For sure I want to measure if the THD is higher because the amp has more THD , and so far it is not measurable.
I will use a very bad amp with a tiny cheap transformer big like a lighter and see if the THD will go up and also I will measure with my most powerful amp and see if there is less THD than the solid state amp.
I will use a very bad amp with a tiny cheap transformer big like a lighter and see if the THD will go up and also I will measure with my most powerful amp and see if there is less THD than the solid state amp.
in gray is more powerful amp but worse power supply and less current, they both sound very good , I will get my pre amp to boost the signal from the dac is not strong enough to blast it out 🙂
The THD then is getting a little lower like 1db..not significant for the gray amp which is way more powerful but it has more noise in the bass as you can see. But mainly the woofers are getting 3x their max excursion, not going to do another test like that, only with the 8 watt tube amp , last comparison.
The poor 6bl7, I probably damaged a tube doing that to it, but whatever. It has more THD if you blast it, It just physically unable to sustain this without damage, the other amps can. And again this THD raise doesn't explain the sound at all. This amp is the sweetest, with great bass, better than the most powerful gray amp. It compares to the very best of solid state. However it has not the impulse\resolution and speed of the other two amps who are astonishing. Beside being all that the 6bl7 is just another PP tube amp. Again even if we see the THD is has zero value to assess how it really sound.
If you gave me the types of amps and their THD profiles and specs I would not be able to predict anything about their sound quality. The bass of the small amp is lush full, not tuby, the other two amps around 40 and 80 watts have definition and accuracy not because they are more powerful at all. It is how they treat the phase and signals. The 80 watts has more flow, but it has more noise in the bass and less composure in the most complex music. All this boils down to the multiple feedback employed in the 40 watt amp + its regulated power supply.
@AKSA, Higher order harmonics are generated by cross-over and switching distortion which can't be eliminated completely. As long the amplifier operates in class A the mentioned criteria of H2+H3+H4 >= 99% of the total distortion which should be < 0.2% is most likely to meet when the amplifier operates in class A, When the input signal crosses the quiescent current limit and transistors enter into switching mode then the distortion pattern changes, and the higher order harmonics are introduced. The signatures are distinct in class A and class B modes and clearly distinguishable.
It occurs to me that if we measured all amplifiers at the output wires WHEN CONNECTED TO A SPEAKER we might see very different THD figures at the usual 1KHz standard.
In that case the measured THD into a non-inductive resistive load might be seriously flawed.......
HD
In that case the measured THD into a non-inductive resistive load might be seriously flawed.......
HD
I've mentioned this multiple times.
I don't know that I'd go so far as to say "flawed", but it is not representative of real-world use except in the narrowest of situations.
There have been a few very nice proposals for a "standard" load that may represent more realistic situations, but I haven't seen them adopted except in individual circumstances. @PMA has shown some excellent examples, IMO.
It could be a problem with low NFB amps that could end being eventualy load sensitive, but with high NFB ones
the speakers wont make any difference.
the speakers wont make any difference.
No on the contrary the amp signal at output into capacitive and reactive wires could shift and then the high feedback sends back this signal at the beginning introducing more THD because now the THD is not 180o and it is now shutting down the power. The amp start to sound sterile and shouts at you.
@wahab - That's the tricky part. Some may argue that a "proper" amplifier shouldn't be load sensitive (within reasonable extents). Others may argue that NFB is the most horrible thing since having to slice one's own bread.
Engineers simply need to know if their design performs 'as-intended'. Other people may want to try to discover / infer some sort of sonic signature from an FFT or distortion numbers.
THD @ 1kHz / 8R / 1W is a single valid test, but (to me) it will never provide enough information to determine how an "amplifier sounds" or even enough information to determine if it performs "well". It's just a test... nothing more, nothing less.
Those that want their amplifiers to have "no sound at all" will likely design into different criteria than those that are intentionally looking for a certain "character".
To me, both are perfectly valid approaches, and the testing to determine their proper operation (or suitability for consumer use with realistic loads) may require different parameters.
Engineers simply need to know if their design performs 'as-intended'. Other people may want to try to discover / infer some sort of sonic signature from an FFT or distortion numbers.
THD @ 1kHz / 8R / 1W is a single valid test, but (to me) it will never provide enough information to determine how an "amplifier sounds" or even enough information to determine if it performs "well". It's just a test... nothing more, nothing less.
Those that want their amplifiers to have "no sound at all" will likely design into different criteria than those that are intentionally looking for a certain "character".
To me, both are perfectly valid approaches, and the testing to determine their proper operation (or suitability for consumer use with realistic loads) may require different parameters.
A low distortion figure implies the harmonics, starting from the second, are small. The formula for calculating the distortion figure is given by:
f1 is the fundamental representing the wanted signal, f2, f3, f4, f5, f6, .... are harmonics.
A low distortion figure is indicative of the cleanliness of an amplifier output. If an amplifier adds something that is not in the input, it will not be just "amplifying" the input, but also putting its own ingredients, whatever they are.
Code:
distortion = square_root( f2^2 + f3^2 + f4^2 + f5^2 + .... )/f1A low distortion figure is indicative of the cleanliness of an amplifier output. If an amplifier adds something that is not in the input, it will not be just "amplifying" the input, but also putting its own ingredients, whatever they are.
That is the 'math' part, however it has to make coils move in linear fashion inside a magnet system through capacitors/resistors/inductors, and capacitance of wires going to XO.
The feedback is 180o of the signal, each gain stage reduces the margin and contribute to THD, at the end of the amplifier placing wires cause a capacitance which is a short to ground for feedback, meaning the feedback is not working at all and gain rise. It means nothing to have low THD.
The feedback is 180o of the signal, each gain stage reduces the margin and contribute to THD, at the end of the amplifier placing wires cause a capacitance which is a short to ground for feedback, meaning the feedback is not working at all and gain rise. It means nothing to have low THD.
There are symmetrical designs mostly used in CFA paradigm where the upper and lower halves cancel out even harmonics leaving only the odd ones in the spectrum. These odd harmonics are said to make the sound unpleasant.
This is my "shoot from the hip" mid-woofer model that serves me well with a nice bass resonance, and a simple LC ladder structure, so the rising impedance at high frequencies is loaded and doesn't rise as quickly as it would without eddy currents.
I've been using it in the context of my personal preference for full-range speakers, but I could see things getting quite crazy with multi-way networks.
I've seen the effect on THD go both ways: when switching between the above and an 8 ohm resistor, sometimes distortion is higher with an inductive load, and sometimes the higher impedance and lighter load seems to offset it. Observing I vs V phase shifts around the bass resonance is interesting. Real-world observations like an 'over-damped' box somehow sounding dull and 'off' may make sense in light of the increased drive current and losses that the amplifier has to overcome.
Why is it that harmonics are summed orthogonally as though they are uncorrelated? If they are uncorrelated it must be largely so just because we discard phase information and only look at amplitudes of harmonic components thus willfully hiding their crest factor from our awareness?
With high GNFB the reactive power is dumped by the amp s very low output impedance, you are talking of low NFB amps here.
I m talking of properly designed amps with robust stability, not to say unconditional.
Often the pursuit of very low THD lead to amps that are too conditionaly stable and wich will mandate an output LR
filter, FI i ve seen many amps with unreasonable amount of return ratio in a TMC network, if you spot values like 1k
for the return resistance then it s not properly designed and is just bad enginering with the LR filter as convenient
stabiliser, and same can be said for exageratly low miller caps values.
I searched for the derivation of the formula but couldn't find it. I tried on books.google.com using the search criteria:
I repeated the search for .edu websites also without success.
.
A derivation would have answered the question.
Code:
derivation of formula for amplifier distortion site:books.google.com
Code:
derivation of formula for amplifier distortion site:.eduA derivation would have answered the question.