Failed firmware update on U1273A and Agilent's Response

[TimothySwieter]

So I've had a Agilent U1273A Multimeter for 11 months now.  I got the USB communication kit a while ago, and when I was on Agilent's site last week I noticed there was a firmware upgrade for the meter.  So I downloaded it and proceeded to flash my multimeter.  But....

...Agilent Windows software crashed, while uploading the firmware to the meter!!     The software just disappeared with no error message or nothing.  Now my meter is stuck with a "D...." on the OLED screen and the firmware upload won't work any more and the meter surely doesn't function. 

I e-mailed Agilent's customer service hoping they could help.  At first they had trouble locating my serial number in their system, but after I got them all the documentation they then tell me that my meter is only under a 3 month warranty and not a 1 year warranty.  How was I to know that?  Further more they state:

Since model U1273A is throwaway product, no Per Incident repairs available.  Please purchase a new unit if not covered by warranty.
Due to Agilent cannot support component repair, I’m afraid you have to purchase a new one if it is faulty.

Really Agilent?  You throw these away?  Let me find your dumpster then.
Surely they had designed a better product than this.  Obviously something is driving the OLED screen and displaying text, so there is some sort of "life" (read: processing) going on in the meter.
I've bought thousands of dollars of Agilent equipment, and there response to my only call for warranty and support is "throw it away, buy a new one".  No!!  I don't throw away tools that cost this much.

Anyone else have similar trouble with firmware upgrades or customer support?
Any tricks for getting inside to do a reset or flashing of the MCU?

This is ridiculous service from Agilent.  I'm even willing to pay a small fee, say US$40, to get my meter repaired, but they told me to throw it away.   

[Orange]

Which country are you located ?
Did you buy it trough an authorized seller ?


The U1273A comes with 3 YEARS warranty and 3 months on accessories, see the product brochure, it is stated clearly....

Unbelievable story

[nowlan]

Surely they were referring to the packaging when they said to throw away. 

I think Europe is 3 years by law.

[nowlan]

Warranty information - link

* 6. Agilent warrants that Agilent Service will be provided in a professional and workmanlike manner. For ninety (90) days from the date of repair, Agilent will replace, at no charge, defective parts used in Agilent’s repair of Products.

* 3. Agilent warrants that Software will not fail to execute its programming instructions due to defects in materials and workmanship when properly installed and used on the hardware designated by Agilent. Agilent warrants that Agilent owned standard Software substantially conforms to Specifications. Agilent does not warrant that Software will operate in hardware and software combinations selected by Customer, or meet requirements specified by Customer.

[Richard Crowley]

So what was so compelling about the "upgrade"?  How do you "upgrade" a basic piece of test gear?
This is exactly why I very rarely "upgrade" anything, software, firmware, or hardware.
Only when there is some overwhelming reason to take the risk.
One man's "upgrade" is another's disaster.

[T3sl4co1l]

Obviously, the software was not "properly installed".  Proof: crash during update.

Shitty...

Tim

[firewalker]

No secret key combination with a boot loader to enter update mode, and reload the firmware?

Alexander.

[TimothySwieter]

Thank you guys for the reply. 

I bought it in the USA, but I am currently living in Hong Kong.  I contact Agilent through their web site, and I've been working with a local customer service rep who has been kind and nice, but is firm that there is nothing to do.  I see on the paper work with the meter the 3 month warranty for the accessories. 

I am just dumbfounded that the rep is stating there is no way to repair this.  I understand "bricking" some pieces when upgrading firmware.  This unit at least has some responsiveness like it is going into a download mode when you restart the meter. 

I was upgrading the firmware because of the smoothing function referenced in the release notes.

[macboy]

The U1273A definitely comes with a 3-year warranty. It was released in late 2011, so every unit ever sold should still be under warranty. Contact Agilent again. If you don't get the answer you need from the person you speak to, ask to speak to their supervisor. Or try to contact Agilent customer relations, or a regional manager or representative. If you do lots of busniess with them, so contact your sales rep and let him know that you don't appreciate the grief you are getting. He has a vested interested in keeping you as a customer. Once people need to start answering to someone else internally, their helpfulness toward you can change quickly.

[tsmith35]

So what was so compelling about the "upgrade"?  How do you "upgrade" a basic piece of test gear?
This is exactly why I very rarely "upgrade" anything, software, firmware, or hardware.
Only when there is some overwhelming reason to take the risk.
One man's "upgrade" is another's disaster.

While there is always a risk associated with updating anything electronic, I avoid letting fear, uncertainty and doubt prevent me from keeping the firmware on my equipment up to date. With most firmware updates, there is typically no overwhelmingly important reason to install them, but if these updates are so insignificant, why would manufacturers even bother to develop and release firmware updates?

Myself, I have updated firmware thousands of times on everything including phones, meters, computers, calculators, watches, PLCs, televisions, streaming media players, game consoles, an automobile (dealer did that one), mp3 players, CD/DVD/Bluray drives, hard drives, SSDs, routers, modems, APs, USB3 chipsets, printers and so on.

How many unrecoverable firmware update failures have I had? I can only think of one: a PC motherboard. How many recoverable firmware failures have I had? Probably half a dozen, but the firmware update software in each case was well-written and allowed me to retry the update until successful.

Fear of updating firmware isn't the issue here: Agilent's failure to honor their three-year warranty is the real issue.

If you're curious about what was fixed with the firmware update, take a look here: http://www.home.agilent.com/upload/cmc_upload/All/U1273A_firmware_release_notes.txt

It's interesting that Agilent believes their multimeters are "throwaway" products, while Fluke understands that their own multimeters are essential pieces of equipment. It appears that Agilent doesn't have a lot of regard for their multimeter customers.

[grumpydoc]

As has been pointed out above the meter itself should have a 3-year warranty.

The 90 days pointed out by nowlan refers to repairs under warranty which are themselves warranted for 90 days (why not the longer of the original warranty or 90 days I don't know).

Still seems a shit response from Agilent

Did you buy the meter new?

[linux-works]

that's a $400 meter!

throw-away?  excuse me??

wow.  unbelievable.  that does not make sense, though.  not one company would 'throw away' a $400 meter in today's economy!  not sure what planet that guy on the phone was from, but I would try again and ask for a supervisor.

tell them you are updating a forum thread with the supervisor's response, and so that might make them take this matter a bit more seriously.  maybe.

also, get it in writing; if that person says it cannot and will not be fixed, get THAT statement in writing.  I bet blogs would love to hear about stuff like this if its really true.

as for firmware bricking, its always a possibility but properly designed gear, TODAY, does not brick due to a failed upgrade of firmware.  I own quite a bit of olympus gear and olympus has done a pretty bad job (over the years) of their firmware upgrade procedure.  many folks on the camera forums had bricked their lenses (yes, they have firmware, too!) and camera bodies.  usually oly will reflash the camera using the service port but its really a sign of sub-par software and hardware design for embedded systems to be brickable due to a failed upgrade.

I would expect this from 1hl companies.  I would NOT expect any firmware fails from HP of all places!

(ok, not real HP anymore, we all know that)

[nanofrog]

Since the meter was purchased in the US, perhaps it would be better to contact Agilent there instead of through the Hong Kong office?

The response received from HK might be the result of cultural differences.  PITA, but worth a shot IMHO.

[macboy]

that's a $400 meter!

throw-away?  excuse me??

It's "throw away" as in, any fault is considered "BER": Beyond Economic Repair. That's fancy talk to say that it is cheaper to buy a new one than to spend time and money trying to repair the fault, or that it is technologically impossible to repair (e.g. glob-top IC).

For the end user, this means that if it fails during warranty, they should give you a new one instead of a refurbished one (since they don't do repairs), and that after the warranty is over, they don't offer for-cost repair services.

[Khashoggi]

Pull the battery out and wait a minute. Check to see if PC software communicates. If not, pull battery and let it sit overnight to drain any onboard memory retention caps and try again with PC in the morning.

[linux-works]

that's a $400 meter!

throw-away?  excuse me??

It's "throw away" as in, any fault is considered "BER": Beyond Economic Repair. That's fancy talk to say that it is cheaper to buy a new one than to spend time and money trying to repair the fault, or that it is technologically impossible to repair (e.g. glob-top IC).

For the end user, this means that if it fails during warranty, they should give you a new one instead of a refurbished one (since they don't do repairs), and that after the warranty is over, they don't offer for-cost repair services.

$30 is a throw-away meter.

$400 is just not.  not even for the US gov that has bottomless pockets

I think this must really be a mistranslation or that the HP rep thought this was an international warranty claim on a US product, or something.  (reminds me of how arrogant nikon is if you buy a greymarket item and want to pay to have it fixed; they will usually refuse you!  in that case, its all about warranty politics and not ability to repair)

[retiredcaps]

You might want to PM eevblog userid mike_kawasaki who works at Agilent.  He participated in this thread about the U1272A crappy soldering and was helpful.  He doesn't regularly login here so make sure to PM him to get his attention.

https://www.eevblog.com/forum/testgear/agilent-u1272a-crappy-soldering/

[jpb]

I was looking at the service manual for the Agilent 53230 counter. Essentially the only replacement parts seem to be the power supply, the whole motherboard and the processor board and most faults led to the "solution" replace motherboard. (I've not included the display and optional 6GHz or 15GHz modules.)

So it seems that even for a very expensive (up to around $9k) units the only repair is to just replace the bulk of the unit. Even the OCXO option is a part of the motherboard - they sell separate motherboard parts for OCXO or TCXO options.

I'm not surprised, given this, that a $400 meter is treated as a throw-it-away-if-it-is-broken indivisible module. I share people's shock at this but it seems to be the most economical approach from Agilent's point of view.

It means that, if like me, you're looking to buy a used item from Agilent on e-bay it is a bit of a gamble as if it becomes faulty replacing the core motherboard will probably cost more than what you've paid for the whole unit.

Of course the main issue in this thread is that the OP's meter should still be under warranty and thus be replaced by Agilent.

[linux-works]

yes, its true that they mostly just swap whole boards, these days.  do they repair the boards and re-use them later?  they should, but maybe the cost of labor is not worth it to them and people who buy NEW agilent generally can afford to eat an occasional loss here and there.

but for us home lab guys, it sounds like most modern gear is not worth owning if you need to count on it AND keep it running yourself.  we pretty much knew this anyway and its why older ebay gear is more appealing to many of us than newer gear.  I won't buy new tek scopes (can't really fix them) and so I'll buy the older ones and keep them running as long as I can.

still, we are talking about a FIRMWARE failure, here.  to not have a way to debrick the module in a cost effective way is pretty insulting.  I would understand a board level failure being too expensive to repair but not a failed firmware flash update!  is there no jtag port at all?   of course there is, how else would they load initial firmware?

[Richard Crowley]

While there is always a risk associated with updating anything electronic, I avoid letting fear, uncertainty and doubt prevent me from keeping the firmware on my equipment up to date. With most firmware updates, there is typically no overwhelmingly important reason to install them, but if these updates are so insignificant, why would manufacturers even bother to develop and release firmware updates?

To my way of thinking the risk/benefit is strongly negative:
RISK: A "bricked" multi-$100 piece of gear, and then having to deal with an intransigent, faceless, soul-less multi-national corp. (Witness the OP's account.)   
BENEFIT: Fixing some problem that doesn't affect me, or adding some feature I don't use or need.

Myself, I have updated firmware thousands of times on everything including phones, meters, computers, calculators, watches, PLCs, televisions, streaming media players, game consoles, an automobile (dealer did that one), mp3 players, CD/DVD/Bluray drives, hard drives, SSDs, routers, modems, APs, USB3 chipsets, printers and so on.

How many unrecoverable firmware update failures have I had? I can only think of one: a PC motherboard. How many recoverable firmware failures have I had? Probably half a dozen, but the firmware update software in each case was well-written and allowed me to retry the update until successful.

I genuinely applaud and envy your luck.  Your upgrade karma is clearly several orders above mine. I have a growing stack of bricked stuff that is probably note even worth a teardown video. Every time I turn on the GPS in my car, TomTom natters at me that my maps are "25 months old", and I should upgrade them.  Now I think that upgrading the maps in my GPS is a legitimate thing, but alas, no computer I have ever tried will recognize my GPS when I plug in the USB cable. 

Fear of updating firmware isn't the issue here: Agilent's failure to honor their three-year warranty is the real issue.

If you're curious about what was fixed with the firmware update, take a look here: http://www.home.agilent.com/upload/cmc_upload/All/U1273A_firmware_release_notes.txt

It's interesting that Agilent believes their multimeters are "throwaway" products, while Fluke understands that their own multimeters are essential pieces of equipment. It appears that Agilent doesn't have a lot of regard for their multimeter customers.

Apparently $400 is in the "disposable" category for HP/Agilent/Keysight. Clearly, they have a significantly different world-view than most of us here.

[jpb]

yes, its true that they mostly just swap whole boards, these days.  do they repair the boards and re-use them later?  they should, but maybe the cost of labor is not worth it to them and people who buy NEW agilent generally can afford to eat an occasional loss here and there.

I would guess that they retreive some components but don't repair the board. I would have thought that they would rescue OCXOs if they still work but perhaps everything else is not worth the cost of recovery (the cost to them of just producing a new board as part of manufacture is probably a lot less than the labour cost of mending boards or even retrieving components from them).

I agree about the older equipment, unfortunately in my case I have very limited space on my "work bench/lab" (really just a modified desk) and old equipment is just too bulky so I have to opt for smaller modern equipment and hope it doesn't fail.

If you have the space you can get some very high spec and well built equipment for very reasonable sums.

[grumpydoc]

It's "throw away" as in, any fault is considered "BER": Beyond Economic Repair. That's fancy talk to say that it is cheaper to buy a new one than to spend time and money trying to repair the fault, or that it is technologically impossible to repair (e.g. glob-top IC).

For the end user, this means that if it fails during warranty, they should give you a new one instead of a refurbished one (since they don't do repairs), and that after the warranty is over, they don't offer for-cost repair services.

http://topdocumentaryfilms.com/light-bulb-conspiracy/

[nanofrog]

of course there is, how else would they load initial firmware?

Easy actually... bulk programming. Either before the parts are ever stuffed onto the boards, or after they've been soldered using a bed of nails jig.  

[linux-works]

right, bed of nails or pogo sticks but STILL a programming port onboard.

I would think an onboard port would exist, if for no other reason that they can do manufacturing testing on the whole board, then.  flashing the ship and saying 'ship it' seems a bit low-service, even for the current company known as keyslight or whatever they call themselves today.

[nanofrog]

right, bed of nails or pogo sticks but STILL a programming port onboard.

I would think an onboard port would exist, if for no other reason that they can do manufacturing testing on the whole board, then.  flashing the ship and saying 'ship it' seems a bit low-service, even for the current company known as keyslight or whatever they call themselves today.

Can be done directly on the pins/pads post soldering if they can be contacted properly with pogo pins, which is why they're desirable for in-circuit testing (and programming). Possible to program and test using the same jig. So adding a port during routing isn't an absolute necessity for all cases (depends on component package/s used). 

And a port certainly isn't necessary if the programming is done prior to stuffing & soldering (even the bottom side of a BGA is accessible at this point).