Poll

How many cycles will the KeySight U1281A's detent spring last?

0-2000
7 (17.9%)
2k-4k
5 (12.8%)
4k-8k
14 (35.9%)
8k-16k
7 (17.9%)
>16k (most rubust meter ever made)
6 (15.4%)

Total Members Voted: 36

Author Topic: Handheld meter robustness testing  (Read 1149867 times)

0 Members and 5 Guests are viewing this topic.

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2150 on: November 26, 2017, 02:41:49 am »
Looking at the the first 7000 cycles from 0.35 to 0.5 ohms for all four meters.  The free meter is off the charts within 500 cycles.  But it's free.   More interesting is I would expect there to be a major difference between the H-PAK and the Kasuntest but show signs of problems right from the start.   The Fluke on the other hand actually improves.  I suspect the parts are seating together.  I know you people like to compare meters with cars so something like breaking in a new car engine.  Once the contacts seat, the are even more consistent.

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 9057
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Handheld meter electrical robustness testing.
« Reply #2151 on: November 26, 2017, 03:28:07 am »
The switch spring failure in the Keysight is unfortunate. Hopefully, their higher-end meters have a better/more durable mechanism. Too much plastic everywhere!
TEA is the way. | TEA Time channel
 

Offline Paul Moir

  • Frequent Contributor
  • **
  • Posts: 926
  • Country: ca
Re: Handheld meter electrical robustness testing.
« Reply #2152 on: November 26, 2017, 06:14:57 am »
It's not too much plastic, it's the wrong plastic.  Notice that fluke makes the spring out of a different plastic than the housing.
 

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2153 on: November 26, 2017, 10:52:04 am »
The switch spring failure in the Keysight is unfortunate. Hopefully, their higher-end meters have a better/more durable mechanism.

"Unfortunate" isn't the word I'd use - it's probably not bad luck.

It'd be worth running another one in batches of 500 cycles just to figure out exactly when the switch breaks. For all we know it might have broken after 1000 cycles (there's obviously nothing left after a "few thousand"). 1000 cycles would only be a dozen measurements per day for 3-4 months.

As for "Hope"? I'd want proof:popcorn:

PS: I wonder if user 'Keysight DanielBogdanoff' reads this thread. Keysight's reputation for multimeters needs some salvage work here. Maybe he could send Joe some meters to test. :popcorn:
« Last Edit: November 26, 2017, 10:58:02 am by Fungus »
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26755
  • Country: nl
    • NCT Developments
Re: Handheld meter electrical robustness testing.
« Reply #2154 on: November 26, 2017, 11:14:18 am »
Regardless, I imagine folks may be up in arms either (1) complaining about you testing a damaged meter or (2) crying foul trying to devaluate HPAK meters solely based on your overkill usage scenario.
IMHO it is good to test stuff until it breaks.The problem is ofcourse that this is just one sample which may be very good or very bad. There is no way to tell other than testing 20 units.
From the graph it seems this one broke after about 6000 cycles. You would have gotten over 4 years of life out of it if you turned the switch 5 times a day on every work day.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2155 on: November 26, 2017, 11:19:02 am »
From the graph it seems this one broke after about 6000 cycles.
The range selector's wipers aren't related to the switch's clicker mechanics, IMHO, so the graph can't show anything about the clicker.

The switch is definitely not clicking at all in the intro video (after "a few thousand" cycles).

Joe: You took a photo of the meter at 5000 cycles. Was that before or after recording the intro video? By how much time, approx? That event might give a ballpark clue as to where it failed.
« Last Edit: November 26, 2017, 11:28:35 am by Fungus »
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
Re: Handheld meter electrical robustness testing.
« Reply #2156 on: November 26, 2017, 11:21:29 am »
I would be surprised if the Keysight higher up meters like the U1273A or the U1253B would behave the same way. I have both and so far they seem reliable.

Yes, may be Keysight will send you some DMMs for testing.
 
There are 3 kinds of people in this world, those who can count and those who can not.
 

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2157 on: November 26, 2017, 11:23:12 am »
I would be surprised if the Keysight higher up meters like the U1273A or the U1253B would behave the same way. I have both and so far they seem reliable.

Could you pop one open and see if the switch clicker is made of decent plastic, like the Fluke?

 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6452
  • Country: hr
Re: Handheld meter electrical robustness testing.
« Reply #2158 on: November 26, 2017, 11:41:45 am »
Broken plastic spring is not a problem. Keysight could send you new one made of better plastic. It is the contact damage, board damage and resulting contact resistance that is problem. Board traces are definitely damaged too soon, and that you can't fix easily. It is obvious that PCB integrated switch is not so robust solution in general, and that tested Fluke seem to does it quite nicely and better than the rest of the tested ones..
This is one of those moments when Fluke actually lives up to it's reputation. I'm really curious now as to how Brymen would stand up to this test..
Like Joe, my BM869S is also my favorite meter. I prefer it to others, and actually bought it instead of Fluke 87/V because it was better for my use, not cheaper.
I wonder if maybe switch related problems might be is it's Achilles heel.. Or maybe it stands up to Fluke on that test too?
Would be really interesting to know.

Also, as far as Uni-T goes, I guess 61E would be the one.. Despite not being very electrically robust, it is a good little, inexpensive meter for electronics desk and many people have one and it would be nice to know how long it's gonna last...

Other choice would be 181, that is fully featured meter good for low power electronics work, but it is not that cheap. Would be nice to know how long it's switch would last..

Regards,

Sinisa
 

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2159 on: November 26, 2017, 11:52:20 am »
Broken plastic spring is not a problem. Keysight could send you new one made of better plastic.

Sure, if you have a local office, a business account with them and the extension number of a support rep.

Good luck if you're in another country or just an individual, eg. Joe couldn't get a reply out of them.
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6452
  • Country: hr
Re: Handheld meter electrical robustness testing.
« Reply #2160 on: November 26, 2017, 12:25:51 pm »
Broken plastic spring is not a problem. Keysight could send you new one made of better plastic.

Sure, if you have a local office, a business account with them and the extension number of a support rep.

Good luck if you're in another country or just an individual, eg. Joe couldn't get a reply out of them.
You are correct in that regard, I didn't think of that that way.

But to be sarcastic, isn't that the reason to pay big bucks to renowned companies, because they don't just sell you products, they give great support...

So far Joe had no particularly good support from Gossen Metrawatt, Keysight... Actually Brymen is known to answer questions... 

My own experience is that big names have great support only if you are a big customer.
I'm bussines customer, but a small one and nobody cares for my few hundred bucks, even few thousand every now and then.. 
6.5 digit benchtop meters are now retail channel products, if you don't need at least 5-10 at a time..

Hence, Brymen for me... Works great, an even it if it doesn't last 10-15 years, at least I payed 30% of the price.. In 15 years, I can buy new one every 5 years, toss old one in the garbage, and still break even compared to "big names", while having excellent instrument that was in warranty period most of the time..
 

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2161 on: November 26, 2017, 12:57:15 pm »
Broken plastic spring is not a problem. Keysight could send you new one made of better plastic.
Sure, if you have a local office, a business account with them and the extension number of a support rep.
You are correct in that regard, I didn't think of that that way.

But to be sarcastic, isn't that the reason to pay big bucks to renowned companies, because they don't just sell you products, they give great support...

I'd rather they made good products that don't need any more support than a worldwide speedy-replacement warranty for DOA devices.

Right now the Fluke 17B+ is looking a lot less overpriced than it did a couple of weeks ago. 

PS: Maybe Joe could send the meter to 2N3055 and 2N3055 could document the process of getting them to fix/replace it. It's easy, right? :popcorn:
« Last Edit: November 26, 2017, 01:00:35 pm by Fungus »
 

Offline HalFET

  • Frequent Contributor
  • **
  • Posts: 512
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2162 on: November 26, 2017, 01:13:10 pm »
Broken plastic spring is not a problem. Keysight could send you new one made of better plastic.
Sure, if you have a local office, a business account with them and the extension number of a support rep.
You are correct in that regard, I didn't think of that that way.

But to be sarcastic, isn't that the reason to pay big bucks to renowned companies, because they don't just sell you products, they give great support...

I'd rather they made good products that don't need any more support than a worldwide speedy-replacement warranty for DOA devices.

Right now the Fluke 17B+ is looking a lot less overpriced than it did a couple of weeks ago. 

PS: Maybe Joe could send the meter to 2N3055 and 2N3055 could document the process of getting them to fix/replace it. It's easy, right? :popcorn:

As a pure hobbyist, don't even try. On the other hand if you work for a company or university and often deal with them, then you can get lucky sometimes.
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2163 on: November 26, 2017, 01:44:58 pm »
Regardless, I imagine folks may be up in arms either (1) complaining about you testing a damaged meter or (2) crying foul trying to devaluate HPAK meters solely based on your overkill usage scenario.
IMHO it is good to test stuff until it breaks.The problem is ofcourse that this is just one sample which may be very good or very bad. There is no way to tell other than testing 20 units.
From the graph it seems this one broke after about 6000 cycles. You would have gotten over 4 years of life out of it if you turned the switch 5 times a day on every work day.
This sort of comment does come up from time to time.     Somewhere in this long pile of posts there is a section on it,  or you could check the FAQ as I believe I cut and pasted it there as well.     I would hope that H-PAK' quality control group has a hand process that is under control but you never know.   

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2164 on: November 26, 2017, 01:59:31 pm »
From the graph it seems this one broke after about 6000 cycles.
The range selector's wipers aren't related to the switch's clicker mechanics, IMHO, so the graph can't show anything about the clicker.

The switch is definitely not clicking at all in the intro video (after "a few thousand" cycles).

Joe: You took a photo of the meter at 5000 cycles. Was that before or after recording the intro video? By how much time, approx? That event might give a ballpark clue as to where it failed.

The first thing I did was pull apart the meter to inspect it.  I took pictures of the switch as well.  Once it was setup, I took a picture of the setup and then shot some video of it.   In the picture you reference, the Cannon is mounted on the tripod to the far right.  Normally for stills, I like using that Sony.   

I never would have guessed the meter would have problems this early on and I was not thinking too much about what the video was even going to show other than the changes I had made to the graphing.   By the time I checked in on the test, it was too late.  This is when I shot the intro.   The meter started that bit of a squeak so I captured that but it slowly went away as the meter continued to cycle.

The cycle time varies.  Windows is not a real predictable OS.   I would guess somewhere in the 2-3 hour range when I noticed it with a fairly low confidence.     

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2165 on: November 26, 2017, 02:15:22 pm »
Broken plastic spring is not a problem. Keysight could send you new one made of better plastic. It is the contact damage, board damage and resulting contact resistance that is problem. Board traces are definitely damaged too soon, and that you can't fix easily. It is obvious that PCB integrated switch is not so robust solution in general, and that tested Fluke seem to does it quite nicely and better than the rest of the tested ones..
This is one of those moments when Fluke actually lives up to it's reputation. I'm really curious now as to how Brymen would stand up to this test..
Like Joe, my BM869S is also my favorite meter. I prefer it to others, and actually bought it instead of Fluke 87/V because it was better for my use, not cheaper.
I wonder if maybe switch related problems might be is it's Achilles heel.. Or maybe it stands up to Fluke on that test too?
Would be really interesting to know.

Also, as far as Uni-T goes, I guess 61E would be the one.. Despite not being very electrically robust, it is a good little, inexpensive meter for electronics desk and many people have one and it would be nice to know how long it's gonna last...

Other choice would be 181, that is fully featured meter good for low power electronics work, but it is not that cheap. Would be nice to know how long it's switch would last..

Regards,

Sinisa

I am not sure what problems the spring could actually cause.  I run a test where I put a full rectified 220V 60Hz wave into the meters and cycle them through all the functions.  The generator detects an over current condition (I think I have it set to 40mA or so) and will shut down to limit any damage.  There were several comments about how I was cycling the meters with power applied.  Without the spring, the switch will not lock to its centered location.  It could even move fairly easily. Could this cause something to happen?  I have no idea.  Certainly, I could see someone reading a 440 bus having the meter  slip off the contacts and start reporting something other than the 440.   Could this create a hazardous condition?     Keep in mind, this meter has no current input.  It uses an external clamp.   It's not something you would want as an electronics hobbyist.   

What bothers me with some of these meters is all that metallic residue around the switch contacts.  I doubt IEC even considers that the meter should be life cycled before running surge.  That residue could play into it.   For us low energy hobbyist, well for me anyway,  the life is more important.   

Your comment about BM869s is true for me as well.  The cost really didn't play into it nearly as much as what the product offered.  I plan to run at least one Brymen product.  It should make for a good video.
 
The following users thanked this post: 2N3055

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6452
  • Country: hr
Re: Handheld meter electrical robustness testing.
« Reply #2166 on: November 26, 2017, 02:37:32 pm »
I am not sure what problems the spring could actually cause.  I run a test where I put a full rectified 220V 60Hz wave into the meters and cycle them through all the functions.  The generator detects an over current condition (I think I have it set to 40mA or so) and will shut down to limit any damage.  There were several comments about how I was cycling the meters with power applied.  Without the spring, the switch will not lock to its centered location.  It could even move fairly easily. Could this cause something to happen?  I have no idea.  Certainly, I could see someone reading a 440 bus having the meter  slip off the contacts and start reporting something other than the 440.   Could this create a hazardous condition?     Keep in mind, this meter has no current input.  It uses an external clamp.   It's not something you would want as an electronics hobbyist.   
I wouldn't presume to speculate what electrical damage might happen to the switch and meter because of switch locking between ranges. I was merely pointing out that mechanically, PCB wear was terminal damage to the meter and not the broken spring. If the PCB was not damaged meter would be repairable by just replacing knob. Also I don't think springs being damaged contributed to PCB and contacts excessive wear, and that those two failures should be considered separately, despite being same switch assembly.

What bothers me with some of these meters is all that metallic residue around the switch contacts.  I doubt IEC even considers that the meter should be life cycled before running surge.  That residue could play into it.   For us low energy hobbyist, well for me anyway,  the life is more important.   
I think that is excellent point, especially if there is lubrication on the switch that will retain metallic filings and particles, essentially making it into conductive paste over time...

Your comment about BM869s is true for me as well.  The cost really didn't play into it nearly as much as what the product offered.  I plan to run at least one Brymen product.  It should make for a good video.
Agree. Looking forward to it!

Take care!
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2167 on: November 26, 2017, 02:49:37 pm »
I'd rather they made good products that don't need any more support than a worldwide speedy-replacement warranty for DOA devices.

Right now the Fluke 17B+ is looking a lot less overpriced than it did a couple of weeks ago. 

PS: Maybe Joe could send the meter to 2N3055 and 2N3055 could document the process of getting them to fix/replace it. It's easy, right? :popcorn:

I have never considered returning a meter that I ran in for repairs as I knowingly run them to failure.   IMO, this is the whole point of doing a review.  If you are just hooking to a few resistors and some batteries, what's the point.    When I was talking with Brymen about transient testing the BM869s, I told them I was going to test it to failure.  Tell this to any other meter company and let me know how they respond.   Brymen's response:

Quote
In case BM869s is damaged out of your transient tests, we will be willing to specially offer free warranty repair or replacement to you.

And when I did finally damage it I offered to return it for a failure analysis.  I also provided them with a little background and offered to do the FA myself.   Their response was:

Quote
If you can not repair it, please send it back to us. We will either repair it or replace it by a new one back to you.

Of course, I was able to repair the meter and have been using it ever since.  Point being that their communications have proven time and time again to be second to none.  At least from my perspective.  When a user had posted about the strange settling times,  I wrote them and we pretty much had an immediate response, including the source for that section of code.   

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2168 on: November 26, 2017, 03:00:12 pm »
I am not sure what problems the spring could actually cause.  I run a test where I put a full rectified 220V 60Hz wave into the meters and cycle them through all the functions.  The generator detects an over current condition (I think I have it set to 40mA or so) and will shut down to limit any damage.  There were several comments about how I was cycling the meters with power applied.  Without the spring, the switch will not lock to its centered location.  It could even move fairly easily. Could this cause something to happen?  I have no idea.  Certainly, I could see someone reading a 440 bus having the meter  slip off the contacts and start reporting something other than the 440.   Could this create a hazardous condition?     Keep in mind, this meter has no current input.  It uses an external clamp.   It's not something you would want as an electronics hobbyist.   
I wouldn't presume to speculate what electrical damage might happen to the switch and meter because of switch locking between ranges. I was merely pointing out that mechanically, PCB wear was terminal damage to the meter and not the broken spring. If the PCB was not damaged meter would be repairable by just replacing knob. Also I don't think springs being damaged contributed to PCB and contacts excessive wear, and that those two failures should be considered separately, despite being same switch assembly.
I am not a plastics expert but the spring does appear to be made of a glass filled plastic.  If the spring would have created some sort of dust, I guess it could contribute to the contacts wear.  All the breaks appear clean and I saw only minimal plastic dust and none of it appeared on the circuit board from what I could tell.  So I tend to agree. 

I don't have a metric for spring failures.  Actually, the only thing I can do is show the before and after for the mechanical parts and leave it to the viewer to decide if its a problem or not.  We certainly saw a fair amount of damage to the detent area on free meter as well as the Kasuntest.   I thought about measuring the torque in real time as another metric but decided it was more trouble than it was worth. 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Handheld meter electrical robustness testing.
« Reply #2169 on: November 26, 2017, 03:16:35 pm »
Great job as always, Joe.  There are very few public tests to show meters live up to their spec as you have.  Switch failure, plus the way you present the data make it fairly clear Fluke does put more into its DMMs, even offshore meters.  I look forward to your Brymen tests in the future.
Best Wishes,

 Saturation
 

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Handheld meter electrical robustness testing.
« Reply #2170 on: November 26, 2017, 04:31:09 pm »
I look forward to your Brymen tests in the future.

Is the BM869 about to take one for the team?  :o
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: Handheld meter electrical robustness testing.
« Reply #2171 on: November 26, 2017, 04:53:47 pm »
Spring is made I guess with a glass filled nylon, probably 30% glass by volume, as that is a pretty common material, and is quite robust and resilient. however for a decent spring that arm should have been thinner, plus the nylon should have been non filled, which would mean they would have had to use 2 moulds to make the outer knob and shaft in a 30% glass filled nylon for robustness, and then the inner spring section as a slip on collar that would have the flex required to survive long term operation as a spring material. In that case I would guess the lifetime would be well past the 50k cycle life, and the wear on the end sections of the spring in the detents would also be a lot less.

Would prefer though that the manufacturers used a thicker gold plate on the switches, along with a thicker gold plate ( or a welded on German silver contact pad) on the wipers, so that the operational lifetime would be longer. However that would require a selective plate on the board, or for them to use a small daughter board that has the contact pads on there, with a set of mousebites or pins to allow the 2 boards to be connected. Might be doable, thin PCB material with double sided etch, lots of small vias and a heavy gold plate, plus plated milled vias on the edges for soldering to the main board, and some solder pins to both locate and carry current. Main board would just need a thin precut insulator to prevent arcing over between traces, and would be cheaper than a bespoke switch, plus frees up board space to route traces on the front and components on the rear of the switch as well. 
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6452
  • Country: hr
Re: Handheld meter electrical robustness testing.
« Reply #2172 on: November 26, 2017, 05:10:02 pm »
My old Fluke 73/III had a real switch, and a separate spring assembly, that spring assembly had separate part number...
I guess they really don't make them as they used to...  :-DD
 
The following users thanked this post: saturation, SeanB

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2173 on: November 26, 2017, 05:18:17 pm »
I look forward to your Brymen tests in the future.

Is the BM869 about to take one for the team?  :o
I don't want there to be any question in peoples minds about the state of the meter prior to running the test and my BM869s has seen a lot of use since I purchased it.  The right thing would be to run a brand new one right out of the box.

I would be fine running the EEVBLOG rebranded Brymen as it has only been used for reviews and should be in near mint condition.  However, this meter has never been damaged and at some point, I may put together some sort of shoot out for the meters that have survived these tests out of the box.   The Fluke 107 will be a tough meter to beat.   

At some point I would like to repeat the tests I ran on the 121GW now that it has been released.   I am a little concerned about the comments Dave made in the last Amphour show about wanting to change the board again.   Because of the time invested in making these videos, I want to make sure that I don't run into a case only to have them say there is a new hardware change in the works.   My plan would be to release the videos I had made of the preproduction prototype proof of concept, or what ever we are calling it now and show how the released meter addressed the problems I had found.   My plan would be to include the life cycle testing as part of the review.    I think as long as I stay away from that half cycle generator until the very end, the switches should be fine to life cycle after the normal testing.   It may be a while for them to go through the 2000 boards they have in stock so don't look for this to happen any time soon. 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter electrical robustness testing.
« Reply #2174 on: November 26, 2017, 06:16:53 pm »
There maybe some confusion what the H-PAK's switch assembly looks like once it's apart.   





Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf