Visualizing EMI Filter Frequency Response with an Oscilloscope

[@Graham_Wideman]

On the positioning of the capacitor: You have to bear in mind that you're not just testing the coil and/or capacitor. You're testing a circuit that includes the 50-ohm output impedance of the HP generator supplying the signal to the input of your L and C components, and they in turn are connected to a 50-ohm load (assuming your terminator on the BNC-tee is 50 ohms). The L and/or C impose a difference on the output only at frequencies where their impedance is some significant fraction of 50-ohms, or more. Evidently, for your L, that frequency is around 100kHz (8:20). For the C, that frequency is apparently around 30kHz. (This suggests L = 80uH and C = 0.1uF.). So when you have the C in the circuit, in either location, you're essentially measuring an R-C circuit (input R = 50-ohm, C is 0.1uF) with a knee at 30kHz. The L plays no role in that knee, because the frequency is below that at which the L has significant impedance relative to 50-ohms. So of course it makes no difference whether you place the C before or a after the L in this case.

[@EverettsWorkshop]

The ability to visualize the signal dampening both with and without the capacitor was fascinating, thanks for sharing this!

[@squelchstuff]

The mystery hump is around 60Hz. I'll eat my hat if it isn't mains coupling you are seeing there James. Try rearranging the cables and monitoring the plot. The oscilloscope lead at the back is laying on the case of the function generator, and the scope lead passes through a loop made up of the handle on it. Lots of little things to watch out for. I may be wrong however :-) Great treatment of the subject once again. Thanks.

[@Clough42]

Yeah, looking closer, you're right. It's definitely at 60Hz. So probably environmental pickup.

[@tetraederzufrequenz7896]

@@Clough42 This kind of filter protect you against the jammer?

[@scrout]

Not an EE and talking Bode plots and filter characteristics and rangling function generators and DSO's? All this and 4 plates? You are killing it man!

[@swolebro]

When I heard you say "4 plates," I checked the video backlog looking for a clip of him squatting 405... I'm a little disappointed to not see that, but I did find is one from ~1 year ago. kzbin.info/www/bejne/l3rQdX-Nqd-rn5o Yup, dude is killing it.

[@MrPatdeeee]

Good show. Thanks James.

[@samvoelkel2046]

Excellent demonstration. Thanks!

[@tejonBiker]

Nice trick with the osciloscope, I will test it!

[@hardkrash]

Our electronics lab teacher back in college taught us to switch to envelope mode for the channel one capture, and we fed channel 2 with the frequency from the generator. I think he also had us perform a math computation so that it gave us the dB loss on the Y axis versus time on the x-axis

[@hardkrash]

I also recall setting up the sweep time to something on the order of 10x the lowest frequency.

[@hardkrash]

For the final run after testing with quicker sweeps.

[@kka10001]

Wow, that was way cool, can't wait to play in a bit.

[@captiveimage]

Nice! Use this technique for looking at filters too, and it's useful for looking at the frequency response of amplifiers as well. Would love a spectrum analyser, but don't do this stuff often enough to warrent the cost, and compared to the virtually useless FFT on the Rigol (I've got the same box), I find using this sweep technique really useful. Nice practical demonstration. 🙏

[@sigurasg]

The 3325A has an X-axis drive, or "ramp" output. With an old style analog scope you'd put the scope in X/Y mode. The X would come from the ramp, and the Y from your measurement. Optionally you could use the Z-axis blank output as input to your Z-axis (brightness). You then look at the envelope of the CRT beam. This gets tricky with very slow sweep speeds with non-storage scopes.

[@Clough42]

Yeah, I tried it, but couldn't get anything useful out of the scope. I could get a single vertical band sweeping left and right, but was never able to get anything with enough persistence to make it easy to visualize. Perhaps I missed something.

[@sigurasg]

@@Clough42 Yeah, digital scopes typically don't have very good X/Y display, and non-storage analog scopes don't have enough persistence for slow sweeps.

[@imajeenyus42]

Now I'm missing my 3325A 😁 Had one several years ago but got fed up lugging it around for simple stuff. The sweep technique is very useful - I often measured frequency responses of ultrasonic transducers. It's neat seeing all the different resonant frequencies.

[@AllenCavedo]

Very cool. For someone who is not an electrical engineer you sure know a lot about electrical engineering. Moving the capacitor makes no difference because the the current in a circuit is the same everywhere in the circuit.

[@DustinWill]

First "like" - YES. I love your work. Keep it coming.

[@danmiller6880]

This was awesome.

[@PeregrineBF]

I've used this technique for years for low-frequency work. Since my spectrum analyzer only goes down to 9kHz anything in the lower audio ranges needs to either get mixed up to a higher frequency carrier or have an alternate analysis mode.

[@digus]

Awesome- You da man! Have you given any more thought to selling us some of those latest PCB top covers for the ELS control? I could sure use one to fix my chicken scratch labels..;)

[@TheDefpom]

@3:00 you could use roll mode on the scope

[@robertfenney]

Interesting!

[@Orbis92]

Really nice video. I wonder, what's your background? I came across your channel when you started to build your electronic lead screw, and it looks like you're a bit more than "just" a talented machinist. Is this all self though? Sorry if there is already a video where you did talk about yourself ;) I didn't test it on my Rigol, but I'm pretty sure you can use the "marker" in the measurement "tab" in FFT.

[@swolebro]

I'd also like to know this.

[@todaysfigures598]

thanks👍

[@Soupie62]

As long as you are just showing a coil and capacitor, this would have been an ideal time to show them in series or parallel. The result is a basic band pass or band notch filter.

[@spokehedz]

Things can force us to stay inside, but they can't stop us from tinkering!

[@Dave-nw3ir]

I just discovered your channel which pretty much covers everything I like doing. I'm just backing into 3d printing from cnc machining. Any chance I could ask you some questions on 3d printer recommendations? You seem to have the 3d printer dialed in with dual hot end extruders. Just what I'm looking for.

[@johannglaser]

As Graham Wideman already said regarding the placement of the C, you are more or less characterizing your measurement setup (source + load), and not the L+C. In other words, please consider the true output impedance of your VFD, and the true input impedance of the motor. This will lead to completely different results. And on a second note: In your sweep you were generating frequencies up to 10MHz, while your scope was sampling at only 5MSps. According to the Nyquist theorem, your sampling frequency must always be > 2x the signal frequency, in your case >20MHz. Or the other way round, the range from 2.5MHz to 10MHz in your curve is not showing you the truth. :-)

[@JohnUsp]

Next time you can use the scope cursors to compare before and after the capacitor.

[@jamespurser357]

I am having excess noise all across the hf band. 1-30MHz coming from my solar charge controller. What size capacitors, would you suggest? Output voltage is as high as 56v input is up to 150v. I have seen a design with 2 capacitors in series across the positive and negative with the center pin of the caps going to ground. Seams like it would work but I am not sure about the size of the cap. I all ready have multiple wraps through a toroid and it helped but it is still noisy. Thank you

[@robertlawson4295]

Ha, I've got the Rigol 1074Z with the 100MHz upgrades ... enjoy the product and it's a great help in working with electronics design. A few years ago I complemented the scope with a spectrum analyzer from signalhound.com ... I bought the USB-SA44B model. It works beautifully with their free companion software called Spike. At a later time, I also bought the USB-TG44A, a tracking signal generator. If you are not aware of Signal Hound, I encourage you to check them out.

[@tetraederzufrequenz7896]

This kind of filter protect you against the jammer?

[@SlamminGraham]

Just get a spectrum analyzer, dude. Some are cheap now.

[@hardkrash]

And he wouldn’t easily get this measurement either. There is a very special spectrum analyzer for the frequency range in use here. E.g. the keysight E5061B-3L5. Or one of the power loop analyzers www.omicron-lab.com/products/vector-network-analysis/bode-100/ or Venable FRA. It comes down to what the equipment was built for with regulars to the frequency injection and monitoring.

[@somebodyelse6673]

He's already got it meeting his needs, and that is a very infrequent need indeed. Why take up more shelf space for something that would largely be gathering dust?