Just bought an Agilent U1733c and trying to measure ESR. I read TFM and used acquired knowledge to engage ESR mode @ 100k, but every measurement seems suspect (high). So I wondered what would happen if I just shorted the test leads. Sure enough, I'm getting an ESR of 0.05 on a dead short. This can't be right, correct? Am I doing something wrong? Do I need to calibrate or zero it out first? Pic included. Thanks fellas!
Even a brand new 33uF cap is reading around 1.7 for ESR @1k, 10k, and 100k.
Just performed the little onboard open/short calibration per TFM. Getting basically the same results.
Looking at some jellybean 160V 33uF parts on Digikey, I am seeing tan D in the order of 0.15 to 0.25 or so (Specified at 120Hz), which I figure makes a ohm or so perfectly believable.
Why are you surprised?
Regards, Dan.
Here is my reference. Your values seem ok.
Sorry, for some reason I was thinking lower value caps should have lower ESR values, when in fact, generally the opposite is true.
Also, just one more thing, I'm noticing my LCR meter (u1733c) tends to read cap values consistently lower than my u1253b DMM when measuring capacitance @ 120hz. Which one should I trust? Is there a setting I should tweak on one of them? Thanks!
Also always try to use kelvin clips for LCR meters. You can zero out the meter as you did but these meters generally come with specific clip leads for a reason. Kelvin clips make it so all the resistance in your pcb traces, banana jacks, and clip leads is neglected. The kelvin clips perform a 4 wire measurement for low resistance values as this is the most accurate way to measure such low impedances.
Thanks! This is very interesting. So it sounds like the LCR is more likely to be accurate with elytics and I should judge whether or not the cap is within spec using that reading, and not so much the DMM?
Other things to know about measuring electrolytics.
The manufacturer's tolerances on them is wide--like -20% to +80%, or some such.
The manufacturer most likely measures the capacitance at 120 Hz with .5 volts excitation.
When measuring capacitors to determine if they meet spec, particularly with electrolytics, the test frequency is very important and can yield very different measurements. Most electrolytics are tested and rated for 120Hz and that meter is surely capable of testing at that frequency. The reason for this test frequency is because most electrolytics will see the output of a full wave bridge which has a frequency of 120Hz as a result of the full wave rectification of 60Hz AC. This will yield much different results than a capacitor tested at 100KHz. The 100KHz is not typically used for electrolytics except to measure ESR and high frequency characteristics for capacitors rated for high frequency switching power supply applications.
This video is a very good resource for information related to this topic.
I hope there isn't any issues posting a video like this. Someone please let me know if this is a no no.
Thanks! Yeah the one downside of the U1733c is that it doesn't have an option for 4 wire leads, so I couldn't really get the full benefit of Kelvin clips. Although if I could get longer leads somehow, and still get it to calibrate with decent measurements, that would be nice!
So it sounds like the LCR is more likely to be accurate with elytics and I should judge whether or not the cap is within spec using that reading, and not so much the DMM?
Indeed - trust your LCR instrument. Capacitor function of multimeters usually is not primary mode but "nice to have" one. Your LCR is 0.2% while DMM is 1% (for 0.2 - 20uF ranges). LCR measures *real* capacitance, ESR and dissipation factor and more, while DMM measures (using charge/discharge) just capacitance and you can only guess - measurement is degraded by poor ESR and dissipation factor or not. Yes - DMM may be way off promised 1% for capacitor with poor ESR&DF. When there's parasitic inductance involved, results of DMM cap measurement may be even unpredictable.
Yeah the one downside of the U1733c is that it doesn't have an option for 4 wire leads, so I couldn't really get the full benefit of Kelvin clips.
Don't bother. You would benefit from Kelvin clips only for last few "small ranges" of R, L, C (2 Ohm, 20 uH, 20 pF).
Thanks! Very helpful! Can I at least get slightly longer test leads? Are there any you can recommend? I know it might make readings a bit more jumpy, but as long as it will still self-calibrate and yield decent ESR readings I'd love to at least have the option if needed. There are instances where it would very much come in handy. Thanks!
There might be a little more to this.
https://www.ittsb.eu/forum/index.php?topic=343.0I would expect to be able to zero the LCR meter on two wires.
When troubleshooting these things I always rule out the leads, make the most direct short possible i.e. the component lead socket, or a bridged double banana plug. Try and zero to see if any difference. It pays to measure several small resistances or even a length of wire can work, shorting directly across to see if you can take various consistent measurements that make sense. Do this before you get to measuring capacitors. You might also want to use new batteries as well.
You can use whatever length leads you like but be aware they do impact the measurement especially when precise measurements are required, this is why shorter the better is generally a rule. While you can't totally eliminate the length of the leads affect on the measurement even after zeroing, the meter should still be usable even with a set of full length DMM leads zeroed out, especially for "ball park" measurements.
Be warned the LCR meter has no input protection against charged caps so is not so suitable (in my opinion) for constant in-circuit testing.
This video demonstrates how to perform a quick consistent measurement test, even though it's 4 wire measurement (modified test hooks) the same principles apply.
Thanks for the info! I've got really good leads on my u1253b, it couldn't hurt to try zeroing them out and see how close they are to the little gators. I don't need mOhm accuracy....95% of the time I'm just checking ESR in vintage studio equipment and need to get a rough approximation of the viability of a lot of caps. Some mixing consoles have thousands of caps. It'd be really nice to be able to sniff out which one(s) might be bad without pulling all of them, then only pulling the ones which seem suspect for closer/more accurate insoection.
I realize you need to discharge all caps before doing this and be especially careful on valve gear, or better yet, just don't use this method on valve gear at all.
You can get the Keysight tweezers for your LCR meter. It has a long lead. Costs around $35 or less in the U.S. But... the tips are huge and the stiff cable has the tendency to move the meter all around the place.
Sent from my Moto Z (2) using Tapatalk
Thanks! I think I'll just do that! I tried my u1253b leads on it today and after zeroing out, they work pretty well! A little jumpy/unstable and probably not good for super sensitive tight spec measurememts, but good enough for 99% of the work I'll be doing.
With this new power comes great responsibility. Thou shalt always discharge thine capacitors!