Measuring Distortion with a HP 339A and a Keithley 2015 THD

[@Frank_K4FMH]

Gerry, a very illuminating video. I'm looking at the Keithley 2015 THD for my bench. This tended to help me immensely understand what it can do...especially versus the HP 339A which is what the ARRL Test Lab uses according to Bob Allison's Test Manual. Thanks again, OM!

[@gerrysweeney]

Thanks for the feedback. Just finished the teardown vid for the HP339A, there is not that much to see apart from the build quality and lots of HP-specific part numbers :) Will publish soon. I think moving from the 339A to the 2015 THD is probably a step dowm, it feels like that but the footprint is so much better. Maybe I will keep it, but store it off the bench - it feels like something I should hang on to I must say. Gerry

[@jamesking7868]

Hi Gerry, one of the deciding factors for me buying the Keithley was the distortion measurement ability. I also like the fact that you can set it to averaging and set the number of measurements it averages from 2 to 100. If you set it to average 100 measurements via it would be slower to respond to changes but would give you far more stable measurement.

[@tmmtmm]

All three types of distortion shown are examples of non-linear distortion. Non-linear distortion where harmonics or additional frequencies are produced. If the sine wave is visually deformed in any way then there is non-linear distortion. Linear distortion would just be a change in amplitude, such as amplification or attenuation. Linear distortion can also affect a specific range of frequencies such as a high/low pass filter or an equalizer. Phase distortion is just a change in time (delay).

[@aptsys]

A video talking about measuring distortion where the audio is distorting due to clipping... Nice.

[@gerrysweeney]

Hi, I am sorry, I missed you comment the last time around, sorry about that. The Chinese rework station has been remarkably good actually. I bought it some four/five years ago and its been reliable the whole time. The soldering iron is really good and the hot air gun works well. I have no complaints at all. Gerry

[@gerrysweeney]

It was not meant to be a competition :) I like to create these crib sheets, it helps me think out what I am trying to day. Even with that prep though it does not stop me calling a Notch filter a Bandpass filter! Oh well, I will persist, thanks for the comments. Gerry

[@fpliuzzi]

Another interesting video, The HP339A is a true classic with a really clean oscillator. It's certainly a step down from the HP339A, but my old HP334A also has the convenient auto null feature.

[@gerrysweeney]

Hi James, the original idea was to reclaim some instrument shelf space by replacing the the HP with the Keithley. The Keithley meter is really a nice meter, I think I might even prefer it to my 33401A meter. However, when playing with the HP 339A I kinda remembered just how nice it still is so I think I am going to hang onto it for now. Gerry

[@jamesking7868]

The Keithley THD measurement allows you to specify how many harmonics you want to measure. The THD + noise measures as many as it can up to the measurement bandwidth limit. Thats why you only got a few % distortion with the THD mode and 50% with the THD + n mode. In addition if you access the Keithley through rs232 or GPIB you can get the measurements of the individual harmonics. Its a pity that functionality is not available through the front panel.

[@hugeshows]

@18:45 when you say "not sure what that does", I believe it is a setting to control the output of the second signal out, and I believe "ISINE" is inverted sine, so it sends the same signal out but at 180 degrees phase to the primary out. I *think*...

[@gerrysweeney]

Hi James, thanks for the comments. Yes you are right, I was only setting this to the default which was 2, I believe this means 2nd and 3rd. I have not managed to play much with it yet, I will definitely at some point, I have some tube amplifier antics to get up to...Gerry

[@jamesking7868]

If you program the keithley via rs232 or GPIB you can also configure it to do sweeps (see pages 2-34 to 2-46. The 2015P (rather than the 2015THD) even does spectrum analysis... but again only accessible through programming. I suspect the THD rms mode of the 2015THD would be closer to the measurements from your HP

[@movax20h]

I think, osciloscope, even USB osciloscope, or spectrum analyzer, is better than dedicated multimeter with THD. You can do THD calculations on osciloscope with FFT or spectrum analyzer much more accuratly than with THD meter. No only that but, osciloscope or spectrum analyzer can do more advanced analysis of a signal and distorsions, especially for high frequency signals. Dedicated THD meter could be useful for audio equipement, if you are testing them every day, and need a standard way of doing so, even if it is less precise, but is faster.

[@florianhofmann7553]

With a 8-bit oscilloscope as most are one can only measure down to 0,4%.

[@nickmcwilliams685]

​@Florian Hofmann agreed scopes are awful for measuring aptitude. Seeing the shape is more useful for troubleshooting distortion, you may know why it's distorted by looking at the wave or FFT. But the resolution is too poor for measurement.

[@aisupporter4775]

Thumbs up for your hand drawing!

[@gerrysweeney]

Aisupporter thank you :) Gerry

[@jamesking7868]

You can also set an upper and lower bandwidth limit for the measurements on the 2015THD - but again only via the programming interface...

[@lazzer408]

Neither tubes nor transistors produce any dominant orders even/odd. The majority of distortion occurs in the output stage. Push-pull will cancel evens. Any differential signal will. A single-ended stage will produce even order due to the asymmetrical nature of the device whether it solidstate or vacuum tube.

[@gerrysweeney]

almost certainly due to the output stage but in my experience tube amplifiers are less prone to dissonant tones when overdriven, which is why guitar players love them so much. The distortion characteristics of tube and transistor amplifiers are different in the frequency domain, that's visible when you compare both. In general, tube amps have a notably high THD figure compared to a well designed transistor equivalent, despite this, tube amplifiers tend to be sonically more pleasing to the ear, The only difference I can see between tube and transistors is the frequency spectrum is under moderate overdrive conditions even harmonic levels are more dominant than odd harmonic levels in a tube amp, so it makes perfect sense to me (as a guitar player too) that my Tube amps always sound better than their solid state counterparts. That being said, nowadays, modelling amps get so close to re-producing these tones there is really no justification for the expense and lugging 60's tube amps around any more, a good quality PA or a FRFR active speaker setup is much better today.

[@lazzer408]

@@gerrysweeneyGuitar guys distort the single-ended stages. Typically a high gain triode like the 12ax7. If they overdrive the output stage it'll start adding odds and becomes very grungy sounding. Too much of that and it's rather unplesent. Transistors have a lot of high-order harmonics and IMD. That can sound pretty nasty. Whip up a class A gain stage with a single transistor. That could be a fun exercise in harmonics.

[@nickmcwilliams685]

The softer clipping of class A tube output vs a class a transistor with a much more abrupt cutoff is probably what creates the warmer tone. But yeah the asymmetrical clipping is 90% of what makes class A overdrive sound good.

[@peterpurpose]

Informative, bang on point and incredibly well presented.... great vid Gerry.

[@erikhovdahl]

Hi... Great video! But you can measure THD within the FFT of the MSOX3024,. Under menu Select measurement. ? I got a MSOX6004A, and it got that possibility.

[@allthegearnoidea6752]

Very nice I'm doing some videos on test gear and some audio measurement and I learned a lot useful stuff many thanks. Regards Chris

[@mjlorton]

Great video...thumbs up and shared. Cheers, Martin.

[@jesterraj]

I checked the circuit before I replced the transistors...and diode circuits before I replaced the outputs.....thou I find 2 of the output risistance reading diff form the ones on the good channel ...could that be a problem?

[@ligius3]

I'm still waiting for an answer :) I'm finding myself working more and more with SMD parts and the iron + strip gun combination will not cut it for much longer.

[@gerrysweeney]

You are welcome, thank you for watching. Gerry

[@googlealainstucki]

thank you gerry, your videos are always very good and very informative. best greetings from switzerland

[@gerrysweeney]

Thanks for the comments. Gerry

[@ligius3]

Thanks for the answer (and the videos as well :) Is is the 802 model? How good is the start-up time and heat recovery on the iron?

[@overtone_audio]

Gerry, there is a 2015 and a 2015-P. what would the difference be?

[@overtone_audio]

figured it out...

[@xyloeye]

Gerry, thanks for the great video. I've been considering a 2015 but have seen some negative comments regarding bad pwr. supply caps. Apparently, Nichicon got a bunch of bad electrolyte for a while. Supposedly they short and can actually the kill meter. Any thoughts on that? Would it make sense to recap it just in case? Thanks, Ken

[@gerrysweeney]

Yeah its a known problem that can kill the meter, its the two big caps, you can simply replace them if you are concerned. The model I have is a newer model that apparently does not suffer form the problem.

[@jesterraj]

hey I need some help I replace some bad tranistor outputs in a America audio amp 3001..but the out put is distorted any ideas where it may be coming form......driver and outputs changed

[@gerrysweeney]

Raj, very difficult question to answer, you need to get a scope on there and have a look, it could be almost anything. Its generally not a good idea to replace output transistors without first making sure the rest of the amp circuit is in good shape. Every scenario is different I am afraid. Gerry

[@andymouse]

Very cool tutorial...cheers,

[@marsattacks7071]

Is there HD bellow the fundamental ? In this case, to measure it, does it require another set of very expensive equipments ? Thank you for the demo !

[@allesklarklaus147]

Eric D. No because you cannot generate a lower sine wave from the "original" sine wave. Well, obviously you could but.. The sine wave is already the perfect shape and perfectly is at the frequency you set it too. If there was another, lower frequency since wave than that would be the new original sine wave and the former original sine wave becomes a harmonic. Or it would just be another tone which would be ignored in this case

[@friedmule5403]

Great video, thanks a lot for sharing!! I hope that you please will answer this noob question. 7:38 Would it be possible to take two probes, one to the output of the functions generator and one to the output og DUT, then substract channel 1 and 2 and then only get the difference (distorsion and noise) displayed on the scope?

[@gerrysweeney]

Yes in theory, but the difference would be so small you would not really be able to tell the difference between the noise you are trying to measure and the noise floor of your scope. There would also be other inaccuracies like the difference in gain calibration and signal delay on the different channels of your scope which would introduce unexpected results, and phase noise etc. So, in theory, yes your thinking is correct, but would require an ideal scope with two perfectly matched channels with a noise floor at least 10x lower than the noise you are trying to measure, so for all practical purposes the answer would be no, that would not work.

[@friedmule5403]

@@gerrysweeney Thanks for a really great answer! So the metode you show, have some sort of amplification to the signal, to get the difference after the filter?

[@gerrysweeney]

@@friedmule5403 metode?

[@friedmule5403]

@@gerrysweeney metode = that what you are doing / describing on your drawing. :-)

[@johnnytacokleinschmidt515]

@@friedmule5403 method I guess

[@InssiAjaton]

Your schematic misses the connection between the junction of the two switches (or jumpers) and the junction of the two diodes. The demonstration implemented the correct effect, though.

[@SteelBlueVision]

It looks like you messed up your explanation of harmonics starting at about 14:24 or so. The first harmonic is the actual fundamental frequency. The second harmonic is the first bump you see to the right of the fundamental. So, for example, starting at 14:24 you are pointing at the 3rd, 5th, 7th and 9th harmonics and not the 2nd, 4th, 6th and 8th as you state in the video. Your example waveform shows a predominance of odd harmonics over even harmonics (i.e., more solid state like rather than tube like).

[@gerrysweeney]

+SteelBlueVision thanks, yes you are right :) thank you for pointing that out. Its funny when you know something instinctively I have never really questioned my understanding of the naming. I know the frequency domain patterns pretty well by sight but not by name obviously :)

[@gerrysweeney]

@peterpurpose thank you for the feedback. Gerry

[@ligius3]

I've seen in a lot of your videos that Chinese rework station.Is it any good, would you recommend it?

[@patricksmith4424]

I find it staggering that the Keithley 2015 is the only multimeter that measures Thd. It costs around 7000 UK pounds new, serious money. If other manufacturers brought out cheaper meters with this on, I am sure they fly off the shelves.

[@GeorgeGraves]

GerryCAD is so much nicer looking then DaveCAD. ;)

[@orvaldurbjarnason4142]

Thanks, great video

[@mattd28]

Can someone please recommend me a scope to start out with, I am in UK if that will narrow down the models I can buy here?

[@gerrysweeney]

If you are just starting, analog would be cheaper than digital and if you want something that would retain its value I would look at something like a second hand TEKTRONIX 2430A, great scopes, affordable (£250 ish) will probably always be worth about £250 in working order. Gerry

[@mattd28]

Gerry Sweeney Many thanks Gerry will google that model now and have a read up.

[@James_Bowie]

Ironically, through my headphones, there is substantial clipping distortion on this video's audio.

[@gerrysweeney]

Thanks for the clarification.

[@gerrysweeney]

Thanks Martin.

[@jesterraj]

what scope is that ur using

[@gerrysweeney]

Raj, thats an Agilent InfiniVision MSO-X-3024A 200Mhz/4GSa/s. Gerry

[@CABohol]

Nice

[@manoharpatil6780]

Reading not show picture the video

[@spesproductions]

Lol an unbearable audio, fix the amplitude distirtion, please