EEVblog #933 - Keysight U1272A EMC Issue

[alouko]

As far as I have understood, the problem is also in some other series, like 1240 and 1280.  They have not addressed that at all, yet.

And LPF doesn't really help.  I just tested with my U1273A and with 10MHz sine wave with 3V amplitude it still gives about -0.2A readings.  I would like to see the new spec sheet where they state +- 0.1% -2000 digits!  10MHz 3V sine is completely feasible voltage in many everyday circuits, not something extreme.

Well, I have my case in local Keysight office and I will probably suggest that they either fix my DMM or replace it with a working one with similar specs, be it either Keysight or Fluke or Metrahit!

If messieurs Hewlett and Packard were still alive, they would probably suffer a stroke because of how Keysight is handling this.

P.S.

At Keysight.com, you can see ad stream, first one stating "Capture Every Signal".  Well, at least that promise holds, even for DMMs 

[Brumby]

Since turning on LPF solves the problem,

That's not how I read it.  Turning on the LPF simply reduces the problem, not solve it - unless I've missed something.

[Brumby]

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.

Classic indeed.

[Fungus]

All firmware released since version 2.04 has the DC low pass filter turned “on” by default.

I wonder why... 

[TheAmmoniacal]

All firmware released since version 2.04 has the DC low pass filter turned “on” by default.

I wonder why... 

So will they send me a free USB-IR cable so I can upgrade to the new firmware? Like they did with the erratic display problem?

[joeqsmith]

Thanks to those on this thread for pointing out a potential EMC issue on the U1272A family of handheld multimeters. We have our best EMC technical experts at Keysight doing a deep dive to replicate the problem and determine its root cause. We expect them to complete the investigation within the week, and we’ll get back to the community on what we’ve found and any follow-up actions we may decide to take.

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.

......  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.

I would be interested in knowing if Keysight's experts were unable to find a solution or if higher up they decided it was not worth going after.

[BravoV]

......  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.

I would be interested in knowing if Keysight's experts were unable to find a solution or if higher up they decided it was not worth going after.

To me, its clear enough, they will not do anything about it, and that direction is clearly came from top management.

The highlighted words above is crystal clear .. its like fridge you bought, the spec and brochure when you purchased it said it was able to chill down to below freezing point, that turns water into ice in the freezer section.

Now, it will be revised "accordingly", its like ... "Sorry, the spec changed, your water will not turn into ice in the freezer, and enjoy chilled drink .. oh yeah ... your ice cream will melt if you put in there" ... go figure. 

[joeqsmith]

......  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.

I would be interested in knowing if Keysight's experts were unable to find a solution or if higher up they decided it was not worth going after.

To me, its clear enough, they will not do anything about it, and that direction is clearly came from top management.

They are doing something about it, they are changing the datasheet.  To be clear, that's not my question.  I want to know if their EMC experts have a solution and the company will not correct the design for reasons like loss of profits.   Is this a technical problem or a money problem?   

[Neilm]

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.

It says it meets "Commercial limits compliance with EN61326-1"

Ether they have to open the spec really wide (it measures current to +/- 5A) or they have to remove the CE mark and withdraw from sale in those parts of the world that require compliance to that standard (Europe certainly, I think Australia and don't know about US).

[joeqsmith]

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.

It says it meets "Commercial limits compliance with EN61326-1"

Ether they have to open the spec really wide (it measures current to +/- 5A) or they have to remove the CE mark and withdraw from sale in those parts of the world that require compliance to that standard (Europe certainly, I think Australia and don't know about US).

Does the CE mark require the meter meet EMC standards or just the safety standards?   I always assumed it was safety only for the meters as most of the ones I have looked at do not reference the EMC standards. 

Would really like to see Dave do an Amp hour on these standards with they interview someone who dares interpret them.   I would guess the commercial limits are 3V/m 80MHz to 1G vs 10V/m for industrial.  The standards cover a wider range.  Even if we use the battery powered Annex A:

Examples for equipment included in the scope of this annex but not limited to, are: equipment covered by the scope of IEC 61326-2-2, digital multi-meters,

it calls out 3V/m. 

Looking at conducted (61000-4-6), this is 3V (150KHz to 80MHz), performance criterion A.   

Criterion A is defined as:

The equipment shall continue to operate as intended during and after the test. No degradation
of performance or loss of function is allowed below a performance level specified by the
manufacturer, when the equipment is used as intended. The performance level may be
replaced by a permissible loss of performance. If the minimum performance level or the
permissible performance loss is not specified by the manufacturer, either of these may be
derived from the product description and documentation and what the user may reasonably
expect from the equipment if used as intended.

The way I read this, they can do anything they want as long as they meet their specified performance level.   I guess I don't see a problem with them changing their spec to solve it.  I actually wonder how bad it really is if the test were done correctly.  It may be fun to run one and just see how bad it really is and what it would take to correct it, assuming it does not pass.

Typically there is a fair amount of variance chamber to chamber.  Just calibrating the RF level is not trivial.  They will need to make sure that what ever they come up with that the product passes with some margin.   Should be interesting.   

[Neilm]

Does the CE mark require the meter meet EMC standards or just the safety standards?   I always assumed it was safety only for the meters as most of the ones I have looked at do not reference the EMC standards. 

The CE marks shows that the product meets ALL relevant EU directives. Low voltage directive (safety) and the EMC directive plus any other relevant directive (Radio Equipment Directive if it has wireless capabilities, Machinery directive if it has moving parts). The Declaration Of Conformity should state all the directives it meets.

[Dr.Trinity]

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.

Please tell me how to fix this problem.
We have many of the U1272A meters for teaching purpose and if this EMC problem is not solved for us, we have to replace all of them.

[FrankT]

My U1272a beeped at me this morning when I switch to the voltage setting and I had the lead in the mA socket.

Could this be related to the problem we are seeing?  Does it rely on picking up stray noise to detect the leads are in the wrong spot?

[TheAmmoniacal]

Poor Keysight..

I've had this meter for 3 years now and only had a measurement issue related to EMC twice, where I've had my function generator leads have been an entropic mess around my DMM and circuit. Now that I'm aware of it it's easy to avoid.

It's really a minor issue, not worth a recall or warranty replacement.. But it does change my impression of the Keysight brand somewhat.

I still think the slow audible continuity measurement is a much bigger issue, especially because it /has/ to be a damn firmware problem! (and fixable)

[TheAmmoniacal]

My U1272a beeped at me this morning when I switch to the voltage setting and I had the lead in the mA socket.

Could this be related to the problem we are seeing?  Does it rely on picking up stray noise to detect the leads are in the wrong spot?

No, this is normal behavior. It's supposed to protect the meter (and you). I can't remember what the U1272A use to detect the leads, some use IR/photodiode I believe.

The U1272A https://youtu.be/YsTZuPN7wQQ?t=9m1s

[joeqsmith]

Does the CE mark require the meter meet EMC standards or just the safety standards?   I always assumed it was safety only for the meters as most of the ones I have looked at do not reference the EMC standards. 

The CE marks shows that the product meets ALL relevant EU directives. Low voltage directive (safety) and the EMC directive plus any other relevant directive (Radio Equipment Directive if it has wireless capabilities, Machinery directive if it has moving parts). The Declaration Of Conformity should state all the directives it meets.

Could you point me to a document that states that the hand held multimeters are required to meet the EMC directive for the CE mark?   I have looked before and have yet to find anything.    I have seen reports for the handhelds from independent labs that did not call out the EMC directive.  The vast majority of the CE marked meters I have looked at do not mention it in their documents.   Even UNI-T's UT181A was certified by Intertek but has no reference to EMC.   

[2N3055]

 EN 61000-6-1 is intended for use by all manufacturers, designers and test houses associated with electrical and electronic equipment for use in residential, commercial and light-industrial  environments. It is used if there is no specific standard..

[joeqsmith]

EN 61000-6-1 is intended for use by all manufacturers, designers and test houses associated with electrical and electronic equipment for use in residential, commercial and light-industrial  environments. It is used if there is no specific standard..

I wonder what document causes them to use 61326 when an EMC standard is called out.  I have never seen 61000-6-1 called out with any of the hand held meters I have looked at.   

[Cerebus]

Could you point me to a document that states that the hand held multimeters are required to meet the EMC directive for the CE mark?   I have looked before and have yet to find anything.    I have seen reports for the handhelds from independent labs that did not call out the EMC directive.  The vast majority of the CE marked meters I have looked at do not mention it in their documents.   Even UNI-T's UT181A was certified by Intertek but has no reference to EMC.

That's simple - the EMC directive itself. Basically it says that everything falls under the EMC directive except a few specific exceptions that are already handled under other EC directives. And there's no way to quote a small, relevant gobbet here as the way the thing is written makes that impossible. It's only, from memory, 28 pages long in the English version but it'll take you about two days to read and properly understand, and it's one of the clearer EC directives to read.

[Neilm]

EN 61000-6-1 is intended for use by all manufacturers, designers and test houses associated with electrical and electronic equipment for use in residential, commercial and light-industrial  environments. It is used if there is no specific standard..

I wonder what document causes them to use 61326 when an EMC standard is called out.  I have never seen 61000-6-1 called out with any of the hand held meters I have looked at.

The first few pages of all IEC standards are usually available from the IEC website. One of the free sections is Scope - which indicates what products fall under the standard.

IEC61326 states that it applies to "test  and  measurement  equipment  which  is  portable  and  powered  by  battery  or  from  the circuit being measured." It also states that "This equipment may also be used in areas other than laboratories." This is relevant as the standard title mentions laboratories.

IEC61000-6-1 states in the scope that it is a " generic  EMC  immunity  standard  is  applicable  if  no  relevant  dedicated  product  or  product-family EMC immunity standard exists.  "

As IEC61326 exists its requirements superseded those of 6100-6-1

Looking at the scope for IEC61326-2-2 it seems to apply to instruments that are covered by the safety standard IEC61557 which covers low voltage insulation tester, RCD testers and the like. I don't believe it covers multi-meters (but I could be wrong. It does reference parts that are behind the pay-wall)

[2N3055]

I'm sorry Joe, I wasn't clear... EN 61326-1:2013 is Electrical equipment for measurement, control and laboratory use – EMC requirements ... But even if it didn't exist, you would have to comply with EN 61000-6-1, general requirements...

Basically, for CE on any electrical device you need to comply with Low Voltage directive and EMC directive... If there is specific directive for your type of equipment, than that is what you use in parts that apply to you..

But as Cerebus said, that whole directive crap was written by lawyers basically... So it's not human readable... That is why people pay so much for certification... Sometimes it's more complicated to figure out what should be tested than to test it...

[Brumby]

Just curious.....

Presuming the schematic isn't generally available, has anybody tried reverse engineering the front end of the troublesome meter to see where the problem might be?

... and to then, perhaps, compare this to a comparable meter without such problems?

[joeqsmith]

I'm sorry Joe, I wasn't clear... EN 61326-1:2013 is Electrical equipment for measurement, control and laboratory use – EMC requirements ... But even if it didn't exist, you would have to comply with EN 61000-6-1, general requirements...

Basically, for CE on any electrical device you need to comply with Low Voltage directive and EMC directive... If there is specific directive for your type of equipment, than that is what you use in parts that apply to you..

But as Cerebus said, that whole directive crap was written by lawyers basically... So it's not human readable... That is why people pay so much for certification... Sometimes it's more complicated to figure out what should be tested than to test it...

Could you point me to a document that states that the hand held multimeters are required to meet the EMC directive for the CE mark?   I have looked before and have yet to find anything.    I have seen reports for the handhelds from independent labs that did not call out the EMC directive.  The vast majority of the CE marked meters I have looked at do not mention it in their documents.   Even UNI-T's UT181A was certified by Intertek but has no reference to EMC.

That's simple - the EMC directive itself. Basically it says that everything falls under the EMC directive except a few specific exceptions that are already handled under other EC directives. And there's no way to quote a small, relevant gobbet here as the way the thing is written makes that impossible. It's only, from memory, 28 pages long in the English version but it'll take you about two days to read and properly understand, and it's one of the clearer EC directives to read.

See what happens when bureaucrats write standards.   If we look at the scope of 61326, it states:

The following equipment is covered by this standard.
a) Electrical measurement and test equipment
This is equipment which, by electrical means, measures, indicates or records one or more
electrical or non-electrical quantities, also non-measuring equipment such as signal
generators, measurement standards, power supplies and transducers.

Which would seem to include handhelds.  It's strange even the companies who do not pay for an independent cert and call out the safety standard do not call out the EMC standard.  Well, I have yet to see a case where they will call it out.  Not that there is any value in it, I just find it funny.   

What about handheld meters that are independently certified and their reports only mention 61010?   It could be as simple as a company only wants to pay of the safety cert and not the EMC.  The labs don't care what tests they run and it all gets called out in the report anyway what was done.   The mfgs are not required to have it independently certified so they are good to go.  The people concerned with safety get their independent cert mark and the few of us concerned with EMC get a worthless CE mark. 

Actually thinking about it, this seems to make sense.  The down side to this is as a consumer, you may purchase a meter with all sorts of markings like the one shown.  This may lead you to believe that the meter is safe and robust.   But really it may have been certified for any combination of EMC and safety.   It's up to the consumer to find the reports and read them.  If EMC testing is cheaper than safety, the company selling the product could pay only for the EMC cert.  You as a consumer have been told over and over to use independent cert for safety and will have no idea they only ran the EMC tests unless you read the report.      

Interesting stuff.  Would really enjoy a video where Dave walks us through the process of having his new meter certified.   

[joeqsmith]

Just curious.....

Presuming the schematic isn't generally available, has anybody tried reverse engineering the front end of the troublesome meter to see where the problem might be?

... and to then, perhaps, compare this to a comparable meter without such problems?

And this is why I asked about is this a money or a technical problem.  Because if Keysight's EMC experts can't come up with a fix, it would very funny to pick one of these up and take a crack at it!   

[Brumby]

Just curious.....

Presuming the schematic isn't generally available, has anybody tried reverse engineering the front end of the troublesome meter to see where the problem might be?

... and to then, perhaps, compare this to a comparable meter without such problems?

And this is why I asked about is this a money or a technical problem.  Because if Keysight's EMC experts can't come up with a fix, it would very funny to pick one of these up and take a crack at it!   

Who said Keysight's EMC experts can't come up with a fix?  For all we have been told, they may well have.

Based on a typical Corporate approach, the difference between what their engineers know and what is being announced is likely to be a gulf that's so wide, it makes the Grand Canyon look like a crack in the sidewalk.  What's being spouted here looks like 'damage control' ... as Dave said, "Classic corporate response".  It's a pretty lame attempt when you consider the affected audience primarily consists of people that understand how the equipment works.  Doing the arse-covering exercise is going to come from the legal and beancounter departments, with the marketing department being brought 'on-board'.  Bugger the engineers.

I would be pretty comfortable to say that there will be at least one or two engineers there that have an idea on how to fix it.  They might even have been allowed to spend some time putting together a fix, just to quantify the cost - but I doubt it.  IF they have worked out something, I reckon they will have done so in their own time.  But the fix (if it exists) will never see the light of day - officially  (IMHO).

Doesn't matter if it's only a matter of changing a component or adding a bodge wire - the cost would be more than Keysight would want to spend.  As soon as such an announcement were made, people would be expecting Keysight to fix it for them.  Even if the fix itself took only 3 minutes, the logistics cost of receiving and returning the meters would not be insignificant.

The only way for something like this to come to light is if it were to come from a totally independent source - and don't expect Keysight to even acknowledge it.

However, the reality of the problem might mean the fix is a redesign of a section of the PCB - with different components and different circuit topology - which means a level of difficulty a couple of orders of magnitude higher.  I tend to think this is closer to the truth than a simple component change ... but that's only a gut feel.