battery voltages (my implementation works down to 6.2V)

I have such a 1273A. I shall investigate.



At 6.0V (this meter uses 4x AAA) the meter pulls 26mA, absolute madness... the first battery warning, an "x" in the battery icon, comes on at around 4.6V, and it's using 32mA. The "Change Battery" text comes on at 4.25V, and the meter shuts down after a few seconds. It will power back up by pressing a button, but shuts off again quite quickly. The current draw at 4.25V is 36mA.

Thanks for your testing work. It looks like your replacement display works fine at low voltage, maybe it's because the U1273 doesn't really care about the data read back from display, or the guy who made the replacement took care of that in the FPGA.

I know that the U1273 is a newer design, using low power devices, and the lower current is to be expected. But considering that most of the power consumption is on the OLED, your test results really shocked me!

For comparison, I made separate power connections for my display module and it draws 20mA at 6V and 27mA at 4.25V.

---

Sorry I made a stupid mistake! The test results above were measured with the brightness set to low. I measured 33mA at 6V and 47mA at 4.25V with the brightness set to high.

If your results are also with the brightness set to high, then there is a big problem with my design.