EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: Timur Born on August 04, 2015, 09:38:45 pm
-
Hello everyone!
I am looking for a multimeter that comes with high sensitivity for frequency measurements, as in low minimum mV/uA levels.
The Fluke 189/289 lists minimum sensitivity levels as low as 5 mV and 25 uA. As comparison the Brymen BM867 lists a minimum voltage of 100 mV, the Gossen Metrahit Ultra lists 10 - 100% of the measurement range, which corresponds to 30 mV and 30 uA respectively (if I interpret this correctly). The 100% of the Gossen also brings us back to the Flukes, which are listed as "maximum 10x measurement range" aka 1000%.
So are the Flukes the ones to look at for low - and varying - signal level frequency measurements? Or are there other contenders I should know about?
Thanks for any advice!
-
I suspect that not many people are interested in frequency readings of low level signals using a multimeter?! What I do is measure noise on ground lines of PC mainboards and connections that is induced by components being under load (especially GPU) and CPU C-states switching.
The multimeter readings are useful for getting mV/uA readings and should provide frequency readings at these low levels for quick and easy analysis. Unfortunately most meters I looked at seem to fail exactly in this department, at least according to their specs. That is, if the specs even list minimum sensitivity for frequency readings or list anything more useful than just "1 Vpp".
Since I am only interested in the audible frequency range I can use an audio interface for more detailed spectrum measurements and dB levels. The latter approach often is not necessary when a quick qualitative measurement provides enough information compared to a more work/time consuming quantitative approach.
-
I suspect that not many people are interested in frequency readings of low level signals using a multimeter?!
Interested many. The issue resides not in the plane - how do you intend to do this? When connecting the meter to the lines of the exchange or frequency drive circuit - will act as a meter shunt. The device stops working. There's nothing you can measure or even destroy the device.
-
I connect the meter to the chassis, ground lines of ports or ground lines of connected cables and measure against earth ground. Nothing stops working and nothing is destroyed.
To give an example of what can be measured by this is the frame-rate of the GPU. It seems to show as noise of similar frequency on all PC ground lines, which then can induce audible noise in connected speakers that come with a 3-prong power-plug (aka earth). Even worse, it can induce audible noise all by itself from the power-section of the mainboard (and/or the power-section of the GPU, still analyzing).
Anyway, this is stuff for a different post, here I am just interested in finding a proper multimeter. :-DMM
-
I connect the meter to the chassis... I am just interested in finding a proper multimeter. :-DMM
Any industrial multimeter would be - in finding a proper multimeter. Its specification describes the capabilities. But any multimeter will have the characteristics of the input impedance and input capacitance. You describe your desire to parallel connection ... Well, maybe you'll get.
-
You don't say what your requirements are. You should quote bandwidth, minimum mV/uA sensitivity , square or sine waveform.
5mV is about the best you will get with a multimeter. If you mostly measure frequency on the current range, decent uA sensitivity can be found with some cheaper meters than Fluke's.
-
The requirements are not easy to state, because that's what I am trying to find out with the measurements. But to give you an idea, we are talking about peak signals in the -100 to -70 dBfs range when measures with an audio interface that offers 110 dB RMS SNR on its inputs. As stated before :P, bandwidth would be in the audible range, around 50 Hz - 20 kHz.
In the volts range sensitivity should preferably be single digits mV. Even the 5 mV of the Flukes 189/289 are already higher to what I am measuring with a Fluke 177 (borrowed) and a VC890 (sent back). Especially because this is a sinusoid rating in Fluke's manual, while the noise signals are modulated (depending on load/content). I suspect that I will mostly have to use current measurements for frequency, because of the low voltages.
In the uA range sensitivities in the two - three digit range might be sufficient. I am not sure yet, because the VC890 disagreed extremely between its A/mA and uA ranges. In the order of measuring 2 mA in the mA range, but only 30 uA in the uA range (even when set to 6000 uA digits). To my surprise frequency was measured somewhat usable in the mA range of the VC890, even though the manual stated only 1 Vpp sensitivity for voltages, and listed no values to current.
The latter is a problem with some other spec sheets I have been looking at, they only least voltage levels as Vpp. Other sheets at least list a percentage range over measurement range (like the 10-100% on the Gossen), albeit this still is a bit vague.
On a side note: Frequency sensitivity and trigger levels are about the only differences I could find between the specs of the Fluke 89-IV vs. 189/289.
-
The requirements are not easy to state, because that's what I am trying to find out with the measurements. But to give you an idea, we are talking about peak signals in the -100 to -70 dBfs range when measures with an audio interface that offers 110 dB RMS SNR on its inputs. As stated before :P, bandwidth would be in the audible range, around 50 Hz - 20 kHz.
Useless figure without specifying the sensitivity of the input (i.e. sine level for 0 dBfs).
You're probably best served by an oscilloscope with a good, possibly differential, pre-amp. A classic would be Tek 7A22. Not sure if much like it is available new, and for what cost. Probably expensive.
/e: First google hit: http://aricorp.com/OSP-1.html (http://aricorp.com/OSP-1.html) ~1200 $. Bandwidth only specified for 1x Gain. Likely less noisy than the fourty year old tek, but less flexible.
-
As stated before :P, bandwidth would be in the audible range, around 50 Hz - 20 kHz.
Again the same. Yes, you can hear an ear. Do you describe noise or research DTL? That transmissions frequency of processor, do you talk? Because not always it spectrum analyzers can. That it is possible to say a multimeter.
you talk about modulation of noise and frequency, after volts simultaneously, after amperes simultaneously. I understand nothing. Or complications of translation. Excuse me, I will be exit.
you are right, this theme not for other position, and other division of forum.
-
+4 dBu or -10 dBV line reference level, take your pick. Anyway, really low levels in the lower mV range, so the question is if there are multimeters out there that can compete with the Flukes (5 mV) or at least Gossen (30 mV)?
-
First of all, I am not 100% shure I understand everything a round what you are to do but i think this could help you on your quest:
http://www.sparkylabs.co.uk/shop/product_info.php?products_id=146 (http://www.sparkylabs.co.uk/shop/product_info.php?products_id=146)
Johan-Fredrik
-
The Fluke 177 mV minimal RMS voltage sensitivity is 5% of range or 30mV. Are you saying you can measure accurate frequencies in the whole audio bandwidth, with just 1mV input with that meter?
-
The Keysight U1272A specifies 3mV or 70uA, might be another option for you. I don't own one so I can't test it.
-
I think you need to consider a good bench meter like the Rigol DM3068. It is $850 but I think it will do the job.
-
Uh, so it's only about frequency measurement?
...
Frequency counter.
-
You may have to run an amplifier with a filter to get the frequency. Most have a sensitivity of 25mV.
-
I think you need to consider a good bench meter like the Rigol DM3068. It is $850 but I think it will do the job.
I'm guessing the OP wanted something handheld for quick and dirty go/no-go testing. =P
But just for kicks, I tested a few bench meters @ 20kHz sine wave. Both the HP 34401A and 3457A started losing the plot at ~5mV, while the mighty 8.5digit R6581 could hold on until 1.6mV. :-DD
-
First of all, I am not 100% shure I understand everything a round what you are to do but i think this could help you on your quest:
Let me try to clarify:
I do not own a (hand) multimeter, so I need one for all the usual benefits that such a (portable) device brings. It will not be used often, but when it's used it may have to measure very low voltages and currents, both AC and DC.
In one of the usage cases measured voltages and currents are usually frequency modulated, depending on their source, and amplitude may change, too. Levels are in the range of mV (to V) and uA (to mA). I don't need extreme precision in these ranges, but a useful number that is close enough to reality.
On top of that I'd like the meter's frequency measurement to be sensitive enough to calculate frequencies of these low level signals. This helps me with analysis and doing quick assessments of different test setups. As I am mostly only interested in frequencies within the audible range down to DC (1 Hz - 20 kHz) I do *not* need the meter for precise measurements. Again, I need useful numbers that are close enough to reality.
I can do precise frequency spectrum measurements using professional audio recording equipment and software. But obviously this involved more work and is not needed in many cases of my tests.
And apart from all that I still do need a multimeter for whatever other stuff people use multimeters for (I also posted another thread about event logging features I need). So my idea is to get one that can help with my specific usage case on top of the usual stuff.
http://www.sparkylabs.co.uk/shop/product_info.php?products_id=146 (http://www.sparkylabs.co.uk/shop/product_info.php?products_id=146)
Thanks for the link. I will keep this in mind should I ever need that kind of current measuring precision. For my current applications it would be more than I need and the money would be better spend on a better meter to begin with.
-
The Keysight U1272A specifies 3mV or 70uA, might be another option for you. I don't own one so I can't test it.
Thanks for the hint. 3 mV is a good number, seems to be 10% of measurement range, so effectively the same as on the Flukes and the Gossen, just that the range is smaller (aka 10% is smaller). 70 uA may or may not be too high already, I cannot say for sure yet. I will have a look at it, especially since I also have some datalogging demands, too (and like BT options).
-
I think you need to consider a good bench meter like the Rigol DM3068. It is $850 but I think it will do the job.
Indeed, but then I would still miss a handheld meter and be stationary. And when I am stationary I can just as well do full audio spectrum analysis with more sophisticated audio equipment. ;)
The idea is to get a handheld(ish) meter for all kinds of stuff, but then get one good enough to also help me with these specific low level noise related measurements.
-
The Fluke 177 mV minimal RMS voltage sensitivity is 5% of range or 30mV. Are you saying you can measure accurate frequencies in the whole audio bandwidth, with just 1mV input with that meter?
No, the 177 cannot measure frequencies at these levels. I used the 177 and VC890 to get an idea of the mV and uA levels I am dealing with. Based on that information I am looking for a better meter that can also provide frequency data at these levels.
-
Uh, so it's only about frequency measurement?
...
Frequency counter.
Nope, it's about getting a meter that is good enough to do low level measurements and then display useful frequency data on top of that. I can do quantitatively better frequency measurements with an audio-interface + computer.
-
You may have to run an amplifier with a filter to get the frequency. Most have a sensitivity of 25mV.
Do these amplifiers come with low enough noise-floor themselves? Sounds like a possible solution, but maybe a bit convoluted and costly one compared to just get a better meter to begin with?!
-
I think you need to consider a good bench meter like the Rigol DM3068. It is $850 but I think it will do the job.
I'm guessing the OP wanted something handheld for quick and dirty go/no-go testing. =P
But just for kicks, I tested a few bench meters @ 20kHz sine wave. Both the HP 34401A and 3457A started losing the plot at ~5mV, while the mighty 8.5digit R6581 could hold on until 1.6mV. :-DD
Indeed, quick and dirty mostly. Albeit sometimes the meter reading could prove even necessary, because doing precise measurements using audio equipment is somewhat limited, and even having (earth grounded) speakers connected may not always be an options while testing.
And thanks a lot for doing the test! :-+ Would your bench meters work better or worse at 150 Hz to 1500 Hz? If not then it sounds like I can just use one of the Flukes or maybe the suggested Keysight instead then.
-
Why don't you use a microphone preamplifier (sensitivity of +- 5mV) and measure the output of the preamplifier...?
-
I don't believe that there is a handheld multimeter that does what you want.
-
I don't believe that there is a handheld multimeter that does what you want.
Well I already know that on paper the Flukes (189/289) offer most of what I want. The question is if there are alternatives to these (somewhat costly bought new) bricks?
-
And thanks a lot for doing the test! :-+ Would your bench meters work better or worse at 150 Hz to 1500 Hz? If not then it sounds like I can just use one of the Flukes or maybe the suggested Keysight instead then.
No problem!
Not much of an improvement I'm afraid, the 34401A started to dance at 4.6mV(1.5kHz) and 4.1mV(150Hz). The 3457A actually performed worse at the lower frequencies. =P
The source was the oscillator output of a Panasonic VP-7722 audio analyser, I would have no idea how the meters would perform in a scenario similar to your intended use.
I really want the Keysight U1272A, can't really justify the purchase though considering I already have an array of bench meters....grumble grumble...
-
FWIW, the Uni-T UT71D can measure frequency with 20mV RMS between 10Hz and 10kHz, on the 4V AC range. Measured voltage will not be accurate though. The Fluke 185 can do it with 15mV up to 15kHz, on the 500mV scale.
-
There more you say the more I get the feeling you won't be happy with a DMM. Get a cheap (5 €) DMM and a scope. Build a suitable pre-amp yourself (e.g. an instrumentation amplifier plus an optional, switchable shunt for current).
-
I am working on a review of a Uni-T UT171A. I just did some tests to see what it could do. It was able to read frequency 20Hz - 20kHz at 5mV and was basically within 1% of the correct voltage reading. It was able to read frequency 20Hz - 20kHz at around 0.9mA and basically within 5% within that range. It is a dual display so you see frequency and the measured voltage or current at the same time. It does not see frequency below 5mV but stills stays relatively flat and within 1% down to to 2mV.
And as a general multimeter it is pretty good, but it is outside its specs on some ranges by a little bit.
-
Actually you might be happier with a handheld scope but you said you were looking for something less of a brick and less expensive than the 289. There are some low price Chinese handheld scopes that might work but it will be up to you if you can live with the lower end stuff for your work. Perhaps something like this?
http://www.ebay.com/itm/MINI-OSCILLOSCOPE-HANDHELD-DIGITAL-SCOPEMETER-POCK-Multimeter-/191647782856?hash=item2c9f1917c8 (http://www.ebay.com/itm/MINI-OSCILLOSCOPE-HANDHELD-DIGITAL-SCOPEMETER-POCK-Multimeter-/191647782856?hash=item2c9f1917c8)
I have no idea if it is any good or not.
-
Well, I don't know. For a scope I would want something with 0.5 mV/DIV vertical sensitivity, because some of my current readings go down to the 1 digit mV range (i.e. 2.5 mV). That would be a flat line on the mobile one linked earlier, unless I use an amplifier. With a scope I wouldn't need mobile at all, because I would only use it for that single purpose. A Rigol DS2072a comes to mind, if I understood the specs correctly.
But then I would have a scope that I use for a single purpose that might even be limited to one time use. Aka, once I'm done with my current analysis I might never need to do these again. And like I mentioned earlier, I can do all this using a laptop + audio-interface, with full (on paper) 24-bit vertical precision compared to the rather bleak 8-bit of affordable scopes.
A multimeter would be useful for other stuff than just the one-trick-pony analysis I am currently doing. Over the years it will come out now and then for all sort of stuff and when it's mobile you can take it along to friends and customers to do measurements. For example, I mentioned in my other thread that I am looking for a DMM that does event logging in order to measure drops and spikes in mains (including duration and amplitude).
So spending money on a DMM gives me less of a long-time worry compared to spending money on a scope that I won't really need. If it has to be one of the large and more expensive Flukes then be it so. If there are alternatives, maybe even better overall ones (including logging, Bluetooth to iOS and whatnot) then I am open to advice. The Keysight/Agilent one looks interesting, I am still looking at the Gossens.
-
The Gossen MetraHit 30M can give accurate frequency measurements up to 20kHz, with a sine wave of 14mV RMS minimum.
-
I already knew the Agilent U1272A is my most frequency sensistive meter and I testet it quick again with a sound gen on my cellphone.
50 and 440 Hz are spot on from 0.25mV or greater (with Low Pass Filter).
10kHz is spot on from about 1.2mV.
With the 0.25mV I was surprised by myself.
-
Thanks for all the feedback and hints. I will have a look at the Agilent and Gossens compared to the Flukes. The Agilent is slightly cheaper than the 289 for the basic unit (about 10%). The Gossen 30M is considerably more expensive (nearly double).
-
Thanks for numbers on the Uni-T, too. 0.9 mA is a bit high though.
-
I watched some numbers with the Fluke 177 today. It hits its lower resolution limit with the mV measurement, but it helps to get an idea. Todays measurements were somewhere between 0.2 - 2.0 mV and 0.1 - 0.6 mA, other measurements will be lower, though (less powerful hardware = lesser noise signals). So according to specs none of the meters would give me frequency reading in the mV range, but several of them should be able to handle the mA/uA range.
Based on specs the Flukes still seem to have an edge here (5 mV, 25 uA), especially because their datalogging and Bluetooth options fit the bill, too. Still looking at the Agilent(s) and Gossen packages, though.
-
Todays measurements were somewhere between 0.2 - 2.0 mV and 0.1 - 0.6 mA, other measurements will be lower So according to specs none of the meters would give me frequency reading in the mV range, but several of them should be able to handle the mA/uA range. :o
Splendidly! I does not know even, that you probe exactly. It is not clear. Acoustic signal, constantly changing on frequency and size. Also well. And it is good - 1.0 % + 3 (45 Hz to 500 Hz) 1. All AC voltage and AC current ranges are specified from 5 % of range to 100 % of range. = 30mv – 600mv
Your measurings unknown what beyond reach reason.
I can assume an error.
-
The UT71D can measure frequency up to 20kHz, with just 9uA RMS on the 400uA range.
-
Thanks, but is that measured or how do you know? I checked it before and now checked the manual again. It only lists 200 mV minimum sensitivity and nothing about current minimums.
-
Measured with both 71D and Gossen MetraHit 30M (100uA range).
-
The UT71D can measure frequency up to 20kHz, with just 9uA RMS on the 400uA range.
Unless the specification is.
Remarks
a) True rms are valid from 10% of range to 100% of range
b) AC crest factor can be up to 3.0 except 1000V where it is 1.5.
c) A residual reading of 80 digits with test leads shorted, will not affect stated accuracy.
d) When frequency is lower than 100kHz, the accuracy guarantee range 10%-100%.
e) When making AC+DC measurment, the accuray need to add (1%+ 35 digits) of reading based on the above table
Remarks Frequency
Input amplitude “a” as follows; (DC electric level is zero)
When 10Hz~40MHz : 200mV ? a ? 30Vrms;
AC Current
400µA - 1kHz~10kHz - (1%+40)
-
You may have to run an amplifier with a filter to get the frequency. Most have a sensitivity of 25mV.
Do these amplifiers come with low enough noise-floor themselves? Sounds like a possible solution, but maybe a bit convoluted and costly one compared to just get a better meter to begin with?!
You could just build a low pas filtered amplifier with a cutoff of 20KHz., battery powered opamp circuit would work well.
-
But then again I could just use one of my audio interfaces that come with 110 dB RMS SNR and use that as amplifier. And while I am at it I can connect this to a laptop and do full measurements using sophisticated software. Of course this wouldn't be quick & dirty anymore. ;)
-
110 dB RMS SNR
Uhm...
-
Yes?
-
Unfortunately the Fluke 289 turns out to be rather useless for my usage case. Blame my demands, not the device. DC mV works for some low (<500 Hz) frequency measurements, but not all. AC mV suffers from too low a signal, even more so when the low pass filter is enabled (2 mV average signal only then).
AC mV and uA for high frequency (>1 kHz) measurements suffers from the noisy source signal (I measure noise) and probably from the sometimes high crest factor (6-10). But my main grief with high frequencies is that the readings are way too high as the 289 seems to pick up the harmonics rather than the fundamental frequency. This comes as a surprise, because the much cheaper Voltcraft I tried earlier was quite able to catch the same high frequency signal close to the correct fundamental.
I obviously don't understand how the uA setting on multimeters works, because the Fluke 289 shows the same huge discrepancy between its mA and uA setting when measuring the same source. I would expect 2 mV to translate to 2000 uA, not to something like 30 - 60 uA. No frequency measurement possible at these levels, probably because the signal includes a large DC offset and isn't close enough to a sinusoid form.
Interestingly the 50000 count display is a hurdle. Some of my measurements average around 50 something (AC) and 500 something (DC) uA, so the 289 is forced to a lesser sensitivity range (as far as frequency measurements are concerned).