Impedance Anyone?

[ttocs]

I finally got around to measuring the impedance of my ML stats. I've looked at various methods for doing this, but decided the easy way out was to get a Current Clamp for my oscilloscope. I'm not expecting exacting results from my low grade equipment, but it's close enough to get an idea of what's going on at various frequencies. The results are what I expected above 8kHz, but I didn't think the impedance would be as low as the measurements say for the range between 6-8kHz. The lower end of the scale was a surprise to me as well.

I wonder if anyone else has tried this with their stats?

I'm really curious how the frequencies below the crossover frequency of 300Hz is reflected back to the power amp? With the woofers being powered by internal amps, I guess there is something in the crossover that shows up as impedance to the outboard power amp?

As I say, these figures are probably off by a bit, but they show a trend none the less. When the impedance falls below 2Ω is where my tube amps began having difficulty with maintaining output, which dropped like a rock above 10kHz, until I began using AntiCables Autoformers connected for a 2X factor.

As was said on the old Malt-O-Meal commercial, "Good stuff Maynard!"

 

[Fidji99]

MartinLogan Montis loudspeaker Measurements

Sidebar 3: Measurements I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the MartinLogan Montis's frequency response in the farfield, and an Earthworks QTC-40 mike for the nearfield and spatially averaged room responses. A note on these measurements is in order:

www.stereophile.com

similar results.

No idea about electrical properties, but [in my case] relieving internal amps from 60Hz down helped a lot through the entire FR.

 

[rpokuls]

Results look good and are very interesting.

ML's spec. sheet itself says 0.7 ohm @ 20KHz so that corresponds with what you see.

Interesting about the jump in impededance between 800 Hz and 1KHz. from 3.6 ohm to 10. ohm.

Some sort of resonance?

 

[ttocs]

With subsequent measurement sessions I'm finding that the measurements below 1kHz are inconclusive, varied from session to session. Some reasons can include the fact that I'm working with such low voltages that they are close to the lower limit of the Current clamp probe, and, the lower frequencies might be giving this probe more difficulty at these lower voltages.

At 5kHz and above I get similar results each time I measure, so these seem to be good and reliable.

To get a better baseline I guess I'll have to use the very reliable resistor method which requires a lot more effort.

 

[Robert D]

MartinLogan Montis loudspeaker Measurements

Sidebar 3: Measurements I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the MartinLogan Montis's frequency response in the farfield, and an Earthworks QTC-40 mike for the nearfield and spatially averaged room responses. A note on these measurements is in order:

www.stereophile.com

similar results.

No idea about electrical properties, but [in my case] relieving internal amps from 60Hz down helped a lot through the entire FR.

60 HZ works best on my Prodigy speakers.

 

[rpokuls]

I found a Stereophile impedance measurement of the Montis. Link is here:

MartinLogan Montis loudspeaker Measurements

Sidebar 3: Measurements I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the MartinLogan Montis's frequency response in the farfield, and an Earthworks QTC-40 mike for the nearfield and spatially averaged room responses. A note on these measurements is in order:

www.stereophile.com

It does indeed show a big rise in impedance around 800 - 1000 Hz.

 

[ttocs]

Thanks for the links. I've seen a bunch of others which all show the impedance rise anywhere from 700Hz to a couple kHz, depending on speaker model.

My issue is with the Current clamp probe I bought which seems to have difficulty below 1kHz with getting absolutely repeatable results, at least with very low voltages. If I pump up the voltage, higher volume, harder on my ears and speakers, it seems to settle into its "Zone" where it works better. I just don't like running sine wave tones into the speakers at anything but very low volume. So . . . I'll need to decide what to do with regards to using this clamp probe. I haven't seen any clamps that are designed to operate specifically at low frequency and low voltage. Most are designed for MHz ranges, not low kHz ranges.

The old school method is probably the best route, plus it's more accurate.

The main question I had going into this exercise has been answered, which is, at what point does the impedance drop and how much? It pretty much follows the frequency response curve I've gotten when I measured my speakers with the Bob Latino tube amps. From 7kHz and upwards the frequency would drop by -4dB at 10kHz, and above that it dropped much more. So this follows the impedance measurements I've gotten so far.

Up to 5kHz pretty much any amp that is happy with a 4Ω Impedance would work fine. It's above 5kHz that the impedance drops to around 2Ω and then lower. Take a look at the graphic I posted above, at 8kHz the Current is double what it is at 5kHz, and at 16kHz it's double what it is at 8kHz. I believe this is when amps have trouble with the load on them, at frequencies above 6-7kHz, all due to how much more current is needed. It's an eye opener for me for sure, and I'm not done, but probably won't be able to do anymore until the weekend. I've got some ideas of what to check that should be enlightening.

 

[DanR]

Thanks for doing the testing and posting the results. It’s very interesting, and should be very helpful for people to understand what to look for in amp specs, this shows just how important it is to have a very “beefy” power supply.

 

[Fidji99]

Bit of theory around (until is descends into usual asr thing)

NC252MP (class D) vs. A250W4R (classAB) burst measurements into 4ohm//2.2uF load

NC252MP (class D) vs. A250W4R (classAB) burst measurements into 4ohm//2.2uF load As I have been doing, quite intensively, wide range of power amplifier measurements, I have decided to compare 2 different power amplifier design philosophies, class AB and class D. The amplifiers chosen have equal...

www.audiosciencereview.com

all in all 40A d class amp is OK.

 

[ttocs]

Well this was fun!
I took a lot more time with this tonight, which by the way I wasn't expecting to have time to do except that because of working late I had to cancel Movie Night! with a bunch of friends, so we'll be doing that in a few days. So, once I got home I had time to play with this.

The first attempt posted above was done by setting up the parameters and then taking a photo of the oscilloscope screen, then moving on to the next frequency to measure, take another photo, repeat, etc. The problem here is that I didn't account for being able to repeatedly check and recheck as I went. The Current clamp probe needs to be degaussed, or, reset, along the way to "clear" the magnetic field it generates. Otherwise it will "read" nothing to measure as having a voltage that seems to be meaningful when in fact it should be very close to zero, which really means a bouncing 5-9mA. Doing the testing "Live" meant that I was able to degauss much more often, as well as Auto Reset the scope every couple of frequency changes to keep things more "in range". I actually rechecked every measurement multiple times after first completing all 20 frequencies.

At 100Hz, the Current was so low that I had to use 800mV instead of 400mV as my target for each measurement. I decided that I needed something to be very consistent and I chose to use the measured Output Voltage of the amp as the unchanging parameter, 400mV. This also meant that I could track how the other parameters were affected at various frequencies, which for the most part were pretty consistent.

I then used the impedance and frequency numbers to create a plot in REW by importing the text files for each set of measurements. Pretty cool! This allowed me to overlay both sessions to see how they compare, or not. It's easy to see that there was some interference yesterday that caused the measurements below 1kHz to be totally wacky. But, other than the "holes" where data was missing in the blue trace, they both are trending in a very similar way as frequency increases. They might differ a little at 6kHz and 7kHz, but otherwise they are close enough to say that I have a great amount of confidence in the results.

 

[ttocs]

I connected a AntiCables Autoformer with the 2X (two times) Impedance multiplier setup and measured to see if it's really two times the impedance. And it basically is, plus or minus a little here and there (sorry for the super scientific language).

Because of the multiplier being involved, I also had to double the target output voltage to keep the Current the same as without the autoformer connected. So the Current is the same, but the output voltage is double, and the input voltage is also double. Now when it came to trying to measure 100Hz and 200Hz, I had to increase the Input/Output voltages so the Current clamp probe could still get a decent reading, otherwise the Current was just too low, down in the noise too much, so that's why you'll see the VOLTS shown as being 1200mV for these two measurements.

The Pink cells are a calc to show how close to 2X the Impedance actually is. Keep in mind that there is a tolerance involved with both sets of measurements, so it's just shown for fun, but it's pretty close overall.

Another thing that's kinda interesting to see is how the Input Voltage sags a bit as the Current goes up with the higher frequencies. I don't know why this would be, but it would be nice to know why it takes less input voltage to keep the output voltage the same from frequency to frequency.

Red is the normal amp connected to speaker, Blue is with the AntiCables Autoformer with the 2X connections being used. And again, I just created a text file to import into REW so I could use the graphing capability, so disregard the fact that it shows SPL at the upper left of the plot. I simply put the Impedance numbers where the dB numbers would normally be in a REW frequency curve. And lastly, the point where Impedance drops to 2Ω is 12.9kHz, not where you see the blue trace crossing the 2Ω line because there are no data points that would hold that line down a bit more so there's a gap between 10 & 16kHz.

 

[Chippieboy]

I love your post. I got to thinking, what about skipping the current clamp and going with some sore of inline shunt?

 

[ttocs]

I love your post. I got to thinking, what about skipping the current clamp and going with some sore of inline shunt?

Using a shunt is more accurate for sure, but also more back and forth to measure with/without the shunt. Also, I'm not quite certain - if using a scope - about the need for Differential Probes? I'm not quite settled on that for this purpose, but I think it's suggested to keep the Ground issues at bay. Otherwise, I could just use multiple DMM's, but the scope is so much easier and fun!

My signal generator is pretty old and is all analog. It has a bit of trouble with some frequencies, like 500Hz for example, really jumps about. I should really give it a good clean inside.

Yellow=Amp Output Voltage
Teal=Amp Output Current
Magenta=Input Voltage and Frequency from Signal Generator

The Current is the Teal colored trace. It's shown as mV, but it's really mA. This is how the clamp probe is sensed, but it's a direct conversion.

Since moving my "test rig" away from the subwoofers, and being more conscientious about what is "near" the Current Probe so external influences are minimized, the measurements have been much more consistent and repeatable day to day. The scope image above I just made today just for this post, and it replicates very well with the measurements from last night.

Tomorrow, or maybe tonight, I may try the shunt method with DMM's just to see how it all works. I've never tried it, but have looked up how it's done many times.

 

[geeji]

Very interesting post, congratulations ttocs !
This should make all tube lovers think twice about the "best" amplifiers for Martin Logan speakers.
Very few amplifiers will remain stable with such ultra low impedances, namely the Benchmark AHB2 and several Class D amplifiers such as the Purifi ones.

Considering the criticality of the impedance curve for the amplifier match, it is a pity that Martin Logan does not systematically publish the impedance curves for all their speakers, since obviously "passive" ones will be different from those with built-in bass amplifiers.