Author Topic: HP E3631A PSU repair- processor programming - need advice (Read 4661 times)

Tracker · « **on:** January 10, 2019, 06:39:28 pm »

Hello folks,

On a E3631A PSU, can the uProcessor N87C51 be copied and programed to a new device?
it's located in the display panel and is used to read the front buttons.

Is this possible to do?
What do I need to do it?

Thanks
Tracker

picburner · « **Reply #1 on:** January 10, 2019, 08:35:50 pm »

If the mcu is 87C51 or even 87C51BHP and the lock bit 1 has not been set, you can read the contents of the internal rom.
But if you have fried just one of the lines that serve for reading there is no way to access the content of the rom.

Tracker · « **Reply #2 on:** January 10, 2019, 09:10:41 pm »

Ok thanks... its bad news I'm faraid.

Port zero is used for programming the EEprom and I blew a line on port zero!

My only hope now is to get another PSU and copy the eeprom from the good device.

Assuming I could get another good device to copy, can you outline what is needed to do it?

Thanks
Tracker

picburner · « **Reply #3 on:** January 10, 2019, 10:03:29 pm »

To read and program a mcu type 87c51 you need an eprom programmer that also manages the mcu, my wellon vp-590 can do it for example.
Remember, however, that if the manufacturer (HP) has set the mcu lock bits 1-2-3 you can not read the rom contents.

Gelf54 · « **Reply #4 on:** January 10, 2019, 10:13:42 pm »

Have you contacted KeySight to see if it might still available as a spare part?

https://www.keysight.com/find/contactpartid

picburner · « **Reply #5 on:** January 10, 2019, 10:59:25 pm »

The Keysight part number is 34401-88804 and unfortunately, as read on other forums, the device is locked and therefore not readable.
It is the same mcu used in some versions of the 34401A dmm vfd assembly.

Tracker · « **Reply #6 on:** January 10, 2019, 11:44:58 pm »

Thanks guys.
What a bummer!
I guess I could always operate it from the serial or GPIB, not ideal but at least it would work.

Contacting keysight is a good thought, I shall call them tomorrow.

And if anyone comes across a faulty or scrap E36xx series power supply, please give me a shout!

Cheers
Tracker

wictor · « **Reply #7 on:** January 11, 2019, 09:02:30 pm »

Hi,

I think the front panel cpu is same as in Hp 34401A. I have also E363x, so I might be able to read the FW. I know, that people have fixed the 34401 front panels by programming new cpu like 8751 or 8951.

I have broken vfd in my supply and I will replace it with OLED display some day. I already have parts for it, but I'm lacking time...

There's one good video about troubleshooting and repair similar cpu and display problem. See https://youtu.be/E3F59nj6Fho

Wictor

wictor · « **Reply #8 on:** January 12, 2019, 08:42:41 am »

Hi,

Here is photo of my broken E3634A and working E3631A front panels.

wictor · « **Reply #9 on:** January 12, 2019, 08:47:50 am »

Quote from: picburner on January 10, 2019, 10:59:25 pm

The Keysight part number is 34401-88804 and unfortunately, as read on other forums, the device is locked and therefore not readable.
It is the same mcu used in some versions of the 34401A dmm vfd assembly.

There is one version of HP 34401 front panel FW here: https://www.eevblog.com/forum/repair/hp-34401a-repair/
And here: http://www.ko4bb.com/getsimple/index.php?id=download&file=HP_Agilent/Agilent_34401A_E3646A_Display_board_MCU_firmware_dump_EPROM_34401-88804_RE-34401-88804.zip

wictor · « **Reply #10 on:** January 12, 2019, 09:10:06 am »

Quote from: Tracker on January 10, 2019, 09:10:41 pm

Ok thanks... its bad news I'm faraid.
Port zero is used for programming the EEprom and I blew a line on port zero!

My only hope now is to get another PSU and copy the eeprom from the good device.

Assuming I could get another good device to copy, can you outline what is needed to do it?

You will need programmer like TL866 and then PLCC44 adapter. Then you either copy the old chip or use binary file of the correct FW and burn it to e.g. AT89C51 version of the MCU.

Tracker · « **Reply #11 on:** January 13, 2019, 02:31:27 pm »

Hi Wictor,
You have just saved me from an enormous expense!
Keysight quoted me 500 Pounds Sterling to repair and said the uProcessor was not available as a part.
So I was about purchsase a damaged/faulty PSU from Israel which would have cost me > £200 plus dutys.
Many thanks for taking to time to reply to my post.
You have another subscriber now on YouTube

I really like your informative style.

Paljon kiitoksia

I have already odered the Microprocessor and I will research the EEprom programing device you have recommended for me.

Best regards,
Tracker - a very happy engineer!

wictor · « **Reply #12 on:** January 13, 2019, 03:57:34 pm »

Quote from: Tracker on January 13, 2019, 02:31:27 pm

Hi Wictor,
You have just saved me from an enormous expense!
Keysight quoted me 500 Pounds Sterling to repair and said the uProcessor was not available as a part.
So I was about purchsase a damaged/faulty PSU from Israel which would have cost me > £200 plus dutys.
Many thanks for taking to time to reply to my post.
You have another subscriber now on YouTube I really like your informative style.

Best regards,
Ray - a very happy engineer!

Hi Ray,

that Youtube video is not made by me

... I'm not feedbackloop. But anyway it's nice to help people here.

What is the actual power supply model that you have? I'm asking, because the front panel FW might still be different in E363x models than what it is in 34401A. I'm wondering, if I should disassemble my supply(E3634A) and try to read the FW for you.

Wictor

Tracker · « **Reply #13 on:** January 13, 2019, 05:34:08 pm »

Hi Wictor,
My model is the triple output E3631A.

My display PCB looks exactly like the photo above.
Although the button assignment may be slightly different from model to model I expect the main board uProcessor will just read the location which was activated and then map that to a function.

I shall try to program it and see how it goes. But if that fails I may well take your kind offer to read the firmware from a good unit.
Don't want you to de-solder the VFD unless absolutely necessary.

Regards
Tracker

Tracker · « **Reply #14 on:** January 13, 2019, 08:31:39 pm »

Hi Wictor,
I compared the circuit diagrams for the 34401A multimeter and the E3631A PSU.
In true HP fashon, they have made the uProcessor as universal as possible.
The difference being that the 34401A has 14 buttons amd the has 13 buttons plus a Rotary encoder.
Both products share the same Drive (Port 0, lines 0,1,2&3) and sense lines (Port 1, lines 0,1,2&3) for the button array which I guess is capable of reading 16 buttons if needed.
The rotary encoder sense lines are unpopulated on the 34401A (port 1, lines 6 &7).

So assuming HP/Agilent designed a common firmware load, which I think makes sense from a manufacturing point of view, then I should be ok with the 34401A firmware.

What do you think?

Tracker

wictor · « **Reply #15 on:** January 13, 2019, 08:49:35 pm »

Quote from: Tracker on January 13, 2019, 08:31:39 pm

Hi Wictor,
I compared the circuit diagrams for the 34401A multimeter and the E3631A PSU.
In true HP fashon, they have made the uProcessor as universal as possible.
The difference being that the 34401A has 14 buttons amd the has 13 buttons plus a Rotary encoder.
Both products share the same Drive (Port 0, lines 0,1,2&3) and sense lines (Port 1, lines 0,1,2&3) for the button array which I guess is capable of reading 16 buttons if needed.
The rotary encoder sense lines are unpopulated on the 34401A (port 1, lines 6 &7).

So assuming HP/Agilent designed a common firmware load, which I think makes sense from a manufacturing point of view, then I should be ok with the 34401A firmware.

What do you think?

Ray

Hi Ray,

I know that there some differencies between 34401 and E363xA as you said. I was thinking that since the original MCU is ROM version, it can't be programmed after it has been manufactured. And if same part code goes to 34401, 363x, 364x and to some other similar models, it should work.

There are at least two different version of the front panel for 34401 and I remember reading that they are not similar. I think there may be changes in protocol and maybe the main board MCU needs to be compatible also.

I know for sure that E3631A and E3634A front panels use common FW, since I have tested E3631A panel in E3634A and it works fine. The VFD is different in each model, but the FW does not care about that. I have decoded the protocol and the displayed text and annunciators are sent by the main board MCU.

If you have ordered the flash version of the MCU e.g. xx8951, you can program it multiple times. I would first try with the http://www.ko4bb.com/getsimple/index.php?id=download&file=HP_Agilent/Agilent_34401A_E3646A_Display_board_MCU_firmware_dump_EPROM_34401-88804_RE-34401-88804.zip and if it for some reason does not work, then I could desolder my 8051 and send you the FW from that. I don't have the VFD in E3634A, since the VFD was broken, when I bought the unit.

When you have programmed the MCU, just solder it back to board, but don't solder the VFD. Connect front panel back to PSU and power it up. If you hear one beep, main MCU should be able to communicate with front panel MCU. After that you can solder VFD like feedbackloop did in his repair video.

What are the symptoms of your device? Do you hear two beeps, when you power up the unit?

Tracker · « **Reply #16 on:** January 13, 2019, 09:39:36 pm »

Hi Wictor,
The symptom is fairly easy to explain.
On the display MCU, output on port 1 , line 2 is dead.
This means there are 4 buttons which are unresponsive and one of these happens to be the 'output enable' button.
Everything else on the PSU is perfect.

I ordered this MCU
https://www.ebay.co.uk/itm/AMD-N87C51-44-Pin-Plastic-PLCC-MCU-8BIT-32K-24MHZ-OTP-New-Lot-Quantity-2/251478619077?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649
N87C51 I beleive the 7 means EPROM so I should be able to re-program it if needed.

I also ordered the TL866II programmer so i will have some learning to do $:-\$

Thanks for the advice on testing it, if all goes well I should be able to enable the output without the VFD present.

Tracker · « **Reply #17 on:** January 13, 2019, 09:52:09 pm »

And I also ordered this as it may arrive faster...
https://www.ebay.co.uk/itm/AT89S51-24JU-Microcontroller-51-Flash4kx8bit-SRAM128B-Interface-UART-ATMEL/201902036492?epid=577130103&hash=item2f024c960c:rk:1:pf:0

wictor · « **Reply #18 on:** January 14, 2019, 04:47:24 am »

N87C51 can be programmed only once. 89xx variants can be programmed multiple times.

Wictor

Tracker · « **Reply #19 on:** January 14, 2019, 12:26:59 pm »

Thanks Wictor,
Glad I ordered the second one so.

Will update you in a week or so , when parts and programmer arrive.

Regards
Tracker

wictor · « **Reply #20 on:** January 22, 2019, 07:38:29 pm »

Hi Ray!

Have you received the MCU already?

It was -30C here in Finland, so I had to do something useful inside... I can now see my E3634A displaying some data

.

Cheers,
Wictor

Tracker · « **Reply #21 on:** January 22, 2019, 07:54:51 pm »

Hi Wictor,
Nice job, it looks excellent!

I have received the processors but the eBay seller of the programmer has let me down.
He claimed it was a UK stocked unit, but it's been over a week now. I suspect it's a Chinese seller pretending to be based in UK and using airmail. I shall see the postmark when it finally arrives.

Hope to get it soon!

Cheers
Tracker

Ordinaryman1971 · « **Reply #22 on:** January 23, 2019, 08:08:52 am »

The LED display looks great... I need two of those for E3646 that got damaged in transit some time ago. Are you planning on making it... available?

wictor · « **Reply #23 on:** January 23, 2019, 06:26:27 pm »

Quote from: Ordinaryman1971 on January 23, 2019, 08:08:52 am

The LED display looks great... I need two of those for E3646 that got damaged in transit some time ago. Are you planning on making it... available?

Most probably I will publish the code at some point. But now I only have first proto and I don't want to publish that. But the same HW that qu1ck made for HP34401 should work fine with E36xx also. See https://github.com/openscopeproject/HP34401a-OLED-HW for more info. I have different display model which has different pinout and stm32 bluepill. There are still unknown things in protocol that I need to solve.

Tracker · « **Reply #24 on:** February 06, 2019, 10:01:03 pm »

Hi Wictor,
Good news to report!

I finally received the TL8600 II Plus programmer, had to order a second one after getting a refund from the first eBay seller.
Programming was much easier than I expected, 15 minutes from box opening to having a programmed processor.

Unsoldering the display is a tricky operation, I would not recommend it to anyone with novice soldering skills, it's incredibly easy to detach a leg from the display. And due to the construction the leg will snap off at the glass so no chance to repair.
But fortunately it went well and I then replaced the processor, I just happened to have a Metcal 44 PIN Quad tip in my toolkit so that part was easy.

After that all went well and I'm not a happy camper with my PSU back to normal.

So many thanks for the advice and to the person who made the firmware available.

Tracker


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: HP E3631A PSU repair- processor programming - need advice (Read 4661 times)

Share me