Best Budget 6.5 Digit multimeter

[ali_asadzadeh]

Hi
This is the first time that I want to buy a 6.5 Digit multimeter, so any feedback is welcome,
My budget is limited so I have these models that are with in my budget.
From Chinese brands we have these models
HDM3065B from hantek that's very similar to keysight 34461A
GDM-9061 from GWinstek
SDM3065X from siglent it's almost a copy of lecroy T3DMM6-5
DM3068 from Rigol

Any other Chinese brands or lower cost brands that you Maye know and suggest?

Which model do you choose? why?

[Jester]

Unless you really need the graphical features I would not purchase any of those, I would find a gently used 34401A. You can get two or three for the price of one new on your list.

I have had a couple of those lower cost DMMs Instek 8251 etc. I found when there was a good Fluke or 34401A on the bench at the same time I would use the wannabe at absolute last resort. I have 3x 34401A on my bench and they report the same measurements +/- the LSD even after not being calibrated for years. The stability over time is just one aspect there are others like noise that make using a first rate DMM(even an old one) a better experience.

Considering the unseasoned and questionable references used in the meters above, don't expect much accuracy a year after you purchase it.   I have a 25 Year old Fluke 8842A with the LTFLU reference and I would not be surprised if it measures 10Vdc more accurately than the above meters a year after you purchase it.

I see your in Canada, Kijiji sometimes has fair deals if your patient, I picked up two of my Agilents locally one for a really good price.

[coromonadalix]

What work do you want to do with this,  at 6.5  digits you have to consider calibration prices too

An 34401a  is an oldie but a goodie and schematics exists, if you can snatch them at a low price, if they work, you do a capacitor recap and re-calibrate them +/- 200$ cad added (160$ calibration)

Same for 34410A  if i remember, not sure for the 34411A

Some Keithley 2000 series had some reverse engineered too

The web site Xdev's  had many repairs and infos ...


For the others, you have to ask the calibration facility if they can do them (Brand and models)  and some of them Graphical and or normal models can have quirks in their firmware,  do a search before ...

You have some meter charts / comparisons here, do a search


At 6.5 digits  they are not "toys"  and when you say "budget is limited"  you have to ponder a few things first

[tooki]

Hi
This is the first time that I want to buy a 6.5 Digit multimeter, so any feedback is welcome,
My budget is limited so I have these models that are with in my budget.
From Chinese brands we have these models
HDM3065B from hantek that's very similar to keysight 34461A
GDM-9061 from GWinstek
SDM3065X from siglent it's almost a copy of lecroy T3DMM6-5
DM3068 from Rigol

The LeCroy is simply a rebadged Siglent. I don’t know if they gave it special firmware, but I kinda doubt it.

[bdunham7]

SDM3065X from siglent it's almost a copy of lecroy T3DMM6-5

Any other Chinese brands or lower cost brands that you Maye know and suggest?

Which model do you choose? why?

I personally don't see the point of having 6.5 digits if the meter is not accurate and stable.  Lower cost meters have been shown to not do quite as well as their A-brand peers in this regard, especially over time.  If you do buy a lower-cost version, make sure you get one that is adequately supported by the manufacturer and seller.  A three-year warranty and a place in-country to send it for service would be the minimum I'd accept.

Just FYI, the Hantek is an unrelated knockoff (appearance-wise) of the Keysight, but the Siglent is not a 'copy' of the LeCroy.  Siglent manufactures them for LeCroy, so they are virtually the same meter.  I don't know if there is any actual performance difference between the two, but I think the published specs are the same.

The Siglent and Rigol meters are fairly popular and I gather that most people are relatively satisfied with them despite a few issues arising.  You can find threads about each if you search this forum. 

Don't overlook the BK Precision 5493C.

https://www.tequipment.net/BK/5493C/Benchtop-Multimeter/?v=65117

[coromonadalix]

[/quote]
 

Don't overlook the BK Precision 5493C.

https://www.tequipment.net/BK/5493C/Benchtop-Multimeter/?v=65117
[/quote]


Look like a Tonghui 1963

[bdunham7]

Look like a Tonghui 1963

It is.  Rebadged with North American support and warranty and maybe BK-specific firmware.

[LordXaos]

While I cannot provide much of an advice regarding the selection of DMMs, I can at least provide some experience with the SDM3065X, which was my first 6.5 digit DMM... Overall, it is not a bad device, and stability is not bad - unless all three of my DMMs chose to randomly drift by exactly the same amount in the same direction, after nearly three years it appears to be still within the 1 year spec (or probably more the 14 days spec), as the overall drift (compared to a Keithley 6500 and 7510) seems to be about 3 counts at the moment over the entire period (in 10V range). Of course that is based on a sample size of just one, so is statistically mostly irrelevant.

Recalibration is reasonably cheap (at least here) - Welectron says 145€, which is not a lot, in that respect most other DMMs of that class are more expensive. What I do like about it is the 10A range at the front, which is quite handy to do recal of the lab PSUs - the Keithleys only support this using the rear inputs or not at all. Less to like is the boot time, although it is not that bad - but the 6500 boots noticeably faster, plus it makes less noise, so it is the usual go-to for some quick measurement. For the stuff where the 6.5 digits or more really matter this is practically irrelevant anyways, as the device has to warm up for at the very least 15 minutes before this makes any sense (and more is more here...), and if working on some actual project where it is on constantly it is not an issue either - really just for when its all about some quick measurement (to stay sane).

There are a few minor issues with the software - I can't understand why it only starts filling the statistics buffer for histogram/trend chart once you first switched display mode, for example, and not just records data there right away, also the scanner card support is a tad rudimentary (if you have the option), though it could be quite useful with only very few minor software-only changes. But overall it works nicely, and in some cases the extended overrange is useful.

Still I will admit that I probably wouldn't buy another one now that I have the pair of Keithleys, but I am not keen to giving it away either - it is still quite useful to have around.

[ali_asadzadeh]

Guys thanks for the feedback.

Unless you really need the graphical features I would not purchase any of those, I would find a gently used 34401A. You can get two or three for the price of one new on your list.

I like graphical features but I consider them only 10% important in my buying decision, the PC connectivity and having at least USB and Ethernet is more important for me, Also I have not a good experience buying second hand stuff, since my last 2465A Tek scope was having loose knobs and was scratched in shipment.

What work do you want to do with this,  at 6.5  digits you have to consider calibration prices too

I need the meter mostly for these works, since my main work is Digital stuff and I do hand assemble complex PCB prototype boards, most of the time, I have unnoticeable solder bridges, I want the meter to find them, since they can measure sub milliohm ranges,
Also I’m developing a smart Grid Protection relay, that uses a 24Bit ADC to measure line parameters (ADS131M08IPBS) I need the meter to calibrate the devices, Though I have found that almost all of the 6.5 Digit meters have horrible tolerances in the order of  0.15%+0.04 for AC current in 10A range, The DC current has it around 0.150+0.010, SO the question is why they have this kind of big tolerance?

I personally don't see the point of having 6.5 digits if the meter is not accurate and stable

I agree with you, I’m mostly toward HDM3065B from hantek, since it has the most options and the specs are copy paste of  keysight 34461A , it has 30K readings per second too, also I have found an 850$ deal on it. It has front and back plugs, and has USB and Ethernet too. Surprisingly in it’s catalog they have used keysight software for PC side connection!
You can check out the review and teardown in here


I wonder if anyone in here has any experience with it and share!

[Kleinstein]

The AC accuracy is often limited by the analog RMS to DC converter, usually a AD637 or similar. So many are quite similar with the AC part and differences are mainly with the bandwidth or accuracy at higher frequency.
A few (e.g. DMM6500 and the modern Keysight meters like 3446x) use digital RMS and may be a bit more accurate for ADC.
For the current ranges the stability of the shunts can be limiting, especially for the lower cost meters. There may also be a limited number of ranges or some that are only a kind of digital scale up. So a closer look at the ranges may be worth it.

[H.O]

One datapoint on one of the "b-brand" meters - for what its worth.
I have a DM3068 that I got brand new in August 2013. According to its paperwork it was calibrated in May 2013. It has never been calibrated or adjusted since. Unfortunately no data was provided, just a certificate.
I also have a 34410A that I bought second hand a couple of years back. It was last calibrated August 18 2022, so less than 60 days ago. The calibration report provided says that at 10.0V nominal it measures 10.00001V.

After a 45min warmup period, with both meters set to 100PLC:
34410A: 2.49900V
DM3068: 2.49903

34410A: 10.00012V
DM3068: 10.00032V

Lets say that the 34410A is spot on. Then the DM3068 is 200uV out which, I admit, might feel like "a lot". But even the 24h specification for the DM3068 says 0.002% of reading + 0.0004% of range which, if I'm not making a fool of myself equals 280uV. It might have drifted, I don't really know that, but its still within its 24h specification, 9 years down the road. Or did I indeed goof the math?

[nctnico]

Hi
This is the first time that I want to buy a 6.5 Digit multimeter, so any feedback is welcome,
My budget is limited so I have these models that are with in my budget.
From Chinese brands we have these models
HDM3065B from hantek that's very similar to keysight 34461A
GDM-9061 from GWinstek
SDM3065X from siglent it's almost a copy of lecroy T3DMM6-5
DM3068 from Rigol

Which model do you choose? why?

When on a budget I'd go for the GW Instek. They have been in the DMM business longer and they are a company targeting the professional market.

But it is likely worth it to spend more and go for a Tektronix / Keithley DMM6500 or Keysight 34661A. Nowadays I'd probably prefer the DMM6500 because it is more versatile.

[bdunham7]

I have found that almost all of the 6.5 Digit meters have horrible tolerances in the order of  0.15%+0.04 for AC current in 10A range, The DC current has it around 0.150+0.010, SO the question is why they have this kind of big tolerance?

You have to look at what is feasible to make a judgment like that--and those specs are not 'horrible', they're typical for a good instrument.  If your plans depend on getting significantly better results than that on any reasonable budget, you are in for a shock.  The two issues are that the shunt can only be so large in a bench meter, so the design has to allow for some self heating and rely on a low-tempco shunt--and then there is a practical limit to how accurately you can measure TRMS AC without specific conditions on the measured signal.  The first just depends on how much space and money you want to allot to the 10A current shunt.  The second comes down to how accurately you can measure an AC signal without any guarantee that it is perfectly repetitive.  The answer for most practical technologies is about 600ppm or 0.06% in the 50/60Hz region that you are interested in.  Even if you got a $5K Keysight 34470A with TrueVolt or a $13K 3458A, you wouldn't do much better than that.  To do better than this typically involves some more advanced (expensive) technology like Fluke's thermal converters or the HPAK synchronous subsampling mode--and then you still have the current shunt issue.

I agree with you, I’m mostly toward HDM3065B from hantek, since it has the most options and the specs are copy paste of  keysight 34461A , it has 30K readings per second too, also I have found an 850$ deal on it. It has front and back plugs, and has USB and Ethernet too. Surprisingly in it’s catalog they have used keysight software for PC side connection!

Its physical similarity to the Keysight and the 'copy-paste' of the specs is no guarantee of actual performance.  I have no personal experience with it and I haven't seen any reviews that would address how accurate or stable it is, but I wouldn't be terribly optimistic TBH. 

[Kleinstein]

The video on the Hantek 6 digit meter shows a MAX6325 reference, which is a bit more drifty than the LM399 used in more classic 6 digit DMMs (and the HP3468/3478). This reference alone makes the accuracy specs questionable. I very much doubt they would do an extensive (e.g. 500 hours) burn in to at last reduce the long term drift. In addition to the reference the 10 V range would also need some divider, as the ADC (though I don't know the type) very likely only does some 2.5 or maybe 5 V range.  So I would not really count on the accuracy specs - maybe add a k factor of 2 ?  .

From the internals the protection against high voltage spikes also looks not that good. So the cheaper design comes at a price.

I would consider the HDM3065 more like a cheaper "copy" of the SDM3065, not so much to original, though it only offers 10 V and not 20 V with high impedance.

[strawberry]

they push hardware to the limits to get 6 digit
even 34401A is basically useful 5 digit

[bdunham7]

The video on the Hantek 6 digit meter shows a MAX6325 reference, which is a bit more drifty than the LM399 used in more classic 6 digit DMMs (and the HP3468/3478). This reference alone makes the accuracy specs questionable. I very much doubt they would do an extensive (e.g. 500 hours) burn in to at last reduce the long term drift. In addition to the reference the 10 V range would also need some divider, as the ADC (though I don't know the type) very likely only does some 2.5 or maybe 5 V range.  So I would not really count on the accuracy specs - maybe add a k factor of 2 ?

My experience with B- and C brand bench meters is that their specs/performance range from 'typical' (k=1?) to 'aspirational' or perhaps even 'fictional' in this case.  Had a 5+ digit GDM-8251 and although it's specs were not great for a 5-digit class instrument, it failed miserably to meet even those.  Any number of 4.5 digit meters were significantly better.  One characteristic that I've seen is that these meters will use up all of their tolerances--or even more--just in tempco within the 18 to 28C range that they are specified for.

The MAX6325 is on par or even inferior to references commonly used in 4.5-digit handhelds, so it seems pretty inadequate in a 6.5-digit model.  I doubt they aged and selected them and I'm confident they didn't splurge on things like high quality divider resistor and shunts.  It would be interesting to see one (or more) of these actually tested by someone with the instruments to show their actual performance.  My guess is that actual performance varies quite a bit between examples.

[Kleinstein]

The max6325 reference is not that bad for a bandgap reference, though in a plastic case. Not good for 6 digits, but should be OK for 5.5 digits. The SDM3055 uses the slightly inferior max6225.  4.5 digit meter usually use way worse ones.

The input divider in the HDM3065 seem to be a Caddock 1776 variant (one of the smaller ones) - so it may not be that bad. However with a usual SD ADC the 10 V range would also have to go through some divider (seperte resistors or maybe part of the same divider). There are a few ADC chips for higher voltage, but these are rare). This makes it odd to have the best accuracy in the 10 V range and how they could reach that accuracy without rather expenside dividers (the 1776 would not be good enough). The 34401 / 34461 and many better 6 digit meters get the 10 V range without a divider before the ADC. So the best accuracy for the 10 V range is natural there.

[bdunham7]

4.5 digit meter usually use way worse ones.

Fluke 289 uses LT1019A 20ppm/1000hr, but that is top-of-the-line.  I suppose for a modern (less-precise) 5.5 digit model with 120ppm basic accuracy, that might be good enough.

This makes it odd to have the best accuracy in the 10 V range and how they could reach that accuracy without rather expenside dividers (the 1776 would not be good enough). The 34401 / 34461 and many better 6 digit meters get the 10 V range without a divider before the ADC. So the best accuracy for the 10 V range is natural there.

Apparently they've simply copied the HPAK specs verbatim and listed them as their own.  I'd give that a 0% confidence interval.

[2N3055]

The max6325 reference is not that bad for a bandgap reference, though in a plastic case. Not good for 6 digits, but should be OK for 5.5 digits. The SDM3055 uses the slightly inferior max6225.  4.5 digit meter usually use way worse ones.

The input divider in the HDM3065 seem to be a Caddock 1776 variant (one of the smaller ones) - so it may not be that bad. However with a usual SD ADC the 10 V range would also have to go through some divider (seperte resistors or maybe part of the same divider). There are a few ADC chips for higher voltage, but these are rare). This makes it odd to have the best accuracy in the 10 V range and how they could reach that accuracy without rather expenside dividers (the 1776 would not be good enough). The 34401 / 34461 and many better 6 digit meters get the 10 V range without a divider before the ADC. So the best accuracy for the 10 V range is natural there.

MAX6325 are buried zener references. There are even 1ppm MAX tempco versions (MAX6325C). Questions are if they use those though...
Biggest problems would be long term drift (30ppm /1000h9 and hysteresis (20ppm) because of the plastic case...

DS:
The MAX6325/MAX6341/MAX6350 are low-noise, precision voltage references with extremely low, 0.5ppm/°C
typical temperature coefficients and excellent, ±0.02% initial accuracy. These devices feature buried-zener technology for lowest noise performance. Load-regulation specifications are guaranteed for source and sink currents up to 15mA

One problem is that they seem to have large spread in quality... And price is not low, so LM399 (or new brethren) start to look good for best results, even with additional work and circuitry needed.

[tautech]

The max6325 reference is not that bad for a bandgap reference, though in a plastic case. Not good for 6 digits, but should be OK for 5.5 digits. The SDM3055 uses the slightly inferior max6225.  4.5 digit meter usually use way worse ones.

Thread title is 'Best Budget 6.5 Digit multimeter' for which SDM3055 does not fit but SDM3065X and its LM399 reference does.

[maxwell3e10]

As a point of reference, here is a comparison of Allan variance for HP3457 and Keysight 34465 and 34470. Even 34465,  a top-brand 6.5 digit meter, drifts too much, in my opinion. Well, it does give 6 digits stability at 10^4 seconds, but not any more than that. In general, I would say modern equipment, even from A-brands, just barely meets the specs these days. Older equipment usually was significantly better than specs.

And unlike say oscilloscopes, which are more WYSIWYG, multimeters have many ranges and obscure effects, it takes a highly experienced and dedicated team of engineers (with few project managers and accountants) to properly design one. So for measurements that really need 6 digits, I would definitely choose an old used meter. Newer budget meters only have an advantage of a graphical display for real time monitoring of short-term changes. But  there are several ways to easily read and plot data over GPIB developed by members of this forum.

[Kleinstein]

The grapf for the Allan variance is not that relevant for the meters accuracy. 1 µV of standard deviation in the 10 V range would already be the limit where one would consider it 7 digit resolution. The noise noise part is one where even the relatively cheap one can be OK and easily meat there specs. This part is easy to check and also easy to come up with reliable specs. 

The more tricky part is the stability with changing temperatur and especially the long term drift. Here it is even hard to come up with good specs for a new meter.  It is only a good guess how much it may drift in the next years to come.

The old HP3457 is one meter that may not be the best when it come to long term drift, because of the somewhat odd internal design with a 3 V prime range.
The general tendency with the more modern meters is that the noise noise got better, but the long term stability not so much or not at all.
The low cost 6 digit meter are based on SD ADC chips and there good long term stability is a bit more tricky.

[maxwell3e10]

The old HP3457 is one meter that may not be the best when it come to long term drift, because of the somewhat odd internal design with a 3 V prime range.

And yet this data clearly show that 3457 drifts less than 34465. As for long term drift in calibration, I find that  A-brand meters are somewhat comparable, i.e. older meters drift similar amount in 30 years as newer meters drift in a few years.
https://www.eevblog.com/forum/metrology/usa-cal-club-round-2/msg3552046/#msg3552046

[bdunham7]

As a point of reference, here is a comparison of Allan variance for HP3457 and Keysight 34465 and 34470. Even 34465,  a top-brand 6.5 digit meter, drifts too much, in my opinion. Well, it does give 6 digits stability at 10^4 seconds, but not any more than that. In general, I would say modern equipment, even from A-brands, just barely meets the specs these days. Older equipment usually was significantly better than specs.

I'm not sure what we're looking at there.  What is the initial condition of these meters at the start of the graph (are they fully warmed up), what are the thermal conditions of the room, etc?  And what specifications are the newer meters 'just barely meeting'?

[coromonadalix]

Thats why sometime getting an old model  can or could be the best  ... they have aged

You have some good software here to do logging

An 3457, 34401a,  8845, 8846 could be useful etc ...   many meters are supported in the software


For 10-20 amps current measurements, simply use a  dedicated shunt or a clamp meter,  more practical and you wont blowup  10-13 $ fuses in case of a problem