Handheld meter robustness testing

[joeqsmith]

This is looking at the power each probe is dissipating for this amount of current.   

[joeqsmith]

And this is the calculated resistance.  Most even with 10A are below 0.1 ohms.
 
I had recently ran the probes that were provided with my Kasuntest ZT102 (AN8002).   The second plot is comparing this probe against my old and new style ProbeMaster probes as I sweep the current from 2A to 20A. 

[joeqsmith]

I played around a little more with the home made lens adaptor trying to improve it.  It's alright but I would guess a cheap microscope would out perform it.

 

[bitseeker]

Likely. However, the mac12 image is probably plenty of magnification and it's nice and clear.

[kalel]

Depends how cheap the microscope is, the $16 USB one has 640 x 480 res, otherwise it's not too bad.

[joeqsmith]

If you have one, post a few pictures off it along with what model number it is.   

To get a decent view of the wiper's contact area I would like something that would focus over a longer depth.  I glued the two worst wipers along with a good one to a chunk of plastic to save for later.  I noticed that the one contact was wore even worse.  A few more rotations and I think the center would have fallen out. 

[Fungus]

Can anybody test the hardness of the metal in the wipers? I'm sure that's a big factor.

A wiper that stays smooth, that doesn't develop sharp edges in use, could make a huge difference to longevity.

[kalel]

If you have one, post a few pictures off it along with what model number it is.   

You didn't specify the subject of images exactly (probably just to see the microscope quality), so I opened a similar model to the "free meter" (DT830) and posted some random images. I didn't get to the wipers as it's easier not to have to put them back on (I managed to do it wrong once, it's good to always keep track where each goes). The LM image is not from the meter.

Random example images:

(Some are "doubled" as I usually hit the button a few times, then choose the clearer image)

Scope:
Don't know the model, but searching USB Microscope I found it from an image:



I used the overhead lighting it has built in which consists of 6 built in LED-s which you can control with a potentiometer on the cable (using some PWM I assume). It would be better to have very powerful ambient lighting and not use the LEDs if possible, to avoid glare. Focus is manually adjustable, but you also adjust zoom at the same time. So you usually need to adjust both height of the microscope and the zoom/focus adjustment to get a clear image at the desired magnification. The plastic stand can crack (near where the rod attaches to the base, which gets most force when you position the microscope up/down), but I've noticed that superglue (cheapest type, no more than $0.4-0.6) solves this, now it's pretty durable with no issues for a few years. The field of view is also not very large (unless you move it far away and use as a webcam), but at 640 480 it would not get great detail over a huge area anyway. I'm not sure I would recommend it to someone, especially if they are able to buy a more expensive unit, but considering the price it is better than I would expect.

[Fungus]

I'm not sure I would recommend it to someone, especially if they are able to buy a more expensive unit, but considering the price it is better than I would expect.

[kalel]

Dave's came with a different, potentially less useful stand. Although, this one is not exactly quality, but after some fixing up it works. It does seem a bit more laggy in the video as well (for me it's relatively fluent when the overhead light is on, and of course quite slow when there is no light - but you can't see much anyway), but this will greatly depend on the light, software and hardware it is being tested on.

As for using it to actually do soldering, I haven't tried that as I don't think the working distance is great (plus the stand is plastic, needs some care or protection). It might be possible, but it doesn't seem right for that task.

[joeqsmith]

The $60 unit may be better than my home made setup.  Personally, I have never used a camera to solder with.  It would be nice to have something where the stage could be remote controlled.   

Looking at the free meter, the contacts really don't look too bad.  The majority of the damage is in the pads where the vias are.  The area of the pads where there is no vias looked fine.  But that meter had less than half the cycles on it than the other two.   

I was able to get both the Kasuntest and free meter working again with a little solder and swapping out the contacts from one of the other Kasuntests.  I will save these two meters for another test.

[joeqsmith]

I had ran this HP/Agilent/Keysight meter some time ago because of a video Dave had made where he showed the front end and made a comment to the effect of GDTs no worries.  The meter was not near as robust as I had hoped for.  I saved the meter in case I decided to do something with it.   I damaged the switch area with the half cycle generator but I looked at it today and it's just one set of pads on  outside ring.   The meter has no signs of grease.  I cleaned the area and removed the one outside contact to prevent it from digging away at the damaged area.  It may be worth cycle test it.  If it does poorly, we would not know for sure if a new would would be better.  At least they were smart enough not to put the vias in the pads.   

This meter uses two differnet spaced contacts.  The contacts are laid out similar to the UNI-T UT61.

[kalel]

Is it possible/how would someone repair a contact like that first image S3B/S4B? I mean if someone found it worthwhile for whatever reason.

[joeqsmith]

Remove the solder mask.  Insert wire into the via. Fill with solder.  Maybe replace the contact.  Good enough for what they will be used for in the future. 

[joeqsmith]

The Agilent has about 5000 cycles on it now.  I did remove the one contact before starting the test and cleaned the PCB and remaining contacts.   The cost for this meter is in the the same ballpark as the Fluke 17B+.   Feature wise, they each have their pros and cons.   The Fluke did much better in the transient tests.  Basically failing at 2X the voltage level of where the Agilent meter was damaged.  Let's see if Agilent can hang with the Fluke in this test.

[HalFET]

Finally managed to sneak one of the DT-830 boards (yellow one) in between a time slot of a colleague.

Since I can't quite be bothered fighting filter settings on a Friday afternoon I'll go for the details. The surface roughness is surprisingly good (Ra of 600-700 nm), but just a few turns already caused noticable wear as you can see (2-3 micron trench).

Good area:


Area where the contacts pass:

[joeqsmith]

Can you scan a larger area to show more than one pad so we can see the thickness of the pads along with the surface finish?

HPAK:  The new face of HP-Agilent-Keysight

The HPAK meter has been cycling for well over 24 hours now.  I plan to make a separate very short video just for this meter to show the changes to the software.  I wanted to try and make the data a little easier to understand.  So I have changed the histograms to display in a bar graph.  The width of the bar represents the size of the bins.   For the log/log I changed the minimum resistance from 5mohms to 1mohms and then added a limit line.   Basically there is now a line drawn for the maximum cycles for the data set being viewed along with it starting and ending on the min/max resistance range.   I did not come up with a reason to view the data as a percentage as the one person requested.  They would need to elaborate on what they were thinking then maybe we could add a new graph if it proved helpful. 

[HalFET]

Yeah, just takes a lot of time to do that scan. It can do about 1x1 mm at once, but for a scan depth of 80 µm as I was using it easily takes 20 - 30 seconds/sq. mm The heavy user of the machine is off to a conference so should be able to use it next week without feeling guilty. 

[kalel]

Looks like a great tool. What technology/sensor type does it use to detect height/depth of a point?

[HalFET]

Looks like a great tool. What technology/sensor type does it use to detect height/depth of a point?

It's based on visible light interferometry. You can find quite a bit on it if you google "Veeco Wyko".    But Veeco seems to have sold it off to Bruker: https://www.bruker.com/products/surface-and-dimensional-analysis/3d-optical-microscopes.html You'll find one of these Wyko optical profiler systems in almost any electronics cleanroom, they have extremely good resolution in all three axis, are reasonably priced, and very easy to operate. The only flaw these things have is that the tilt/tip system of the sample stage can get stuck.

[joeqsmith]

I figured it would take a while but you may be able to change some of the settings to scan a wider area with less accuracy just to get us in the ballpark. 

48 hours on the H-PAK meter now.  I wonder why Bill got the short stick in the rename.  We can tell they were trying to spell Dave's last name.  Seems like the next two company names should start with "E" and then "W" to make thing fair.   

Anyway, it's in the home stretch now.   

[joeqsmith]

Warning, it's not a good video for the H-PAK fan boys to watch.   For the rest of you wanting to see it, here's the U1231A.

https://youtu.be/z9732OYPRx8

[rsjsouza]

Joe, that does not look good at all, especially the mechanical wear of the position stops. It makes me wonder if operating the switch with the thumb would wear it even more due to the off-centered pressure applied.

I suspect the more expensive HPAK meters also use the same position stop mechanism.

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.

[joeqsmith]

When I was looking for a good handheld meter to use on the bench, I had looked at Keysight.   I wrote them a few times but no one responded.  Had they, I would ask them about the spring failure.  It really took me by surprise.   

It has not came up yet but if you watch the very first part of the video, the switch had a nice clean and sharp click to it.   This was right after I had setup the test.  The meter had ran a few thousand cycles by the time I decided to make the intro.  By then the damage had been done or at least the click was not near as pronounced.   I was glad I had captured some video what it first sounded like just to give people some contrast. 

You could be right about the thumb wheel causing the parts to wear even faster.  I suspect the spring would fair at the same rate.  I wonder if they had a bad batch of plastic parts and their quality department didn't catch it or someone wrote a deviation to allow them to be used.  Hard to say.  I doubt we would ever know shy of running more of them and seeing if they fail the same.

I get complaints all the time for one reason or another.   This would not be the first time I have been called out for demoting or devaluing a brand.  The problem with this mind set is that it somehow suggests I have something to do with the end product which I certainly have no part in.  I just test them on a level playing field and report my results.   To your other point about it being an overkill usage scenario, that could very well be.   I have yet to find any documentation where the manufacture calls out the number of life cycles the switch is rated for.   Of course, if I am designing a product that uses a switch, the life cycle is something I am looking at.  I find it hard to believe they don't know.  Perhaps their marketing departments do not feel the life of the product is a selling point.

[joeqsmith]

In the attached plot showing the resistance, it looks like it really started to get bad about 6500 cycles into the test.   I thought about aborting the test but the switch started to make contact again and I wanted to see the visual damage with the same number of cycles as the Fluke.    Note the Fluke is the RED plot at the very bottom.  Now that is what a switch should act like.     

I think more than one person would want to see a UNI-T ran on this test.   I currently have two 210s, 90, 181 and a 61E.  None of these are a very good fit to run.  The 210s would be difficult to attach to without some modification to the knob.  The 90 has been repaired more times than I care to count which includes heavy damage to the switch.  Both the 61 and 181 have had circuit board modifications to them.   

I believe the two AMPROBE branded meters I have were made by UNI-T.  Both meters are still in excellent condition as they only are used for comparisons.  Both use a screw and fiber washer that holds the switch assembly to the PCB.  So it's not like the UT61E and other UNI-T meters I have looked at.    It does appear they did not place the the vias in the pads.  After seeing how quickly the H-PAK meter failed, it's not a guarantee the switch will survive.   

Another option is to buy one more UNI-T branded meter and run it.  May be time for a new poll.