What are - for you - the 3 biggest disadvantages of your benchtop DMM(s)?

[PushUp]

As I - and probably some others - plan to buy a benchtop DMM or further benchtop DMMs, this probably helps, to filter and watch out for your own needs just by elimination?!



You don't need to mention three disadvantages, when there is only one for you...just put it in a nutshell.

Here are some examples - I hope they are right: 

Although the price can be a disadvantage, it is better to skip this point, to focus on your personal preferences, for example:

Keysight 34465A
1. price   no "AC+DC" true RMS value
2. 100uF max is a joke
3. ...

Rigol DM3068
1. a green (and blue) LED lights up, but does not show the voltage drop in diode mode

The SnapShot is out of this video, which is 8 years old: https://www.youtube.com/watch?v=eR0TM3eitNk

Correction:
In order to update the outdated video footage...

It is also possible to see the voltage drop with different current values now having specific upper voltage limits:



...in this case with "100uA = Measure Range: -0.2V to 4.2V" for a white diode:



2. too small Display for trend chart
3. hardly any useful 2nd display combination, such as AC vs. DC

Rohde & Schwarz HMC8012
1. maximum current load above 5A only for 30 sec., followed by a pause of 30 sec.
2. no auto ranging for 2nd display
3. 500uF max is better than Keysight but still useless

Keithley DMM6500
1. quite long dimensions like an old analogue Scope
2. mainly touchscreen (might be a benefit for others)
3. ...

If someone is very disappointed from very old and known firmware bugs, it is probably useful to name them, but all under its own category "FW-Bug".

Siglent SDM3055
1. a blue LED lights up, but does not show the voltage drop in diode mode

The SnapShot is out of this video, which is 5 years old: https://www.youtube.com/watch?v=rmTP-aR3CxQ

Correction:
As mentioned by tautech, it is solved: https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/msg3028294/#msg3028294



2. ...
3. ...
FW-Bug: https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/msg3328326/?PHPSESSID=e8ja7i0ragru6auspt39agp3m1#msg3328326


It is useful to mention a missing option, although it is known from the beginning, but reveals that it would have been nice to have with hindsight, for example:

Keysight 34460A
1. no trend chart
2. DCI & ACI measurements only up to 3A
3. ...


ThanX & Cheers! 

[rvalente]

No bip in continuity mode, like fluke 8012a. Why fluke, why?

[Electro Fan]

After 7 years with only handheld DMMs and now having a bench DMM I’d say other than not being portable there aren’t many (any?) disadvantages to a bench DMM other than cost.  You can’t buy a bench DMM for $25 (like an Aneng 8008) but if your budget can justify a Siglent 3055 it might be hard to justify a Fluke 87V for a similar price (unless you need portable).  I think a Siglent 3055 and a good (enough) handheld would be a good first two DMMs (everyone who wants to learn/do electronics should have at least two DMMs).  So the entry is two 8008s and where you go from there is up to the you - but one handheld DMM and one bench DMM might be a good objective.

[coromonadalix]

1  totally  hate any vfd display based meter ---- if they have unobtanium parts
2  lack of a simple acquisition software  if they have a serial / lan / gpib or usb port ...  unless someone has done a great job of building one   
3  repairability / schematics

4  At least a good diode test around 5 volts would be perfect loll     unless you have an eevblog 121gw  who could do a 15v diode test ??



But the list would go on and on  loll           

My dream would be a nice meter say at leat a 5 1/2 digits  with an added lcr meter in it like an De-5000, franken meter  loll

My problems is :  i have the Fluke 189 habits, totally love the way they work, display resolution, good accuracy, some data logging, and i have their bp-189 battery pack, sure they get old  and some parts are hard to get, and i loooove the fuses acess if needed.

I have the same fun with  An Amprobe am-140  but i hate the fact, i have to open the meter if i blow a fuse (wich never happened)

For my 34401a,  well  they get old, but are rock solid, schematics are available,   now you have some vfd clones available, you can extend their lives ... i did it for one of my meter, aaannnd  no fans 

[Electro Fan]

another thread that might be relevant to this thread:

https://www.eevblog.com/forum/testgear/aim-tti-1908-5-12-digit-benchtop-multimeter-(preview)/

[tautech]

..................
Siglent SDM3055
1. a blue LED lights up, but does not show the voltage drop in diode mode

Oh yes it does !!!
White diode test from this post:
https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/msg3028294/#msg3028294

[rvalente]

Reduced input protection when compared with a industrial class handheld can be a pain, specially if you're from development or repair of power electronics.
The diode measurement is really a thing, no reason why ir could not be as high as 40v, so many zeners could be tested

[Kleinstein]

I currently have not Bench meter.  For use at work :

HP3457:  only up to 3 V with high impedance, relatively low contrast LCD, limited current ranges
Prema5000: relatively noisy, some visible DNL, only limited current ranges

With both meters I liked the internal scanner.

[coromonadalix]

5   Damn forgot this one,  having  giga ohm impedance input(s) up to 20v  like some meters ...

[HKJ]

Disadvantage with bench meters are often high voltage in ohm range and no AC+DC reading. Continuity may also be as fast a fastest handheld.
I have 34470A, 34465A and DMM6500 on my bench. The Keysight wins for AC, but the Keithley wins for fast sampling/charting.
For logging I mostly use 34461A, I also have a DM3068, but do not like the display on it and when I got it, I could not get logging to work (After some software update I believe it works fine, but I have not needed it for logging).
I do also have a couple of cheaper bench meters, I have done reviews of them (https://lygte-info.dk/info/DMMReviews.html ), but are not using them much.

[joeqsmith]

I bought the HP34401A because I had no way to control the HP3468A (HPIL) with a standard PC.  For my hobby, both have served me very well. 

The HP34401A's fuse is located in the rear.  Other equipment is stacked on it, it has a thick HPIB cable going to it. 

Not too long ago, we had a tree in the front yard that was struck by lightning which then coupled into the home through the coax, the phone lines, the sprinkler system and the mains.  It caused enough common mode voltage in the lab that it damaged some of my equipment that uses GPIB.   I recently went full in porting my old XP code to Windows 10 and when working with the VNA, was reminded just how slow that bus is.    I would like to have Ethernet integrated on all my test equipment.   Smaller cables, unlimited devices, faster, much high common mode immunity....   

[Wytnucls]

Keithley 2000 and Rigol DM3058E.

Very happy with both meters.
Minor gripes:

Keithley:
DC nA accuracy should have been better for a 6 1/2 digit meter. (10mA range)
No capacitance measurement
No USB

Rigol:
No uOhm resolution for 4 wire measurement
Continuity beeper faint
No auto hold
 

[bd139]

3478A:

1. Horridly bad LCD display
2. No continuity / feeper
3. Terrible front panel legend meaning I always stick my probes in the wrong holes.

That’s about it.

[maginnovision]

Prema 5017:

1. Somewhat slow for highest accuracy (min 2s for 7.5 digit)
2. Despite not having a fan/internal airflow it's fairly sensitive to temperature changes.
3. Parts can be tough to get.

For the most part though it's been working really well for me and it uses rs232 which I prefer over most alternatives.

[bdunham7]

Fluke Bench DMMs from 8840A to 8846A (which are otherwise outstanding meters) have terrible current measurement implementations.  Just give up and get external shunts.  If I had to measure current very accurately on a frequent basis and didn't want to use shunts, my 8846A would be gone in a flash.

And, of course, the later "6.5 digit" meters are 1.2M count instead of 2.0xM counts, which is very common but still something to grumble about.  I really would like my 10V high-impedance range to go to 20V instead of just 12. 

[floobydust]

Agilent 34461a
1. Cooling Fan
WTF it has a Cray or ENIAC inside? Even a RPi does not need a cooling fan. The fan is noisy, howls like a coyote. Myself and others find large temperature swings/reference drifts from the room A/C temperature changes. Deal breaker for me.

2. 6-1/2 digits but only shows 3sd on diode-test (5V).

3. Kill me with zeros
000.0005  the leading zeros are bad practice and contain no information.

4. Unknown capacitance to earth-ground, sometimes bench DMM's have large Y-caps.
You can get strange side-effects when connecting probes to gear referenced to earth-ground.

I went back to the 34401a's - silent with no boot time, no F/W bugs, no SMPS hash it has a linear power supply so it's just CPU EMI, no wailing fan, and just a smoother drive.
edit: removed bashing of the power supply EMI

[EEVblog]

4. Unknown capacitance to earth-ground, sometimes bench DMM's have large Y-caps.
You can get strange side-effects when connecting probes to gear referenced to earth-ground.

Yep, I've seen that, particularly on fussy low current systems.

[wizard69]

It is pretty easy to come up with 3:

• No Cat rating or CAT ratings that don't reflect needs on a repair bench.   The reality is that even rather small servo drives run at 500 VDC sometimes 650 VDC bus voltages.   There is enough happening on repair benches that having the higher safety ratings would be of real value.
• No buzzer or one that is effectively silent.   Frankly I'd rather see a knob on the front panel for volume adjustment of the buzzer.It would probably be too much to ask for multiple tones for example a high frequency buzz for exceeding the upper limit and a low frequency buzz for the low limit.
• You have to pay way too much to get RTD and thermocouple interfacing.

Other things that drive me crazy in a multimeter.   for one they seem to want to replace a scope with some of the functions, Id rather that a multimeter to multimeter things instead of trying to be a scope.   For example some multimeters have capacitance functions but why not go a step further and offer more LCR functionality?    The same goes for inductors.    It should be possible to put some basic LCR functionality, testing up to 100kHz.

Another feature that should be built into a bench meter is a real frequency / counter subsystem with a real BNC input and at least 10 MHz input range, {nothing fancy here}.   This should run completely independent of the DVM inputs and frankly with modern silicon should cost a few cents.   In a nut shell put some real meat into the "multi".

For the most part modern multimeters have not shrunk much at all compared to a modern scope.    So in simple terms modern multimeters need to go on a diet, especially with respect to meter depth .   I'd prefer a higher profile to be honest.

Why not a built in programming language like MicroPython or even a Swift implementation.   You need a programming langauge that makes it simple to automate tests directly on the meter.   This kinda implies a keyboard / mouse input, USB or a decent interface over Ethernet.

Meter handles that don't break!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Add some digital I/O so that you can use MicroPython or Swift to automate your test.   Again we have a lot of CPU horsepower available these days, why not do as much as possible locally.   Todays meters have some I/O but it is dedicated in most cases.    Even if we only have a dozen or so of I/O, it could make for a far more flexible meter.

How about synchronized clock calendars for data time stamping.

Why are these meters battery free.   A snap on battery option would be nice for field work, generally calibration.   Sometimes a bench meter makes sense in the field.   Further even scopes come in battery powered version so why not?

After all these years why are we stuck with one input channel.   Two channels that can sample in a synchronized manner, should be a snap with todays tech.   Maybe even three.   They don't need to be 4 wire capable either.

Finally (I could go on forever), I'd rather see a 4.5 or 5 digit meter that works well and is fast, than to have a bunch of useless digits on the meter.   So is this a disadvantage?   Well yeah excessive resolution in the meter can be a problem if it creates a high cost device that doesn't support other features that would be nice to have in a meter.

[KaneTW]

Agilent 34461a
3. Kill me with zeros
000.0005  the leading zeros are bad practice and contain no information.

The leading zeros are likely on purpose to indicate the range.

[tautech]

It is pretty easy to come up with 3:

• No Cat rating or CAT ratings that don't reflect needs on a repair bench.   The reality is that even rather small servo drives run at 500 VDC sometimes 650 VDC bus voltages.   There is enough happening on repair benches that having the higher safety ratings would be of real value.
• No buzzer or one that is effectively silent.   Frankly I'd rather see a knob on the front panel for volume adjustment of the buzzer.It would probably be too much to ask for multiple tones for example a high frequency buzz for exceeding the upper limit and a low frequency buzz for the low limit.
• You have to pay way too much to get RTD and thermocouple interfacing.

Other things that drive me crazy in a multimeter.   for one they seem to want to replace a scope with some of the functions, Id rather that a multimeter to multimeter things instead of trying to be a scope.   For example some multimeters have capacitance functions but why not go a step further and offer more LCR functionality?    The same goes for inductors.    It should be possible to put some basic LCR functionality, testing up to 100kHz.

Another feature that should be built into a bench meter is a real frequency / counter subsystem with a real BNC input and at least 10 MHz input range, {nothing fancy here}.   This should run completely independent of the DVM inputs and frankly with modern silicon should cost a few cents.   In a nut shell put some real meat into the "multi".

For the most part modern multimeters have not shrunk much at all compared to a modern scope.    So in simple terms modern multimeters need to go on a diet, especially with respect to meter depth .   I'd prefer a higher profile to be honest.

Why not a built in programming language like MicroPython or even a Swift implementation.   You need a programming langauge that makes it simple to automate tests directly on the meter.   This kinda implies a keyboard / mouse input, USB or a decent interface over Ethernet.

Meter handles that don't break!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Add some digital I/O so that you can use MicroPython or Swift to automate your test.   Again we have a lot of CPU horsepower available these days, why not do as much as possible locally.   Todays meters have some I/O but it is dedicated in most cases.    Even if we only have a dozen or so of I/O, it could make for a far more flexible meter.

How about synchronized clock calendars for data time stamping.

Why are these meters battery free.   A snap on battery option would be nice for field work, generally calibration.   Sometimes a bench meter makes sense in the field.   Further even scopes come in battery powered version so why not?

After all these years why are we stuck with one input channel.   Two channels that can sample in a synchronized manner, should be a snap with todays tech.   Maybe even three.   They don't need to be 4 wire capable either.

Finally (I could go on forever), I'd rather see a 4.5 or 5 digit meter that works well and is fast, than to have a bunch of useless digits on the meter.   So is this a disadvantage?   Well yeah excessive resolution in the meter can be a problem if it creates a high cost device that doesn't support other features that would be nice to have in a meter.

You need look harder at modern offerings.

[AVGresponding]

3478A:

1. Horridly bad LCD display
2. No continuity / feeper
3. Terrible front panel legend meaning I always stick my probes in the wrong holes.

That’s about it.

Ooer missus!   



For me, it's size (just can't get away from those double entendres),   

The ones that aren't exactly full or half rack width,   

The LCD display ones with no/poor backlighting.   

[joeqsmith]

It is pretty easy to come up with 3:

• No Cat rating or CAT ratings that don't reflect needs on a repair bench.   The reality is that even rather small servo drives run at 500 VDC sometimes 650 VDC bus voltages.   There is enough happening on repair benches that having the higher safety ratings would be of real value.

I would imagine most bench repair work is CATI/II.   If the safety standards are only there to keep the operator safe and not prevent damage to the meter, then the CAT ratings would seem to have little value.  If the intent of the safety standard is only to keep the operator safe, which is what many here suggest (I don't know personally)  there is little need for MOVs, GDTs, PTCs, surge rated resistors, high speed clamps and such. 

Even some of the larger supplies I have worked on don't have enough storage to pose much of a risk if a meter were to breakdown.  I'm not suggesting there are no safety risks, but that the meter would have little to do with it.  If we understand the risks and how to mitigate them (remove rings, one hand rule...) then to me what becomes a real value is meters that are robust enough to survive some low energy transients that you may see on the bench (repair, experimenting....).    This is where the EMC standards come in to play and there should be little doubt the need for added protection circuits to allow the meter to survive.   

How do you quantify how robust a meter is?  Who knows.  I've presented my test methods and the data I have collected.  I would say the actual data carries less weight than opinions.   If your goal is to sell meters and the products you sell perform well against my tests, you may praise the results.   If they do poorly, you may talk about how the tests don't represent real world conditions.   It becomes more about the culture we have today where everyone's an expert after a quick browse of wiki.   

I'm currently reading a pretty good book on this topic titled "The Death of Expertise: The Campaign against Established Knowledge and Why it Matters"
https://www.amazon.com/Death-Expertise-Campaign-Established-Knowledge/dp/0190469412

Sorry for the long post.  If you were trying to make a different point with your comment, if you would'nt mind explaining further, I would like to understand it.

****
first post of the day.  cleaned up some of the wording.

[Fungus]

It is pretty easy to come up with 3:

• No Cat rating or CAT ratings that don't reflect needs on a repair bench.   The reality is that even rather small servo drives run at 500 VDC sometimes 650 VDC bus voltages.   There is enough happening on repair benches that having the higher safety ratings would be of real value.

CAT ratings have very little to do with voltage, they're to do with energy. I doubt your repair bench can supply enough energy to create/sustain a decent arc flash.

(and you should probably be upgrading your fuse box if it can)

[bdunham7]

I would imagine most bench repair work is CATI/II.   If the safety standards are only there to keep the operator safe and not prevent damage to the meter, then the CAT ratings would seem to have little value.  If the intent of the safety standard is only to keep the operator safe, which is what many here suggest (I don't know personally)  there is little need for MOVs, GDTs, PTCs, surge rated resistors, high speed clamps and such. 

I'd speculate that some people may have lost sight of exactly what those CAT ratings are actually intended to protect us from.  I work on some fairly high voltage and moderately high energy systems, on the bench, but the typical bench meter that is rated CAT 1/1000V and CAT II/300V (or more in some cases) is more than adequate, IMO.  Unless  you have a service panel in the middle of your bench, those are the actual levels of exposure you have.
   

How you quantify how robust a meter is, who knows.

I just took apart an old 8842A that apparently was connected to excessive voltage--at least 1500V---with some power to back it up.  The entire front end protection--4 MOVs and 2 fusible resistors--were incinerated.  I removed the burnt parts and subbed in 2 regular 1K resistors (to be repaired properly later, of course) and the meter works properly and is in cal.  No evidence of any external damage, the circuit board is unharmed (but sooty)  and I'm sure nobody died.  I'd say that is 'robust' although I can't quantify that. 

[ferdieCX]

My dream would be a Fluke 8050 with LED Display and:

1) Service documentation
2) No unobtanium parts
3) 4-wire resistance measurement would be a plus.

[joeqsmith]

I just took apart an old 8842A that apparently was connected to excessive voltage--at least 1500V---with some power to back it up.  The entire front end protection--4 MOVs and 2 fusible resistors--were incinerated.  I removed the burnt parts and subbed in 2 regular 1K resistors (to be repaired properly later, of course) and the meter works properly and is in cal.  No evidence of any external damage, the circuit board is unharmed (but sooty)  and I'm sure nobody died.  I'd say that is 'robust' although I can't quantify that.

As I have said, the first DMM I ever owned was a brand new Fluke 8000A bench meter.  Fairly costly and on par with what you may find free at Harbor Freight.   This meter has little to protect the sensitive custom ICs.  A simple mistake and you're changing $70 or so (back then) for new ICs and a handful of precision matched passive parts.   Could it be repaired? Sure, I repaired it more than once.   But the downtime and cost drove me to retire the Fluke and move to HP3468A.   An old story.   

While I would guess every meter I have evaluated could have been repaired, given enough time and money,  once the main controller is damaged I scrap them.   In very rare cases, I have gone so far as to replace ICs to recover them.   Only the 121GW prototype, Summit/TPI 194II, Yokogawa TY720, and UNI-T UT181A all fell into this category.   

There have been some really low end meters, like the UNI-T UT90A that have such poor clearances, they sustain a lot of damage but I don't test with enough energy to finish them off.   Not that I am suggesting these are robust meters but they have been repairable.   The UT90A may be the best example of that.   

Without looking at the data, I would guess somewhere around 60% of the meters were scrapped.  I am not suggesting that these products wouldn't be a fine choice for some people.  I'm not here to push any brands or sell meters.  However, personally, I don't have a use for a meter that is damaged as easily as my first Fluke.  The transients I would encounter in my early days are not representative of what I experiment with now days.    For me, the more punishment the meter can take on the bench without downtime, the better.  It's not about safety for me but uptime.   

I've had a few people comment on my degrading the MOVs over time.   After a member here also made the same suggestion, I ran a long term high voltage test on my Fluke 87V where I applied 1.6KV DC.   This meter certainly fulfills my needs for a robust meter but its lacking in other ways.   

Today, I use the 34401A if I need that performance or am running some experiment that I have automated or want to log the data.   After discovering the BM869s and the Fluke 189, if I am working on anything that I have any concerns about HP, I use these handheld meters as I have no doubt that they would survive levels that the old HP would not.



***
SP

[floobydust]

Agilent 34461a
3. Kill me with zeros
000.0005  the leading zeros are bad practice and contain no information.

The leading zeros are likely on purpose to indicate the range.

My rant about it here.


34461a manual has Cat. II 300V rating, Input protection: 1,000 V on all ranges.
The common-mode vs differential-mode voltage limits are not spelled out, for the in-guard isolation barrier. I frequently go near or over 750-1000V, which leads me to my next disadvantage:

5. RIGHT TO REPAIR
If you damage the front-end, no schematic is available. Keithley has some kindergarten-grade block diagrams but you will never find that leaky JFET or mux.
Very costly at almost 1/2 the price of the instrument to repair it, which is the going rate. I find many bench DMM's are too fragile for real world use, they need to up the front end protection although pA measurement would suffer I guess.

[Fungus]

I just took apart an old 8842A that apparently was connected to excessive voltage--at least 1500V---with some power to back it up.  The entire front end protection--4 MOVs and 2 fusible resistors--were incinerated.  I removed the burnt parts and subbed in 2 regular 1K resistors (to be repaired properly later, of course) and the meter works properly and is in cal.  No evidence of any external damage, the circuit board is unharmed (but sooty)  and I'm sure nobody died.  I'd say that is 'robust' although I can't quantify that.

Nowhere in CAT documents does it say the the meter should survive.

[Fungus]

How about this:

https://www.aliexpress.com/item/4000343031770.html

[PushUp]

As a "little" interruption, I want to thank all of you, who shared their personal knowlege and experience so far, when dealing with their equipment at home or at work!!! 

In addition to this, it is always good, to discuss certain aspects in a forum - please go on with it, as this is always helpful and mostly purposeful! It is clear that experts discuss on another level than beginners and this is good, as it helps others to improve!

As a consequence of this (or nevertheless), I want to encourage all users of all level to share their thoughts, as only this creates an overall picture. You cannot post something pointless, as the question "What are - for you - the 3 biggest disadvantages of your benchtop DMM(s)?", is that I wanted to hear your personal opinion, which of course should be different to others, depending on the field you are working on and how experienced you are and your know-how in general.

It is clear, that anybody has different preferences depending on the area of responsibility, which makes it so valuable to share, so that anyone can filter for his/her own needs.


Thank you very much!!! 


...don't want to disturb the flow - over and out... 

[bdunham7]

Nowhere in CAT documents does it say the the meter should survive.

The 8842A isn't actually 'CAT rated' as far as I know, at least neither CAT xx/yyyV or any reference to IEC1010 appears anywhere on it.  I wonder when that started to be a thing--my earliest product that I can lay hands on that has a CAT rating insignia is my old Scopemeter--CAT III/600V.

However, I don't think this was any type of CAT event.  I think it was subjected to a severe overvoltage, not a transient.   I think it is open for discussion as to whether the meter 'survived'.  IMO it did survive--just a little more repair needed than replacing a fuse. 

[bdunham7]

I frequently go near or over 750-1000V,

You might want to rethink routinely running near-limit voltage through your very best DMM (if that's the case).  I either use an HV probe or a less valuable meter when I'm testing unknown-ish high voltages.  The only time I'll run 1000V+ into a 6.5 digit precision DMM is something like adjusting a calibrator (which is what I'm doing right now) where I need that last digit or at least the one before it.

5. RIGHT TO REPAIR
If you damage the front-end, no schematic is available. Keithley has some kindergarten-grade block diagrams but you will never find that leaky JFET or mux.
Very costly at almost 1/2 the price of the instrument to repair it, which is the going rate. I find many bench DMM's are too fragile for real world use, they need to up the front end protection although pA measurement would suffer I guess.

The days of being able to repair fried meters are mostly over.  Even the OEMs can't fix a lot of this stuff and practice 'repair by replacement', either the board or the unit.  But I think they are actually protected remarkably well considering their sensitivity.  Try designing a circuit that responds to 10 nanovolts but doesn't complain when you suddenly stuff it with a kilovolt.

[tautech]

Nowhere in CAT documents does it say the the meter should survive.

The 8842A isn't actually 'CAT rated' as far as I know, at least neither CAT xx/yyyV or any reference to IEC1010 appears anywhere on it.  I wonder when that started to be a thing--my earliest product that I can lay hands on that has a CAT rating insignia is my old Scopemeter--CAT III/600V.

However, I don't think this was any type of CAT event.  I think it was subjected to a severe overvoltage, not a transient.   I think it is open for discussion as to whether the meter 'survived'.  IMO it did survive--just a little more repair needed than replacing a fuse.

No it didn't survive, it needed to be repaired before further use.
You saved it from going in a dumpster that's all.

[bdunham7]

No it didn't survive, it needed to be repaired before further use.
You saved it from going in a dumpster that's all.

That is a perspective that has changed since the time that this meter was originally built.  Damaged but repairable is now no different to most than destroyed and unrepairable, so there is no point in designing things to fail in a repairable way.  It wasn't always that way.

[tautech]

No it didn't survive, it needed to be repaired before further use.
You saved it from going in a dumpster that's all.

That is a perspective that has changed since the time that this meter was originally built.  Damaged but repairable is now no different to most than destroyed and unrepairable, so there is no point in designing things to fail in a repairable way.  It wasn't always that way.

You repaired it to a condition that works for you, congrats and I mean that however for most that rely on a bench meter for trusted best possible accuracy they would bin it.
Obtaining the equipment to verify its performance will cost more than replacing it.

[ferdieCX]

That is a perspective that has changed since the time that this meter was originally built.  Damaged but repairable is now no different to most than destroyed and unrepairable, so there is no point in designing things to fail in a repairable way.  It wasn't always that way.

Not for me. For this reason I didn't bought a benchtop DMM and got a Fluke 87V instead. It is cheaper, more robust, and 4 1/2 digits are enough for me

[bdunham7]

You repaired it to a condition that works for you, congrats and I mean that however for most that rely on a bench meter for trusted best possible accuracy they would bin it.
Obtaining the equipment to verify its performance will cost more than replacing it.

Calibrating it would be part of a proper repair and I will certainly do a complete check at some point.  I understand that the 'average user' doesn't have the means to check the performance, but my point is that in its day, this product would be sent in for service and calibration, not binned.  New stuff, well that does get binned.  But as far as trusting it, at least on the DCV ranges that I've checked so far, it is still more accurate toasted than any product your company sells brand new, so there's that. 

[ResistorRob]

I just have a hobby level bench meter, but it's very accurate, looks great, and and has everything I need at a great price.
However, it's missing just one feature that I wish it had....

- Data Logging!

If it had that I might cross the Keithley off my lust/dream meter wishlist :-)
(ok, probably not, lol)

[joeqsmith]

I just took apart an old 8842A that apparently was connected to excessive voltage--at least 1500V---with some power to back it up.  The entire front end protection--4 MOVs and 2 fusible resistors--were incinerated.  I removed the burnt parts and subbed in 2 regular 1K resistors (to be repaired properly later, of course) and the meter works properly and is in cal.  No evidence of any external damage, the circuit board is unharmed (but sooty)  and I'm sure nobody died.  I'd say that is 'robust' although I can't quantify that.

Nowhere in CAT documents does it say the the meter should survive.

Again, 61010-2-033 section 14.101 states:  "The component shall not rupture and shall operate as intended during the test."   

IMO, operate as intended means just that, the DMM continues to read properly. 

One of the tests I have shown several times is from section 101.3.2.   The criteria, "During and after the test, no damage to the equipment shall occur."  IMO, that's not the test equipment being used to test the DMM they are referring to, it's the DMM.

"After the voltage of 4.4.2.101 has been applied to the METER, the METER shall continue to be
able to indicate the presence of HAZARDOUS LIVE voltages up to the maximum RATED voltage.
NOTE The METER is not required to maintain its normal accuracy. A maximum deviation of 10 % is acceptable."

Of course there are many other cases as I have mentioned before.  But again, just to be clear, this is outside of my area and I am providing my interpretation of what the standard states.    I would find it hard to believe for example that the pass fail criteria for an impulse test would be if the combo generator was damaged or not.    Fungus may see it otherwise.   

That said, I would imagine that some of the meters I have looked at would pass both the safety and EMC standards with a wide margin but there are far more that wouldn't come close.   

[zitoune]

I have the Keysight 34461A, and for me the booting time is a disadvantage compared to any handheld DMM.

You cannot do a quick measure on the spot, so most of the time I keep it running for long periods without necessarily using it.

[joeqsmith]

Nowhere in CAT documents does it say the the meter should survive.

The 8842A isn't actually 'CAT rated' as far as I know, at least neither CAT xx/yyyV or any reference to IEC1010 appears anywhere on it.  I wonder when that started to be a thing--my earliest product that I can lay hands on that has a CAT rating insignia is my old Scopemeter--CAT III/600V.

However, I don't think this was any type of CAT event.  I think it was subjected to a severe overvoltage, not a transient.   I think it is open for discussion as to whether the meter 'survived'.  IMO it did survive--just a little more repair needed than replacing a fuse.

No it didn't survive, it needed to be repaired before further use.
You saved it from going in a dumpster that's all.

We may think of a loved one who lived through all the treatments of cancer as surviving.  However, I don't see it this way when it comes to test equipment.  If it was exposed to an event that required repairs, it didn't survive.  Rather I would say it was repairable or salvaged. 

[wizard69]

I just took apart an old 8842A that apparently was connected to excessive voltage--at least 1500V---with some power to back it up.  The entire front end protection--4 MOVs and 2 fusible resistors--were incinerated.  I removed the burnt parts and subbed in 2 regular 1K resistors (to be repaired properly later, of course) and the meter works properly and is in cal.  No evidence of any external damage, the circuit board is unharmed (but sooty)  and I'm sure nobody died.  I'd say that is 'robust' although I can't quantify that.

Nowhere in CAT documents does it say the the meter should survive.

CAT is about keeping the user safe.   The meter may not be functional but it shouldn't harm the user when an accident happens.

[bdunham7]

If it was exposed to an event that required repairs, it didn't survive.  Rather I would say it was repairable or salvaged.

OK, fair enough.  I did say it was a debatable point.  Would you say the same for a meter that required the replacement of an internal fuse after exposure to excess current?  If not, then why would you have a different opinion when the items that 'failed' (actually they worked) are protective devices specifically intended for this purpose?  Of course, I'll grant that the MOVs might not fall into the same category as the fusible resistors, incineration isn't their normal mode of operation. 

My gripe here isn't about the definition of 'survived', it is that the aversion to repairability is causing people to completely shun 'repairable' failure modes.  Why have MOV protection when they could fail?  Just let the CPU take the hit--after all dead is dead.  I suppose that's true for cheap consumer products and things we send into space, but I don't think it should be for expensive instruments--at least where it is practical to build them otherwise.

[tautech]

You repaired it to a condition that works for you, congrats and I mean that however for most that rely on a bench meter for trusted best possible accuracy they would bin it.
Obtaining the equipment to verify its performance will cost more than replacing it.

Calibrating it would be part of a proper repair and I will certainly do a complete check at some point.  I understand that the 'average user' doesn't have the means to check the performance, but my point is that in its day, this product would be sent in for service and calibration, not binned.  New stuff, well that does get binned.  But as far as trusting it, at least on the DCV ranges that I've checked so far, it is still more accurate toasted than any product your company sells brand new, so there's that. 

Yes well I could tell you a horror story of a mate that recently bought a spanking new DMM6500 but I shouldn't. 

Let's just say it got sent back to the US to get sorted and arrived back here in NZ still not right. 

[joeqsmith]

If it was exposed to an event that required repairs, it didn't survive.  Rather I would say it was repairable or salvaged.

OK, fair enough.  I did say it was a debatable point.  Would you say the same for a meter that required the replacement of an internal fuse after exposure to excess current? If not, then why would you have a different opinion when the items that 'failed' (actually they worked) are protective devices specifically intended for this purpose?  Of course, I'll grant that the MOVs might not fall into the same category as the fusible resistors, incineration isn't their normal mode of operation. 

My gripe here isn't about the definition of 'survived', it is that the aversion to repairability is causing people to completely shun 'repairable' failure modes.  Why have MOV protection when they could fail?  Just let the CPU take the hit--after all dead is dead.  I suppose that's true for cheap consumer products and things we send into space, but I don't think it should be for expensive instruments--at least where it is practical to build them otherwise.

In the case of the fuse, we are talking about a user serviceable component.   Of course, you may consider all component user serviceable but that's not what I am referring to.  The manuals for a meter will commonly detail how to test and replace the fuses, along with providing details about what to replace them with.   In some cases, meters like my HP34401 and Fluke 8000A, the fuse (for the current measurement) can be accessed without opening the case.   The fuse was designed to prevent damage to the user (say an arc flash) as well as protect the meter.  Other parts, like MOVs, surge rated resistors.... are soldered in place.  They are not called out in the manual as something the user should be servicing.   If the only damage was the fuse, then I would say the fuse blew but the meter survived.   Say some idiot uses the wrong fuse or even better, jumps the fuse they blew and then the meter is exposed to something that damages other components, I would say you not only damaged your meter but you were an idiot for not following the manufactures requirements. 

Once the parts are soldered into the board, they become part of a larger assembly and require special skills to service.  Of course this doesn't prevent some people from attempting to do repairs they may not be qualified to do.   Again, as we have seen in many posts, there are people who are not qualified to change a fuse.   As to your comment about repairability,  I imagine in cases where we have unqualified people working on safety equipment, we run into cases where they create an unsafe condition (say the person jumping the HRC fuse).  If something were to happen, I would assume the manufacture will be named on the suit.   We may see a time when fuses are no longer user serviceable. 

As I said, the MOVs, PTC, GDTs, surge rated resistors, high speed clamps that you find are there to try and make the meter more robust so it survives various events the user exposes the DMM to.    If we exceed the limits of what any of these combined parts can handle, we have damaged the meter.    If we can salvage or repair it, comes down to the cost/time.  If these parts save the downstream ICs, then I will typically repair them.  If not, they are normally scrapped.

[Rasterist]

Keysight 34461A:  takes too long to boot up.  Especially annoying since it's never completely off; there's an LED on the front indicating it's only asleep.  Nevertheless, it's 29 seconds from hitting the power button until being able to make a measurement.  For this reason, I always reach for my Fluke 87V unless I need the precision or some other unique feature of the 34461A.

[floobydust]

I frequently go near or over 750-1000V,

You might want to rethink routinely running near-limit voltage through your very best DMM (if that's the case)...

I understand benchtop DMM's are fragile. Cat. II 300V is 2,500V/12R impulse. Will a BBQ lighter kill it? These kinds of small transients (not mains spikes) are commonplace to a multimeter when doing SMPS work for example. You get a fast dV/dt and it's a little hit. I don't want to find out the hard way the benchtop DMM is too fragile for practical use.
I've also made test benches for products using benchtop DMM's and a mux for bed-of-nails automated testing. Sometimes there is a failure (connection, pogo-pin, software etc.) and the DMM gets overloaded. I think it's a reasonable use-case. I see a couple MOV's, couple gas-tubes but no PTC in common benchtop DMM pics. Their relays are like 1,000V or 1,500V isolation. They have weaker front-end protection and questionable transient testing 61010 certificates.

"Keithley products are designed for use with electrical signals that are measurement, control, and data I/O connections, with low transient overvoltages, and must not be directly connected to mains voltage or to voltage sources with high transient overvoltages. Measurement Category II (as referenced in IEC 60664) connections require protection for high transient overvoltages often associated with local AC mains connections. Certain Keithley measuring instruments may be connected to mains. These instruments will be marked as category II or higher."

[0culus]

If it was exposed to an event that required repairs, it didn't survive.  Rather I would say it was repairable or salvaged.

OK, fair enough.  I did say it was a debatable point.  Would you say the same for a meter that required the replacement of an internal fuse after exposure to excess current?  If not, then why would you have a different opinion when the items that 'failed' (actually they worked) are protective devices specifically intended for this purpose?  Of course, I'll grant that the MOVs might not fall into the same category as the fusible resistors, incineration isn't their normal mode of operation. 

My gripe here isn't about the definition of 'survived', it is that the aversion to repairability is causing people to completely shun 'repairable' failure modes.  Why have MOV protection when they could fail?  Just let the CPU take the hit--after all dead is dead.  I suppose that's true for cheap consumer products and things we send into space, but I don't think it should be for expensive instruments--at least where it is practical to build them otherwise.

Lack of repairability is largely pushed by the vendors, and it really started in the 1980s and has only gotten worse. Why? Well, now when you break it out of warranty they can sell you a new one. They all like to maintain their trade secret designs too. 50 years ago this was not the case. Crack open the manual for a Tektronix 500 series oscilloscope...it is plainly obvious that they intended the end user to service their own instrument. They even gave you a roll of silver bearing solder! Then go download a manual for the 24xx series scopes and it's obvious how things had changed even by 1990 or so. 

That doesn't mean new stuff is unrepairable; the company just won't help you out at all. You have to reverse engineer it and understand how it works and the failure mode, and maybe have a few special tools. A perfect example of this is the infamous "red ring of death" failure for Microsoft's Xbox 360 game console. Many of them got thrown away, but it's really a failure of the BGA under the CPU, and it can be easily fixed if you have the capability to reflow. Also "The Signal Path" on youtube; his expert knowledge allows him to repair test gear that would otherwise get tossed.

Personally, I don't do very much work around truly high energy stuff (beyond single phase wall voltage), but if I did I certainly would want NEW equipment and not repaired equipment. It's not worth gambling your life when arc flash hazards are involved. For everything else I play with on my bench, repaired old equipment all the way.

[joeqsmith]

I understand benchtop DMM's are fragile. Cat. II 300V is 2,500V/12R impulse. Will a BBQ lighter kill it? These kinds of small transients (not mains spikes) are commonplace to a multimeter when doing SMPS work for example. You get a fast dV/dt and it's a little hit. I don't want to find out the hard way the benchtop DMM is too fragile for practical use.

I never thought I would see the small lighter I use damage any meter but here we are.  Countless UNI-Ts dead and even the holy grail, the 121 was damaged to where it threw off the measurements a fair amount.   It's hard to believe as that ignitor is so weak.   

While I would assume most people would take steps to reduce ESD in their labs, its the other transients that I am also concerned with.  When I was young, I built a power supply like many beginners.   I then proceeded to hook it to all sorts of things.  One was an electromagnet.  I wasn't educated to know what BEMF was.  40 some odd years later, I have MOVs and TVSs internal to all my supplies.     They still are not bullet proof.

Imagine the beginner wanting to measure the resistance change of their large electromagnet.  They attach their meter across the coil, measure the resistance, then power up the coil with their 9V battery, let it heat and then remove the power by disconnecting the leads.  The beginner who swears they never play with HVs has just damaged their new meter.      Most of us do dumb things like this at one time or another.  It's how we learn.   

IMO, cheap meters have their place.  But a beginner who buys a meter that can't survive even that stupid grill starter and other basic transients may find themselves buying another meter.   Steering beginners to buy robust meters that may survive all the thing they expose them to, may make learning electronics more enjoyable but stressing about the meter you just blew up because of some stupid mistake is part of it.   I can't see robbing a beginner of that experience..     

[SilverSolder]

3478A:

1. Horridly bad LCD display
2. No continuity / feeper
3. Terrible front panel legend meaning I always stick my probes in the wrong holes.

That’s about it.

A DYMO labeler could be your new best friend! 

[bdunham7]

I understand benchtop DMM's are fragile. Cat. II 300V is 2,500V/12R impulse. Will a BBQ lighter kill it?

They aren't that fragile, at least properly made ones.  Your CAT II/300 test requires clamping 100+ amps, the BBQ lighter is just a bit of ESD.  The two are not similar events in any way.

"Keithley products are designed for use with electrical signals that are measurement, control, and data I/O connections, with low transient overvoltages, and must not be directly connected to mains voltage or to voltage sources with high transient overvoltages. Measurement Category II (as referenced in IEC 60664) connections require protection for high transient overvoltages often associated with local AC mains connections. Certain Keithley measuring instruments may be connected to mains. These instruments will be marked as category II or higher."

That's a lot of words to simply say you need a CAT II rating to measure non-isolated mains-derived sources.  Keithley DMMs will all meet at least CAT II AFAIK, but there are scanner cards and so on that do not.

[mansaxel]

Obtaining the equipment to verify its performance will cost more than replacing it.

Makes perfect sense. I'd totally do that. Because I would learn something! And, GAS.

[bson]

I wouldn't pay extra for any cat ratings whatsoever, for a bench instrument whose purpose is things like measuring the linearity of a high resolution DAC or ADC.

A dog rating however, that I'd pay for!

[floobydust]

They aren't that fragile, at least properly made ones.  Your CAT II/300 test requires clamping 100+ amps, the BBQ lighter is just a bit of ESD.  The two are not similar events in any way.

It's that I want hard proof this very expensive DMM is impulse-tested for differential and common-mode transients to some value say at least 1,500V. Not someone's "opinion", not "designed to" (engineers with good intentions) or "meets" (who says?) or "CE" sticker or "spec'd to" etc.

An ongoing problem with multimeters is fake or misleading 61010 approvals.
"Cat. II" on the front panel means nothing - was it evaluated to 61010-1, 61010-2-033 ? And even then, a certification agency can make mistakes. Remember the Keysight 600V fiasco it was just terrible.

For a bench DMM it's not safety approvals so much as a guarantee it can take a BBQ lighter so I'm not looking at a massive repair cost due to normal use.
i.e. DMM6500 has Intertek approval ETL marking for 61010 but Keithley 2000, 2001, 2002 has nothing- which is probably why they frequently get damaged.

[joeqsmith]

They aren't that fragile, at least properly made ones.  Your CAT II/300 test requires clamping 100+ amps, the BBQ lighter is just a bit of ESD.  The two are not similar events in any way.

It's that I want hard proof this very expensive DMM is impulse-tested for differential and common-mode transients to some value say at least 1,500V. Not someone's "opinion", not "designed to" (engineers with good intentions) or "meets" (who says?) or "CE" sticker or "spec'd to" etc.

An ongoing problem with multimeters is fake or misleading 61010 approvals.
"Cat. II" on the front panel means nothing - was it evaluated to 61010-1, 61010-2-033 ? And even then, a certification agency can make mistakes. Remember the Keysight 600V fiasco it was just terrible.

For a bench DMM it's not safety approvals so much as a guarantee it can take a BBQ lighter so I'm not looking at a massive repair cost due to normal use.
i.e. DMM6500 has Intertek approval ETL marking for 61010 but Keithley 2000, 2001, 2002 has nothing- which is probably why they frequently get damaged.

61010-1 is "for electrical equipment for measurement, control, and laboratory use –"

-2-033 is for "Particular requirements for hand-held multimeters and other hand-held meters, for domestic and professional use, capable of measuring mains voltage"

"It's that I want hard proof ..."
I doubt you would ever find it.    BBQ starters will very a LOT.   I have characterized the one I use, presented the data for it and have shown how it compares with IEC waveform.   I would expect completely different results if I went to the hardware store and bought a new one.   BBQ starters are going to output several KV, far beyond the 1.5KV you mention.  Their rise times can be well under 1ns and pulse width is going to be <100ns.    I think the one I used can put out close to 5A peak, far from what the standards call for.   

If say you pick some magic number (1.5K) as your criteria.  Let's not focus on how you came up with it but assume you know this is what you require.   What does the waveform look like?  What's the voltage rise time into and open circuit?  What's it's decay look like?  What's the current look like into a short? ....     These are the questions I was pondering when trying to come up with a way to benchmark handheld meter.   Being unable to locate any sort of existing standards, my proposal was to somewhat follow the surge open circuit voltage waveform, greatly limit the energy available and not superimpose it onto the line.   

If I applied this type of waveform at 8KV peak to your favorite bench meter and it survived, it may give you a great feeling of comfort but it will not mean you can connect the same meter to 2KVDC or a MOT.    Hard proof requires hard requirements.  From there it's possible you could have these tests performed at some outside lab.  The cost  may far exceed the cost of the meter but you would have your hard proof.   

A beginner may suggest if a meter survives +/-1.5KV DC, it would handle every transient up to that.  At one time I made an attempt to provide some insight on how to harden a cheap UNI-T UT-61E and show how a fast edge can be a problem.  Sadly, most of the people who watched the video didn't seem to care about the approach to solving this sort of problem.   The wiki experts just need to know how to mod their cheap meters. 

I just finished the chapter on colleges.   An excellent book.
https://www.amazon.com/Death-Expertise-Campaign-Established-Knowledge/dp/0190469412

[Doctorandus_P]

I've thought long and hard about buying a benctop DMM, and decided not to.
The deciding factor was that the boxes are too big. They are some 25 cm deep.
If they had the same depth as a digital scope I very likely would have bought one.

I had a look at the Siglent ...3055. and though about sawing the box in half, maybe relocate the transformer.
The box is almost empty, with just the front panel PCB and a PCB at the bottom which is about half the size of the box. But Also decided against that. I'm not going to spend EUR500 on a DMM to cut he box in two.

Maybe, some day I'll design my own custom benchtop DMM.

[Fungus]

I've thought long and hard about buying a benctop DMM, and decided not to.
The deciding factor was that the boxes are too big. They are some 25 cm deep.
If they had the same depth as a digital scope I very likely would have bought one.

I had a look at the Siglent ...3055. and though about sawing the box in half, maybe relocate the transformer.
The box is almost empty, with just the front panel PCB and a PCB at the bottom which is about half the size of the box. But Also decided against that. I'm not going to spend EUR500 on a DMM to cut he box in two.

https://www.aliexpress.com/item/4000343031770.html

[bdunham7]

"It's that I want hard proof ..."
I doubt you would ever find it.    BBQ starters will very a LOT.   I have characterized the one I use, presented the data for it and have shown how it compares with IEC waveform.   I would expect completely different results if I went to the hardware store and bought a new one.   BBQ starters are going to output several KV, far beyond the 1.5KV you mention.  Their rise times can be well under 1ns and pulse width is going to be <100ns.    I think the one I used can put out close to 5A peak, far from what the standards call for.   

A beginner may suggest if a meter survives +/-1.5KV DC, it would handle every transient up to that.  At one time I made an attempt to provide some insight on how to harden a cheap UNI-T UT-61E and show how a fast edge can be a problem.  Sadly, most of the people who watched the video didn't seem to care about the approach to solving this sort of problem.   The wiki experts just need to know how to mod their cheap meters. 

Meters (well made and well specified) typically have a dV/dt limitation of a few million V*Hz, so your BBQ lighter will greatly exceed that, of course. However, I'd consider it poor design if that limitation was anything other than a thermal limitation on whatever absorbs that high frequency energy. 

I'll drag out my old fried 8842A again, here's a picture of the damaged unit with the incinerated parts removed.  Do you see anything in the picture that might help with that fast edge?  Do you think there's room in a hand-held DMM for this type of protection?  Is there something else that can do the same thing in less space?

[bdunham7]

https://www.aliexpress.com/item/4000343031770.html

Did you look at the comparison photos where they are demonstrating it with the probes into a power strip and then connected to a 9V battery.  Perfect for the target market!

[tautech]

I've thought long and hard about buying a benctop DMM, and decided not to.
The deciding factor was that the boxes are too big. They are some 25 cm deep.
If they had the same depth as a digital scope I very likely would have bought one.

I had a look at the Siglent ...3055. and though about sawing the box in half, maybe relocate the transformer.
The box is almost empty, with just the front panel PCB and a PCB at the bottom which is about half the size of the box. But Also decided against that. I'm not going to spend EUR500 on a DMM to cut he box in two.

Maybe, some day I'll design my own custom benchtop DMM.

The SDM3065X is deeper again !
They can't be too compact without compromising safety clearances or the size for TH components used to withstand the voltages they are rated for.
As for being almost empty cases, I can assure you when the 16ch scanner card is installed there isn't much free space left.
Instruments like SDM's are made to be stacked so while you desire something smaller this also compromises stability of anything stacked on top of it.

[Fungus]

https://www.aliexpress.com/item/4000343031770.html

Did you look at the comparison photos where they are demonstrating it with the probes into a power strip and then connected to a 9V battery.  Perfect for the target market!

I don't own one myself but all the threads on it have concluded that it's quite a good meter if you can get past the silly photos and alarm clock part.

Plus you can take it with you on a picnic and listen to music.

[floobydust]

Most models have a 500Vpk CM voltage rating and are using slow gas-tubes for protection of that.
If you have a 750VDC source that is earth-grounded, what happens when you first connect one DMM probe? You are trying to fast lift the entire in-guard section (capacitance) to 750VDC.
Compare that with nicely connecting both leads to the (off) HV power supply under test, where the DMM (-) gets 0V potential and upon PSU power-on there is no overload beyond the auto-ranging having gone to the mV range and having to clank relays up-ranging.
A common-mode choke at the input does help against a fast transient - but they don't use those anymore. Very rare to see one, it's extra manual labour. Nowadays, it's tiny smt inductors with low breakdown voltage, 10 miles after the input jacks, pic related.
My position is these multimeters need to have been well tested, that they really withstand ESD and fast transients at their rated voltage because repairs are super expensive.

[bdunham7]

"Cat. II" on the front panel means nothing - was it evaluated to 61010-1, 61010-2-033 ? And even then, a certification agency can make mistakes. Remember the Keysight 600V fiasco it was just terrible.

"Nothing" is a bit harsh.  Sure, certifications aren't infallible, but what else can you do unless you do your own test?  The Keysight deal was indeed a shitshow and I'm not sure what I would do if I had one of the implicated devices.  A bit more transparency would be reassuring.

Most models have a 500Vpk CM voltage rating and are using slow gas-tubes for protection of that.

Which ones?  I'm not seeing that on mine except on a very old 8505A, which is a reference meter not really intended for bench service use--it has the 500V CM limitation.  Everything else I have is at least 1000VDC/700VAC to ground, and the 8846A has to be at least 1500V common and differential because it allows 1000Vrms inputs--and doesn't go into overload even slightly over that.  I haven't disassembled it to see what it looks like.

My position is these multimeters need to have been well tested, that they really withstand ESD and fast transients at their rated voltage because repairs are super expensive.

I'd prefer repairability, but that ship sailed long ago.

[tautech]

I'd prefer repairability, but that ship sailed long ago.

Hardly any surprise why in this day and age.

[HKJ]

Bench meter voltage on the front:

Keysight 34461A/65A/70A: CAT II 300V, 500Vpk to gnd
Keithley 6500, 7510: CAT II 300V, 500Vpk to gnd
Fluke 8846: CAT II 600V, CAT I 1000V
Rigol 3068: CAT II 300V, CAT I 1000V
Siglent 3045X: CAT II 300V, CAT I 1000V, 500Vpk to gnd
East Tester ET3240: CAT II 300V, 500Vpk to gnd

The "to gnd" is from the black terminal.

The CAT II 300 is nearly universal, only Fluke will handle more.

[bdunham7]

The CAT II 300 is nearly universal, only Fluke will handle more.

OK, didn't realize that those meters all had specific 500V-pk to ground ratings as I don't have any of them.  But CAT II/300 doesn't imply that--my modest BK Precision 2831E (by TongHui) is CAT II/300 and yet 1kV max to ground.  Of course it is also CAT I/1000V, so I suppose those meters that omit the CAT I/1000 rating really are fragile in that you can't measure 500VDC or 300VAC without being sure of how it might be ground referenced--so even floating or presumed floating above those limits would be off-limits.  That does seem weak tea.

And then there's the Siglent.  Not sure how you can be CAT I/1000 and still have the 500V-pk to ground limitation.  Perhaps someone with knowledge of the CAT standards could clarify?

[tautech]

And then there's the Siglent.  Not sure how you can be CAT I/1000 and still have the 500V-pk to ground limitation.  Perhaps someone with knowledge of the CAT standards could clarify?

Specifically this is how it's rated:
Measuring Method and other Characteristics
DC Voltage
CMRR 120 dB ( For the 1 KΩ unbalanced resistance in LO lead, max ± 500 VDC )

[bdunham7]

Specifically this is how it's rated:
Measuring Method and other Characteristics
DC Voltage
CMRR 120 dB ( For the 1 KΩ unbalanced resistance in LO lead, max ± 500 VDC )

I'm not clear on exactly what that means, but does that mean that there is some other non-protection related reason that the 500V-pk limitation is posted?  I suppose you could design a meter that passes CAT x, but then you otherwise greatly limit its maximums for other reasons.  What confuses me is that although I can't find a specific cite, the CAT voltages are V-rms, so 1000VAC => 1.4kV-pk.  Unless the  IEC procedures allow you to specify the the black lead must be ground referenced--no floating or split supplies-- the math doesn't work.

[SilverSolder]

The 3 biggest disdvantages of a benchtop DMM are Size, Size, and Size.

But, that is also their advantage...  Large displays that can be read from across the room in some cases, plenty of knobs and buttons brought out on the front panel for convenience, many options for connectors front and rear, including remote control...   Room inside for better / more stable electronics with more features.  They are heavy enough not to slide around on the table in use...

[tautech]

Specifically this is how it's rated:
Measuring Method and other Characteristics
DC Voltage
CMRR 120 dB ( For the 1 KΩ unbalanced resistance in LO lead, max ± 500 VDC )

I'm not clear on exactly what that means, but does that mean that there is some other non-protection related reason that the 500V-pk limitation is posted?  I suppose you could design a meter that passes CAT x, but then you otherwise greatly limit its maximums for other reasons.  What confuses me is that although I can't find a specific cite, the CAT voltages are V-rms, so 1000VAC => 1.4kV-pk.  Unless the  IEC procedures allow you to specify the the black lead must be ground referenced--no floating or split supplies, the math doesn't work.

I take it to mean the LO (negative) lead has a max +500VDC rating WRT mains ground.
Which implies for HV or floating measurements with a 500+VDC offset relative to mains ground connect the positive (Hi) lead first.
CAT ratings will be in addition to this.

I suspect it's due to internal clearances primarily as a question I sent to Siglent's tech support recently about series ganging their PSU's had the same 500V limitation in that you can series several units together as long as the 500V max was observed.....which is also specified in their PSU datasheets.

[HKJ]

The 3 biggest disdvantages of a benchtop DMM are Size, Size, and Size.

That depends, if you also have power supplies, arbs, electronic loads, etc. they all stack nicely.

I have 15 devices stacked in a couple of stacks, I would never have had space for them if they had each had their own shape.

[Electro Fan]

It's only a matter of time until either the budget or the space is maxed out.

[Electro Fan]

The 3 biggest disdvantages of a benchtop DMM are Size, Size, and Size.

But, that is also their advantage...  Large displays that can be read from across the room in some cases, plenty of knobs and buttons brought out on the front panel for convenience, many options for connectors front and rear, including remote control...   Room inside for better / more stable electronics with more features.  They are heavy enough not to slide around on the table in use...

+1 on advantages...

After years with only handheld DMMs the bench DMM has become the go to DMM.  It takes up more space than the handhelds but it wasn't space I was using, and I never have to look to see where it might be; and it holds still (hands-free) - just grab the leads and measure (although you still have to change out leads based on the connectors needed). 

And without a doubt a large and multifaceted (graphical) display is way better than the relatively small and dim handheld DMM.  Both have their uses - probably should have at least one of each, but unless you have no handheld DMMs it's worth seriously considering a bench DMM.  If you already have a handheld DMM, I don't see any disadvantages in a bench DMM.

[joeqsmith]

"It's that I want hard proof ..."
....

Meters (well made and well specified) typically have a dV/dt limitation of a few million V*Hz, so your BBQ lighter will greatly exceed that, of course. However, I'd consider it poor design if that limitation was anything other than a thermal limitation on whatever absorbs that high frequency energy. 

I'll drag out my old fried 8842A again, here's a picture of the damaged unit with the incinerated parts removed. 

Do you see anything in the picture that might help with that fast edge? 

It's hard to say when it comes to something like ESD.  The edges are very fast and we are dealing with RF.   I addressed the weakness of the UNI-T UT181A by changing the layout.  It should be obvious that even wire will have some inductance and will effect the signal at these speeds.  Surely the large inductor on the front will play into it.   

Do you think there's room in a hand-held DMM for this type of protection?

Based on the number of hand-held meters I exposed to not only the BBQ lighter but also my home made gun which produces waveforms more in-line with the IEC standards,  I would say there are certainly ways to design the front ends to survive it. 

Is there something else that can do the same thing in less space?

Again, I went into detail the problems with the UT61E and one possible way to solve such a problem.

Most models have a 500Vpk CM voltage rating and are using slow gas-tubes for protection of that.
If you have a 750VDC source that is earth-grounded, what happens when you first connect one DMM probe? You are trying to fast lift the entire in-guard section (capacitance) to 750VDC.
Compare that with nicely connecting both leads to the (off) HV power supply under test, where the DMM (-) gets 0V potential and upon PSU power-on there is no overload beyond the auto-ranging having gone to the mV range and having to clank relays up-ranging.
A common-mode choke at the input does help against a fast transient - but they don't use those anymore. Very rare to see one, it's extra manual labour. Nowadays, it's tiny smt inductors with low breakdown voltage, 10 miles after the input jacks, pic related.
My position is these multimeters need to have been well tested, that they really withstand ESD and fast transients at their rated voltage because repairs are super expensive.

We have seen companies like Gossen successfully utilize GDTs in their meter.   I wouldn't suggest that a product that uses them would be any less robust.  The low capacitance and leakage would seem to make them a good fit for higher end products. 

I think my old HP 34401A uses a combination of GDTs, R,C and MOV for the clamp.  The nice thing about this technique over using the GDT alone is that when working with DC, once the transient dissipates and the voltage falls below the MOVs trip point, the circuit resets.   With the tests I run it doesn't really matter as the transients are never superimposed on another waveform.  They always start at 0V and and at  0V, allowing a GDT to clear.     


***
For got to say that your point about the power supply being active and clipping onto it is certainly a valid case and I wouldn't be surprised to hear of meters being damaged this way.   I look at my bench meters pretty much how I look at any other sensitive equipment.  I have learned to treat them with caution.    Now these new handhelds I have started using, not so much.  These things have seen some fair abuse in my hands.

[JimKnopf]

In spring I bought a UT 8803e. I was very unhappy with it because it measures very, very inaccurately. In diode mode you can only measure down to 16 Ohm. Below that Err (short circuit) is displayed.

Today my Keithley DMM6500 arrived.
The first tests are very positive. Diode Mode I can measure very accurately down to below 1 Ohm. Continuity test is very fast. The only drawback is that capacitance is only up to 100µF or 20% above (all measurements 20% above max range), i.e. 120µF.

[TheUnnamedNewbie]

I realize that this question is probably asked from a 'disadvantages compared to other bench/handheld DMMs' and my answers don't reflect that, but for me one of the big disadvantages are:

1) Display is attached to the meter, and sometimes I want to be able to look at the meter value and the probes at the same time (though I solved this with a SCPI script on my laptop)
2) To my knowledge, pretty much every bench DMM is HIGH-Z (usually >1 GOhm) input impedance only on the 10V range and below. If you want high impedances at higher voltages you need either electrometers, or SMUs.
3) While bench DMMs can often measure quite small voltages, measuring small currents is hard. A TIA-based front-end such as some electrometers or pico/femto-amp meters would be awesome (and is still one of my projects but haven't finished it yet)

[bingo600]

1: A bit pricey
2: Have to replace Ram's every 10 years
3: Calibrating is $$$

Model # ends w. 58A

/bingo

[HighVoltage]

Biggest disadvantage ... you like them so much, you have to get more, many more

This was 4 years ago ...
Some more have been added to the benches since then.
So, don't start with a benchtop DMM, they multiply.

[Keysight DanielBogdanoff]

Biggest disadvantage ... you like them so much, you have to get more, many more

This was 4 years ago ...
Some more have been added to the benches since then.
So, don't start with a benchtop DMM, they multiply.

Dang that's a lot of DMMs...

[Joel_l]

DMM6500

1 - A little slow too start up, don't know how it compares to others
2 - Can't do higher current measurements from front jacks, have to use rear which is not convenient.
3 - No continuity beep

[rvalente]

Biggest disadvantage ... you like them so much, you have to get more, many more

This was 4 years ago ...
Some more have been added to the benches since then.
So, don't start with a benchtop DMM, they multiply.

Thats when the need has been superseded and becomes a collection

[maginnovision]

It happens.

[Ground_Loop]

I have a stack of HP meters 3455, 3456, and 3457. The first two don't have current measurement. But they do have math functions and I have a bench mounted current shunt. The 57 has that awful LCD display, but at least has fresh lab cal. The 55 has a fan. None have diode drop measurement. And worst of all they never all read exactly the same. Kind of like having more than one watch: you never really know what time it is. My handhelds are Fluke 289 and 787. Neither sees much bench use.

[JimKnopf]

DMM6500

1 - A little slow too start up, don't know how it compares to others
2 - Can't do higher current measurements from front jacks, have to use rear which is not convenient.
3 - No continuity beep

You don't have "Cont" in the bottom middle of the menu? It beeps.

[Fungus]

So, don't start with a benchtop DMM, they multiply.

You're saying that handheld meters don't??

[HighVoltage]

So, don't start with a benchtop DMM, they multiply.

You're saying that handheld meters don't??

Oh, they do too!
But they are easier to hide in a drawer under the bench.

A representative of Keysight Germany visited my lab last year and he was slightly exited to see so many of their DMMs in action.

 

[rvalente]

I've a 8010 in my bench, but always use the 87V. I believe I have not yet been infected with the bench multimeter plague.

[Kleinstein]

The boot time is not so much an issue for precision meters - they need some 20-100 minutes to warm up for a precision reading anyway. Some of the boot time could also be some extra self tests - battery powered meters tend to spend less time on these. I don't think boot time makes a lot of difference between 5 digits and 8 digits, its usually the graphics they may need a little more boot time. It may take longer warm up to get 8 digit stability of cause, but 5 digit precision should be reached fast.

A CAT IV protection is in conflict with very low leakage and very low noise (where series resistance becomes an issue). Also the bench is usually CAT 2 or less. The screw terminals to directly connect bare wires are also conflicting with higher CAT ratings.

Still better protection (less chance to damaging) is always welcome if it does not conflict with the main task.

[rvalente]

It is just me or you also find the VFD displays sexier than this TFT thing?

[SilverSolder]

I'm happy with only one 34461A and one PXIe-4081 on my bench.

Now to the questions: they don't like high surge transients that are perfectly absorbed by a proper CAT IV handheld DMM. Also they tend to boot slower. Yes, I have OCD and I can't withstand keeping my gears powered on for entire days. That's why I didn't bother with 8.5 digits -- they need to warm up for longer than my OCD can tolerate.

Just put them on a timer, so they start up 3h before you arrive! 

[AVGresponding]

It is just me or you also find the VFD displays sexier than this TFT thing?

Nope, it's not just you!

[ferdieCX]

It is just me or you also find the VFD displays sexier than this TFT thing?

They age and are unobtanium. I prefer plain 7 segment LED displays

[HKJ]

VFD may look nice, but their lifespan is fairly short (I have two 8846A that is less than 10 years old, but the VFD are rather aged).
I hope that graphic color LCD will last much longer, I have a lot of meters with them.
LED are very reliable and easy to replace if the die, but they are fairly limited. Generally I like them, but a DMM can do a lot more with a graphic color LCD (See new Keithley or Keysight).

[Shock]

Are those diy GW Instek kelvin clips with the screws on them or a different clip with fixed cables? Am curious if the connections to the clips are low emf, serviceable and what cabling/shielding was used.

[SilverSolder]

VFD may look nice, but their lifespan is fairly short (I have two 8846A that is less than 10 years old, but the VFD are rather aged).
I hope that graphic color LCD will last much longer, I have a lot of meters with them.
LCD are very reliable and easy to replace if the die, but they are fairly limited. Generally I like them, but a DMM can do a lot more with a graphic color LCD (See new Keithley or Keysight).

I like connecting the DMM (or any other instrument) to a PC, with a gigantic monitor, if I need anything more than basic information.  So I've never cared that much about graphic displays (except on oscilloscopes obviously! Even there, piping the output to a gigantic monitor suitably suspended over the bench is better).

[tautech]

Are those diy GW Instek kelvin clips with the screws on them or a different clip with fixed cables? Am curious if the connections to the clips are low emf, serviceable and what cabling/shielding was used.

Shock, these are Pintek 4 wire leads that we got a couple of years back. The leads ends to the Kelvin clips were disappointingly stiff and they promised later manufacture units would be improved however we've not got more since then to know for sure they have. Nice leads in every other way although not cheap. 
http://www.pintek.com.tw/product_detail/landersound/index.php?Product_SN=232908&PHPSESSID=5ubfkn4vbbie2hnjkpk2r07am6&Company_SN=6002&Product_Site_Classify_SN=17075

[AVGresponding]

VFD may look nice, but their lifespan is fairly short (I have two 8846A that is less than 10 years old, but the VFD are rather aged).
I hope that graphic color LCD will last much longer, I have a lot of meters with them.
LCD are very reliable and easy to replace if the die, but they are fairly limited. Generally I like them, but a DMM can do a lot more with a graphic color LCD (See new Keithley or Keysight).

Some last better than others I guess. My 8840A is still nice and bright, and it and a recently acquired Keithley 2000 are my go-to bench meters.
I love the look of the panaplex on my 7075, and the nixies in my 8300A, but in all honesty I don't use them much (though they work perfectly) and as often as not it'll be a handheld I reach for, just for convenience.

[Shock]

Shock, these are Pintek 4 wire leads that we got a couple of years back. The leads ends to the Kelvin clips were disappointingly stiff and they promised later manufacture units would be improved however we've not got more since then to know for sure they have. Nice leads in every other way although not cheap.

I don't mind if the cables are not so flexible as long as there is a good reason for it. If it lacks strain relief, has molded over clip connections and rotating lantern style banana plugs it's probably signaling a waste of money.

Hopefully they haven't made a connection where the cable splits in two, that would be unforgivable. They have another set of kelvin leads with BNC connectors, seems neither them (unfortunately for me) use shielding.

[colorado.rob]

I own a Siglent SDM3065X.  I do a bit of lab automation with PyVISA.  I use this instrument quite a bit for automated programming and diagnostics.

This is my first bench meter so I don't have reference for comparison with other brands or models.

My only gripe with the meter is that the optional scanner card cannot be added after purchase.  I didn't realize at the time that this is something that would become useful for me to have until a couple of years after owning and using the meter.  I am now considering selling the one I have on the second-hand market and buying a new one with a scanner card.  I would prefer being able to send it in to Saelig (where it was purchased) or Siglent NA and have them retrofit the option.

[tautech]

I own a Siglent SDM3065X.  I do a bit of lab automation with PyVISA.  I use this instrument quite a bit for automated programming and diagnostics.

This is my first bench meter so I don't have reference for comparison with other brands or models.

My only gripe with the meter is that the optional scanner card cannot be added after purchase.  I didn't realize at the time that this is something that would become useful for me to have until a couple of years after owning and using the meter.  I am now considering selling the one I have on the second-hand market and buying a new one with a scanner card.  I would prefer being able to send it in to Saelig (where it was purchased) or Siglent NA and have them retrofit the option.

Sadly not possible.
On one I looked at a few years ago was missing the physical mounting HW and connection looms and componentry for the SC.

[skullan]

I can say a reason why I won't have one at home, is purely due to the "Omg, where the hell do I plug THIS in".

[bikeNomad]

I have two Fluke 8840A and one Fluke 8842A meters.
These are remarkably accurate, even many years out of calibration (I've compared with in-cal bench meters at work).
Their disadvantages:

  * VFD displays (I'm making an LED replacement display)
  * no AC+DC RMS readings
  * no continuity buzzer (my LED replacement display will provide this)
  * Can't leave the 4-wire ohms test leads in for other uses (SENSE terminals aren't disconnected in non-Ω ranges)
  * auto-ranging sometimes gets confused
  * current measurement isn't that great
  * AC measurement and IEEE-488 comms are optional (though my meters have both)

I end up using my Fluke 287 handheld DMM for a lot of general purpose work but its start-up time and lack of an AC power supply annoy me.

[Wytnucls]

I just took apart an old 8842A that apparently was connected to excessive voltage--at least 1500V---with some power to back it up.  The entire front end protection--4 MOVs and 2 fusible resistors--were incinerated.  I removed the burnt parts and subbed in 2 regular 1K resistors (to be repaired properly later, of course) and the meter works properly and is in cal.  No evidence of any external damage, the circuit board is unharmed (but sooty)  and I'm sure nobody died.  I'd say that is 'robust' although I can't quantify that.

Nowhere in CAT documents does it say the the meter should survive.

Again, 61010-2-033 section 14.101 states:  "The component shall not rupture and shall operate as intended during the test."   

IMO, operate as intended means just that, the DMM continues to read properly. 

One of the tests I have shown several times is from section 101.3.2.   The criteria, "During and after the test, no damage to the equipment shall occur."  IMO, that's not the test equipment being used to test the DMM they are referring to, it's the DMM.

"After the voltage of 4.4.2.101 has been applied to the METER, the METER shall continue to be
able to indicate the presence of HAZARDOUS LIVE voltages up to the maximum RATED voltage.
NOTE The METER is not required to maintain its normal accuracy. A maximum deviation of 10 % is acceptable."

Of course there are many other cases as I have mentioned before.  But again, just to be clear, this is outside of my area and I am providing my interpretation of what the standard states.    I would find it hard to believe for example that the pass fail criteria for an impulse test would be if the combo generator was damaged or not.    Fungus may see it otherwise.   

That said, I would imagine that some of the meters I have looked at would pass both the safety and EMC standards with a wide margin but there are far more that wouldn't come close.

You are taking statements out of context.

'The intent of the standard is to ensure the safety of the operator and surrounding area.
Performance / functional requirements are not covered by the standard'. (1)

101.3.2 deals with the fuse protection on the A/mA ranges. Of course no damage should occur, as a conductive path or spark for 1 minute with a 2000V test applied, would cause havoc and affect the safety of the operator.

4.4.2.101 refers to the voltage (1100V AC) applied to voltage measuring terminals in single fault conditions 4.4, (with either MOV or current limiting resistor shorted). The meter must be able to display the presence of a hazardous voltage in a damaged condition (safety consideration).

As for 14.101 (10 high voltage transients), the standard requires that no hazard shall arise due to the operation of overvoltage limiting component.
The component (MOV, sparkgap) shall not rupture and shall operate as intended during the test.
If the component is heated as a result of this test, it shall not heat other materials to their ignition points.
There is no concern for the serviceability of the meter after the test. (It may or may not survive)

(1) Claudia Sirch is the global chief engineer for medical, laboratory, measurement and control equipment and laser products at Intertek. She has over 25 years of experience in the testing and certification business, and is a lead and technical assessor in the IECEE CB Scheme.  She holds a Bachelor of Science in Engineering Physics from Munich University of Applied Sciences.

[TiN]

Keithley 2002

* Shrouded banana jacks. Would prefer low-thermal 5-way binding posts so much...
* Hot
* No electronically controlled front/rear switching

HP/Agilent/Keysight 3458A

* Poor resistance performance compared to competition
* No electronically controlled front/rear switching

Datron 1281

* No NPLC control => poor speed/perf flexibility
* Not really a disadvantage, but complex calibration procedures

[AVGresponding]

Keithley 2002

* No electronically controlled front/rear switching


HP/Agilent/Keysight 3458A

* No electronically controlled front/rear switching

Depending on your POV this can be an advantage; I recently picked up a Keithley 2000 quite cheap, as it was showing errors 600:1 and 600:2, which turned out to be just the switch needing cleaning (calculated risk I know).

[JoeRoy]

I have 34470A, 34465A and DMM6500 on my bench. The Keysight wins for AC, but the Keithley wins for fast sampling/charting.

Please, could you do reviews about the 34470A and 34465A?

Thanks.

[KungFuJosh]

I have a Siglent SDM3045 upgraded to SDM3055X-E, and an SDM3055. My only problem is that on a rare occasion, I want a 3rd bench DMM so I don't have to grab one of my other DMMs out of a drawer when two meters aren't enough.

[tautech]

I have a Siglent SDM3045 upgraded to SDM3055X-E, and an SDM3055. My only problem is that on a rare occasion, I want a 3rd bench DMM so I don't have to grab one of my other DMMs out of a drawer when two meters aren't enough.

The 16ch Scanner card model will fix that. 

[HKJ]

I have 34470A, 34465A and DMM6500 on my bench. The Keysight wins for AC, but the Keithley wins for fast sampling/charting.

Please, could you do reviews about the 34470A and 34465A?

I have done the 34461A, the other two can be called improved models of that meter:
https://lygte-info.dk/review/DMMKeysight34461A%20UK.html

[KungFuJosh]

I have a Siglent SDM3045 upgraded to SDM3055X-E, and an SDM3055. My only problem is that on a rare occasion, I want a 3rd bench DMM so I don't have to grab one of my other DMMs out of a drawer when two meters aren't enough.

The 16ch Scanner card model will fix that. 

Yeah, along with the 3rd DMM that comes with the scanner card. 🤣

[BlownUpCapacitor]

1. Not portable
2. Large power consumption compared to a handheld
3. pricey

[HKJ]

1. Not portable
2. Large power consumption compared to a handheld
3. pricey

1. Many people have a bench where they use their meter.
2. Mains powered, the large power consumption is probably cheaper than batteries.
3. Not really when comparing specifications.

I use both and do see som advantages to bench meters, but they also have a lot of disadvantages. I.e. You need both in many cases.

For the first DMM a handheld would be best, for the second it is about equal (Depending on what you do), for the third your definitely need a bench meter.