Handheld meter robustness testing

[joeqsmith]

Consider the newer small Fluke multimeter input protection circuit, instead of desoldering things.

I was thinking I could maybe take the spring contact out of the range selector - it breaks the circuit, it's much easier!

You're firing on all eight cylinders today I see.   Still, it was a good opportunity to try out that heat gun. 

[AVGresponding]

Consider the newer small Fluke multimeter input protection circuit, instead of desoldering things.

I was thinking I could maybe take the spring contact out of the range selector - it breaks the circuit, it's much easier!

Just don't lose the bugger!

[bdunham7]

I would expect that my first Fluke DMM was highly susceptible to being damaged by RF (or looking at it sideways).   I suspect as the front ends of these handhelds evolved, the need to derate them was less important.  Of course, you still end up with meters like the UNI-T UT61E+ having some sort of frequency counter speced at 220MHz with a couple of PTCs in series, but I suspect most modern robust meters would not have a problem.   Like I had shown with the BM78x, with the same setup that damaged the UT61e+, the PTCs had very little heat.   As I demonstrated, that resistor was fairly stable to 50MHz and a 1kohm just doesn't present much of a load to the small amplifier. 

I don't mind running other tests if you had something in mind.  I could take that same resistor/PTC we just used and run them any way you like.  The same with that Brymen BM78x prototype.   Let me know.

I have 3 of the 8000A models in my junkpile, they're all toasted and they all have evidence that they've been soldered on during their lifetimes.  But those were early days in the whole field, not just for Fluke.  Meters may be more durable nowadays, but they still have limits and I, for one, would like to see those limits specified in detail--but they usually aren't. 

I haven't see the video of you giving any other meters the jqs MeltCal^TM treatment, I'll have to go look.  I think it might be interesting to try to determine what parts are used in the meters and then characterize and test those--and look at datasheets if possible--to see who is using what and how the expected component performance matches up with the claimed performance of the meter.

To start, for Fluke I'm fairly sure those big PTCs in the larger CAT III/1000V meters are something like this:

https://www.mouser.com/datasheet/2/18/amphenoladvancedsensors_YS4020-1157120.pdf

and the green input resistors are Bourns WS5M:

https://www.mouser.com/datasheet/2/54/ws-778275.pdf

Those Bourns resistors have a nice surge rating chart:

[joeqsmith]

I have 3 of the 8000A models in my junkpile, they're all toasted and they all have evidence that they've been soldered on during their lifetimes.  But those were early days in the whole field, not just for Fluke.  Meters may be more durable nowadays, but they still have limits and I, for one, would like to see those limits specified in detail--but they usually aren't. 

I haven't see the video of you giving any other meters the jqs MeltCal^TM treatment, I'll have to go look.  I think it might be interesting to try to determine what parts are used in the meters and then characterize and test those--and look at datasheets if possible--to see who is using what and how the expected component performance matches up with the claimed performance of the meter.
...

I wouldn't have spent any time rebuilding my 8000A had it not been my first DMM.  Beyond it's sentimental value and being the catalyst for running these tests,  it has no other redeeming qualities.   

After the attached comment,  I demonstrated the BM78x at over 100MHz during Part 2 of the UT61E+ review.  At these higher frequencies, the majority of the voltage drop will be across the resistors, not the PTCs.   I'm guessing this would have been obvious to all the radio hobbyist.   

[bdunham7]

After the attached comment,  I demonstrated the BM78x at over 100MHz during Part 2 of the UT61E+ review.  At these higher frequencies, the majority of the voltage drop will be across the resistors, not the PTCs.   I'm guessing this would have been obvious to all the radio hobbyist.   

OK, so I see you actually did some characterization of components.  I'm not a radio hobbyist per se, but this would not have been so obvious to me because I wouldn't have known what the parasitic capacitance of the PTCs was.  MOVs I know about, PTCs I never thought about.  Of course it makes sense in hindsight.

Is that resistor a Bourns WS5M or similar?  I wonder what its reactance to your piezo igniter would be.  Perhaps we need a MacroVNA, with a 1kW output.

[joeqsmith]

I suspect the OP was not thinking about the AC characteristics which is a bit odd as we were working with AC.     Then again, if you never work with AC, you may think of a resistor as a resistor, not a complex device with RLC components to it.  You may also think of a wire as a short.   When I was playing with that fastest breadboard oscillator contest, basically I made the whole circuit from wire + transistor.

We have some idea on the current waveshape of the grill starter.  The little pocket VNA will work to about 300MHz.  The other low cost one, a bit over 4GHz.  It would not be a problem to look at it at higher frequencies but I would ditch the breadboard.   

I've looked at a few different brands/types of resistors.  That particular one was made by Ohmite.  Have a look at the OX/OY series.   

[Fungus]

You're firing on all eight cylinders today I see.   Still, it was a good opportunity to try out that heat gun.

I'll try it on one of the "cheap" ones first. 

Just don't lose the bugger!

OK, I took out the spring from the selector switch.

Results:
a) VDC/VAC ranges work just fine.
b) mV AC range doesn't work. Shows a small reading which changes if I wave the leads around. Doesn't go to zero when I short the probes.
c) All the other ranges show 0.00 on screen.
d) Continuity range beeps continuously if I select it.
e) Diode range still lights up LEDs, resistance range still outputs a test voltage.

[joeqsmith]

OK, I took out the spring from the selector switch.

Now that you know that the PTC side can be used for an output as well as an input, and you know the other leg is an input only.  How do they clamp  both legs when the resistance function is selected?     Are you going to trace it out and provide us some drawings?

Yesterday my entire day consisted of making a few minute video.   Today should be a bit more exciting.

[bdunham7]

Perhaps we need a MacroVNA, with a 1kW output.

The 1kW VNA might not be so funny.  I got the calibrator (the actual project I'm working on here) working for the moment and I tried testing meters again.  In troubleshooting the calibrator I had set up one DMM, the trusty F27, to monitor the output current, so I left it in place. 

The F116 in AutoV/Lo-Z mode (or probably any of the PTC-protected modes) would draw about 1.8mA @ 600VDC or 600VAC/60Hz.  At 600VAC/400Hz, this climbed to about 3mA, and at 1kHz, 5mA.  I tried the same thing with the F27, but it had much higher currents--20mA +-- and at its 1000V level, the calibrator simply could not supply enough current without tripping off.  I did observe the same thing--current climbing with frequency. 

I have no idea how much power is being dissipated and how much is imaginary, but I suppose another experiment is in order.  I'll have to set up some resistors and a scope to see what the angle is between current and voltage at 50-1000Hz--so not a 1kW VNA, but at least a 1kV version!

[joeqsmith]

I assume these are all after the PTCs have switched.   Otherwise for LowZ, I would expect the current to be much higher.   At a kHz, I doubt you will see much change.   

For the F27, you could for example, try using the AC line to get the PTCs to switch, then quickly attach it to the calibrator.  Maybe just a DPDT switch to prevent it from cooling.   

[joeqsmith]

In part 4, the UT61E+ is temperature cycled, dropped and transient tested.   I also use the NanoVNA to compare the 61E+'s input impedance with a few other meters. 

[bdunham7]

For the F27, you could for example, try using the AC line to get the PTCs to switch, then quickly attach it to the calibrator.  Maybe just a DPDT switch to prevent it from cooling.

The calibrator is now up to snuff.  It will put out more current at lower voltages, up to about 250 volts, above that it is limited to 8-10mA continuous.  For the F27, when I monitor current, I can see that the PTCs do take some time to heat up, they don't 'switch' quickly at all.  At 250V it is 'over the hump' so to speak, but the huge PTC in the F27 is still beyond what the calibrator can handle in one go.  If I give it a 1000VAC jolt cold, the current goes to a bit over 20mA, as would be expected given it's ~4.6K cold combination of the 3.5K resistor and PTC--but this is too much current and the calibrator trips off.  The two series PTCs of the F116 are a bit easier for it, but even at 600V the current takes some time to settle.  Anyway, I can now energize both meters at their rated voltage, although the F27 has to be brought up in stages.

I assume these are all after the PTCs have switched.   Otherwise for LowZ, I would expect the current to be much higher.   At a kHz, I doubt you will see much change. 

Well, there's the problem--it does change, and a lot.  I tested them at low and high voltages with different frequencies.  I used 6V and 600V for the F116, then 15V and 1000V for the F27.  I needed to use the higher voltage on the low end for the F27 because it doesn't have a Lo-Z mode, so I have to overwhelm the low voltage clamps in whatever circuit I'm using (OHMS in this case) so that I see mostly the PTC/resistor characteristics.

At 6VAC in 'AUTO V/Lo-Z',  the F116 had a current of about 2.0mA @ 50Hz, and that didn't change much at 1kHz or even 10kHz.  However, at 600VAC, it had about 1.9mA at 50Hz, but that rose to 5.1mA @ 1kHz. 

The F27 at 15VAC stayed in the 2.1-2.3 range, not totally steady because the OHMs circuit was reacting a bit.  At 1000VAC, it showed about 2.0mA @ 50Hz, but raising the frequency to 1kHz caused a surge to about 7mA and then it settled back down to 5.4mA.

Those are huge differences from 50Hz to 1kHz, but only at higher voltages.  I have to wonder if the trend would continue if you kept going--the calibrator is maxed out.  I'm only at 10⁶ V-Hz but some of these meters are supposedly good for 10X or 20X more.  The next time I place an order maybe I'll get a few of those big dark grey monster PTCs that you find in the larger CAT III/1000 Flukes and try to characterize them.  It looks like they pick up 500-1000pF when they warm up. Maybe they should be called thermisto-varactors. 

[joeqsmith]

I assume these are all after the PTCs have switched.   Otherwise for LowZ, I would expect the current to be much higher.   At a kHz, I doubt you will see much change. 

Well, there's the problem--it does change, and a lot.  I tested them at low and high voltages with different frequencies.  I used 6V and 600V for the F116, then 15V and 1000V for the F27.  I needed to use the higher voltage on the low end for the F27 because it doesn't have a Lo-Z mode, so I have to overwhelm the low voltage clamps in whatever circuit I'm using (OHMS in this case) so that I see mostly the PTC/resistor characteristics.

At 6VAC in 'AUTO V/Lo-Z',  the F116 had a current of about 2.0mA @ 50Hz, and that didn't change much at 1kHz or even 10kHz.  However, at 600VAC, it had about 1.9mA at 50Hz, but that rose to 5.1mA @ 1kHz. 

The F27 at 15VAC stayed in the 2.1-2.3 range, not totally steady because the OHMs circuit was reacting a bit.  At 1000VAC, it showed about 2.0mA @ 50Hz, but raising the frequency to 1kHz caused a surge to about 7mA and then it settled back down to 5.4mA.

Those are huge differences from 50Hz to 1kHz, but only at higher voltages.  I have to wonder if the trend would continue if you kept going--the calibrator is maxed out.

"Not much change" meaning Amps vs mA.  At these low frequencies, Xc is not going to be a problem.   As you move up in frequency, Xc will dominate and we can damage the parts with a very low voltage, unless there is something else to limit the current.   Perhaps a resistor...

I'm only at 10⁶ V-Hz but some of these meters are supposedly good for 10X or 20X more.  The next time I place an order maybe I'll get a few of those big dark grey monster PTCs that you find in the larger CAT III/1000 Flukes and try to characterize them.  It looks like they pick up 500-1000pF when they warm up. Maybe they should be called thermisto-varactors.

  

[bdunham7]

"Not much change" meaning Amps vs mA.  At these low frequencies, Xc is not going to be a problem.   As you move up in frequency, Xc will dominate and we can damage the parts with a very low voltage, unless there is something else to limit the current.   Perhaps a resistor...

Yes, the resistor is key at low voltage and much higher frequencies.  But take the case of the F116 running at 600V and continuing to ramp up the frequencies.  If you just assume that the PTC goes to a state of very high resistance but a capacitance of 0.5nF, the current will go high enough (~70mA) to exceed the rated power of the resistor (which I'm presuming is 5W for now) at around 40kHz.  So unless there is some other change, by 100kHz it has likely unsoldered itself if the PTC hasn't come undone first.  This is probably a bit of a far-fetched example, but perhaps not when it comes to VFD drives and such.  In any case, what caught my eye was the reactance change at higher voltages that wasn't there at lower ones.

[joeqsmith]

"Not much change" meaning Amps vs mA.  At these low frequencies, Xc is not going to be a problem.   As you move up in frequency, Xc will dominate and we can damage the parts with a very low voltage, unless there is something else to limit the current.   Perhaps a resistor...

Yes, the resistor is key at low voltage and much higher frequencies.  But take the case of the F116 running at 600V and continuing to ramp up the frequencies.  If you just assume that the PTC goes to a state of very high resistance but a capacitance of 0.5nF, the current will go high enough (~70mA) to exceed the rated power of the resistor (which I'm presuming is 5W for now) at around 40kHz.  So unless there is some other change, by 100kHz it has likely unsoldered itself if the PTC hasn't come undone first.  This is probably a bit of a far-fetched example, but perhaps not when it comes to VFD drives and such.  In any case, what caught my eye was the reactance change at higher voltages that wasn't there at lower ones.

0.5nF would give us 7957.75k at 40kHz.  In series with a 1k, 8957.75k.   With 600Vrms, (assuming the low voltage clamp is active an has no loss) the current would be 66.98mA.  The power dissipated by the resistor would be 4.49 Watts.    Changing to 100kHz would cause the resistor to dissipate 20.57 Watts. 

So, I think your math is correct.  Now the question is the capacitance at these higher frequencies once the part has switched.   Where is the 500pF coming from?  If it were 50pF we would be talking under half a Watt.   

[joeqsmith]

Section 6.2 Figure 7 provides some insight.

So use of the PTC thermistor in the AF and RF ranges is not possible, meaning that applications are restricted to DC and line frequency operation.

https://www.tdk-electronics.tdk.com/download/408374/d78540dfe0589d2bd90cabef477c90b9/pdf-general-technical-information.pdf

[bdunham7]

Section 6.2 Figure 7 provides some insight.

https://www.tdk-electronics.tdk.com/download/408374/d78540dfe0589d2bd90cabef477c90b9/pdf-general-technical-information.pdf

Excellent find!  That explains what I'm seeing.  And 6.1 refers to the hot PTC having a high voltage coefficient as well, which just makes things even more complicated.

Where is the 500pF coming from?

My off-the-cuff guesstimate of what (in capacitive reactance terms) would account for the extra 3mA of current going from 50Hz to 1kHz.  If/when I get some samples, I'll try to determine the relationships between voltage, current, frequency and phase.  Now it's back to my bulk tantalum capacitor replacement project....

[joeqsmith]

Keep in mind, using your 500nF, 40kHz, 1k resistor that the power dissipation for the PTC is 36 Watts. 

[bdunham7]

Keep in mind, using your 500nF, 40kHz, 1k resistor that the power dissipation for the PTC is 36 Watts.

Well, they did unsolder themselves in your case!  But we don't know the power dissipation because we don't know what the phase angle between I and V across the PTC is.  The resistor can't escape the heat so easily. This is one of the things I'd like to test on part samples, rather than cooking a nice meter.

[joeqsmith]

Now these are some classic comments!   There's just so many good points but that last sentence of Capture5 is priceless.   I never knew I was a 121GW fanboy!       

[joeqsmith]

The viewer's have spoken. 

[Fungus]

In part 4, the UT61E+ is temperature cycled, dropped and transient tested.   I also use the NanoVNA to compare the 61E+'s input impedance with a few other meters. 

it survived the low voltage zapper?

You know what that means, joe? You're going to have to buy another UT61E+ so you can repeat the tests on an "unmodified" one. 

[rsjsouza]

Now these are some classic comments!   There's just so many good points but that last sentence of Capture5 is priceless.   I never knew I was a 121GW fanboy!   

Man, those are precious! He clearly did not watch the video.

It is very easy to throw as many digits as possible on a screen if you don't care about accuracy - also, if you rely on a multimeter to have a reliable capacitance, inductor or transistor test, you are doing it wrong.

Also... parts operating from 0~50C?!? Perhaps in Shenzhen's specials.

[joeqsmith]

In part 4, the UT61E+ is temperature cycled, dropped and transient tested.   I also use the NanoVNA to compare the 61E+'s input impedance with a few other meters. 

it survived the low voltage zapper?

You know what that means, joe? You're going to have to buy another UT61E+ so you can repeat the tests on an "unmodified" one. 

Videos up.   

We will see how it holds up to cycling the switch.  That has been a weak point for the UNI-T meters I have seen getting a little use.  I looked at the free HF meter and those cheap ANENGs.  Surely it will hold up better than them and the detent spring won't crack like the Keysight meter.  My expectations are low for the UT61E+.     

[joeqsmith]

...
Also... parts operating from 0~50C?!? Perhaps in Shenzhen's specials.

That was the first comment I saw from them.  Painfully lost.  I was going to provide them a few links but after reading the second post, knew they are already an expert and would never read them.