Great post @erik_squires.
Impedance - The most ignored and useful measurement tool
I’m constantly reading about audiophiles diagnosing their speakers or attempting to mod their crossovers with expensive new parts. The one tool I wish they’d all get and rarely do is an impedance graphing tool. These are either nearly free or affordable.
They let you produce impedance charts like Stereophile does, as well as measure capacitors and inductors with ESR/DCR respectively.
In the nearly free variety you can build one and use Room EQ Wizard. In the affordable is the Dayton Audio V3. Either one does an excellent job of measuring a driver, crossover parts and the entire speaker as well. Completely irreplaceable tools.
Diagnosing your speakers with the help of others on the internet is made so much easier when you have one of these. Even if your speakers are fine, measure them and keep the charts handy for when they do go bad, it will make it much easier to understand what is and is not working. Replacing a cap/coil? Measure them and the speaker before and after when you are done to make sure everything came back together properly.
You’d be surprised how many speakers have a woofer or tweeter that has stopped working but the listeners don’t even realize it. These are immediately visible problems in the impedance plot.
Of course, it's just a tool, but when a driver goes bad, or a solder joint fails the impedance charts will go wildly off track. It's up to you to dig in and diagnose further.
29 responses Add your response
Interesting! Kudos to you @erik_squires. Is there extra functionaity in the Dayton piece that makes it worth it. I already have REW but have not used it for this. In fact, I was not aware of this capability. |
YES @erik_squires !!!!!!! |
I was talking about this yesterday with @decooney -- the usefulness of knowing speakers' impedance loads. We were discussing how much class A power would be optimal for speakers. |
@sns Great point. In both the case of Fritz's speakers and my others, Ascends, the builders (Fritz and Dave Fabrikant) wrote to me and explained about the kindness of the phase angles in their designs and how their speakers would work with a wide variety of amps without much trouble. On the other hand, when I was auditioning some other speakers, a couple years ago, @tvad helped me realize that the Focals were not sounding great because of a terrible, sharp dip down to 3ohms. I could not have understood it without the phase angle concept. |
To be honest I have only used the DATS V2 and really love how well and easy it is to use, I’ve not tried the REW, but if I recall it requires you to use 2 audio channels at once? I'm not sure. It’s really a matter of convenience. Also, AFAIK DATS is Windows only, so if you are a MacOS only home REW is your only option. There are other professional impedance measurement tools, BTW, these are just the two least expensive I know of. |
@lanx0003 Suggest you hit up the very active REW forums! but I believe that's not negative impedance, it's phase angle. Ask in DIYaudio for help with that, but it's related to how capacititive or inductive an impedance is. These cause the current and voltage to be out of phase compared to a resistor. |
Negative impedance must cause those pesky electrons to speed up. Possibly a battery? @lanx0003 If you look carefully at your graph, you will see a second vertical scale on the right hand side, calibrated in deg (degrees). The bottom curve is indeed the phase angle, centred around zero, and swinging plus and minus about 35 degrees. Ever wonder why VA is often used instead of Watts? If the Voltage and Current in an AC circuit do not peak in unison, they have a time shift which can be measured as a phase angle. Useful power in Watts is V A cos(phase) which is only equal to VA for resistive loads (the phase angle is zero). The amplifier has to provide additional power to make up for capacitive and inductive losses. |
@erik_squires @ddlux @richardbrand Thank you. You're correct. The lower chart should represent the phase angle, which varies from -90° to +90°. I prefer Stereophile's presentation, as shown below, where the impedance and phase angle charts are superimposed on the same chart with the left axis ranging from 0 to 20 ohms for impedance and the right axis ranging from -90° to +90° for phase angle, making it easier to interpret. On another note, Stereophile began reporting EPDR (Equivalent Peak Dissipation Resistance) around 2020. As you may know already, EPDR represents the lowest effective resistance, accounting for both impedance dips and reactive phase angles. The exact formulation used by Stereophile remains unknown to me, but thanks to Mr. Jack Oclee-Brown, his derivation appears accurate enough to reproduce the values reported by Stereophile. As shown in the example below for the Wilson Alexia V, winner of Stereophile's 2023 Speaker of the Year award, the nominal impedance is approximately 2.9 ohms, which may seem manageable. However, when factoring in the phase angle, the corresponding effective impedance drops to just 1.15 ohms, indicating that the speaker is actually quite current-hungry at low frequencies.
|
I wonder how many of us try fixing something we hear as undesirable using other methods in/around the system and while not understanding the speaker impedance first. Re-tracing old steps, I ran into a situation last week about this, also realizing the different amplifiers I simply should have avoided trying prior. |
@lanx0003 Here is an easy way to remember what happens in an LC circuit: Remember the phrase ELI the ICE man. Voltage leads current in an inductive circuit and Current leads Voltage in a capacitive circuit. Helpful tool to remember. When you first apply current to an inductor the coiled wire resists current flow so voltage rises before current starts flowing. In a capacitive circuit, the current flows heavily at first and then slows as the voltage rises. ELI the ICE man...simple no? |
@livinon2wheels Got it. Easy to remember. Thank you. |
Not long ago I went down that rabbit hole. Replaced a set of Focal 807Ws as I wanted to downsize. It's impedance curve if I remeber correctly show two humps in the bass regions and a rising impedance in the mids to relatively high ohms although it's sensitivity was measured an avg of 88.2 or thereabouts. What I replaced them with were Audel Magika II as a smaller option that has worked well for me. It also shares a similar impedance curve. Some notable brands/reviewers have stated (without assessing phase) that these kinds of impedance were an old design designed to be more tube friendly and modern designs should be more level after bass and closer to 4 ohms to maximize the power output and ability to play the speakers louder. Wont mention who but many should know. Probably right for some, not for all. Whether it's better SQ, specs, whatever is the question I suppose. My Dali Opticon 1s have a relatively flatter impedance, also decent measured sensitivity at around 88 dB. My main amp plays both well although my smaller amp could not bring the Audel to perform as well while it did with all others so current might be a factor also. If I went tubes, then impedance may be more critical. Also, the Audels hit a low of 7 ohms in the mids whereas many speakers can dip below 4. In the higherst frequencies it dips to 5 ohms but there it is inconsequential regarding power output of the amp. essentially it acts like a 9 ohm or higher speaker which relies on current output. Perhaps the high ohms prevent my 40 watt amp form delivering enough current? My CA-1 performs great and the combo is very transparent but just avoids being too much in being described as analytical as well as being overbearing in the upper mids like many speakers can be. Ideally I suppose tolerances and curves would be tighter but those are how things measure versus sounds to an individual. Learned alot from Erin's Corner but never bought based soley on his recommendations. I live with what I buy and like. Not perfect but good. If it works well, I'm good. Getting a better something here typically meant giving up something there all too often. If a speaker, amp, dac, etc. performs well enough, then enjoy.
|
I haven’t read this entire thread, but so far, I’m finding it to be a very smart conversation, just the kind of thing that I wish was posted more often on platforms like Audiogon. The only thing I want to add is that anybody seriously interested in this topic should try to search out a copy of Joe D’Appolito’s long-out-of-print book "Testing Loudspeakers," published in a 1988 small-print run by Audio Amateur Press (which at the time was publishing Audio Amateur and Speaker Builder magazines). Yes, this book was writtten by THAT D'Appolito. It’s an outstanding reference, a bit technical for some, but written from an audiophile / speaker-designer’s perspective. It took me several years to find a copy in the early 90s, and it cost me over $150 from a rare-book seller. But it was worth the cost & effort to me b/c at the time, I was testing speakers for mainstream magazines. Today, though, even Amazon has copies of a 2018 reprint in the $40-50 range and I believe it's also on Google Books. Highly recommended.
I wonder if Joe is still alive. Last I heard, he was at Snell, but that was 15 or 20 years ago... |
@erik_squires +1 "You’d be surprised how many speakers have a woofer or tweeter that has stopped working but the listeners don’t even realize it. “ I will be not! speakers impedance is good start, but it will not tell us whole story! whole amp+cable+speakers system need to be characterized on frequency and time domain, incl. distortions etc., using precise measurement microphone in listening room, using fuzzmeasure software or audio analyzer. I wouldn’t be surprised to see large number of setups will show >5% distortions, +/-10dB amplitude deviation in freq. response, and bad resonances! Tests should be done at different power level, starting 50mW. |
Nor do I ever claim it is.
Why? I was only posting this to show how an important tool can help audiophiles diagnose their problems faster, fix broken systems and mod safely. I never claimed it was a replacement for full speaker analysis, which in the use cases I spoke about is not needed. When you say need I think you are thinking far beyond the limited situations I was discussing.
|
So I emailed Mr. John Atkinson regarding what exactly the fomula they have used in Stereophile for reporting the EPDR. He kindly repsond promptly as follows:
Here Vdiss represents voltage dissipation. As a further analysis, this linear equation for Vdiss provides two exact boundary values: one at 0 degrees (which is obvious) and the other at 90 (-90) degrees, as derived by Mr. Jack Oclee-Brown, shown below. However, the linear equation overestimates Vdiss values in between, which in turn results in even lower reported effective impedance. One possible remedy is to fit the data range with an exponential term, as illustrated in the attached figure. Instead of using 1 + 4.2 * abs(phase angle)/90, nonlinear fitting yields 1 + 4.196 * (abs(phase angle)/90)¹.¹⁷⁸, which provides a much closer approximation to Jack’s formulation.
|
@re-lar-kvothe , “ 2 Ohm stable” while better than nothing is really more market speak than of much value. It only suggests that the amp won’t go into oscillation when presented with a 2 Ohm load. It does not suggest how much power and/or at how much distortion will be provided at 2 Ohms. Much more useable would be an actual spec. of ? Watts into 2 Ohms +/- dB. And in the case of stereo amplifiers with both channels driven. |
