I have known John for years. And he is a member on ASR. So I know how to contact him. And he is a nice guy. I am not reviewing his personality however, I am reviewing audio gear.
Anyway, you seem to have missed the point I was making. JA provided measurements that mean nothing due to lack of specificity. The fact that you believed them and are asking me to replicate them, shows lack of knowledge of that. Here is Audio Precisions stating the same thing I explained: THD and THD+N – Similar, but not the same - Audio Precision
"When THD+N measurements results are stated or specified, to be meaningful, details such as the stimulus level and frequency (or frequency range), the measurement bandwidth and the DUT gain, should be included. For example:
Correct: THD+N less than 0.01%, 1 Vrms, 20 Hz – 20 kHz, unity gain, 20 kHz BW
Incorrect: THD less than 0.01%"
If you want to do some good, ask John to improve his measurements. I already follow proper practices for these measurements. Notice for example how the gain is specified in my dashboard as is the bandwidth (at the bottom):
If you don't understand the topic then here is a tutorial I wrote on what all of these measurements mean: Article: Understanding Digital Audio Measurements | Audio Science Review (ASR) Forum
Debating tactics only go so far without knowledge....
In this specific 1kHz example, THD and THD+N would be almost same, even for 90kHz measuring BW (you measured with 22kHz BW), because of quite high H3 that dominates above integrated noise.
In case of very low noise preamps, like H2/H3 about -110 to -120dBr, noise may dominate above THD in the THD+N measurement, so the value of THD+N without measuring BW stated is pointless.
Personally, I do not care about the 1kHz plot with H3 of -90dBr. But I do care highly about the fast distortion rise with frequency. This results in intermodulations of all kind and possible colouring the sound.
In case of very low noise preamps, like H2/H3 about -110 to -120dBr, noise may dominate above THD in the THD+N measurement, so the value of THD+N without measuring BW stated is pointless.
Personally, I do not care about the 1kHz plot with H3 of -90dBr. But I do care highly about the fast distortion rise with frequency. This results in intermodulations of all kind and possible colouring the sound.
Last edited:
JA’s THD and noise plot for JC’s pre was done with the 80 kHz AP BW.
I don’t understand why the stimulus in the ASR tests is so high (genuine question). I use very high level stimulus when I test (QA401) if I want to drive the distortion products out of the noise floor In preamps. However, most power amps are at full power with 1V RMS input.
So, my preference would be to test the output at a standard 1 V RMS output and then, if the harmonics are very low or buried in the noise, to raise the input level and look at the profile as they come out of the noise.
Looking at JC’s design, you can probably conclude that it’s balanced (low 2nd’s because they partially cancel, higher thirds).
I don’t understand why the stimulus in the ASR tests is so high (genuine question). I use very high level stimulus when I test (QA401) if I want to drive the distortion products out of the noise floor In preamps. However, most power amps are at full power with 1V RMS input.
So, my preference would be to test the output at a standard 1 V RMS output and then, if the harmonics are very low or buried in the noise, to raise the input level and look at the profile as they come out of the noise.
Looking at JC’s design, you can probably conclude that it’s balanced (low 2nd’s because they partially cancel, higher thirds).
Last edited:
A complete set of measurements will usually make use of different input voltages but if sources are delivering these levels as a standard output it is good choice to measure it that way too.
Just to see if some issues might exist with input stages.
Amirmaj isn't correct about stereophiles measurements if characterizing them as "meaningless" because of missing measurement parameters; it needs often some basic calculations and reading to get these informations, though.
And of course, one should know about the AP model stereophile/JA used for the specific measurements.
Obviously you want to justify using 4 volts. When using balanced inputs the reference level is typically the voltage produced at a level equal to 0 dBm or .001 watt into 600 ohms. If you put a real VU meter on your test signal it would be pinned off the top of the scale even using the recommended 6 dB pad!
But do you use 2 volts when a unit is unbalanced using RCA inputs? Do you understand 4 volts RMS out of a preamp would with a normal 26 dB gain amplifier provide 800 watts into an 8 ohm loudspeaker? At that point the distortion of the electronics are pretty much meaningless.
The curves shown clearly indicate the onset of clipping, or do you not understand that?
Do you think there is any CD out there that has full scale output at 20,000 hertz?
But do you use 2 volts when a unit is unbalanced using RCA inputs? Do you understand 4 volts RMS out of a preamp would with a normal 26 dB gain amplifier provide 800 watts into an 8 ohm loudspeaker? At that point the distortion of the electronics are pretty much meaningless.
The curves shown clearly indicate the onset of clipping, or do you not understand that?
Do you think there is any CD out there that has full scale output at 20,000 hertz?
Last edited:
Your statement is correct that the measiring BW should be mentioned, in JA’s case always 80Khz AFAIK.
It it also correct that noise can corrupt THD figures in case all harmonics are close to the noise level.
But when measuring with a Fixed BW, noise is the same all the time and only THD+Noise can change the Outcome of the measurement.
Ergo a fast rising THD with frequency can never be caused by noise but only by THD. Whatever JA did, in his measurement there was no sign of this.
My guess is that your preamp malfunctioned.
Hans
As it is both channels such a malfunction would most likely be the primary power supply voltage. There are active power supply conditioners on each channel.
He does show 60 hertz artifacts but lost in the noise at 120 hertz. So that would seem to hint at that, as full wave rectification would increase 120 hertz. However the BT had very robust DC buss filtering. It more likely is leakage into his setup.
But a line voltage measurement might be useful to check on things that don't require opening the case. Of course with an outboard powèr supply that is not hard to check.
My experience in center city Seattle is that the AC power line voltage is well controlled.
He does show 60 hertz artifacts but lost in the noise at 120 hertz. So that would seem to hint at that, as full wave rectification would increase 120 hertz. However the BT had very robust DC buss filtering. It more likely is leakage into his setup.
But a line voltage measurement might be useful to check on things that don't require opening the case. Of course with an outboard powèr supply that is not hard to check.
My experience in center city Seattle is that the AC power line voltage is well controlled.
Last edited:
You talking to me?
If so, Bonsai was asking about the input signal and so, yes I wanted to justify using "4 Volts" for the reason I've mentioned before.
Using the "4 Volts" as standard output voltage for the tests isn't reasonable, but I considered that as obvious.
And they do. Most digital sources are delivering 2Vrms (or more) in SE mode and 4Vrms (or more) in balanced output mode. And we definitely do not want input stage clipping or distortion, so it is wise to make a test at this input level (I hope everyone understands I speak about the level at the input of the preamplifier).
I find many statements at diyaudio more political than technical and it makes me sad. Though I sometimes do not agree with Amir, I think he made a good job this time. I agree that John Atkinson's measurements are incomplete and again are rather politically correct, with respect to the designer or company, than revealing.
Most of all I like to have the amplifier at my own workbench. As I had Halo A21 here.
Last edited:
But what about output? Do many people have their gain setup so unity through the preamp is a normal operating point. We don't know if the distortion is in the input stage or output stage, but it does get higher above 3.5V out as seen.
A secondary question is over if you test at expected operating point or worst case with signals that cannot exist in music. Purely personally I think both should be measured.
It's got better after recent departures. But this thread still has a long leash!
More to the point would be why you would couple a DAC with that output level (Benchmark DAC3 goes even higher to partner with their power amp) to a line stage with 14dB gain. Even my miniDSP can put out 8V on the diff output but I can use that to my advantage. The pro world is encroaching into the living room, which is no bad thing as few of us use 500mV sources these days.
AFAIK that level is withing the unit's spec - so it should perform ok with it regardless of any opinion on the level used, surely.
As far as I can see, what is shown is that many much lower cost devices perform better. So, if performance is your goal, rather than the look / feel / nametag, the review has provided what's needed.
As far as I can see, what is shown is that many much lower cost devices perform better. So, if performance is your goal, rather than the look / feel / nametag, the review has provided what's needed.
Someone at Philips Medical / Max Planck Inst. has done at least 400 for a receiving
coil of a spin resonance machine. They proved there was no way to circumvent that
by matching.
Last edited:
Still no clearer to me why 4V RMS was used as the input stimulus on a piece of home stereo gear.
If there is no input attenuation possible before any active circuitry,but otoh "home stereo gear" sources with this fixed high output level is (quite) common, then it is interesting to know if some kind of input overload happens.
As said before, using the preamplifier at unity gain (via adjusting the volume potentiometer accordingly) for an output level of 4V RMS isn't reasonable at all.
Especially when used that way for this kind of measurement.