Author Topic: EEVblog #933 - Keysight U1272A EMC Issue  (Read 146015 times)

0 Members and 1 Guest are viewing this topic.

Online PA0PBZ

  • Super Contributor
  • ***
  • Posts: 4133
  • Country: nl
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #100 on: October 29, 2016, 07:47:46 pm »
It probably meets all published specifications.

Keysight engineers found the majority of U1272A measurement modes and ranges meet published specifications in the presence of an RF field...
...
...and will revise the product’s data sheets accordingly.

Volkswagon anyone?
Keyboard error: Press F1 to continue.
 

Offline madires

  • Super Contributor
  • ***
  • Posts: 4781
  • Country: de
  • A qualified hobbyist ;)
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #101 on: October 29, 2016, 08:06:31 pm »
Keysight engineers found the majority of U1272A measurement modes and ranges meet published specifications in the presence of an RF field...
...
...and will revise the product’s data sheets accordingly.

Volkswagon anyone?

VW cheated on purpose. Keysight didn't, I'd assume.
 

Offline FrankBuss

  • Supporter
  • ****
  • Posts: 2280
  • Country: de
    • Frank Buss
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #102 on: October 29, 2016, 08:20:08 pm »
Right, I meant it probably meets the requirements of the RF field immunity test standards for all modes and ranges. I might be wrong, but if you read it carefully, what Keysight said was just that it doesn't always work in the presence of an RF field (I guess a strong field as tested by Dave was meant). So the meter sucks sometimes, but technically there is nothing wrong, no need to fix it. The datasheet gets an additional note that it doesn't work with very strong external electrical fields, case closed.

Dave, can you confirm that the Keysight Rep account is genuine? I have Keysight gear, too, a very nice DSO oscilloscope and wouldn't have expected such a solution. Or maybe Keysight could publish the analysis and the datasheet solution on their webpage.
So Long, and Thanks for All the Fish
Electronics, hiking, retro-computing, electronic music etc.: https://www.youtube.com/c/FrankBussProgrammer
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 3437
  • Country: gb
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #103 on: October 29, 2016, 08:34:56 pm »
If some of the results reported, and what we saw in Dave's video, are 'in specification' behaviour then I'm flabbergasted. It's quite clear from the observed behaviour that this meter is capable of giving highly misleading readings in circumstances where the average user would not be aware of the EMI causing the problems. Remember, this is a portable handheld device that one can reasonably expect to be used in less than ideal circumstances.

Dave's mentioned 'measurement confidence' before. If you can't trust your meter to give reliable, reasonably accurate readings in all predictable circumstances then you don't have that 'measurement confidence'. It's a short step in going from "I don't trust this meter" to "I don't trust this brand" and a brief announcement  that boils down to "It's all good. It's in specification. You'd better turn the low pass filter on for all your readings" is a good push in the 'don't trust this brand' direction.

It's also foolish to make that announcement late on a Friday* and then not follow up on people's concerns as they unfold over the weekend.

I'm not in the market for a new handheld meter at the moment. If in the future I am then Keysight's handling of this is definitely going to push me away from them towards the competition. Contrast with Fluke and the EMC issues with the 87V, they fixed it and fixed the meters that were in the field.

*A traditional time to publish announcements you want to bury. I'm not saying that was the intent here. If it wasn't then a 'late on Friday' publication was a particularly foolish thing to do.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 
The following users thanked this post: MK14

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 7890
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #104 on: October 29, 2016, 08:55:15 pm »
Right, I meant it probably meets the requirements of the RF field immunity test standards for all modes and ranges. I might be wrong, but if you read it carefully, what Keysight said was just that it doesn't always work in the presence of an RF field (I guess a strong field as tested by Dave was meant).

Yeah, can't really tell from what they wrote whether the susceptibility is compliant with the intended standards or not:

Keysight engineers found the majority of U1272A measurement modes and ranges meet published specifications in the presence of an RF field, but the product is susceptible to RF fields in certain measurement modes.

So the meter sucks sometimes, but technically there is nothing wrong, no need to fix it. The datasheet gets an additional note that it doesn't work with very strong external electrical fields, case closed.

It really boils down to whether or not the devices still pass the tests that they're supposed to. If they do, then adjusting the datasheet is one way to rectify the problem. However, doing so also implies that the product was never able to deliver on the performance that you expected and paid for. :-//
I TEA.
 

Online PA0PBZ

  • Super Contributor
  • ***
  • Posts: 4133
  • Country: nl
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #105 on: October 29, 2016, 09:32:32 pm »
VW cheated on purpose. Keysight didn't, I'd assume.

I totally agree, but I guess that VW by rectifying their little problem will have to adjust their spec sheet too. It makes no sense to cheat on purpose, then correcting it and then it having no influence on the engine specs. The similarity I see here is that you buy something and then after that the manufacturer changes the specifications.
« Last Edit: October 29, 2016, 09:34:18 pm by PA0PBZ »
Keyboard error: Press F1 to continue.
 

Offline retiredcaps

  • Super Contributor
  • ***
  • Posts: 3328
  • Country: ca
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #106 on: October 29, 2016, 11:07:46 pm »
How many people who have purchased this meter with a *datasheet footnote* stating such limitations?  Once discovered, word spreads quickly and many people who have never purchased this meter.

One thing Keysight can do is to offer a replacement meter of equal value that doesn't have this problem/limitation or give full refund.

PS.  Delivering this post on a Friday at business close is also poor PR.  I see too many people fired, let go, bad news delivered on a late Friday around 4PM.
 

Offline Dwaine

  • Frequent Contributor
  • **
  • Posts: 265
  • Country: ca
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #107 on: October 29, 2016, 11:09:19 pm »
Using the LPF helps the situation but it does not correct it completely.   The other meters Dave tested did not experience the same results as the Keysight meter.  The "high EMI" condition,  what exactly would define that?  Maybe a "low EMI" condition in the right situation will cause the same behaviour?

I'm a person that just bought a Keysight u1272a meter.  I paid a lot of money.  Thinking it would be the instrument to help me in my Heathkit SB100 restore project.  Dealing with high voltage during this project, the Keysight meter will be shut off and on the shelf. 

Back to the meter everyone loves to hate.  My trusty Rigol DM3068.   During my little restore project, the Rigol has done the business while the Keysight meter has failed.   Keep the Keysight meter away from those tubes or it will dance around.

It's not about the fancy spec sheets or chatter/excitement from people.  It is a piece of equipment on your bench in your lab and it's performance and the confidence it's right.  Using it with your stupid project , however big or small it is.

Sorry Keysight.  I wanted to really like this line of meter, but based on my experience it's not something I want to use in my lab. 

Oh well, a learning experience.  Life goes on.
 
The following users thanked this post: MK14

Offline FrankT

  • Regular Contributor
  • *
  • Posts: 151
  • Country: au
Updated specifications?
« Reply #108 on: October 29, 2016, 11:20:51 pm »
Is this the new specification...


 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 3529
  • Country: ca
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #109 on: October 29, 2016, 11:38:19 pm »

It's also foolish to make that announcement late on a Friday* and then not follow up on people's concerns as they unfold over the weekend.
...
*A traditional time to publish announcements you want to bury.

No no no, it was a shear coincidence that FBI released Hillary Clinton's use of private email server investigation docs on Friday afternoon on Labor Day long weekend.
Facebook-free life and Rigol-free shack.
 

Online EEVblog

  • Administrator
  • *****
  • Posts: 29866
  • Country: au
    • EEVblog
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #110 on: October 29, 2016, 11:41:01 pm »
Thanks again to the EEVblog and those on this thread who pointed out the EMC-related finding with Keysight’s U1272A family of digital multi-meters. Your input is very helpful and we value it. 
 
As mentioned in our earlier post, we asked our EMC technical experts to look into this – and they were able to replicate the symptom. Keysight engineers found the majority of U1272A measurement modes and ranges meet published specifications in the presence of an RF field, but the product is susceptible to RF fields in certain measurement modes.

You can significantly improve the immunity performance by using the U1272A’s low pass filter, especially during DC measurements. Based on the insight you provided, we are characterizing the accuracy of the U1272A’s performance in the presence of an RF field, and will revise the product’s data sheets accordingly. We will let you know via this blog when we have posted the revised datasheets on Keysight’s website.

For those of you who use a U1272A, please be aware that since we released firmware version 2.04 in 2013, you can independently configure the low pass filter for DC and DCAC measurement modes. You can refer to pages 131 through 133 in the U1272A product user guide for instructions and filter options.
 
And if you haven’t done so already, you can download the latest U1272A product firmware, Version 3.03, which was released in 2015. All firmware released since version 2.04 has the DC low pass filter turned “on” by default. To get the latest product updates as soon as they are available, sign up at myKeysight.
 
Again, we thank the EEVblog and everyone on this thread who have been involved in this productive conversation. This exchange reinforces Keysight’s belief in the important role your feedback and social media plays in improving product quality and increasing customer satisfaction.

The worst possible response would have been no response at all, so kudos for that.

But the 2nd worst possible response would have been "it meets specs", take it up the arse, and here's a kinda-sorta workaround to minimise it. Completely ignoring the fact that no other meter has reported doing this, and doing this whilst sitting next to standard test gear like a function gen on the bench or being used with typical modern circuitry like NFC.
How about a refund offer for those people who have found it "not fit for purpose"?
 :--
 
The following users thanked this post: MK14

Online EEVblog

  • Administrator
  • *****
  • Posts: 29866
  • Country: au
    • EEVblog
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #111 on: October 29, 2016, 11:43:29 pm »
Again, we thank the EEVblog and everyone on this thread who have been involved in this productive conversation. This exchange reinforces Keysight’s belief in the important role your feedback and social media plays in improving product quality and increasing customer satisfaction.

You have neither given any hint you'll improve the product quality, not have you increased customer satisfaction. In fact you've decreased it, well done :clap:
 
The following users thanked this post: madires, dr.diesel, WackyGerman, mtr, KE5FX, TheSteve, MK14

Online EEVblog

  • Administrator
  • *****
  • Posts: 29866
  • Country: au
    • EEVblog
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #112 on: October 29, 2016, 11:46:42 pm »
Dave's mentioned 'measurement confidence' before. If you can't trust your meter to give reliable, reasonably accurate readings in all predictable circumstances then you don't have that 'measurement confidence'. It's a short step in going from "I don't trust this meter" to "I don't trust this brand" and a brief announcement  that boils down to "It's all good. It's in specification. You'd better turn the low pass filter on for all your readings" is a good push in the 'don't trust this brand' direction.

+1
Bang on.

Quote
I'm not in the market for a new handheld meter at the moment. If in the future I am then Keysight's handling of this is definitely going to push me away from them towards the competition. Contrast with Fluke and the EMC issues with the 87V, they fixed it and fixed the meters that were in the field.

That's the thing. Fluke fixed the GSM issue which was an order of magnitude more niche and unlikely to happen, and didn't actually upset the reading.
 

Online EEVblog

  • Administrator
  • *****
  • Posts: 29866
  • Country: au
    • EEVblog
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #113 on: October 29, 2016, 11:51:13 pm »
Dave, can you confirm that the Keysight Rep account is genuine?

Yes. They use an @keysight.com social media email address, so would have had to have got forum account setup confirmation through that address.
 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 1074
  • Country: us
    • KE5FX.COM
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #114 on: October 29, 2016, 11:53:08 pm »
As someone with a significant amount of HP/Agilent/Keysight gear this is a pretty disappointing response to the problem.

Agreed.  If it weren't for all the cheap crappy Chinese competitors who performed well in the same test, it could almost be excused.

This response fails the "What Would Bill and Dave Do?" test, badly.
 

Offline blueskull

  • Supporter
  • ****
  • Posts: 12403
  • Country: cn
  • Power Electronics Guy
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #115 on: October 30, 2016, 12:37:22 am »
A hint for Keysight, if you want to cross pin point where the bug arose: U1461A does not suffer from this EMC issue. So the clue is, what's the difference in current measurement frontend between U1461A and U125/7x.
 

Online Brumby

  • Supporter
  • ****
  • Posts: 9300
  • Country: au
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #116 on: October 30, 2016, 04:04:00 am »
As referenced many times above, I'd like to underline one outstandingly prominent point....  As far as has been reported here - this problem is not found on any other meter - even the cheaper ones.

Turn on a filter?  Change the specs? ... and live with it???

Don't give an explanation? ... or offer a solution?


This is disappointing.

And after all the effort in getting the Keysight name out there with 'Scope Month', this little turd lands right in the middle of the arena - and it smells just as bad.
 

Offline blueskull

  • Supporter
  • ****
  • Posts: 12403
  • Country: cn
  • Power Electronics Guy
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #117 on: October 30, 2016, 04:11:39 am »
As referenced many times above, I'd like to underline one outstandingly prominent point....  As far as has been reported here - this problem is not found on any other meter - even the cheaper ones.

Turn on a filter?  Change the specs? ... and live with it???

Don't give an explanation? ... or offer a solution?


This is disappointing.

And after all the effort in getting the Keysight name out there with 'Scope Month', this little turd lands right in the middle of the arena - and it smells just as bad.

Probably this is a real hardware issue that cannot be solved by firmware unless the new firmware forcibly turns on LPF.
So, this demands a real product recall. Considering the quantity sold, especially for lower-mid range U123x and U125x, the recall will cost quite a lot.
If they really decide to do so, that requires the decision makers to have quite hard balls. The DMM department may run negative profit for months if not longer.
I don't think they will make such a decision so quickly. "We have a design bug that can't be fixed." "Let's just throw away our one year's revenue!"
 

Offline FrankT

  • Regular Contributor
  • *
  • Posts: 151
  • Country: au
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #118 on: October 30, 2016, 05:17:27 am »
Has anyone tried the recommended fix?

Measuring a 1mA reference, and holding a cable carrying a 10Mhz signal still gives the wrong answer.
 

Offline blueskull

  • Supporter
  • ****
  • Posts: 12403
  • Country: cn
  • Power Electronics Guy
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #119 on: October 30, 2016, 05:21:21 am »
Has anyone tried the recommended fix?

Measuring a 1mA reference, and holding a cable carrying a 10Mhz signal still gives the wrong answer.

Since turning on LPF solves the problem, I would say adding a non-bypass-able LPF capacitor to the PCB, setting it at a very high cut off frequency, above specified bandwidth, but lower than affecting RF frequency.
 

Offline MasterTech

  • Frequent Contributor
  • **
  • Posts: 891
  • Country: 00
Re: Updated specifications?
« Reply #120 on: October 30, 2016, 05:25:20 am »
Is this the new specification...
Am I reading correctly ? ten thousand percent (+.3) in accuracy ?

On a side note Ive been using a Vici 99, 30€ multimeter, to measure microamps in a 8kV line with great accuracy. Its the cheapest one I have and I use it here as it is expendable
« Last Edit: October 30, 2016, 06:52:02 am by MasterTech »
 

Offline KE5FX

  • Super Contributor
  • ***
  • Posts: 1074
  • Country: us
    • KE5FX.COM
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #121 on: October 30, 2016, 05:53:06 am »
Probably this is a real hardware issue that cannot be solved by firmware unless the new firmware forcibly turns on LPF.
So, this demands a real product recall. Considering the quantity sold, especially for lower-mid range U123x and U125x, the recall will cost quite a lot.
If they really decide to do so, that requires the decision makers to have quite hard balls. The DMM department may run negative profit for months if not longer.
I don't think they will make such a decision so quickly. "We have a design bug that can't be fixed." "Let's just throw away our one year's revenue!"

They'd only be expected to do that for a safety issue, which this isn't likely to be.  (Admittedly, anything that causes a DMM to display a grossly incorrect value under reasonable operating conditions has the potential to be a safety issue.) 

What will probably happen is that they'll develop a fix internally, and issue a service bulletin -- which may or may not be public -- that says "Apply this fix on customer request."
 

Offline WackyGerman

  • Regular Contributor
  • *
  • Posts: 209
  • Country: de
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #122 on: October 30, 2016, 07:11:31 am »
My 1241B doesn t have any filter so I have to live with that issue . The response of keysight is really disappointing even if I don t use the current measurement often .  :-- .
What would be happen if they find a safety issue ? Will they update the manual as follows : " Before measuring please ensure : Keep a fire extinguisher , your last will and a coffin within reach !!! " ???
You re right Dave , they decreased the customer satisfaction . My 1241B was the first and the last item I bought from Keysight . On the market there are more than enough contender who do their job better than Keysight .
 

Offline blueskull

  • Supporter
  • ****
  • Posts: 12403
  • Country: cn
  • Power Electronics Guy
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #123 on: October 30, 2016, 07:18:38 am »
You re right Dave , they decreased the customer satisfaction . My 1241B was the first and the last item I bought from Keysight . On the market there are more than enough contender who do their job better than Keysight.

But on the meantime, their price had dropped quite a bit. For instance, now thanks to Scope Month, you can get a touchscreen 1GHz scope for a price cheaper than similar product from Rigol.
I got my U1461A insulation testing DMM with more UI bugs (thankfully, none severe, just UI annoyance) than features, but hey, for the same feature set (programmable isolation test output voltage and cut off current, as well as bluetooth and USB interface, plus all features of U1273AX), a Megger would have costed me 3x the price if not more.
You cannot really expect competitive price plus good quality plus decades of experience plus good customer service.

I think the entire T&M market is moving from building "permanent asset" to building "3-year lifetime consumable".
 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 3529
  • Country: ca
Re: EEVblog #933 - Keysight U1272A EMC Issue
« Reply #124 on: October 30, 2016, 07:41:00 am »
Yes please go get yourself a Rigol, who is incompetent in RF design. I can't recall so much bashing  of Rigol when I published my horrific findings about their piece of crap DS2072A in my Project Yaigol thread. Where is your professional integrity? Rigol... Give me a break  :palm:

Facebook-free life and Rigol-free shack.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf