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.
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.
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.
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.
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.
Is this the new specification...
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.
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"?
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
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.
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.
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.
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.
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.
Has anyone tried the recommended fix?
Measuring a 1mA reference, and holding a cable carrying a 10Mhz signal still gives the wrong answer.
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
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."
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 .
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
I think the entire T&M market is moving from building "permanent asset" to building "3-year lifetime consumable".
For sure, Agilent (pre Keysight) did it long time ago , not a big deal though, but just look at the label at my 3 yrs old U1242B, no harsh nor field used, only at the lab all the time.
Heck, even screen printing at common cheap 2 dollars DMM like DM803's label don't wear of that easily.
-keysightagilent-uses-cheap-paint-for-label/?action=dlattach;attach=168660;image)
Maybe they dropped the price for some other stuff , but here in Germany the price for the U1272A is close to the Fluke 87V . Fluke did do their work and fixed the emc issue , Keysight didn t . When buying a multimeter for 460 € you can expect more customer service that that
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."
Yep, I think they will quietly fix this in a future design revision.
I suspect they came to the conclusion that it's better to give the nothing response they gave and take the heat for that, then issue a notice admitting a major issue and risking any legal consequences. When they basically said "It meets the specs", that covers their arse fairly nicely legally, and gives them a long time to fix it on the quiet.
Classic corporate response.
Really ? I think that s a personal opinion . For me the F289 plays in it s own ligue with its dot matrix display
Really ? I think that s a personal opinion . For me the F289 plays in it s own ligue with its dot matrix display
+1, the logging and capability of displaying it at the meter, this alone made the F289/287 is more like a hand held ultra high resolution oscilloscope, not high speed though.
+1, the logging and capability of displaying it at the meter, this alone made the F289/287 is more like a hand held ultra high resolution oscilloscope, not high speed though.
While Keysight meters cannot plot on meter, but they can display individual data points on meter.
Also, there are a few OLED models that are dot-matrix based.
Dot matrix ? Yeah, still no where near plotting this kind of graph on the spot "at the handheld".
Example -> ->
F28X log example Its temperature log for 8 hours that can be zoomed down to seconds, made this log while ago to check the newly installed air con unit.
Imo, the dot matrix model you mentioned is not even the same league, c'mon, you own F289 as well, you have to agree F287/289 is like handheld scope, but with 1000V capability and the whole shebangs built in it.
You cannot really expect competitive price plus good quality plus decades of experience plus good customer service.
For nearly EUR 500 I expect good quality plus decades of experience plus good customer service. After Keysight's response I wouldn't pay more than EUR 200 for the U1272A. Somehow I got the feeling, that Keysight just discounted everything they sell by 50%.