Siglent Bench DMMs SDM3065X, SDM3055 and SDM3045X

[klausES]

Beeper on medium...
Tautech, you either seem to have hearing loss or on a street with a lot of traffic ?!?   

[klausES]

SDM3055 (1.01.01.20R2) just measured the diode test mode.

Clarity.
Maximum voltage at the output (i.e. for a measurement object) of the 3055 is limited to 6.3V (e.g. if the object is too high or open),
the measuring current of the 3055 from the smallest setting up to max. 4V is kept at exactly 0.998 mA.

Conclusion:
The 3055 definitely uses a constant current of 1 mA for diodes, which is maintained up to the maximum setting of 4V
(if the internal resistance of the measurement object allows this) and an additional voltage limitation of max 6.3V.

[OoVoO]

Great news that can check the LED up to 4v, thanks everyone!

I was based on the yotube video, and http://siglentna.com/wp-content/uploads/dlm_uploads/2017/10/SDM3055_DataSheet_DS06035-E03F.pdf but apparently the documentation has not been updated for a long time.

It remains to wait for a solution to the issue of the speed up of the buzzer when a short circuit is dialed

[CDN_Torsten]

One work-around for the sluggish continuity beeper is to set threshold to around 1k.  This gives a virtually instant response.

[klausES]

SDM3055. Now measured continuity.
The measuring current of the 3055 is also regulated at a maximum of 1 mA and the voltage is limited (too little or no load) to a maximum of 6.2V.

The beeper sounds quickly (in my opinion immediately).

At the moment I don't know how to measure it well (I'll try to measure the real time afterwards).

In the end, that's a very subjective thing.
One feels e.g. a quarter of a second as fast, the other may be upset by a delay of a tenth of a second... 

[klausES]

One work-around for the sluggish continuity beeper is to set threshold to around 1k.  This gives a virtually instant response.

I have to try that later if this can have an impact.

I use this not only to check for continuity but often in the opposite case,
whether e.g. a test setup nowhere has an unwanted bridging in the PCB without having to constantly look at the display.
And especially for checking for good continuity, I appreciate the possibility of a very low-resistance setting until the beeper sounds.

[CDN_Torsten]

I think for most uses, the speed of the continuity is fine.
The very fast response time is useful occasionally when 'looking' for a connection and the DMM leads are quickly dragged across many fine-pitched pins.  In this case, the default response time is a tad too slow.

[klausES]

...when 'looking' for a connection and the DMM leads are quickly dragged across many fine-pitched pins...

This is finally a good sensible objection to insist on a quick acoustic response. 
I didn't even think about this at this point.

I have already done this, with an old test buzzer, but never with the Bench DMM.

[Electro Fan]

The price on the 3055 definitely makes it attractive but I'd like to make sure it checks off all the basics, and I'm having a hard time figuring out what to expect from it on resistance measurements. 

For sub Ohm measurements either in Auto range or the minimum 200 Ohm range it could certainly be improved with shifting the decimal point to the left so to allow for more digits below the decimal point.
The 200 Ohm range needs be lowered to 100 Ohms where 99+ Ohms would be shown as Overload.
Latest current firmware.

Here a 4W measurement of a 5W 0.2 Ohm current sense resistor in the 200 Ohm range:

Thanks for posting the 000.204 ohms photo.

Good idea about shifting the decimal point to the left.  Any chance you can suggest this to Siglent?  Is it something that can be fixed with a firmware release?

Any chance you could post another showing a low dBm reading?  What’s the lowest dBm reading the 3055 will display?  Does it show a minus sign?

Thanks again

[edigi]

Thanks for posting the 000.204 ohms photo.

Good idea about shifting the decimal point to the left.  Any chance you can suggest this to Siglent?  Is it something that can be fixed with a firmware release?

A lot of other things would be more important to fix before this.
Stability (and accuracy) of the 4W resistance measurement is not very good for very low resistance (at least with 3065X) so for that reason it's better to use a different setup for this kind of measurement.

Let me show this via an example.
A very low resistance (~1.5mOhm, just a short piece of wire) is measured with 4W way and alternatively using a current limiting lab PSU set to 1A (just for the sake of easy calculation) and measuring the voltage at the same time (with 4W measurement basically the same is done but without control over the current).
The resulting trend chart I think tells it all and I trust the millivolt measurement more (that corresponds to mOhm) just looking at the stability....

[Kleinstein]

The price on the 3055 definitely makes it attractive but I'd like to make sure it checks off all the basics, and I'm having a hard time figuring out what to expect from it on resistance measurements. 

For sub Ohm measurements either in Auto range or the minimum 200 Ohm range it could certainly be improved with shifting the decimal point to the left so to allow for more digits below the decimal point.

There probably just is not more resolution to show. The test current is likely 1 mA and thus 1 µV voltage resolution for the last digitt. An additional digit would probably be noisy and far from accurate anyway. After all the 3055 is 5.5. digits only.

[tautech]

The price on the 3055 definitely makes it attractive but I'd like to make sure it checks off all the basics, and I'm having a hard time figuring out what to expect from it on resistance measurements. 

For sub Ohm measurements either in Auto range or the minimum 200 Ohm range it could certainly be improved with shifting the decimal point to the left so to allow for more digits below the decimal point.

There probably just is not more resolution to show. The test current is likely 1 mA and thus 1 µV voltage resolution for the last digitt. An additional digit would probably be noisy and far from accurate anyway. After all the 3055 is 5.5. digits only.

alexvg and her interesting research thinks otherwise:
https://www.eevblog.com/forum/testgear/hacking-the-siglent-sdm3055/
Unfortunately on hold while she helps with the current health crisis.

[klausES]

...Stability (and accuracy) of the 4W resistance measurement is not very good for very low resistance (at least with 3065X)...
...A very low resistance (~1.5mOhm, just a short piece of wire) is measured with 4W way...

In this very low ohm range, "even with 4-wire operation", the quality of the contacting of the terminals is extremely important.

A second pass with only slightly changed clamps (just plugging in and reconnecting) can produce completely different results.
And also the stability, even during one and the same measurement (you can almost say not cough, don't hit a door, don't wobble at the table...)
depends extremely on the quality of the contact.

When it gets very low-resistance I have started to connect loose, not built-in objects with screw terminals.

[edigi]

In this very low ohm range, "even with 4-wire operation", the quality of the contacting of the terminals is extremely important.

I've used exactly the same probing for the 2 measurements of my 2 screenshots. Actually I didn't even touch the probes that contacted the "resistor" to be measured as I plugged to the PSU the other end of the cable.

The only difference was the the higher current used with the measurement. With that the DMM was no longer forced to measure voltage near to the noise floor. Btw. the stability of the DMM is probably better and when I write noise floor I mean the entire setup. So the noise floor of the DMM is probably better, but my setup (mostly cables) were picking up noise. I could have used better probes but that's the hard way (and expensive as well) to improve measurement accuracy and stability.
The easy (and cheap) way in this case is to use higher current.

Talking about improvements, some ideas to Siglent related to screenshot creation.
First: It would be better if I don't need a deep dive in the menu system just for creating a screenshot. If a dedicated button is not possible (because of the current labelling) a 2 button shortcut would be great still. As a (non-negligible) side effect, the screenshot would be probably showing the measurement setup menu and not the storage menu that is always shown now (and bears little value to me).

Second: The confirmation request for screenshot saving is really annoying. Is that really such a huge risk that by accident an extra screenshot is created? This resembles me to the age of Windows 3.1 or so when everything was asking confirmation that if I really really really want to do some trivial and low risk operation.

Third: If I plug in a USB memory stick (or storage) I usually do it on purpose. I would be so great if in this case the device defaults to that so I don't need another deep dive each time in the menu system to switch from internal storage to external. Your mileage may vary, but for that could be another option what should be the default behaviour.

[pipe2null]

I received my SDM3065-SC a couple days ago.  So far I haven't had time to use it for anything real, but I've been playing with it a bit.  I did a quick measure the resistance of a short, just clipping a couple leads together so there was something to measure and watching the trend chart scroll by with the statistics displayed underneath.  It varies more than I thought it would, 0.0411 to 0.0449 ohm for a couple bran new (literally first time connected to anything, still has that fresh silicone smell) test leads clipped together.  I hope the uA range doesn't have the same variance.

General first impression is the UI has some fixable awkwardness.  The depth of menus to get to the screen shot feature is a good example.  There is that nice bright blue, tangible, physical "Shift" button on the keypad, but it is under utilized, no effect on any on-screen menus or under-screen button usage.  The numbering and ordering of "Probe Hold" measurements displayed is counter intuitive.  It makes sense up to the 8th measurement, but then it keeps the "1st/2nd/3rd measurement" numbering but shifts the measured values up for the 9th and beyond measurement.  If they changed to ordering of the measured values to always have the most recent first on the list, or otherwise increment the measurement numbering to reflect the actual Nth measurement, it would make much more sense.

The layout of the scanner card seems odd to me...  I was planning on making a breakout box with a bunch of banana jacks and using ribbon cable or something similar to keep it all nice and neat going from breakout-to-DMM, but with the board's layout and terminal block choice I'm thinking using ribbon cable will be a PIA.  Also, would be nice if front panel measurements were exposed as "channel 0" when in scanning mode so you have one channel with full V and A ranges.  Not sure if the front panel technically counts as 1 channel plus a second dedicated current channel, so perhaps "channel 0" is not sufficient...

The custom sensor feature is pretty nice, for my purposes anyway, but it needs work.  Same UI awkwardness, and you have to navigate and reload your custom sensor file every time you switch to any other function (like DCV/DCI/etc).  For built in supported temperature sensors, the last one selected is always the default, but not your custom sensor.  You can use arbitrary current, voltage, or resistance based sensors, specify what custom unit you want displayed, and it will automatically prefix magnitude (m/u/n/etc).  I'll be trying it out more in the next couple weeks.


But I'm happy so far.  At the price with a 16 channel scanner card, my nitpics are easy to work around.

[Kleinstein]

With very low resistors it is just a very low voltage to measure. With likely one 1 mA the 0.1 mOhms digit corresponds to 0.1 µV and at this level some noise is normal. It is not just the meter, but also changes in external thermal EMF. So one may have to protect the resistor to measure from air drafts and not hold the cables in the hand.

More current helps, but it needs quite some extra HW, especially the protection gets more complicated.

 

[klausES]

...just clipping a couple leads together so there was something to measure and watching the trend chart scroll by with the statistics displayed underneath.  It varies more than I thought it would, 0.0411 to 0.0449 ohm for a couple bran new (literally first time connected to anything, still has that fresh silicone smell) test leads clipped together.  I hope the uA range doesn't have the same variance...

To see what really happens with a short-circuited input, it is better to use a direct bridge (no cable) with well-contacting resilient surfaces.

Congratulations on the SDM with scanner card. 
Since it is not bought so often with SC, you don't hear very much experience with it.
I wonder what they will report with the handling of this card.
(I am thinking about buying a second 3055 with SC in the distant future...).

[Electro Fan]

Still interested in case anyone with a 3055 can post a photo showing the lowest dBm reading that will display on a 3055.
Thanks

[klausES]

Screenshots SDM3055 on the fast.
I don't know whether this is the lowest possible dBm value ?

[flash2b]

Ok, just tested with 1.01.01.16R2 and with 1.01.01.19 (unfortunately I didn't have 1.01.01.20 and 1.01.01.20R1).

With these two older versions, this problem does not exist.
Saving works perfectly with these.

So clearly a bug in 1.01.01.20R2 

A big thanks for your efforts on this. 
High importance email going to Siglent with your findings. 

Any news on this, since I like the logging function (1.19) and also the new bigger font in dual mode (1.20R2).

[klausES]

Unfortunately so far nothing new whether results will come soon.

Since I can not do without some functions of the 1.01.01.20R2, I have been living with the errors since then.
Logging and screenshots so long, unfortunately, only via the computer.

[pipe2null]

@klausES:  I'll post scanner card info when I build my breakout for it. I have some printing experiments to do before I finalize my faceplate, so prob a couple weeks-ish.  I'm not sure what "best practices" exist for such a thing.  I'm debating "installing" a many-position connector mounted on the back of the DMM (the scanner card has an external aluminum handle/wire tie ring on the back) but I'm not sure how much negative effect that will have on measurements.  But then I wouldn't need to pop out the scanner card if I needed to move or maneuver the breakout around and reroute wiring in and around my bench/shelves...  Or use different breakouts for different uses/projects/fixtures.  TBD I guess.

[klausES]

As long as you don't have to measure in really low µV ranges in which thermal voltage plays a role, the other 4 or 2 mm laboratory standard sockets will not be a problem.

How about routing the cable from the card into another small flat housing with the sockets, cable lengths just so that you can e.g. could stand next to or under the DMM.
The socket housing may snap into a socket or rail that is screwed to the table / shelf so that it cannot slip away when the plug is inserted / removed.

(I'm building something like this as an external scanner for my 3055 without SC).

[pipe2null]

Has anyone messed with the new "Custom Sensor" feature of the 3.01.01.07 firmware for SDM3065X?  I haven't found documentation on it yet.  I think I've figured it out but I'm getting some odd behavior, so I'm not sure if it's just me or if I've found a bug.

I'm measuring a photodiode in photovoltaic mode, so zero applied voltage.  Spec says 0.4A/W response with 1.2mm^2 active area.  I measure it with regular DCI mode on 3065 DMM and get +/- 0.8nA when the sensor is in a dark box, and around 1.5uA with sensor pointed at ambient lighting.  Those numbers seem reasonable to me with a little noise around zero.  Turning on statistics with averaging puts "zero" at about 0.1nA, which is decently accurate in my book.

With the custom sensor mode, you can select DCI as the sensor type, which I did, and you enter a few Measurement->Result data points into a table for the DMM to interpolate, which I did.  But the sensor results I'm seeing appear to have 2 separate issues that might be bugs.  First, the interpolated displayed result of a current sensor measurement in ambient light seems close but not accurate to what I calculate it to be.  This might be a misunderstanding on my part, but I havent found docs on this feature yet.  The other issue is when you change Temp mode (with custom sensor selected) to display both measured value and interpolated result, the current reading from the sensor is showing around -170nA for "zero light", which I already measured with the same DMM at 0.1nA.  Changing back to normal DCI mode instead of Temp mode shows the expected 0.1nA.  The other issue I found is not enough significant digits in the table entry, +/-000.000001 is the smallest you can enter, so I can't just manually force an offset adjustment to work around the wierd "-170nA" reading issue.  I'm not positive, but auto-zeroing seems like it might be turned off in Temp mode with a DCI mode sensor.

So...  Am I hitting bugs, or am I just not doing something right?

[Tjuurko]

Firmware bug, with the newly added PT1000 thermistor the SCPI command "CONF:TEMP RTD,PT1000" is missing

TEMP:MDEF:RTD:TRAN PT1000

TEMP:TRAN?
>> MANU DEFINE,RTD,PT1000