Siglent Bench DMMs SDM3065X, SDM3055 and SDM3045X

[rfclown]

The trigger is staying in the immediate state (verified with TRIG:SOUR?). If I use "INIT" followed by "FETCH?" it does the same thing. I see the display update when I send the INIT, but the display stays constant afterwards. Yes, like the trigger is stopping.

[tautech]

Does anyone know how to programmatically put the DMM back to LOCAL after making a remote reading? I'm using LAN right now. I searched the manuals and Googled to no avail. We just got some SDM3045Xs at work. When automating measurements I like my instrument displays to keeps updating when I'm not making readings, which for some instruments requires sending a command to go LOCAL after making a measurement. Only way I've seen to do this so far is by sending a *RST (which resets everything).

Please quote the firmware version installed ?
Strongly advise to update to latest if not already:
https://int.siglent.com/upload_file/zip/firmware/A_multimeter/SDM3045X_5.01.01.07R1.zip

Also what SW you're using for remote control ?

[rfclown]

SW Ver: 5.01.01.01.05. I'll look at updating this to .07. Thanks for the link. I just downloaded it.

I'm using NI VISA to communicate. For high level language I use LabVIEW and C.

[tautech]

SW Ver: 5.01.01.01.05. I'll look at updating this to .07. Thanks for the link. I just downloaded it.

I'm using NI VISA to communicate. For high level language I use LabVIEW and C.

You need to.
There have been a couple of versions since 05 that have SCPI bugs fixed including the latest 07:
Release notes
2. Optimize flash R&W design.
3. Fix some UI & SCPI bugs.

Both those fixes are important to have installed.

[rfclown]

updated firmware to 7R1. Same. Some instruments just don't have the ability to programmatically go back to LOCAL. One of my pet peeves.

One thing I discovered earlier with my Siglent SDS-1104X-E at work was that in LabVIEW I needed to specify "Synchronous I/O Mode" on the VISA vis. I've been remotely controlling equipment since about 1988 (mostly HP gear) and I'd never run across the Synchronous/Asynchronous thing before. I've had instances where I had to fiddle with the GPIB Mode for some old gear. When I use the NI VISA calls in C, I don't have to specify this, but in LabVIEW, the default for VISA write and read vis is Asynchronous. I guessed correctly that I have to use Synchronous on this DMM just like the scope. I think there was something in the Siglent scope documentation that mentioned this.

[Johnny B Good]

 Hi one and all,

 Ever since I acquired an SDL1020X-E (now upgraded to 1030X spec) just over a week ago, I've developed a sudden and keen interest in K type thermocouples and how well Siglent have implemented their "Isothermal" cold junction reference in the SDM3065X.

 Failing to find any in depth tear down videos or even just write ups, I started a more general DDG search and came upon the service manual here:

https://int.siglent.com/u_file/document/SDM3065X_ServiceManual_SM06036-E02F.pdf

 Which offered this little nugget hidden away on page 14 of this 36 page pdf (read it and weep).

====================================================================================
[3] Relative to cold junction temperature, accuracy is based on ITS-90. Built-in cold junction temperature
refers to the temperature inside the banana jack and its accuracy is ± 2.5 °C.
====================================================================================

 That's a pretty lousy performance and suggests they're using a cold end temperature sensor well removed from the banana jacks of which they speak.

 I mean  If a cheap Chinese knock off of the Hakko FG-100 can track well within 0.5*C of three separate "Weather Station"'s indoor thermometers, you'd think Siglent might have taken just a modicum of care to get this right and at least place the NTC resistor sensor in between the relevant banana jack sockets to keep the cold end junction compensation error minimised.

 That statement strongly suggests they've placed it well away somewhere on the mainboard or, Gawd forbid! resorted to the one built into the main cpu with a compensatory fudge factor applied ("fudge" being rather apt if such a Half Arsed method has actually been employed).

 The fact it appears to be under-reading only by about 0.6*C as best as I can tell, using the K type TC that came with a BSIDE ACM91 I'd bought primarily for its DC clamp meter's 1mA resolution, would appear to be down to 'dumb luck' in my example.

 I think most thermocouple based thermometers manage to sense the cold junction temperature to well with 0.5*C so it's rather a disappointment to see Siglent take such a lackadaisical attitude to the temperature measurement functions. This is made all the more egregious by the fact that the temperature display is irritatingly cursed with some two to four meaninglessly dancing digits to the right of the decimal point.

 Right now, I've only got the one K type TC (fortuitously terminated with banana plugs) and a bunch of spare TCs for that FG-100 soldering iron temperature calibration aid. I've got a 5 metre coil of fibreglass and stainless steel insulated TC wire to make my own K type probes along with a couple more K types (another banana plug ended probe like the BSIDE one and the other with the mini TC polarised plug supplied with a cheap TM-902C Meter on order from eBay sellers. Hopefully, I'll be able to run more tests with another two K type TCs by the start of next week (the 5 metre coil of TC wire is on a slow boat from China).

BTW, what had prompted this sudden interest in K type TC probes was the fact that when I was stress testing the SDL1030X at 309W, I was reading heatsink temperatures of 90 to 92 deg C with no hint of the OTP kicking in, despite Siglent specifying the threshold temperature had been set to a mere 85*C, leaving me to ponder just what this OTP trip out temperature is actually set to (poor calibration or a design decision to go some ten degrees higher and not bother the user guide's copy writer with an update?).

[2N3055]

Yeah sounds not really good at first sight.. but..
Thermocouples are not accurate at all, most of them. They will have lousy absolute accuracy unless you shell out for really good ones.

For more accurate measurements at moderate temps RTD and thermistors is a way to go.

Keithley DMM6500 doesn't even have any junction compensation built in in base meter.

Keysight 34461A doesn't have thermocouple measurements at all. You need to upgrade to 34465A to get it, and then yeah it goes like this:

K, J, T, E, N thermocouples Probe accuracy + reference junction accuracy + 0.3 °C
R thermocouples 14 (250 to 1760 °C) Probe accuracy + reference junction accuracy + 0.5 °C
The internal reference junction uses the U1180A or equivalent adapter. This has a typical performance of ± 1.0 °C. This internal reference junction can be adjusted for better accuracy. An external reference junction can also be used.

So however, you need to calibrate for the thermocouple you'll be using because those will vary few °C between each specimen.

What matters is how meter measures after you calibrate for the thermocouple, i.e. after you adjust how good is relative measurement. And that seems to be more than OK in SDM3065X.

None of them are really brilliant for temp measurement. That is one measurement that benefits from highly specialized instruments. Also, nowadays, you'll pretty much get better and from the factory calibrated guaranteed specs by using one of the several high accuracy integrated temp sensors. At which time it would be a good idea to get humidity and barometric pressure too..

[Kleinstein]

====================================================================================
[3] Relative to cold junction temperature, accuracy is based on ITS-90. Built-in cold junction temperature
refers to the temperature inside the banana jack and its accuracy is ± 2.5 °C.
====================================================================================

 That's a pretty lousy performance and suggests they're using a cold end temperature sensor well removed from the banana jacks of which they speak.

 I mean  If a cheap Chinese knock off of the Hakko FG-100 can track well within 0.5*C of three separate "Weather Station"'s indoor thermometers, you'd think Siglent might have taken just a modicum of care to get this right and at least place the NTC resistor sensor in between the relevant banana jack sockets to keep the cold end junction compensation error minimised.

The +-2.5° C has to be seen as a spec limt, while the 0.5° C seem to be a measured deviation. Spec limits are often 2 to 3 times the measured variations and may be more conservative in some cases. So the difference is not really large.

No matter what a thermocouple reading with banana plugs is never really accurate. The cold junction measurement is just not working well with these plugs. 
If you need a accurate measurement with a thermocouple, use a seprate thermometer with proper therminals.
These special thermocouple plugs are however not safe for higher voltage. So there is no easy way to have that with a normal DMM. The best one can find is a slot for a scanner card that than includes a good cold junction compensation. External cold junction compensation is another option, though this would need support on the side of the meter or manual conversion from voltage to temperature.

[tautech]

In the Temp menu there is a feature to add custom sensors and tune them for accuracy however at this time there is little documentation about this but we've asked Siglent to prepare an App note on it.
Some info here:
https://www.eevblog.com/forum/testgear/sdm3045x-versus-dm3058e-sensor-input-4-20ma/msg3757376/#msg3757376

[Johnny B Good]

 Thanks Guys!

 I'm all too familiar with the practice of showing much slacker worst case tolerance values in datasheets than the end customer is more typically likely to see. It's known as "Covering your backside".

 I assumed this was the case here but even so, that +/- 2.5*C seems stupendously wide imo. However, late last night before retiring for a night's sleep, I noticed that the Siglent had gotten to within 0.2*C of the cheap inside/outdoor thermometer which was itself within 0.5*C of the two "weather station" indoor temperature readings.

 This morning, all three indoor temperature readings were within 0.1*C of each other (18.1*C) and the temperature readout after a minute or so of switching the Siglent meter on, was showing just 0.2*C higher (on average - pesky digit dance!) but, of course, after the obligatory 30 minute warm up, it had drifted upwards (18.7ish*C versus the cheap indoor/outdoor thermometer's 18.2*C right by the TC tip but another two hours later it was adrift by -1.4*C on that indoor/outdoor thermometer's reading.

 The ITWorks Wx station is now showing 23.0*C whilst the cheap 'n' cheerful Atech (doesn't show barometric data like the ITWorks unit) shows 23.9*C (the indoor/outdoor thermometer shows 23.0*C). The room has warmed up some 5*C over the past two and a half hours due to the XYL wishing to have the CH turned on and the temperature readout differences are simply due to the variation in thermal inertia between the Wx stations and that indoor/outdoor thermometer.

 Over the years of ownership of the ITWorks and Atech Wx stations, I've become resigned to the accuracy limits of temperature sensors (particularly the outdoor ones due to actual temperature differences) and don't expect any of these to be better than +/- 1*C at best (none of them are meteorological grade units).

 That very generous +/-2.5*C tolerance Siglent gave themselves made me wonder just exactly how they were implementing the "Cold End" compensation in the SDM3065X. From the way it drifts during changes of room ambient, it looks like the cold end sensor isn't where it should be (between the banana sockets) and might even be inside the cpu with a compensatory 'fudge factor' applied.

 If it's the latter, then "It is what it is". Otoh, if they've simply used a badly placed thermistor, that would give me the option to relocate it close(r) to the banana sockets in three years time when the warranty runs out.

 In the meantime, I've got a TM-902C Meter with K type probe winging its way to me from a UK based seller. I rather think this will provide better accuracy and stability than the SDM3065X has to offer with K type probes. I'll see whether this will be the case though soon enough.

[HKJ]

As already stated banana plugs are not the best way to connect thermocouplers, but neither is cheap thermocoupler connectors and I would not expect cheap thermocoupler meters to be good either (The few I have looked inside do not place the compensation sensor near the connector).
Generally the problem with cheap thermocoupler connectors is that the two pins on the plug is the same metal and not thermocoupler alloys.

I made me own thermocoupler devices where I had that problem: https://lygte-info.dk/project/ThermoSensor%20UK.html

[Kleinstein]

The relatively cheap (e.g. yellow plastics for type K, the color corresponds to the elements) flat pin thermocouple connector are actually quite good, as they use proper alloy pins. They are available in a small and large typ.   I would definitely prefer them over the rather expensive Lemo ones with gold plated copper alloy contracts and meatal shield.  One can do the test, joining a TC cable with a connector pair and measure the voltage, with both end of the TC keept at a constant temperature and than warm up the connector: this usually gives very little voltage (e.g. 1 µV range).

[HKJ]

The relatively cheap (e.g. yellow plastics for type K, the color corresponds to the elements) flat pin thermocouple connector are actually quite good, as they use proper alloy pins.

Brand name models probably do, but the 1$ Chinese do not and they are the ones supplied with cheap meters.

[Johnny B Good]

 Regarding cheap K type thermometers such as that TM-902C I mentioned, I came across this YT test and teardown video (about the best of the many YT video reviews of this ubiquitous meter!).



 Looking at the PCB in that vid, there doesn't seem to be a discrete NTC near the socket but with a power consumption likely less than 100mW (possibly much less - something I'll be checking out on the one I've ordered) using a reference junction temperature sensor integrated into the COB ic is probably going to introduce less error than that of SDM3065X.

 I've since done some more testing and I now have reason to believe the SDM does use a sensor close to the banana jacks - it's just unfortunate that the 13W's worth of heat (fan notwithstanding) creates a thermal gradient of some 2*C between the banana socket's solder tags and the external ambient. However, this should be a relatively fixed temperature offset which could be adjusted for if only Siglent had included an option to calibrate for this effect.

 As has already been suggested, a K type 4mm banana to mini adapter might prove an effective solution (with mini jack terminated thermocouples, of course!). I'll have a better idea as to whether such expense is warranted once I have a couple more K types to compare with (and there's always that 5m coil of K type cable to experiment with when that finally lands on my door mat).

 The trouble is not that there aren't any teardown reviews (video or text only) which delve into this aspect of the SDM3065X so much as the fact that no one has posted any tear downs of any kind on this bench meter.

 You'd think YT would be awash with such tear down vids by now but so far so absent.

[tautech]

New SCPI command:
https://siglentna.com/operating-tip/sdm3000-remote-scpi-command-to-request-a-screen-image/?pdf=23982

SDM3000 remote SCPI command to request a screen image

The SDM3045X, SDM3055, and SDM3065X can return bitmap images of their displays by sending the following command:
SCDP

Also applies to all SDG AWG's.

[TheDefpom]

I actually wish that my SDM3065X didn't lock the front panel when using the remote interface at all.

I wanted to control and send the readings from "Test Controller" to display on screen when I live streams, but also I need to be able to control the meter from my desk as I am working on something, as it stands right now I have to choose between using the meter at the bench OR having the readings in test controller (or any other SCPI control interface).

Rob can you please put a request through to Siglent to have them allow local AND remote control simultaneously, so it just responds to controls from whichever it gets a command from, whilst also updating the screen and sending SCPI readings at the same time.

[oz2cpu]

+1 for that !!
scopes dont do this, i can pc remote, AND adjust on the scope at the same time,
but multimeters, powersupplies, signal generators, dont alow this, when in remote, they are all locked,
but why ??

[tautech]

I actually wish that my SDM3065X didn't lock the front panel when using the remote interface at all.

I wanted to control and send the readings from "Test Controller" to display on screen when I live streams, but also I need to be able to control the meter from my desk as I am working on something, as it stands right now I have to choose between using the meter at the bench OR having the readings in test controller (or any other SCPI control interface).

Rob can you please put a request through to Siglent to have them allow local AND remote control simultaneously, so it just responds to controls from whichever it gets a command from, whilst also updating the screen and sending SCPI readings at the same time.

Done.
They go on CHY soon so it might be a couple weeks before we get any reply.
2022, the year of the tiger !

[2N3055]

+1 for that !!
scopes dont do this, i can pc remote, AND adjust on the scope at the same time,
but multimeters, powersupplies, signal generators, dont alow this, when in remote, they are all locked,
but why ??

That has been industry norm for many years. It comes from fact that if you have some remote control script that is in control of instrument inside some automated procedure (test for instance) you want front panel locked so nobody can do screw with your test protocol by mistake. Scopes don't do that by default, but you have command to lock them too....
Or rack mount scopes that don't even have front panel and screen for dedicated test...

Pretty much all of them have "Local" button for that purpose. If you press "Local" you get your control back.

[TheDefpom]

+1 for that !!
scopes dont do this, i can pc remote, AND adjust on the scope at the same time,
but multimeters, powersupplies, signal generators, dont alow this, when in remote, they are all locked,
but why ??

That has been industry norm for many years. It comes from fact that if you have some remote control script that is in control of instrument inside some automated procedure (test for instance) you want front panel locked so nobody can do screw with your test protocol by mistake. Scopes don't do that by default, but you have command to lock them too....
Or rack mount scopes that don't even have front panel and screen for dedicated test...

Pretty much all of them have "Local" button for that purpose. If you press "Local" you get your control back.

I want it to be selectable, right now it is one or the other, not both, there is no reason why the *LOCK command cant be used when needed to lockout the front panel in those situations as has been done on other gear.
Some of my other Siglent gear allows you to choose whether or not you lock the front panel when using remote, I cant remember if it was the SDL1030X or the SPD1168 that did it, if you didn't lock it you could use both local and remote at the same time, if you lock it then it is remote only unless you force it into local.

[oz2cpu]

EXACTLY Scott,  going local, will loose your remote data feed/control = sad

[tautech]

Run/Stop is the one free button that doesn't have a Shift function, maybe some innovative function can be applied to it. 

[2N3055]

+1 for that !!
scopes dont do this, i can pc remote, AND adjust on the scope at the same time,
but multimeters, powersupplies, signal generators, dont alow this, when in remote, they are all locked,
but why ??

That has been industry norm for many years. It comes from fact that if you have some remote control script that is in control of instrument inside some automated procedure (test for instance) you want front panel locked so nobody can do screw with your test protocol by mistake. Scopes don't do that by default, but you have command to lock them too....
Or rack mount scopes that don't even have front panel and screen for dedicated test...

Pretty much all of them have "Local" button for that purpose. If you press "Local" you get your control back.

I want it to be selectable, right now it is one or the other, not both, there is no reason why the *LOCK command cant be used when needed to lockout the front panel in those situations as has been done on other gear.
Some of my other Siglent gear allows you to choose whether or not you lock the front panel when using remote, I cant remember if it was the SDL1030X or the SPD1168 that did it, if you didn't lock it you could use both local and remote at the same time, if you lock it then it is remote only unless you force it into local.

Scot,
I'm not opposed to idea, just commenting on state of the art...
User control is a good thing, except where it is obviously dangerous.
Best,
 Sinisa

[TheDefpom]

That is fine, I can see exactly where you are coming from, I just think the ability to choose needs to be there to allow for more usage scenarios.

[tautech]

Scott's use case is unusual however not unreasonable in that he's needing a SCPI graphic on his PC while filming of his repairs and such for his YT channel where the DMM is within reach to change its settings (within reason) and still have them displayed on his video capture.

He and I talked about this requirement a while back and Scott started writing SW to meet his requirements but discovered HKJ's excellent TestController saved him a lotta work especially when HKJ also tweaked some features to satisfy his needs.
https://www.eevblog.com/forum/testgear/program-that-can-log-from-many-multimeters/