Handheld meter robustness testing

[joeqsmith]

Interesting post from a viewer attempting to show their U1282A will read up to 44MHz.  I am not sure what they are using for a termination, cables or connectors.   Nor do they do any sort of sweeps of the voltage or frequency so I really have little to go on.  They don't show the power up but the claim is it has the same firmware.  Maybe just luck of the draw, older hardware.   Still it's  a long way off from the 100MHz printed in the manual.   

[skander36]

The meter is calibrated in January 2020. FW 01.05.
I have used a 50 ohm (50 cm long) cable. You can view attached the cable used. Worth mentioned that with normal cables I wasn't able to get more than 35 MHz.
The measurement is quite unstable and I need more trial and error rounds to achieve these value. Definitely not practical. 
I have obtained also max.  40 MHz from first try with an ADF4351 generator, rectangular waveform this time at about 950 mVpp (fixed value).
For me this meter was a little disapointing, because of long time needed when change between measuring domains. It was a surprise because the Agilent 1252A is one of the fastest multimeters that I have. 1282A became comparable faster if you switch to fast from settings but with a lower resolution.
This frequency measurement also show a weird behavior and a discrepancy beyween declarations (100 MHz) and reality.
Overal is a good meter because precission and resolution, long life of batteries, shock proof, but for me it is not a "daily driver" due to its slowness.
Better resolution for pics here : https://ibb.co/album/f4r0dp

[rsjsouza]

Do you find that the system will glitch as I saw for longer captures?

Glitch? I haven't seen any issues. Long captures were not a problem as shown below (albeit with my U1273A)

As in that first shot.  Keep in mind you are only three minutes of data.  I was showing 7 hours.   That glitch was after 3 hours with a 10 second sample rate.

That was an older capture, but the total was more than twenty minutes of data without a glitch (not three). I can retry with the U1282A, but I won't promise a runlength of hours.

Ok, I put it to run while doing other things around the house. Indeed there was a glitch at around 40min mark for the U1282A.


I will probably test the U1273A and see if similar things happen, which might help isolate the issue to either the meter or the software.

[joeqsmith]

So no termination.  With the Arb set to 50 ohm, I am guessing the readout is then off by a factor of 2X,  or really 240mVpeak or 0.169 RMS.   

Still, with the one I have, using no termination, 200mm of  RG400, setting the RF generator to even 35MHz and changing the amplitude, I can find a sweet spot where it comes in.   This is why I was asking about you changing the frequency and amplitude in the video to give me some idea how well yours detects the signal. 

With  your second source, is that 950mVpp without the termination?   When I tried to square it up with the LVDS and PECL drive, it made no difference in how my meter behaved. 

Both are not good compared with what is stated but I'm surprised yours behaves differently, especially after rsjsouza's post showing the exact same error in the frequency I see with mine.   

[skander36]

Second source - no termination.
The "video" is not a video, I just insert the picture in a video format and save as mp4 because youtube delete any post that contain a link to image hosting sites.
At 38 MHz it read 36,53 MHz.
For 44 MHz at 380 mvpp it show a variable value between 57-66 MHz.
I have no intention to make a video. I just wanted to show that it can read a higher value. In  my opinion this meter simply cannot read 100 MHz.

[rsjsouza]

So, the U1273A shows no glitch.

[joeqsmith]

I will probably test the U1273A and see if similar things happen, which might help isolate the issue to either the meter or the software.

Thanks for double checking.  With HKJ and myself using two additional programs to talk with the meter and still seeing the problem, I doubt it's the software causing it.  Maybe the driver.  To help answer this, I used a dumb RS-232 IR adapter and sniffed the Keysight adapter while trying to connect.  I get the *IDN?   Trying it with LabView, I see the 0D0A terminator.  So nothing magic going on.  The old time IR adapter was then stuck on the U1282A meter and I started the logging.  I'm using an FTDI RS232-USB adapter so no Keysight software or driver.   

I have the same 40 ohm resistor attached and again collecting at 1sps.  I'll let it run for the afternoon and see if the glitch is still there.  If so, we are down to firmware and hardware.   I would be embarrassed if I had customers trying to solve my problems...

[joeqsmith]

So, the U1273A shows no glitch.

You may want to continue to let it run for the afternoon as well.   Doesn't hurt to make sure. 

[joeqsmith]

Well, that was quick.  It's not a large dip but similar to the small one that I showed the last time.   I'll let it run but it's down to the meter now.  Good job Keysight! 

[HKJ]

It probably depends on range.


A 100 ohm resistor (0.01%) and there is no huge glitches.

Only a small one, where it dropped to 96ohms.

TC will not always show short glitches on the chart, there is not enough dots for it. but searching will find it and the range page list the absolute minimum value in the data.

I switched to a 10ohm resistor and got a large glitch again:

Here it is -0.4164 ohm

[TheAmmoniacal]

Defpom's review is up.

Part 1

https://youtu.be/bFyJlMd35v0

Part 2

https://youtu.be/UgbDNE0KhhI

[Chance92]

Well, that was quick.  It's not a large dip but similar to the small one that I showed the last time.   I'll let it run but it's down to the meter now.  Good job Keysight!

Hi Joe. Do you think if you could run it again with the min/max function enabled on the meter, so that we can see if this glitch would show on the display? This way we may be able to isolate the problem further.

[joeqsmith]

Well, that was quick.  It's not a large dip but similar to the small one that I showed the last time.   I'll let it run but it's down to the meter now.  Good job Keysight!

Hi Joe. Do you think if you could run it again with the min/max function enabled on the meter, so that we can see if this glitch would show on the display? This way we may be able to isolate the problem further.

So you're thinking just let the meter sit stand alone with min/max and see if it happens?  It does seem like a good idea.     

[TheDefpom]

Defpom's review is up.

Part 1

https://youtu.be/bFyJlMd35v0

Part 2

https://youtu.be/UgbDNE0KhhI

LOL, yes up, but not publicly published yet, I wish youtube would hide unlisted videos from public playlists until the video gets made public, oh well.

Enjoy the early access I guess !

[joeqsmith]

Defpom's review is up.

Part 1

https://youtu.be/bFyJlMd35v0

Part 2

https://youtu.be/UgbDNE0KhhI

LOL, yes up, but not publicly published yet, I wish youtube would hide unlisted videos from public playlists until the video gets made public, oh well.

Enjoy the early access I guess !

Good to see you also see a similar problem with the counter.   Outside of what you have shown, did you try logging any data with the meter?   If so, are you seeing this same glitch we are showing above? 

[TheDefpom]

There are a few features I haven't tried using yet, logging is one of them, I have a lot on at the moment so I don't know when I will get to look at that.

[joeqsmith]

There are a few features I haven't tried using yet, logging is one of them, I have a lot on at the moment so I don't know when I will get to look at that.

Lucky for me, logging requires very little of my time.  Basically set the meter up and walk away.  Well, I spend maybe an hour to put that simple program together, test their interface cable, then sort get it attached to the meter.    Watching grass grow, paint dry, kettle boil.... all much more exciting. 

[TheAmmoniacal]

Defpom's review is up.

Part 1

https://youtu.be/bFyJlMd35v0

Part 2

https://youtu.be/UgbDNE0KhhI

LOL, yes up, but not publicly published yet, I wish youtube would hide unlisted videos from public playlists until the video gets made public, oh well.

Enjoy the early access I guess !

Oops, sorry about that - didn't even see they were unlisted before you mentioned it. Part 2 appeared in a recommended playlist on my TV, when I clicked to watch it, Part 1 was suggested just below it. Then just copied the links from my watch history.

[joeqsmith]

I let the U1282A run for 25 hours using my software with the simple RS232 dumb IR cable and FTDI USB controller.  After that initial small glitch, the setup remained stable.   

Not being able to reproduce the glitch, I wonder if we are dealing with multiple problems.    I have increased the sample rate from 1sps to 4sps 5sps to see if it has any effect.   

***
correct brain fart...

[joeqsmith]

Well, how things change....

So after changing the sample rate I captured another small glitch which first appeared to have an identical amplitude as the first.   Zooming in, they are close but not the same.   Then, shortly after I captured another large glitch.     

I want to point out that the 4sps is a mistake and should have been five.   Obviously Windows and Labview are going to have a timing error.  The measured jumps around +/-1mS.  So basically 5sps sample and the meter seems to keep up just fine.    After the large glitch, it's been stable.  I plan to continue to let it run. 

It's starting to look more and more like the meter has a problem (hardware and or firmware).   

[NoMoreMagicSmoke]

Well, how things change....

So after changing the sample rate I captured another small glitch which first appeared to have an identical amplitude as the first.   Zooming in, they are close but not the same.   Then, shortly after I captured another large glitch.     

I want to point out that the 4sps is a mistake and should have been five.   Obviously Windows and Labview are going to have a timing error.  The measured jumps around +/-1mS.  So basically 5sps sample and the meter seems to keep up just fine.    After the large glitch, it's been stable.  I plan to continue to let it run. 

It's starting to look more and more like the meter has a problem (hardware and or firmware).   

Joe just curious when u think it might be a hardware issue do u suspect the chypset at all?  Do any other meters based on the HY3131 have similar glitches?

[joeqsmith]

Well, how things change....

So after changing the sample rate I captured another small glitch which first appeared to have an identical amplitude as the first.   Zooming in, they are close but not the same.   Then, shortly after I captured another large glitch.     

I want to point out that the 4sps is a mistake and should have been five.   Obviously Windows and Labview are going to have a timing error.  The measured jumps around +/-1mS.  So basically 5sps sample and the meter seems to keep up just fine.    After the large glitch, it's been stable.  I plan to continue to let it run. 

It's starting to look more and more like the meter has a problem (hardware and or firmware).   

Joe just curious when u think it might be a hardware issue do u suspect the chypset at all?  Do any other meters based on the HY3131 have similar glitches?

I think the only other meter I looked at with that Hycon chipset was the 121GW.  It's been so long since I looked at that meter, I'm not sure if they ever sorted out all the problems with it or not.   There were certainly problems when logging to the SD card but I don't remember there being any problems when using BLE.  I would think if its hardware, it is something in how the chipset is being used, not the IC itself.   

If old age has taught me anything when it comes to solving problems, stay data driven.  Don't let feelings cloud your judgement.    We know the U1282A sends up ASCII data.   We are using a 40 ohm resistor and the meter sends up  something like: +4.00108000E+01.  There's no missing data that would get us to what we are seeing.    So far it appears random and can be reproduced without Keysight's software, drivers or interface cable.   

I changed the test program to switch between two sample rates every 10 samples to see if that has any effect.   

Personally, I blame this problem on those that voted for this meter!!  You knew dam well I wouldn't just unbox it and give it five stars!!   


 

[25 CPS]

Defpom's review is up.

Part 1

https://youtu.be/bFyJlMd35v0

Part 2

https://youtu.be/UgbDNE0KhhI

I'm in the process of watching Part 1 now.

Right off the top though, he says that his U1282A was supplied by Keysight.  If there's differing performance, could it be possible the sample unit specifically selected from a group of meters to make sure a good one was sent out for review vs. whatever you'd end up with by purchasing through regular retail channels?

The second question is why?  That one I truly don't understand.  If Keysight's going to be so painful to deal with for individuals and small businesses, why send meters out for review on YouTube?

Maybe I'll lash up my U1282A which was a used eBay purchase up to a laptop since I've got the Keysight software running on one and several of the interface cables and let it rip for a few days when I'm on some crappy shift work and then look at the results after.

[joeqsmith]

Right off the top though, he says that his U1282A was supplied by Keysight.  If there's differing performance, could it be possible the sample unit specifically selected from a group of meters to make sure a good one was sent out for review vs. whatever you'd end up with by purchasing through regular retail channels?

That's always been a concern for me as well and why I will normally try to procure them the way anyone else on this forum would.   It's also why I will not use my professional background when attempting to open a dialog with a company.  I want people to know as a hobbyist how the company may treat them.   Of course, there have been cases where I have looked at preproduction meters.  In these cases, you get to see things as they evolve which I find rather interesting.   

The second question is why?  That one I truly don't understand.  If Keysight's going to be so painful to deal with for individuals and small businesses, why send meters out for review on YouTube?

With the timing, I still think it's damage control.       In my case, there's no income from the channel.  I don't ask viewers to subscriber, ring some bell or consider becoming a Patreon.  I don't even run ads.  There's nothing even setup to take in money if I wanted to.   A total loss on my part.  On the flip side,  don't expect to see any basic unboxing 5-star review videos anytime soon!   

[NoMoreMagicSmoke]

I think the only other meter I looked at with that Hycon chipset was the 121GW.

I expect that the BM78x series is the same chipset, but alas it also has no data collection abilities.

If old age has taught me anything when it comes to solving problems, stay data driven.  Don't let feelings cloud your judgement.   

So very true. It's really hard to keep feelings from clouding judgement... Especially in the current atmosphere where anyone can start a youtube channel and instantly be an "Expert" on whatever... Data is all that more important to be able to sort the truth from the opinion.

Personally, I blame this problem on those that voted for this meter!!  You knew dam well I wouldn't just unbox it and give it five stars!!   

I think I am one of those to blame then.  . I recently was making the decision between a BM789 or a U1272A... I ended up picking the Brymen and based on your reviews and the current discussions about Keysight not caring about non-business customers I think I made the correct choice.