Bench multimeter

[Anding]

Hello,

I'm torn between the Silgent SDM3045X and SDM3055 bench multimeters.  I read in another post that the 3045X is fanless (hence silent!), but the specs tell me that the 3055 has a much better range at low voltages (200mV threshold rather than 600mV). 

No annoying noise is important or I just won't use it, but - for reference - my Rigol MSO5104 is quiet enough for me.  How much noise does the 3055 make in comparison?

600mV seems a high threshold, maybe too high for reliably checking the bias on transistors, hence my interest in the mid-range option.  The price difference between the two will not affect the decision on this occasion.

Also, any other competing units I should consider at this level?

Many thanks for your thoughts!

[Kibabalu]

Hi,

I'm using the SDM3065X and I'm quite satisfied with it. I bought it, because I 'needed' the 6 and ½ digits.

I don't see the benefit of buying a 4 and ½ Digit bench multimeter compared to buying a handheld one. They are much cheaper and more practical.

Thus, skip the SDM3045X and go for the SDM3055!

[Kleinstein]

AFAIK in China there is no SDM3045, but a SDM3055E instead: essentially the same hardware but with software essentially like the 3055. From the HW side the 3045 and 3055 are very similar - mainly lower grade parts (especially the HV divider) with the 3045 but in theory the same ranges. The 3045 is more like an afterthough with odd ranges not directly reflecting the hardware side. As a result there is additional range switching with the 3045 that can cause problems.
With the very similar HW side and essentially the same power consumption it should be OK to disable the fan in the 3055: this will however effect the temperature distribution and thus void the calibration. The time for warm up to a stable state also gets longer.

The SDM3045 still offers 4 wire ohm mode, that handheld meters usually don't have. For longer time use (especially to log data to the PC) the mains supply is also a plus with a bench meter.

[slavoy]

If price isn't an issue, there's no point even considering the SDM3045 as it has a lower specification. I have an SDM3065X (very similar in terms of functionality to 3055) and the fan noise is about the same as my Rigol DS1054Z. You need fans to distribute heat evenly across the board to ensure the stability of the measurements.

[alm]

You need fans to distribute heat evenly across the board to ensure the stability of the measurements.

There are plenty 6.5 digit (and better) meters without a fan (like the venerable HP 34401A). It's just a matter of thermal management: how much heat the instrument produces and how the engineers decided to get rid of the heat. Actually if you watch Dave's video about the HP 3457A tear down, you'll find him saying the instrument does not have a fan because the air currents would introduce thermal gradients. You can also find good 6.5+ digit meters with a fan, though.

[Kleinstein]

The SDM3045 should have essentially the same power consumption as the SDM3055. So there is not absolute need for the fan.  There are pros and cons with the fan.
On the positive side the overall temperature could be a bit lower, but the DMM is not really using a lot of power, even with the graphics display.
The main positive aspect I see is a shorter time for warm up (e.g. 30 min compared to some 2 hours) as the heat exchange to the environment is faster.
A point that may happen without a fan is thermal instability - that is quite some thermal fluctuations from turbulent and possibly oscillating convection. This may depend on instrument stacking, especially with something stacked on top.  A fan can avoid / reduce this.
The thermal gradients in side are not so much a point - here it is more that they are different. So when looking for details it is not a good idea to add or remove a fan, but use the same fan speed for normal use and calibration. So a thermal regulated fan is not a good idea and AFAIK no DMM uses this.
At the 5.5 digit level the thermal gradients are usually not that critical - this is more a thing for sub 1 µV or 7 digits and up.

The main downsides of a fan are noise and getting more dust / dirt inside.

[KungFuJosh]

I have the SDM3045 SDM3055X-E. I bought the SDM3045 and upgraded the firmware to the SDM3055X-E, which is a 5.5 digits, 240,000 counts DMM with identical hardware. If you want a fanless 5.5 digit DMM, go for it. The upgrade was really easy.

You read about doing the hack here: https://www.eevblog.com/forum/testgear/siglent-sdm3045x-enough-is-enough/msg3909188/#msg3909188

[mawyatt]

You need fans to distribute heat evenly across the board to ensure the stability of the measurements.

There are plenty 6.5 digit (and better) meters without a fan (like the venerable HP 34401A). It's just a matter of thermal management: how much heat the instrument produces and how the engineers decided to get rid of the heat. Actually if you watch Dave's video about the HP 3457A tear down, you'll find him saying the instrument does not have a fan because the air currents would introduce thermal gradients. You can also find good 6.5+ digit meters with a fan, though.

The two 34401As we have, one is an old HP and the other a AG version, are the ones we use the most. They are quiet, quick, stable, and precise. They've had years for the nice selected LM399s to stabilize, and probably one of the finest, if not the finest, electronics instruments ever created, stunningly good instruments that have held up amazingly well over time

All toll, we have and use seven 6 1/2 digit DMMs, 3 KS34465A, DMM6500, SDM3065X, and mentioned two 34401As (a few handhelds). Long story how all these were acquired (recently last KS34465A by accident), but some of our previous use cases demanded extreme accuracy in various parameters and results reproduced "blindly" in areas we weren't allowed to attend, so "hand off" if you will

The SDM3065X is a nice DMM, but not in the same league as these others, however it costs about 1/2 the 34465a or DMM6500, so a good value in that respect. The fan is not too bad, little louder than KS or DMM, and the general performance is OK. Would like to operate longer to help "age" the LM399 (understand it's not selected nor aged), but fan is loud enough to prevent continuous operation and we generally only operate the SDM3065X when other instruments with fans are operating.

We've thought about disabling the fan, but as others mentioned this might disrupt the calibration and introduce thermal issues, and our various instruments are stacked, which introduces other thermal effects, so haven't attempted to disable the fan.

Which brings up an interesting point about having a SDM3065X power down mode where the display & fan are off, as is most of the circuitry, but having a scavenged power mode that keeps the LM399 reference biased on to assist in "aging". Understand why Siglent doesn't use a preselected and aged LM399, too expensive for this price range DMM, but this could help the user "age" the reference without keeping everything else active, including the fan.

Anyway, just a thought.   

Best,

[Joop01]

Which brand depends greatly on what you want to measure. I'm busy with sound/audio and in that area Siglent is not the brand to have. Not mentioned in the specs is the input resistance which is going down with climbing frequency. With 10kHz already 160kOhm. That will introduce a big error. You can find this somewhere in the back of the manual, but not in the datasheet.

[BillyO]

The SDM3055 has considerably higher specifications.  Probably due mostly to the precision ceramic voltage divider network and better voltage reference as opposed to some SMD resistors like the SDM3045X.  I personally went through this very same decision nearly a year ago and chose the SDM3055.  I don't find the fan in it very noticeable but then again I'm quite hard of hearing.  However, I'm enjoying the additional precision and accuracy.

[mawyatt]

Yeah, think this is a little different than "Enabling" a SDS2000X+, SDG2042X, or SDP3303X, since apparently there are some hardware differences between the SDM3045 and 3055.

Even if it's "just" the voltage reference and some resistors; precision, stable references and resistors can get expensive, and real precision, gets real expensive, real quick

Best,

[zepto]

Which brand depends greatly on what you want to measure. I'm busy with sound/audio and in that area Siglent is not the brand to have. Not mentioned in the specs is the input resistance which is going down with climbing frequency. With 10kHz already 160kOhm. That will introduce a big error. You can find this somewhere in the back of the manual, but not in the datasheet.

You might find this in the back of a 34401a manual as well.

I like some of the offerings from Siglent and Rigol, but I have never considered purchasing any of their multimeters (or power supplies). Too many great 5.5/6.5+ meters were produced in the last 30 years to warrant buying new as a hobbyist.

[KungFuJosh]

Yeah, think this is a little different than "Enabling" a SDS2000X+, SDG2042X, or SDP3303X, since apparently there are some hardware differences between the SDM3045 and 3055.

Even if it's "just" the voltage reference and some resistors; precision, stable references and resistors can get expensive, and real precision, gets real expensive, real quick

Best,

SDM3055 and SDM3055X-E are two different models. SDM3055X-E is identical hardware to SDM3045X. The SDM3045X is crippled by the firmware, and swapping model firmware gives you the full 5.5 digits / 240,000 counts. There are a couple small hardware improvements for the SDM3055 version, but the question is if that's worth the extra $$ to the OP.

I've been running the SDM3055X-E firmware on my SDM3045X for over a month with no issues. 5.5 digits, and no issues with the calibration or accuracy.

Thanks,
Josh

[tautech]

Yeah, think this is a little different than "Enabling" a SDS2000X+, SDG2042X, or SDP3303X, since apparently there are some hardware differences between the SDM3045 and 3055.

For western models yes, however we have proved a long standing hunch I had that western SDM3045X = eastern SDM3055X-E HW and only the SW/product ID is different.

Study of the thread linked earlier by KungFuJosh reveals the issues western SDM3045X has with ranging whereas eastern SDM3055X-E doesn't have such issues due to how the SW/OS has it operate....yet it's the same HW !
One only has to study/compare datasheets of SDM3045X and SDM3055X-E to see SDM3045X is a bit of an oddball with different measurement ranges to the other 2 western SDM3k models.....yet eastern SDM3055X-E has those same 2,20,200 range steps too.

[Anding]

Thank you all.  I learned a lot from these replies.  I will be placing an order for the SDM3055

[mawyatt]

I've been running the SDM3055X-E firmware on my SDM3045X for over a month with no issues. 5.5 digits, and no issues with the calibration or accuracy.

Thanks,
Josh

How do you know about the relative accuracy? Do you have a verifiable reference to compare against, or a known good higher resolution/accuraracy DMM to compare?

Extra digits are fine, but they need to "mean" something, rather than justing "being there"!! As we've said, "Real Precision, gets Real Expensive, Real Quick"!! If you wander over to the Metrology section you'll find a number of folks have investing significant time and $ in such.

Ultimately the "Goodness" of a DMM can be traced to the voltage reference and divider resistors, why some of the better ones employ LM399 or LTZ1000 references, some even aged and selected for low noise. The reference utilized in a given bench DMM is a good indicator of what to expect regarding performance in measurements (of course this can be hosed up with an overall poor or sloppy design).

The historic HP34401A actually has more digits available than shown on the display, some folks have addressed these "extra" digits and utilized such to extend the 34401A resolution beyond 6 1/2 digits, and from the results they've posted these aren't just "random digits"

Best,

[KungFuJosh]

It feels accurate. 🤣

I'm sure somebody with the resources and desire could test and/or calibrate this better than I could. I don't see any odd behavior or inconsistency from it, and it's not less accurate than my other meters. I could throw some kelvin probes on it and compare the 4W Ω test mode with my TH2830, but who knows if either of them are properly calibrated.

[mawyatt]

There are some reasonably priced voltage references like the one's from voltagestandard.com (VREF10-001 R9) that are characterized with a certified Keysight 3458A. They even offer a recalibration at a charge of $15 + shipping.

Best,

[KungFuJosh]

There are some reasonably priced voltage references like the one's from voltagestandard.com (VREF10-001 R9) that are characterized with a certified Keysight 3458A. They even offer a recalibration at a charge of $15 + shipping.

Best,

I want one of those, but they've been out of stock for a while.

I tried getting a couple reference devices from other sources, and they were crap. Either not marked with target numbers, or inconsistency across the board.

[Kleinstein]

The 3045 / 3055E use cheaper parts for the reference (Ref5025 vs max6225) and also cheaper single resistors for the divider and Ohm part.  So one can can expect more TC and more aging from those parts. However the noise from the amplifier and ADC, e.g. for measuring small voltages should be close. If one is after accuracy the relatively small price difference to the real 3055 can be worth it.

Checking the dift for the 3055E voltage and resistance can make sense, as the reference parts are relatively poor compared to the ADC. The ADC itself can give 6.5 digits and possibly even more. AFAIK this is available via RS232.

[tautech]

The 5/5 digit SDM3055 has been pushed to 7.5 digits....
Still a work in progress:
https://www.eevblog.com/forum/testgear/hacking-the-siglent-sdm3055/