Three Ways to Measure the Output Impedance of a Circuit or Device (066d2)

I had to measure the output impedance of the circuit I designed in the video on the multi-stage amplifier example. It occurred to me that not everyone might know how to make this measurement. So, in this video I am going to show you how we go about doing this. I will also explain the basis of these three methods. This is not just a "do as I do" video. It is an "understand what I do so you can do it, too" kind of video.
WAY#1: The Direct Loading (Single Point) Method ... the simplest, but doesn't always apply.
WAY#2: Modified Single Point Method ... this works for all the WAY#1 cases and covers most of the rest.
WAY#3: Two-Point Method ... this takes care of the remainder...works in ALL cases, but takes more time.
I have created a "go-along-with-the-video" sheet which steps through the whole process exactly as I am presenting it here. The link to this sheet is below.
I've provided links to all of the videos in this series below the time markers.
=============== LINKS =================
LINK to the promised formula sheet:
drive.google.c...
=======================================
Time Markers for Your Convenience
----------------------------
00:05 Introductory Comments
00:59 What is Output Impedance?
01:13 The Output Model
02:18 WAY #1: Direct Load/Single Point Method
05:09 WAY #2: Modified Single Point Method
08:59 WAY #3: Two Point Method
12:48 Concluding comments and Toodle-Oots
==== This video Series ====
Video#1: Introduction to the Junction Diode (How PN junctions work)
• Introducing PN Junctio...
Video#2: Introduction to the Bipolar Junction Transistor (BJT)
• Introducing Bipolar Ju...
Video#3: Basic Bipolar Junction Transistor (BJT) Metrics for the Experimenter
• Basic Bipolar Junction...
Video #4 Basic Bipolar Junction Transistor Analysis: The Model
• Basic Bipolar Junction...
Video #5 Basic Bipolar Junction Transistor Analysis: The Common-Emitter Circuit
• Basic Bipolar Junction...
Video #6 Basic Bipolar Junction Transistor Analysis: The Common-Collector Circuit
• Basic Bipolar Junction...
Video #7 Basic Bipolar Junction Transistor Analysis: The Common-Base Circuit
• Basic Bipolar Junction...
Video #8 A Beta Stabilized Common-Emitter BJT Circuit: Analysis & Design
• A Beta-Stabilized, Com...
Video #9 A Beta Stabilized Common-Emitter BJT Circuit: Part Selection, Reverse Engineering and Bench Results
• A Beta-Stabilized, C-E...
Video #10 A Multi Transistor Example Ciruit
• A Multi-Transistor Exa...
Video #11 Three Ways to Measure the Output Impedance opf a Circuit or Device
** This video **

[@maurosobreira8695]

Best usage of a decade box I've seen so far...Thanks for the video!

[@eie_for_you]

🙂 You are very welcome! 😀

[@fullwaverecked]

I dig your channel and your content. Cheers!

[@eie_for_you]

Thanks! 🙂

[@Steve_LSP]

Was searching the net to find practical ways to measure the output impedance and found you. Obviously subscribed and will watch all of your videos. please keep on. Yuo have such a nice clear way to explain things that are only generally theoretically taught in schools. Can't wait for the part 2 series where you will for example practically measure the real output impedance of an op-amp configuration known as Howland Current Pump. Whenever possible. Thank You so much.

[@eie_for_you]

Thank you! You can use this procedure to measure the output impedance of that op amp at lower frequencies. You are welcome! 🙂

[@PixelSchnitzel]

I'm definitely saving this for reference. This is an excellent expansion on W2AEW's also-excellent video on this topic.

[@eie_for_you]

Thanks, man!🙂

[@nagyandras8857]

just a quick notice. if ya went to public school learning electronics... what this fine gentleman summed up in just a few minutes, would take there about a quarter year. give a few thumbs ups and subscribe. time these lecutres saves, are way more valuable than you think.

[@eie_for_you]

@@nagyandras8857 Thanks so much, man! 😁

[@W1RMD]

Excellent as always! Thanks and have a great one.

[@eie_for_you]

Thanks, man! Enjoy your day! 🙂

[@RensePosthumus]

Very lucid presentation, learned something. Many thanks.

[@eie_for_you]

Thanks! And you are very welcome! 🙂

[@StingerSecSol]

I like the way you explain everything. Subscribed.

[@eie_for_you]

Thank you! 🙂

[@brucenewengland]

Loved seeing John 3:16! May our God bless you and your loved ones brother.

[@eie_for_you]

Thank you! He has blessed us abundantly! 🙂

[@gregwmanning]

That was very enlightening, I can use this knowledge, thanks

[@eie_for_you]

You are very welcome! I'm glad it was helpful. 🙂

[@pault6533]

@11:20 Additionally for case #3 you can calculate the theoretical open circuit voltage using either side of the equation shown at this location of the video. Thank you Mr. Gable!🎉

[@eie_for_you]

True that! You are very welcome! :-)

[@randyanders]

Great video Ralph Thank You 😊😊😊

[@eie_for_you]

Thanks, man! And you are very welcome! 🙂

[@BobMellor1954]

Thank you so much for this, it makes everything so clear; 73 Bob

[@eie_for_you]

You are very welcome! 🙂

[@jluke6861]

Thank you for the Great Video.

[@eie_for_you]

You are very welcome! :-)

[@nevmarr]

Thank you. One assumption you disclosed was using a "low frequency". At what frequency does a method start to give misleading results? Why are the methods frequency dependant?

[@eie_for_you]

Well, this depends on the particular method of termination. Using my old Heathkit decade box with all of its leaded resistors inside, this frequency is pretty low...below 100KHz. If you have a "higher end" decade box with all surface mount components, you might get away with higher frequencies. If you are using the third method and have two coaxial termination resistors to make the measurement with, you might be able to use even higher frequencies. The limitation is due to stray inductances and capacitances leading to possible self-resonances and other RF weirdnesses. The rule of thumb with most every measurement is to use the lowest applicable frequency to avoid all of the effects of these strays. Hope this helps. 🙂

[@glasslinger]

So we have a perfect signal source (OP AMP) generating the 1 volt signal. We connect a capacitor between the OP AMP output and the output terminal.. We connect a 50 ohms resistor to the output. We adjust the capacitor to give 1 volt. We hide this in the signal generator cabinet and hand it to you. Will your method and arithmetic still give the proper value? I suspect these simple impedance measuring techniques will only work in a purely resistive circuit.

[@eie_for_you]

Nice question! You have defined a frequency dependent output impedance. This method will predict the output impedance at a given frequency. By taking similar measurements at one or two frequencies additionally and using some math, you could predict the value of the capacitor and the output impedance at any given frequency. 🙂

[@AzCrawler00]

Given the proliferation of affordable hobby grade VNAs, it would be interesting to host a demonstration covering the considerations necessary to properly and safely measure the output impedance of a black box device (e.g. maybe it is an amplifier with or without DC offset). Two port measurements of a filter or non-active components are fairly straightforward and well documented. There are already many videos covering input impedance of an arbitrary device using a VNA. Thoughts? Maybe the topic has been covered well elsewhere.

[@eie_for_you]

Interesting question ... Yes, two port measurements are perfect for filters and even amplifiers provided that one takes proper precautions to protect port 2, the input with proper attenuation between the amplifier output and the VNA input. Using a VNA to perform a one port measurement of the output impedance of an amplifier is quite a different matter. Output impedance is a matter of an *active* device. For instance, the output impedance of a common collector circuit is tied to the transistor itself and the quiescent emitter current. You cannot measure the output impedance of this circuit without the circuit being in an active, powered up, state. I would be very wary of connecting a VNA directly to the output of an amplifier to measure its output impedance because a VNA generally uses and measures very small signal levels. Amplifiers generally provide anything *but* small signals. Connecting a VNA to the output of an amplifier would almost certainly result in the frying of the VNA input at worst and, at best, probably give results that are not altogether commensurate with reality. Afterall, even with no input to the amplifier, it still will have some sort of residual output which will compete with the VNAs anticipated response to its own signal. My take. 🙂

[@AzCrawler00]

@@eie_for_you That makes sense and confirms what I was envisioning. Even if the output port of the DUT was decoupled to eliminate any DC offset, and the appropriate attenuation added to satisfy the input restrictions on Port 1 (as opposed to Port 0), would that even measure the true output impedance of the circuit? Seems the reactive component (DC block / capacitor) and the attenuator (resistor network) would have an impact on the impedance measured by the VNA. Seems whatever that measurement is would be of the whole circuit and not the actual output port of the device under test, leaving the original question. Maybe there's a way to calculate away the effects of the DC block and attenuator on the captured S parameters to derive the actual impedance of the DUT output. Wonder if a SIMSmith model can help there. I like the two point general method since the other two are really just special cases of that general solution. The other two (very useful) methods are readily derived if needed.

[@bretfuzz925]

If you perform OSL calibration on VNA with the DC block attached to the VNA, you have nullified the reactance from your measurement. Also, you must terminate the input of the dut. Then you can perform an s11 measurement to the output of the dut using port 1 of the VNA (technically this is an s22 measurement of the dut). Port extension may need to be performed to null out any added cable length. Most people will get confused by the results using smith chart due to rotation from added conductor lengths but it is an accurate method when done properly.

[@bretfuzz925]

For performing s21 to understand frequency response/gain, ensure your input signal (VNA RF signal level) is sufficiently low that you will not violate the maximum input level to port 2 of the VNA. Rumor has it you can perform through calibration with an attenuator, then ensure that attenuator is used between the output of the dut and port 2 of the VNA. If VNA RF output is too high and you suspect compression, place attenuator at RF output of VNA (port 1).

[@eie_for_you]

@@bretfuzz925 If we are doing an S21 measurement of an amplifier, then it is very easy to have a high enough output of the amplifier to fry the port 2 input of the VNA. Depending on the expected input level of the amplifier in question will determine if the attenuator should be between port one and the DUT or between port two and the DUT. Yes, performing SOLT calibration with everything BUT the DUT in place is a must. Than way, the only thing we are measuring is the DUT. If the attenuator is to be placed between port one and the DUT, then the SOL portion should be done with the attenuator in place and the standards attached to the end of the attenuator that will connect to the DUT. This might not be possible if the required attenuation is exceptionally high. The object in all cases is to be sure that, when we make our measurement of the DUT, [a] we ONLY are measuring the characteristics of the DUT and [b] we are not damaging our VNA n the process. This all takes careful planning *before* any attempt is made to perform the measurement. 🙂

[@mario57j72]

Hi Ralph, I have seen the video and I have a doubt. I have to measure the output impedance of an AC Power Supply for professional purposes. I think that the idea of using a signal generator is not applicable in this case as que are in AC so I was wondering if you have done a similar case and you can give me some advice. Thanks for the video, it was very instructive for me!

[@eie_for_you]

The principle would be the same. I would use the two point method, though, because you do not want to load down the output to half its voltage. 🙂

[@rtybn2012]

Excellent presentation. Using #3 equation for a common collector you will find that Zo is less than 2 ohms.

[@eie_for_you]

Thank you! Actually.. when it comes to a common-collector configuration the actual Zo is dependent on the quiescent emitter current and it is approximately equal to V(T)/I(EQ) where V(T) is the thermal voltage (0.026 Volts at room temperature) and I(EQ) is the quiescent emitter current. I did use method #3 to determine the output impedance of a common-emitter circuit and it came out to be 12 Ohms. :-)

[@rtybn2012]

You are right about the Zo is very dependent on Ie. I remembered that after I set that message.

[@eie_for_you]

@@rtybn2012I've done that! The best editor is the "send" button! 😀

[@rtecha.m9648]

Thank you Sir. How does Earth Loop Impedance ( live and Ground) - High and low loop work? I am wondering if you can make a video about it. My mean interest is the circuit diagram not the actual test itself. I need to know for example how Megger tester works or Mega-ohm meter works. Thank you again very appreciated.

[@eie_for_you]

A megger is a special, high resistance ohmmeter. The resistances are so large that the usual low voltage testing cannot be used because the resulting current is too small. The original models (eons ago - before my time) had a simple analog meter and a crank operated high voltage supply. Newer ones can be very, very sophisticated in setting controls and resistance ranges. These will have limits as they ramp up the voltage ... if they high the limit, they could shut down the supply and terminate the test. In the end, they simply provide a voltage that it applied across the high resistance and then measure the current flowing. A little math knowing both voltage and current and we know what the resistance is. We used to have to use this very kind of device to test conductivity of all of the "furniture" and the floors in the operating rooms at the hospital I worked at so that static couldn't be generated to cause a spark and ignite possible flammable gasses being used. This is much like the "conductive" mats used on electronics benches and the "conductive" bags used for static sensitive devices like transistors and ICs. 🙂

[@rtecha.m9648]

@@eie_for_you Thank you 😊

[@eie_for_you]

@@rtecha.m9648 You are very welcome, my friend! 🙂

[@michaellybarger8119]

Outstanding presentation Ralph, as usual. KC9PFH

[@eie_for_you]

Thank you so much Mr. Mike! 😀

[@johnfitzpatrick2469]

G,day from Sydney Australia. You measured the open circuit V applied resistance to it using an appliance (decade box) and turned the top until half V drop witch was 5 on the knob (5 Ohms). Q1. Is the "decade box" within its specifications? I'd look at the label on the decade box or seek owners manual technical specifications list? ➖💫❌

[@eie_for_you]

One does have to be very careful not to apply too much current to one's decade box to avoid frying it. As my one bud used to say, "It'd be cactus!" Regarding the setting of the decade box, that was the x10 Ohm dial which was set at the 5, making it 50 Ohms. My decade box is a very ancient Heathkit Model IN-11. I did not build this. I "inherited" it from some forgotten source. I have checked it to some extent (DVM to the terminals), but not extensively (every position of every switch). :-)

[@jameswarren1831]

Excellent video! Can you follow up with input impedance?

[@eie_for_you]

Thank you! I was actually thinking about that. I can add this to my list of videos to come. :-)

[@peterayearst23]

I'd be most interested in a video on this subject. At the moment, I've been trying to understand input impedance of a dual gate mosfet. Looking at LTspice simulations, I was surprised at how much the impedance changed over a range of frequencies. In particular, between 1 - 30Mhz. Question now is ho to accurately measure the circuit confirming the simulation results? Thank you for your most informative videos.

[@eie_for_you]

@@peterayearst23This topic is now on my list of videos to create. The short answer is ... series resistance (Rs) between the signal source and the circuit. Measure the voltage on each side of this series resistance (V1 = source side, V2 = load side). Then use this formula: Rin = (V2*Rs)/(V1-V2). 🙂

[@jameswarren1831]

While the calculations are straightforward my brain doesn’t work well without a physical demonstration, it’s been broken from the start :)

[@eie_for_you]

@@jameswarren1831 Aaaah! I see! Well, like I said, it is in the list of requested topics. It will be out ... sometime. Cannot say when, but I will do it! 🙂

[@ornithopterindia]

👍Thank you sir.

[@eie_for_you]

You are welcome! :-)

[@JanJeronimus]

Great video. However i miss some information as impedance is not exactly the same as resistance. Also how frequency dependance is this. This as signal sources can also have some capacity or inductance.

[@eie_for_you]

True ... resistance and impedance are related, but not the same. The frequency dependence is highly dependent on the load you are using. My old Heathkit decade box has a LOT of leaded resistors in it and so I wouldn't want to go above 100KHz with it. However, if you are using a really good decade box with all surface mount parts, you could go higher, depending on the the box. If you are using coaxial terminations, then higher yet. Nonetheless, the general rule of thumb for all such measurements ... the lower the frequency the better to avoid self-resonances and other RF weirdness. :-)

[@bobodyuknow]

Loved it

[@eie_for_you]

I glad you liked it! 🙂

[@farisikhmal4868]

sir, is there any method to find short circuit in pcb using this analysis ?

[@eie_for_you]

If you are talking about locating the short on a known trace, I'm afraid that this won't do that for you. Finding shorts on traces is a tough job on a two layer board and **WAY** tough on a multilayer board. With the latter PCB companies often use special X-Ray machines. There are other devices that shoot signals down a trace and you use a special sensor-type instrument to find where the signal stops (short spot).🙂

[@yanliu1060]

Hello Sir, can you make a video about how to safely measure the output impedance of a RF Amplifier circuit with attenuator, there are already many videos covering how to measure input impedance of a RF Amplifier circuit using a Nano VNA. Especially for non-50ohm amps, how do I measure the output impedance on the VNA?

[@eie_for_you]

First, I would so totally NOT recommend trying to measure the output impedance of an amplifier using a nanoVNA (or any VNA). This would have to be determined sort of forensically in a method similar to what I've done here only with a combination of resistive and reactive components. I'd have to give the process some thought as to how to do that safely and effectively at the experimenter level. I've added this to my queue of videos to create. 🙂

[@yanliu1060]

@@eie_for_you Thanks so much !

[@eie_for_you]

@@yanliu1060 🙂

[@anlpereira]

Hi Ralph, in case if I wanted to measure an output impedance of a microcontroller pin that only can generate square wave frequency. For example, if I have a 100khz pwm in a pin output and want to measure its impedance, how should I do that? I know it will generate harmonics through many multiple frequencies. Another question is if I also want to measure the output impedances of this pin in the differents harmonics like 3rd, 4th and 5th. Thank you very much again.

[@eie_for_you]

Thinking out loud here ... The main difficulty here will be the inductance and capacitance associated with the resistor decade box. But, if you can keep your leads *SHORT*, at only 100 KHz you might be OK on that count. You can use the same process used in this video using a scope for the overall output resistance. But, if you are interested in impedances at specific frequencies, not you have to graduate to a signal/spectrum analyzer which will report amplitudes in terms of dBm and the problem here is that they will have a 50 Ohm input impedance. To avoid skewing the results, you would have to put a large value resistor between the pin and the input of the analyzer. This will form a voltage divider with the input impedance of the analyzer (50 Ohms) being the "output resistor" of the divider. Once you get all set up, the process is still the same as the video except now you are looking at frequency pips on the analyzer and the half voltage point is -6dB. Hope this helps. 🙂

[@anlpereira]

@ thank you once again. I don’t know if I understood well. I have a tinySA ultra in my lab, a nanoVNA and an Oscilloscope for 100Mhz. The output voltage of the microcontroller is 3.3V.

[@anlpereira]

Can you make a small video showing how to do with a SA?

[@eie_for_you]

@@anlpereira This might help ... I did this this morning (zipped Excel spreadsheet). This was a FUN experiment this morning: drive.google.com/file/d/1VUFedOKYQOkBsQDpsLz_FienDHG_0-HW/view?usp=sharing

[@eie_for_you]

@@anlpereira See my reply to you next post 🙂

[@rtybn2012]

How about a video using PIN diodes?

[@eie_for_you]

Hmmm ... I'll have to give that some thought. :-)

[@Jnglfvr]

Actually you already had 2 data points based upon the previous 2 methods namely RL1 = 50 ohm V1 = 0.51 V and RL2 = 300 ohm, VL2 = 0.874 V. Plugging these into the third formula gives 49.95 ohm.

[@eie_for_you]

But, each method was presented as a stand alone method as opposed to using all of them. So, it doesn't matter what I determined in the previous 2 methods. 🙂

[@POLMAZURKA]

show us the schematic and how do they work.....................

[@eie_for_you]

If you look in the description, you will find a link to the "go along with the video" sheet that I provided which has the schematics and the formulas in it. 🙂