Seeing I have this blog space which I promised to put some element14 related content on, and I have about half an hour that I can spend, I thought I'd take a look back on some of the RoadTests I've been involved in and see just what my opinion of the product is, one year after I started using it. After all, some products are all fine and great in the "honeymoon" phase, but after a bit of usage, you really get a feel for the developments, limitations and annoyances of the product and they're worth noting as well. It'd also be a good chance to provide a "where are they now" snapshot of how it's been used since reviewing the product.
This post will be looking at the Keithley 2110 5.5 digit Benchtop Digital Multimeter. This was supposed to be Keithley's lower-cost entry level broad-purpose multimeter, and in the original review, things were mostly good. Is this still the case?
During the review, I didn't have that much time to do teardowns and research, and back then, I really didn't want to break something I had only just gotten. While I still haven't torn it down, some trawling online led me to this discovery - the PicoTest M3510A is a dead ringer for the Keithley Model 2110! It's actually a 6.5 digit version of it, complete with the same "slightly small" LCD screen. In fact, the software for it is name PT-TOOL as opposed to Keithley's KI-TOOL. Co-incidence? I think not.
Further digging reveals that the product is likely to be OEM'ed by a company called Array Electronic Co. Ltd, which doesn't have this unit on their site anymore.
Does this diminish the "Keithley" feeling? Maybe a little. But in the end, many products at the low end of the price tiers are rebadged or compromise products, so this is not really a surprise. Ultimately, the reliability and performance is more important. What is surprising is that the 2110 doesn't seem to be that popular around here ... despite my initial thoughts that it would.
Since the review and its firmware update to 1.02-00-00, it seems three new firmware releases have been provided for the multimeter. The latest version of the firmware is now Version 2.01-02-01. Interestingly, when you visit the product page, you won't find the firmware update linked in the Software section anymore. Instead, you need to go to the support pages and search for 2110, where you will find the release notes and firmware, along with the software.
You will find that if you're running Firmware 1.02, that the upgrade for 2.00-01 is still available. This is because of the upgrade considerations table, which seems to suggest you should upgrade to 2.00-01 first.
Upgrading to 2.01-02 doesn't seem to be directly supported, based on their release notes, so I wouldn't try it.
On the whole, the firmware fixes issues with digit flickering (which didn't seem to affect my unit's hardware), and some SCPI command performance anomalies. No major "calibration" affecting bugs have been seen, which is good.
An update to the KI-TOOL software seems to have been released as well, which doesn't change the spartan look of the software, but does make some slight functionality improvements. One much appreciated improvement is the DMM's settings are read from the DMM on the software starting, so if you took the effort to pre-configure the DMM through the front panel, the settings are not lost. It also means that the settings page isn't "defaulted" to some random value, but instead, it takes the values presently programmed into the DMM.
The appearance and performance did remain the same. It was noted that only 10 recording slots were available, and there was a suggestion there was a limitation to the number of recorded values. It seems that I've been able to record over 100k values into a single slot with no difficulty, so that's a bit of a bonus. Exporting and deleting the data still takes a little bit of time. It seems that the recording is done fairly frequently too, as we have had a power outage during a test run and the data recorded was intact until about 10-seconds before the loss of power, which was a nice discovery. The trigger delay is still in software, so unfortunately, depending on the setting and the integration time, the actual sample rate is widely off expectation (e.g. 10PLC voltage + 10PLC current in dual measure at 50Hz + 100ms trigger delay would have you expecting 500ms per sample right? In reality, it's closer to 800ms!).
Unfortunately, I never really had the chance to try it with NI LabVIEW or MATLAB or anything else. Instead, it's spent most of its life running standalone or with the KI-TOOL software.
What is it doing now?
Since the RoadTest review, it has been heavily used as a data acquisition device for testing the capacity of USB power banks. These products are ubiquitous, but their true capacity sometimes deviates markedly from the named rating even when conversion efficiency is taken into account. This has led to my personal blog being one of the primary reference sources, especially for cheap and nasty power banks, along with ripple voltage testing with the Picoscope 2205A I RoadTested as well.
Over 15 different models of power bank were tested, with 15 runs each, a test time averaging about 4 hours per run ... so already in that set, we're seeing over 900 hours of usage there.
The Keithley has also been the star of an endurance experiment with two different Lithium Polymer battery banks - the Orzly (which fared terribly) and a Hillo Power (which is still running). A total of 190 runs were made, for a total test time of about 570 hours (due to smaller capacity cells). It is amazing to see the stability of the unit, which provided very very solidly correlated results - in most cases, subsequent cycling of the power bank and integration of readings taken >1/s over periods of 3-4 hours showed <100mAh of difference, and much of that can be attributed to the power bank charge termination itself. It makes a very good meter for checking up on small variations in capacity, especially when treated with care. I have always kept it powered on for at least 30 minutes before taking readings (warm-up) and kept it in my air-conditioned room (to keep the temperature stable). It's also not subject to vibration or movement, which helps.
In power bank testing alone, we're seeing over 1000 hours of usage, and the meter has been rock solid. Never did it drop-out once. Never did the software fail to do the recording. Never did it error out. Only when the electricity had a significant dip, did it once go into error (likely due to loss of synchronization with mains cycle), which is perfectly reasonable. The same can't be said by some PicoTest users who were complaining of problems, however, I did not use the GPIB interface as I don't have any interface cards (and USB is perfectly sufficient for me).
Aside from that, it formed the reference readings for my how accurate is that 3.5 digit multimeter? article, which helps provide some insight into just how hit and miss cheap meters are. The Keithley is rock solid though, and the readings it has provided has been transformed into knowledge for the greater community. How great is that?
It's been a year of use, and a tough year at that, but this multimeter has been nothing but solid. Despite the slightly sub-par LCD display instead of a VFD display, when used as a data acquisition unit, it doesn't really matter. The accuracy, overall, appears to have been maintained, and the USB SCPI link was rock solid through almost continuous usage. It never got any rest, and yet, it's still as bright and sharp as day 1. The software provided isn't perfect, but it's still adequate for some very involved experiments. For the price, it's a solid buy, especially with the dual-measure facility which I have made extensive use of.