Keysight 34460A/34461A/34465A/34470A Maximum Allowable Voltage change notice.

[nctnico]

I'd been thinking about buying a 34465A for my bench, but this spec change would be a deal-breaker for me so I'm glad I haven't pulled the trigger.  I'll be curious to watch Keysight's response to the issue - anything less than a free replacement for affected users would seriously damage their credibility in my eyes.

The problem right now is that Keysight has nothing available even to replace these meters with (except, perhaps, 3458A but that would be a bit much to expect  ), so if you need to measure voltages over 600V DC and 440AC accurately, you have to look elsewhere or resort to some band-aids like an external voltage divider.

Let's wait and see what Keysight comes up with. It is not like the DMMs will catch fire when you apply 1000V. Maybe Daniel from Keysight can chime in?
An alternative could be a DMM from Tektronix and it wouldn't surprise me if they have a compatible command set.

[Alex Nikitin]

It is not like the DMMs will catch fire when you apply 1000V.

We don't know that yet.

Cheers

Alex

[agdr]

Geezzz, this was a close call.  I almost bought a 34470A two weeks ago, but got sidetracked when the R&S RTB scope deal came up.  The other choice was the Fluke 8846A.  I just took another look and sure enough the Fluke is (still!) rated at 1000V AC and DC.  So along with all the other reasons, Keysight needs to get this fixed to stay up with the competition.

[Cerebus]

This is beginning to smell like the U127xA all over again.

Their response to that was to change the published specifications and it was, in retrospect, clear that they had no plan to recall/replace these until they got given some stick on here. You'd think they would have learned something that time around and once they discovered this problem would have put a complete plan in place and made complete, cogent and clear details available.

The absence of those complete, clear and cogent details a second time around does immeasurable harm to the Keysight brand. When you make instruments that people the world around rely on you have to act proactively in a situation like this to maintain people's confidence and failure to do so yet again is inexcusable.  If a couple of the right stock analysts catch wind of this repeated incompetence after the U127xA debacle then Keysight's stock is in for a pounding.

A half-arsed de-specification of the product is not enough. You need to:

1) Proactively tell all affected customers about the problem.

2) Explain the problem honestly, fully and clearly. People with this kit in production use have to be able to evaluate what action they have to take. People can't just turn production lines off if they are affected by this. At the same time, they also need to know of there is a genuine safety issue or measurement confidence/reliability problems if they do keep equipment in production. And they shouldn't have to waste time and money reactively chasing you to get an explanation.

3) Tell people what you are going to do to fix the problem that you created for them. Are they getting replacements? When?

And, given that we're doing this again after the U127xA affair:

4) Explain what you're going to do to make sure that:
(a) Other recent products in the field aren't going to turn out to have similar problems because you've spent too much time and money changing the company name, and not enough time and money on designing and manufacturing reliable multimeters.
(b) Future products aren't going to be similarly laden with defects that show up in the field.
(c) When or if this happens again you'll move expeditiously to inform customers, promptly explain the problem fully and honestly, and replace any affected products as quickly as humanly possible.

As it stands at the moment I have zero confidence with anything with a Keysight badge on the front of it unless it once bore a pukka HP badge and I know the product hasn't changed in the interim.

Keysight management ought to have figured all of this this out for themselves, but if they can't then I and "Mr. Cluebat" are available at our usual extortionate consultancy rates to explain it to them. Oh, and if you decide to change the company name yet again (after this you may need to) may I suggest "HalveTheSpec".

Yes, I know that amongst the serious stuff I've been a bit sarcastic, but having to say this stuff again can make you feel a bit edgy.

[d-smes]

What's USA warranty for "defect"?  I know this is a "design defect" and most warranties only state "manufacturing defects", but thought I'd ask.  I'm outside the one year mark, so suspect I have no warranty claim.  Can anyone confirm?

[TheSteve]

Warranty on the 3446x/70A series is 3 years.

This is quite disappointing. I bought a U1273A and it turned out to have the conducted interference issue. I bought a new 34461A and now I've lost 40% of the potential range it can measure voltage(albeit at the high end where I rarely use it).

[wraper]

What's USA warranty for "defect"?  I know this is a "design defect" and most warranties only state "manufacturing defects", but thought I'd ask.  I'm outside the one year mark, so suspect I have no warranty claim.  Can anyone confirm?

3 year warranty. Also on previous instances if there were any widespread issues, Keysight fixed them for free even out of warranty period.

[GlowingGhoul]

Keysight has always provided me excellent support in the past, I have no doubt they'll take care of this issue.

Besides, this is such a cut and dry manufacturing defect, they must know that in the case of a class action lawsuit, they'll not only end up paying up, but they'll  take a permanent hit to their reputation.

My guess is we'll see a proposed resolution this week.

[EEVblog]

The problem right now is that Keysight has nothing available even to replace these meters with (except, perhaps, 3458A but that would be a bit much to expect  ), so if you need to measure voltages over 600V DC and 440AC accurately, you have to look elsewhere or resort to some band-aids like an external voltage divider.

What's the big deal?
How many people need to measure between 600V and 1000V regularly in practice?

[EEVblog]

This is beginning to smell like the U127xA all over again.

Their response to that was to change the published specifications and it was, in retrospect, clear that they had no plan to recall/replace these until they got given some stick on here. You'd think they would have learned something that time around and once they discovered this problem would have put a complete plan in place and made complete, cogent and clear details available.

How do you know they don't have such a plan?
They would have had no choice but to change the specification, as they have sold countless units already. The specifications should be changed to reflect the actual units that have shipped and are currently in stock around the world.
For example, a user who just bought a unit would download the spec sheet to find out what the maximum specs are before they go use it for such a purpose. You can't leave those people hanging, so changing the specs right away was the right thing to do.
Don't read any more into that.

Coming out with a revised design, fix, recall, refund offer, or whatever is another issue entirely.

I have no doubt they know it's a big deal and are working on a solution.

[Alex Nikitin]

The problem right now is that Keysight has nothing available even to replace these meters with (except, perhaps, 3458A but that would be a bit much to expect  ), so if you need to measure voltages over 600V DC and 440AC accurately, you have to look elsewhere or resort to some band-aids like an external voltage divider.

What's the big deal?
How many people need to measure between 600V and 1000V regularly in practice?

I suspect many more than just me and HighVoltage, who described the problem very well.

Cheers

Alex

[Messtechniker]

How many people need to measure between 600V and 1000V regularly in practice?

Me, for one, not ever. Because I will never ever shoot more than 100V up its sockets.
In my case to start with it would suffice if KS would send me stickers for my 3465A adapting/correcting the font
and rear panel labelling as well one to stick on the manual.

Thereafter followed in due time by a firmware version which deals with the restrictions as far as measurements and calibration for the high voltage ranges are concerned.

[0xfede]

The problem right now is that Keysight has nothing available even to replace these meters with (except, perhaps, 3458A but that would be a bit much to expect  ), so if you need to measure voltages over 600V DC and 440AC accurately, you have to look elsewhere or resort to some band-aids like an external voltage divider.

What's the big deal?
How many people need to measure between 600V and 1000V regularly in practice?

The deal is that they cannot downgrade specification and ratings by 40% when they want.
I have measured voltage in excess of 600VDC just occasionally with it and now I don't know how much reliable the instrument is.
And the fact that I discovered the addendum because TheSteve pointed it and I was not warned directly by Keysight nor Microlease.

Best,
0xfede

[FivePoint03]

I DO need to measure very stable voltages up to 1000V (power supplies and amplifiers).  It is of course preferable not to use dividers unless you have to.  I'm really not happy about this.... I'd return my 34465A but I've spent money on the MEM and DIG options and have BenchVue as well.  What a mess!

Keysight should be given a chance to do the right thing.  For me that would be a complete refund of meter and license options or provide a swap for a meter that meets the spec. I bought it for.

As for whether it's the PCB itself or some component causing the issue I think it could be either.

[Alex Nikitin]

I've just received an email with the latest Keysight Infoline Support Web update. Nothing in it on this issue  .

Cheers

Alex

[HighVoltage]

What's the big deal?
How many people need to measure between 600V and 1000V regularly in practice?

Well, for me that is a huge deal.

I have many 34461A in the field with customers, integrated in to test stands and they are all certified and being used for the DC range up to 1000 Volts. The certification is within the automotive TS16949 certification and several ISO specifications. These instruments are listed with serial numbers and cal certificates in those specifications and can not just be removed and replaced. This would cause a chain reaction of complete new certification of the full test stand.

It really is a nightmare that I see coming here and right now there is no solution. I might have to replace them all by a 34401A.
Hmm, but where will I get those, since they are out of production?

I am looking forward to an answer by Keysight.

[mikeselectricstuff]

I think an interesting question is what the nature of the "damage" is.
If it something which is an obvious failure, as opposed to a possible loss of accuracy with no indication, that's probably not actually a huge deal, assuming they deal with any failures under warranty. 

[Alex Nikitin]

It really is a nightmare that I see coming here and right now there is no solution. I might have to replace them all by a 34401A.
Hmm, but where will I get those, since they are out of production?

Both 34401 and 34410 are still available (until stock is exhausted) here in the UK from RS and Farnell.

Cheers

Alex

[alm]

Keep in mind that it took Keysight a couple of years to identify this, even though many of the instruments will be on yearly calibration cycles using a quality management system that keeps track of the number of out of spec results for each range. If 50% of the units were out of spec in the 1 kVDC range after the first year, they would have responded sooner. So it is either a problem that will only occur in a small fraction of instruments (but not limited to a particular serial number range), takes a couple of years to develop, or something that only happens in rare cases (like instruments connected to > 600 VDC for prolonged periods of time).

[HighVoltage]

I had my four meters hooked up to 1000 V DC over night and I see no problem at all.
May be not all meters are effected?

 

[Alex Nikitin]

I had my four meters hooked up to 1000 V DC over night and I see no problem at all.
May be not all meters are effected?

It is not an instant failure, otherwise it would have been noticed a long time ago  . It appears that the probability of a "damage" (what kind of a "damage"/failure is still unknown) if you go over 600V DC 440V AC is sufficiently high to change the meters rating. First and foremost Keysight should explain the nature of this problem in detail, otherwise it is not even clear what kind of a risk we are looking at and how high that risk is. And there are more questions to ask. For example, if one of your meters would fail that test at 1000V DC - would they repair/replace it if you now knowingly went over the maximum allowable voltage? 

Cheers

Alex

[mikeselectricstuff]

We definitely need some clarification on the exact nature and extent of the "damage".
It is quite possible they don't know the exact cause or circumstances that maks it happen, but they must know the result and affected component(s).

[TiN]

Thanks for the thread.

Also BTW 3458A is NOT rated to 1000V AC, only special version 3458-H01 option is, which has own DCV A1 and A11 (switch/protection) PCBAs.

[GlowingGhoul]

I had my four meters hooked up to 1000 V DC over night and I see no problem at all.
May be not all meters are effected?

What do you expect to see for a problem described as causing damage "over the life" of a unit used at these voltages?

On another note, as to whether warranty claims for damage done by applying high-voltage post Keysight spec change....well, they'd have to prove you knew and did it anyway. Which would be a difficult thing to prove, unless you did it, took pictures, and posted in the forum discussing the notice from Keysight. 

[HighVoltage]

I had my four meters hooked up to 1000 V DC over night and I see no problem at all.
May be not all meters are effected?

What do you expect to see for a problem described as causing damage "over the life" of a unit used at these voltages?

On another note, as to whether warranty claims for damage done by applying high-voltage post Keysight spec change....well, they'd have to prove you knew and did it anyway. Which would be a difficult thing to prove, unless you did it, took pictures, and posted in the forum discussing the notice from Keysight. 

I have been using these meters a lot over the years at just below 1000V, so they have aged over their life so far.
This was the first time I hooked them all in parallel to see if they show any sign of differences and they don't.

On the other hand I can not stop my customers from using them right now at just below 1000V, no matter if Keysight releases an updated note or not, this would mean production STOP !