EEVblog #489 - Agilent 34461A Multimeter Review

[EEVblog]

Who else but Dave can take an hour and a half to review a multimeter!
Agilent's new 34461A/34460A 6.5 digit bench multimeter. A replacement for the venerable 34401A
Quick links:
Startup: http://youtu.be/SObqPuUozNo?t=6m42s
Labels: http://youtu.be/SObqPuUozNo?t=15m52s
Ratio: http://youtu.be/SObqPuUozNo?t=19m53s
ProbeHold: http://youtu.be/SObqPuUozNo?t=20m58s
Continuity: http://youtu.be/SObqPuUozNo?t=22m35s
Probes: http://youtu.be/SObqPuUozNo?t=23m20s
Statistics: http://youtu.be/SObqPuUozNo?t=26m07s
Overshoot: http://youtu.be/SObqPuUozNo?t=29m20s
Overload: http://youtu.be/SObqPuUozNo?t=32m09s
Diode: http://youtu.be/SObqPuUozNo?t=33m39s
Current: http://youtu.be/SObqPuUozNo?t=34m22s
Trend Plot: http://youtu.be/SObqPuUozNo?t=37m04s
Histogram: http://youtu.be/SObqPuUozNo?t=46m28s
Res Standard: http://youtu.be/SObqPuUozNo?t=50m35s
Bargraph: http://youtu.be/SObqPuUozNo?t=52m40s
Cal Check: http://youtu.be/SObqPuUozNo?t=55m27s
Cal Drift: http://youtu.be/SObqPuUozNo?t=58m30s
Wrapup: http://youtu.be/SObqPuUozNo?t=1h4m55s
Fan Noise: http://youtu.be/SObqPuUozNo?t=1h6m52s
Comparison: http://youtu.be/SObqPuUozNo?t=1h7m58s
AC Testing: http://youtu.be/SObqPuUozNo?t=1h8m40s
Specs: http://youtu.be/SObqPuUozNo?t=1h11m47s
Competition: http://youtu.be/SObqPuUozNo?t=1h12m46s
Software: http://youtu.be/SObqPuUozNo?t=1h14m45s
Conclusion: http://youtu.be/SObqPuUozNo?t=1h22m19s

[c4757p]

I was rather confused for a moment (OK, stupefied - it's after midnight here and I'm tired as all hell...) to see my desktop background in a YouTube video.

Oh. Right. I got that photo from you.

[Mike Warren]

No, Dave stole your computer when you weren't looking.

Just shows how deep this meter is that even after 1.5 hours showing it there are still things not covered.

Actually, I was disappointed, but not really surprised that the software was not as well thought out as the hardware. This seems to be so common in modern equipment, no matter how expensive or well made.

Still, a very nice meter.

[Smokey]

Hmmmmmm.... I wonder why the 34461A showed a lower voltage measurement when the HP and the Keithley pretty much agreed.... Maybe because it wasn't re-calibrated after someone took the shielding can off during the teardown even after they were warned not to... twice... 

I'm just messing around.  I'm actually curious though, since it sounds like you've been in contact with the Agilent guys since you did the teardown because of some software stuff, if anyone commented about you removing the shield.  Do they ever have anything to say about your teardowns of brand new gear (that you can tell us about at least)?

[retiredcaps]

Hmm, possible give-a-way prize for eevblog #500?  Everyone must have at least 25 "useful" posts? 

Excellent review so far and I'm only 33 minutes into it.

[c4757p]

I'm quite accustomed to using older test equipment. Actually, excluding a couple cheap-ass eBay specials and one $120 multimeter, nothing currently on my bench was made this decade. So, I typically don't get too excited about new equipment. But holy crap do I want one of those. Sadly, I'm not so skilled at mental gymnastics to justify a $1000 multimeter for a home workbench...

[casper.bang]

For those who don't know (Dave had to read up on it?), whether a comma or a dot is used as a decimal point (and what's used for digit grouping) is entirely a localization issue - Europe, South America, Greenland etc. observes comma; Asia, North America and Australia observes dot. Good on Agilent to cater to both camps.

[duskglow]

They could have added a GPS and automatically set up the localization based on where you are.   

Maybe for something that costs 20 times as much...

[EEVblog]

Hmmmmmm.... I wonder why the 34461A showed a lower voltage measurement when the HP and the Keithley pretty much agreed....

Coincidence most likely.
I'd rather trust a brand new Agilent meter than an unknown Keithley from ebay and an old 3457A, albeit one that one cal checked in 2010.
The cal cert reading for the 34461A (forgot to show it) shows a measured error of -0.0001% for a 10V DC on the 10V range.

I'm just messing around.  I'm actually curious though, since it sounds like you've been in contact with the Agilent guys since you did the teardown because of some software stuff, if anyone commented about you removing the shield.  Do they ever have anything to say about your teardowns of brand new gear (that you can tell us about at least)?

They haven't mentioned it.

[EEVblog]

For those who don't know (Dave had to read up on it?), whether a comma or a dot is used as a decimal point (and what's used for digit grouping) is entirely a localization issue - Europe, South America, Greenland etc. observes comma; Asia, North America and Australia observes dot. Good on Agilent to cater to both camps.

No I didn't have to read up on it, I knew some countries used the comma, but I didn't know which ones or how many.

[mikeselectricstuff]

When I saw those ridiculous ranges on the limits (GV etc.) I assumed this was because you could add a maths function to allow use of external shunts, dividers & amplifiers to display actual values taking the scaling into account, which would be really useful, but looking through the Maths menu it appears you can't  - this seems a bit of a silly limitation.
 

[lewis]

How does the AC measurement range work at such high bandwidth? It uses digital RMS, so it must(?) capture and store many millions of samples per second to get reliable measurements at high crest factors. That seems pretty damn hard to do with an integrating ADC at 6-digit resolution which is inherently relatively slow. Even equivalent time sampling seems like it would be to slow.

I must be missing really something simple here...

EDIT - found this: http://cp.literature.agilent.com/litweb/pdf/5988-6916EN.pdf

Awesome review, gonna order one. Dave, you've cost me a fortune. I hope they release updated firmware soon, there do seem to be some small niggles, and I hope they make the firmware user updateable. Mike - have you tried updating yours?

[John Coloccia]

For those who don't know (Dave had to read up on it?), whether a comma or a dot is used as a decimal point (and what's used for digit grouping) is entirely a localization issue - Europe, South America, Greenland etc. observes comma; Asia, North America and Australia observes dot. Good on Agilent to cater to both camps.

No I didn't have to read up on it, I knew some countries used the comma, but I didn't know which ones or how many.

Pretty much Europe (except the UK) and South American use a comma.  Everyone else uses a dot.

The continuity test on the Agilent was seriously disappointing.  Fluke's been getting it right for decades.  I don't understand why this seems to be so difficult for everyone else.

[EEVblog]

The continuity test on the Agilent was seriously disappointing.

I wouldn't say that, it was quite usable.

[EEVblog]

EDIT - found this: http://cp.literature.agilent.com/litweb/pdf/5988-6916EN.pdf

IIRC at one point you can see in the video the "rolling average" like display change in the RMS measurement which shows it collecting data over a long time period.

[AlfBaz]

I wonder if the continuity tester would miss a hit if you decreased NPLC?

[laborratte]

When I saw those ridiculous ranges on the limits (GV etc.) I assumed this was because you could add a maths function to allow use of external shunts, dividers & amplifiers to display actual values taking the scaling into account, which would be really useful, but looking through the Maths menu it appears you can't  - this seems a bit of a silly limitation.

I think it could be useful when doing measurements on thunders on planet Jupiter, but the leads are too short. And for nearfield measurements during atomic bombs explosions the beeper is not loud enough...

[ttp]

Battery voltage measurement - auto range vs manual range. I suspect in manual 34461A displays average of last 10 (or whatever the setting is) samples, when you connect the battery it may have sampled few samples close to 0V then some samples of battery volts and first displayed average is low, next is correct as all samples are of battery voltage. In auto range it ranges first and it knows to discard the samples as useless, once it is in the right range then it takes required number of samples and averages those, because all samples are battery voltage so the first reading is spot on.

[Rasz]

@9:20
No dave, this is not a "very good display let me tell ya". This was a good display in 2004 when Sony PSP came out.
This is the CHEAPEST TN screen they could find. In Tablets this is considered garbage bin quality. Difference between weak 480x272 TN screen and 1024x600 IPS is about $15. Agilent decided they like $15 better than perfect viewing angles. How much do you value proper (no discoloration/negative colors) viewing angles?
This smells like a bean counter decision, or no boots on the ground in Asian parts supplier market.
Yes, one can argue this is the best display available in bench meters, but its like saying sand flavored ice cream is the best ice cream in Sahara Desert because its way better than camel shit ice cream.

Win CE booting 50 seconds - you can see it is going through detecting hardware phase - this is most retarded on fixed hardware design. Its not like its going to boot up one day on a different motherboard with less ram and bigger screen. Shows how lazy person doing Windows integration was. Firmware like this should boot same way suspend to disc works, with preselected batch of drivers and plug&play disabled.

Other than that great meter.


btw: Chris was right once again - IPad app for bench instruments Oscilloscopes are next

[EEVblog]

@9:20
No dave, this is not a "very good display let me tell ya". This was a good display in 2004 when Sony PSP came out.
This is the CHEAPEST TN screen they could find. In Tablets this is considered garbage bin quality. Difference between weak 480x272 TN screen and 1024x600 IPS is about $15. Agilent decided they like $15 better than perfect viewing angles. How much do you value proper (no discoloration/negative colors) viewing angles?

Is your wizz-bang hi-res consumer IPS display still going to work in 20 years time?
This is a piece of quality industrial test gear, it is not a consumer phone designed to be replaced after a couple of years.
Do you really need that high a view angle. The one supplied is more then good enough for it's intended purpose.

This smells like a bean counter decision, or no boots on the ground in Asian parts supplier market.

No, there could be very good technical engineering reasons for choosing this display over a consumer IPS type display.

Yes, one can argue this is the best display available in bench meters

Bingo. It is a nice display, end of story.

btw: Chris was right once again - IPad app for bench instruments Oscilloscopes are next

99% of users of this meter will never use a tablet to view the display.
Chris loses again 

[mikeselectricstuff]

The continuity test on the Agilent was seriously disappointing.

I wouldn't say that, it was quite usable.

..but as it _can_ sample much faster it seems silly not to give the option. Also silly not to be able to set teh threshold.
However you could probably set up limit test on the ohms range to get a more flexible continuity, though you really shouldn't have to.

[mikeselectricstuff]

@9:20
No dave, this is not a "very good display let me tell ya". This was a good display in 2004 when Sony PSP came out.
This is the CHEAPEST TN screen they could find. In Tablets this is considered garbage bin quality. Difference between weak 480x272 TN screen and 1024x600 IPS is about $15. Agilent decided they like $15 better than perfect viewing angles. How much do you value proper (no discoloration/negative colors) viewing angles?
This smells like a bean counter decision, or no boots on the ground in Asian parts supplier market.
Yes, one can argue this is the best display available in bench meters, but its like saying sand flavored ice cream is the best ice cream in Sahara Desert because its way better than camel shit ice cream.

I'd slightly agree with this - If the meter is on a shelf, sitting on top of something else, so you're looking up at it, you do get some colour distortion, and it's not far away from being harder to read. However angling it downwards improves things a lot. An IPS display may be slight overkill...

[mikeselectricstuff]

@9:20
No dave, this is not a "very good display let me tell ya". This was a good display in 2004 when Sony PSP came out.
This is the CHEAPEST TN screen they could find. In Tablets this is considered garbage bin quality. Difference between weak 480x272 TN screen and 1024x600 IPS is about $15. Agilent decided they like $15 better than perfect viewing angles. How much do you value proper (no discoloration/negative colors) viewing angles?

Is your wizz-bang hi-res consumer IPS display still going to work in 20 years time?

I doubt there's any difference in lifetime or production longevity - I'd be surprised if the display is any different from a typical consumer one, except maybe specced for long backlight life using better LEDs, or under-running them.
Agilent are big enough that they will be able to get displays made to their spec for continuity of supply.
Even if they do hit a supply problem, worst-case  it would be a minor redesign of the front panel PCB, which wouldn't be that a big a deal.

[moemoe]

You're criticizing the trend chart space @0:39:xx, but I'm pretty sure the graph has to keep it's width.

You can only make the display (readings/min) pixels wider, as long as you want to keep them equally wide. So there's just space for some additional bargraph of the current value or sth like that, but  I can't figure out anything really usefull.

[madires]

They could have added a GPS and automatically set up the localization based on where you are.   

Maybe for something that costs 20 times as much...

Yeah! Even my cheap EUR 250 radio scanner supports GPS. Any smartphone has also one built in. And it would give you a nice frequency standard for measuring frequencies. Actually that idea is not so bad at all :-)