The Keithley 2700 multimeter is not a data logger until you add a switch card. If you've been searching for a 'Keithley data logger,' that's the sentence to remember. After five years of buying lab equipment—roughly $400,000 a year across 15 vendors—I'd also tell you that the 2700 is a great choice for some labs and a poor choice for others.
Let me rephrase that. The 2700 is a 6.5-digit multimeter with a slot inside for a multiplexer card. The base unit measures DC voltage, AC voltage, current, resistance, and frequency. Add a card such as the 7700, and it can scan thermocouples or voltage signals across many channels. Without that card, it's just a very good bench meter.
What most buyers miss is that the scanning architecture matters more than the accuracy spec. A 6.5-digit meter with a 20-channel relay card is not the same as a 20-channel data logger. Each channel has to settle before it is read, so the scan rate is a real constraint. If you need all channels sampled at the same instant, you need a different instrument entirely.
Data logger vs. data acquisition: why it matters
It's tempting to treat 'data logger' and 'data acquisition system' as the same thing. They aren't. A data logger is usually a self-contained instrument that reads channels at a set interval and stores values locally. A data acquisition system is usually modular, with more channels, higher speed, and software for real-time analysis.
The Keithley 2700 sits in the middle. It is a meter first and a logger second. It can scan through relay cards, but the scan speed depends heavily on the card configuration and the measurement range. That's fine for a battery discharge test or a temperature soak. It's not fine for capturing a switching transient.
What I check before buying a Keithley data logger
I'm not an engineer, so I can't speak to the noise floor or the fine points of offset-compensated ohms. What I can tell you from a purchasing perspective is to read the order configuration before you sign. When I took over purchasing in 2021, I became the person who reports to both operations and finance. That means I care about completeness, not just price.
In 2023, I approved a quote for a 'Keithley 2700 multimeter' and didn't notice that the switch card was not listed. The engineer who requested it assumed the card was included. It wasn't. The system couldn't do what we needed until we spent more money and waited another two weeks for the card. That error is now a checklist item.
Here's something vendors won't tell you: the first quote for a Keithley data logger is often for the mainframe only. The line item may say 'model 2700' and look complete. It is not. Ask for the model number of every switch card, terminal block, and cable. The card part number is as important as the mainframe part number.
Per FTC advertising guidance (ftc.gov), claims have to be substantiated. I read Keithley accuracy specs the same way. A figure like '0.002% of reading' means nothing without the temperature range, warm-up time, and calibration interval that follow it in the spec sheet.
Where the Keithley 2700 multimeter still makes sense
The 2700 has been around for a long time, and that's a feature. The 7700-series cards are fairly easy to find, and the programming commands are well documented. If you have a small test rack with a few thermocouples, or if you need 6.5-digit DC voltage logging at one or two points, the 2700 is a solid, boring choice.
The way I see it, the 2700 is the lab equivalent of a workhorse pickup truck. It's not flashy. The front panel is clunky. But it starts up, reads accurately, and doesn't lose calibration every time someone looks at it. For a small lab that needs reliable logging without a dedicated DAQ, that's enough.
I'd also choose it if you need a unit that can double as a bench multimeter when the logging task is done. In our lab, the 2700 spends most of its time logging battery discharge curves. When the test is over, we use it as a plain meter.
Put another way: the 2700 works best when your application matches its strengths—low-channel-count, direct-current electrical measurements, thermocouples, and long unattended logging at moderate scan rates.
When I'd skip the 2700
If you need high-speed transient capture, or if you plan to scan more than 60 channels, skip it. A dedicated data acquisition system with per-channel ADCs is a better fit. I'd also skip it if you need simultaneous sample-and-hold: relay scanning can't do that. And if your engineers expect a modern touchscreen interface, the 2700's front panel will feel dated. You can control it over GPIB or serial, but don't expect app-style usability.
If you're starting from scratch and want a Keithley data logger without old switch cards, I'd look at the DMM6500 with a 2000-SCAN card. It's a newer 6.5-digit meter, and the scanning is easier to set up through the touchscreen. The 2700 still makes sense if you already own 7700 cards or if you want the older, widely documented programming commands.
Inductive sensor M12, CMM, and other requests I get
Once people know you buy test equipment, the requests get more varied. The same week I ordered a Keithley data logger, an engineer asked for an inductive sensor M12 and a CMM probe. These are not Keithley accessories, and they don't belong on the same PO.
An inductive sensor M12 is a cylindrical proximity switch used in industrial automation. M12 refers to the mounting thread diameter, not the sensing distance. Buy it based on rated sensing distance, output type (PNP or NPN), and whether you need normally open or normally closed contacts. I've made the mistake of ordering by thread size alone.
For an M12 inductive sensor, a flush-mounted version has a shorter but more stable sensing distance. A non-flush version can reach farther but needs more clearance around the sensing face. If you install it wrong, the sensor may trigger when it shouldn't. The data sheet will give you the derating factors for steel, aluminum, and other targets.
CMM—coordinate measuring machine—is a different world. CMM styli have specific thread sizes, lengths, and sphere materials. If you order a cheap set without checking the machine's probe head, you'll end up with calibration errors. I'm not a metrologist, so I can't recommend specific stylus specifications. I can recommend asking the CMM manufacturer for the exact part number before you buy.
One more CMM note: don't assume 'CMM' means a computer-controlled machine. Some shop-floor CMMs are manually operated. The stylus thread size still matters. If the requester can't find a part number, ask for a photo of the probe head. That sounds obvious, but it saves a return.
When to change your columns: HPLC Agilent and other brands
The other question I keep seeing is 'when to change your columns HPLC Agilent?' The answer is: change it when the column no longer meets your system suitability criteria. Don't change it just because someone says it should last a certain number of injections.
For an Agilent column, or any reverse-phase HPLC column, the typical warning signs are an increase in backpressure of roughly 15% from a clean baseline, a tailing factor above 1.5, retention time shifting by more than 0.5% between injections, or a visible drop in theoretical plate count. Agilent's column care guides list similar indicators. I use them as a starting point, but the final decision belongs to the chemist running the method.
I should add: 'when to change your columns' is not the same as 'when to clean them.' Many reverse-phase columns can be flushed with a strong solvent to remove contamination. But if the inlet frit is blocked or the packing bed is voided, cleaning won't bring it back.
Since I'm not a chemist, I can't tell you which column chemistry to use. What I can tell you is that replacement columns are one of the biggest line items in a lab supply budget. That's why I ask for the method suitability report before writing a PO. If the report shows pressure or tailing problems, the answer is yes. If the report is clean, the answer is probably no.
Bottom line
If your lab needs a proven Keithley data logger for low-to-moderate channel counts, the Keithley 2700 multimeter is still a reasonable choice—if you buy the switch card and budget for the setup time. For a larger system, look at a newer scanning DMM or a dedicated data acquisition system.
There's no perfect instrument. There's only the one that fits your channels, your accuracy, and your budget. That's the one I'd put on the PO.
One more boundary: I'm not telling you to avoid other brands. Fluke, Keysight, and Tektronix all make excellent equipment. I wrote about Keithley because that's what our lab uses for low-level electrical measurements, and because the 'keithley data logger' search term deserves an honest answer.
