Distortion Analysis: I've been doing it all wrong

[@bill_p2074]

This is what I like, and appreciate, about you - you don’t let pride stop you from sharing things you have had to re-learn.

[@raygianelli3612]

Bill, I have no pride when it comes to learning. I used to play guitar, and if I saw some 10 year old kid in Guitar Center using some technique I wanted to learn I'd ask him to show me. Does make me wonder what else I've been doing wrong all these years. Might be time for a frank talk with the little woman. 😝

[@scrappy7571]

Good info. I will try this with my keithley 2015. Conrad is a great asset to AK. Very knowledgeable, and can explain things in simple terms.

[@fichambawelby2632]

Thank you, Ray. That’s the way to do it, indeed. It’s the method I normally use. We all are using resistors as dummy loads. It’s the best solution we have to test a big amp to a max volume. The problem is resistors are resistive (!) loads, and speakers are inductive loads, with all the differences this implies (you say it in at the end of your video). And the cable itself has its own impedance. Putting open, two wire cable in series with shielded (maybe some kind of RJxx) will vary the impedance too. So, at the end we have a sum of factors which affect readings. All in all, you are right too when you say that nobody is capable of distinguish between 0.03% and 0.004% of THD. As ever, excellent video. And BTW, I still love your Panasonic meter…here in Spain is impossible to obtain a single unit (not three!!!😂). And I hope your eyes are fine. A big hug from Madrid.

[@raygianelli3612]

Appreciate your comments, Fichamba. Thanks for the well wishes!

[@danhorton6182]

Hey Ray, great video. I purchased the QuantAsylum QA403 and it was my thread in question where I was having trouble with measurements. People over at the forum like Connor saved me a lot of trouble. I would say the biggest issue I was having was with speaker cables, connecting the load that were way too long, and I was measuring at the load. I shortened up my cables, got better cables, and I am now reading distortion measurements that are extremely low. I still need to finish your video so you may have addressed it, but another thing I was dealing with was the input voltage to the amp, as you increase the output voltage of the analyzer, the distortion of the amp gets measurably lower. I will continue the video and comment again, but I learned real quick there are wrong ways to do it lol. On a sidenote, I thought there was an issue with my load, and it was causing distortion. But appears that we use the exact same loads, and I am no longer having an issue. I measured my Kenwood KA-8006 that I restored and at rated output it was measuring 0.009% THD and a little bit higher when you add “+N” into the equation.

[@Stereosandotherfun]

Oscillator noise floor is a lower percentage of the signal as the signal increases

[@herbward5240]

For those of you that have had a similar experience, any loose or oxidized connection or a tired amplifier output speaker relay can be a significant contributor to a high THD reading.

[@Douglas_Blake_579]

Excellent tech tip. Thanks for posting it. 1) By measuring at the load you are also measuring your banana plugs, cables, load reactance, load temperature and a whole bunch of other stuff. It is counter-intuitive but at those small numbers even dust matters. 2) You are absolutely correct that nobody can hear distortion at those levels. Heck, most speakers run at 2 and 3% and we think they're perfect!

[@raygianelli3612]

1) LOL dust! Yeah, even the phase of the moon matters the further down you test, be it level or distortion. 2) Rest assured, someone will claim they can hear the difference. 😬

[@Douglas_Blake_579]

@@raygianelli3612 Of course. There's a whole clan out there who claim super-human hearing abilities. I would very much like to introduce them to Mr Dunning and Mr Kruger ... 👭😀 Also keep in mind that the lowest number is not always the correct number in these measurements.

[@Edisson.]

The overall distortion on the connected speakers can be even greater, it depends not only on the actual construction of the speaker cabinets but also on their connection, signal distortion can be reduced by connecting BI-WIRING, or BI-AMPING for more expensive devices. A bass driver can be a pretty decent noise generator. Nice day 🙂 Tom

[@Stereosandotherfun]

I would say .1% is acceptable distortion

[@Stereosandotherfun]

But I still use an amp with .0005% distortion lol

[@frankscarano4708]

Great video and very very informative find and it makes perfect sense! I am gonna have to do some modifications to my connections as well. I really like the way you source and make your cabling! Also thanks for the shout out and am glad you now have a spare display! Always good to have backups to keep this fine analyzer running!

[@raygianelli3612]

Thanks Frank. I'd have never guessed that you could get an off the shelf display for the analyzer, elt alone that cheaply! Although I did the same with the display in my HP 8904A, which stock is barely viewable but amazing since replacement.

[@Edisson.]

Hi Ray, I have my workplace set up so that I can use both options. If I need a quick connection of the meter to the amplifier under test, I can use the rear input of the 2015 and extend it to the connection points next to the load, I connect the load and the connection point with a short jumper - that way, another cable does not get in the way in my miniature workplace. However, I made a separate shielded cable for direct connection of the amplifier to the 2015 front inputs - I deduced this connection method from a four-wire measurement. All my connecting signal cables with RCA / RCA or RCA /BNC termination are wired as directional to minimize possible interference in the signal path. Thank you for sharing this information and thus confirming my reasoning 👍 Nice day 🙂 Tom

[@raygianelli3612]

Thanks Tom. Good to hear from you!

[@Edisson.]

@@raygianelli3612 I like to hear and see you 😁 but I still envy you that Panasonic analyzer 🤣 When I save some money, I plan to buy a QA403 - nothing bigger will fit in my workplace 😂 I like the Sencore PA-81, for example, but it's too big for me.

[@sandroreis5593]

Interesting for sure. As far as the question of "Can you hear very low THD", I believe it's a sign, a stand in for other factors we don't measure, either because we can't at all (subjective qualia) or because we can't YET.

[@raygianelli3612]

At the end of the day, it's just a measure of faithful reproduction of an audio signal. In the big scheme of things, as others have pointed out, speakers are a larger contributor of distortion by far.

[@poormanselectronicsbench2021]

I think most distortion analyzer manuals tell you to measure directly at a signal source, but who among us reads directions! 🤣 XraytonyB recently found this out about using his Keithley 2015 for distortion level measurements as well. You are right about the differences, if they were real, still being mostly inaudible, but, a harmonic not generated is a harmonic not heard. It's kind of like scoping a fiber end, seeing that little speck of crud on the cladding, but thinking it's OK and not cleaning it ;)

[@TB-nz4kf]

I found this out about six months ago. Removed all but the dummy load and readings were much lower.

[@ohmbug10]

Good to know. Thank you for passing this info to us.

[@glasslinger]

The real joke is when Mr. smartie hooks his .002% amp onto a speaker system with 15% distortion!

[@poormanselectronicsbench2021]

The harmonic distortion level generated by speakers vary by the frequency, and also by the design as well, A lot of speakers will be in low single digits until they go into bass frequencies, and even then some are still single digit below 100 Hz.The time honored Acoustic Research AR3a had a measured 5% or lower down to 32Hz @ 20 Watts of power input ( DIYaudio has a link to company provided specs and graphs) One of the main faults can be the hearing of the listener, as some among us are deaf ad doorknobs, while some can hear nuances that most of us miss.

[@danhorton6182]

On a sidenote about something, I learned when I was dealing with this whole thing was that I got a much cleaner graph if my laptop was unplugged from the wall. If I was running on the laptop battery, my noise for was far far lower than if I had my Laptop powered on the wall wart.

[@raygianelli3612]

Dan, when we're testing distortion as low as we are, apparently everything matters. Even barometric pressure and the phase of the moon. Well, maybe not those things. but damn near everything else!

[@Douglas_Blake_579]

@@raygianelli3612 I suspect it's reading the resistance of your jumper cables as distortion.

[@raygianelli3612]

Douglas, I'm not sure what's at play here. It may well be resistance; I've noticed that the power readings are higher when configured this way.@@Douglas_Blake_579

[@Douglas_Blake_579]

@@raygianelli3612 I have my dummy loads wired with 16ga lamp cord, less than a foot long. The readings I get are virtually -- but not perfectly -- identical at both ends. Edit: Scrap that... moving to the source reads better by about 0.05% ... So now I'm with you trying to figure out what's going on.

[@philsanalogbasement]

Wow, thanks for the video. I’ll have to try that tonight. I’ve been trying to understand a different issue - some amps I test seem too good to be true.

[@raygianelli3612]

Interesting. I haven't run across that yet.

[@andymouse]

Awesome ! more by luck than judgment I have been measuring like this for sometime (at source) but fascinating to see your set up. Who Knew 'eh ? gonna enjoy the screen replacement !....cheers.

[@James_Bowie]

6:14 ... maybe that should be 'there's more to that than meets the ear'.

[@raygianelli3612]

True. And beauty is in the ear of the behearer. 😁

[@frankscarano4708]

Hey Ray I am not sure if i mentioned this to you. The issue you mentioned with the isolation ground being defeated by connecting it to a scope thru the rear monitor output of the 7725 in this video, I found if i use a set of BNC isolation baluns between the monitor out and the scope i can pass the signal thru to the scope adequately and isolate both the ground and signal between and resolve this issue and still keep the scope and the 7725 earth grounded. One drawback though that cannot be defeated is if you are testing a stereo amp that has a hot negative and positive signals that the input ground on the 7725 is all tied together so there is no way to test amps like this driving both channels at the same time. You would only be able to do one channel at a time. trying to use a balun like this on the input side will interfere greatly the ability to read both levels and THD properly though i havent actually tried this but i assume it would have that affect.

[@1320pass]

I really like the way you explain things. Complete novice here wanting to learn this stuff. I have the 2 channel hifi bug and I love vintage gear...so..I'd better get to learning, right? My start in the basics was RC cars, planes, CB radios and automotive. Thinking about building a dedicated test/repair bench. Thanks for sharing your knowledge, Ray. On edit; new subscriber as well👍

[@raygianelli3612]

That's great! I was very fortunate to have people take the time to show me how to diagnose and repair this stuff. I know I tried their patience on more than one occasion! Now I'm just trying to pay it forward.

[@1320pass]

@@raygianelli3612 In my opinion you are doing an exemplary job. Not like I know the trade, just, it's one thing to learn a discipline and then something quite extraordinary to explain it universally and concisely.

[@leiferickson3183]

Wow! Thanks for posting this!

[@alvalle1206]

Many years ago in an article in the old StereoReview magazine it was said that specs are great but will never determine how any given piece of audio equipment will sound ( voiding just pure garbage ,mind you ) over the years I've come to agree . Somethings just sound beautiful and the specs are ehhhh on paper . Some have great specs all around but sound not so good. Odd how that works out .

[@Edisson.]

Paper can take anything, it's worse with the ear 🤪

[@nancy4don]

This reminds me of the old horsepower debate in automotive things. It used to be they would put the engine on a dynamometer, and measure the output directly from the crankshaft. Around about 1972, they changed it so that it would be from the drive wheels. That meant a huge reduction in numbers. This feels like the same kind of thing, except in reverse. You would hear distortion from your speaker; so shouldn’t the wires that connect the speaker to the amp be included in the measurement? Sure do appreciate your videos; I’m learning an awful lot…

[@raygianelli3612]

Makes a good case for powered speakers, where the amplifiers are mounted within inches of the drivers. I literally just learned this days ago. Just passing it on!

[@roroseppel2077]

Nice video Ray ! What about the Luxman LV105U ?

[@raygianelli3612]

It's kicking my butt pretty well. I have some video I've shot that will show what's going on. Hope to get that posted soon!

[@Trucam2020]

Dear Ray, you didn't show the comparison with and without?????

[@raygianelli3612]

Uri, Channel A is with, Channel B without.

[@Janktzoni]

What is your load made of? Is it a perfect resistor without inductive nor capacitive component or is it made out of wire-wound resistors that is also an inductor? If you have a vector analyser or a very good LCR meter you might want to try measuring the loads behavior at all frequencies of the audio spectrum. You might be surprised what you find. The inductive and capacitive component will greatly influence the amplifier output stage and thus its THD and other capabilities. THD of a single quality device of the modern era is not very important, the numbers will be far better than what the best ears can distinguish. The 'compatibility' between devices on the other hand is important. Therefor you have to measure the complete setup, not just one component: As speakers are inductive loads, sometimes also with a capacitive component depending on eventual filters in the speakers, these will influence the amplifier differently from your load. Consequently you will measure yet again different values. It is important to perform the THD measurement end-to-end including both the amplifier and the speakers. If the combination speaker/amplifier is not well chosen you can actually end up with a system that does not sound good because of 'incompatibility' of the components. And there is no way for you to know it in advance as these specs are not available unless you have tried the combination on the shop. There is also no amount of equalizing or delay or cable or whatever that you can add to make it sound good. Far more important, in my opinion, is what you also stated in your video: your ears can't distinguish between these numbers behind the decimal point. In other words, don't focus all too much on that THD number, its contribution to the feelings that are generated by listening to music is negligible.

[@Edisson.]

I allow myself to slightly disagree, there is a functional device that compensates for the different passage time of high-frequency signals with the "BBE" device, and I can say for myself that it works perfectly (I have it at home) and the spatiality and detail of the musical expression will greatly increase. However, it never occurred to me to measure THD after this correction, my ears like it and it is the main thing for me.