Amplifier measurements may require improvement

[restorer-john]

I have on several occasions suggested to use three test jigs derived from data we already know from certain speaker reviews and tests.
"typical 2 way", "typical 3 way", "The Ruthless"

There's no such thing as a 'typical' 2 way or 3 way anymore. Back in the day when crossovers were simple first or second order, with off-the-shelf values, things were somewhat more predictable. But now we have higher order Xovers, lots of tweaking to iron out repsonse variations and I would argue a 2 way crossover of any type won't be too much of a problem.

It's the multiway systems with multiple similar drivers set to cross in and out at different parts. These 2.5 and 3.5 way systems with twin chambers tuned differently etc.

This is one of my speakers in currently in the system (Jamo 507ii). You wouldn't call that typical.

 

[GXAlan]

Your sampling rate or settings are making a mess of the harmonics above 8kHz.

I no longer have the speaker or the PM-11s2. It’s inadequate FFT size for the sweep length, but the interesting part was the midrange difference or that resistor vs speaker was different.

Personally, I think the Salon2 and Revel M16 are both fair choices for standardized speaker loads. The M16 because it’s cheap and modern and readily available and the Salon2 because it’s a gold standard.

There is no right answer for the ideal “typical speaker load” but certainly a real speaker is worthwhile if it shows clear differences between amps which don’t show up in a resistor. At some point, you bite the bullet pick something and hope it works out ok.

The problem as I said earlier is that it makes amplifier testing annoying because it’s loud. With resistors, you can do it at night without waking anyone up…

 

[Deleted member 48726]

There's no such thing as a 'typical' 2 way or 3 way anymore. Back in the day when crossovers were simple first or second order, with off-the-shelf values, things were somewhat more predictable. But now we have higher order Xovers, lots of tweaking to iron out repsonse variations and I would argue a 2 way crossover of any type won't be too much of a problem.

It's the multiway systems with multiple similar drivers set to cross in and out at different parts. These 2.5 and 3.5 way systems with twin chambers tuned differently etc.

This is one of my speakers in currently in the system (Jamo 507ii). You wouldn't call that typical.
View attachment 312545

I know. But some average could be modelled from the collection of tests. Of course in the evaluation should also be the popularity of the speaker. Not in detail but just so, that the edge cases and more special designs are not defining for the jig. -Like your example.
I know this Jamo. It was quite popular here in Scandinavia and still is sold regularly on the used market for as low as 140 USD for a working pristine pair.

 

[Randyman...]

I know. But some average could be modelled from the collection of tests. Of course in the evaluation should also be the popularity of the speaker. Not in detail but just so, that the edge cases and more special designs are not defining for the jig. -Like your example.
I know this Jamo. It was quite popular here in Scandinavia and still is sold regularly on the used market for as low as 140 USD for a working pristine pair.

View attachment 312547

MTM with an integrated isobaric 5th order sub. Neato!

 

[peng]

I have on several occasions suggested to use three test jigs derived from data we already know from certain speaker reviews and tests.
"typical 2 way", "typical 3 way", "The Ruthless"

That would be a lot better than nothing, but the question is, can someone like Amir afford to spend more time on doing this when he practically have been testing at least one device a week. So, I would suggest he do this to selected amps such as something that costs >$2,000 and manufacturers claim the amp is designed to drive 4 ohms speakers know to be highly reactive, based on the phase angle curves. There is no point doing it to the likes of the $80 Fosi V3 amp (or even mid range AVRs) because anyone who want to spend less than $100 to drive their full range tower speakers to high spl should just keep on dreaming lol..

 

[Deleted member 48726]

That would be a lot better than nothing, but the question is, can someone like Amir afford to spend more time on doing this when he practically have been testing at least one device a week. So, I would suggest he do this to selected amps such as something that costs >$2,000 and manufacturers claim the amp is designed to drive 4 ohms speakers know to be highly reactive, based on the phase angle curves. There is no point doing it to the likes of the $80 Fosi V3 amp (or even mid range AVRs) because anyone who want to spend less than $100 to drive their full range tower speakers to high spl should just keep on dreaming lol..

You are not wrong. The thing is I can well imagine someone finding e.g. the Fosi V3 on a shop, advertised as 2x280 W @ 4 ohm (with the 48 V 10 A PS) and google it to see reviews, sees the nice numbers and the "recommended" conclusion of course would think that it's perfect for said "full range tower speakers" and thinks it's the bargain of the century.

 

[SIY]

It's trivially easy to do a direct electrical measurement. Why this wasn't done and fifty new uncertainties introduced baffles me. It's not even worth discussing until this basic measurement- fully documented- is performed and presented.

 

[peng]

Another thing I don't know if mentioned already is that in that YT video, the REW appears to have no smoothing applied, in that case I don't think the numerous dips of the Fosi vs Yamaha are valid, except for the wide dip in the deep bass range. At the higher frequencies, the combination of mics such as the Umik mics and REW may not reliably record the FR in terms of repeatability. One can do the sweep many times and will see that those sharp dips will vary a lot each time.

Below are my sweeps for the AVR-X3400H, a 100 W AVR, versus the A21, a 250 W power amp paired with a separate preamp, and using the same conclusion the YT reviewer seemed to have drawn, the tiny little AVR might win, but I know that is not the case. My graphs obviously have not much to do with comparing Fosi V3 and Yamaha A-S2100, because in this case it is class AB to class AAB (more like A because of the low SPL at 10 ft), but it does show when compared at low enough output such that both DUTs are operating at well below their limits, there aren't significant differences in FR. If my measurements were taken at 95 dB, that is, the AVR might have been in trouble in the range 50-400 Hz.

With smoothing, in this case, use REW's psychoacoustic:

 

[peng]

You are not wrong. The thing is I can well imagine someone finding e.g. the Fosi V3 on a shop, advertised as 2x280 W @ 4 ohm (with the 48 V 10 A PS) and google it to see reviews, sees the nice numbers and the "recommended" conclusion of course would think that it's perfect for said "full range tower speakers" and thinks it's the bargain of the century.

Of course, some people would believe anything they read or are told. On the opposite end though, this YT review is also misleading, for example, the OP is right that measurements can, or even as he puts it, need to improve, but he was lead to believe Amir's test failed to predict the issue shown in the video, when in fact Amir tests were sufficient in this case as it clearly showed the output limits of the little amp regardless of the advertised 2X280 W 4 ohms. By the way, in that video, was it the 48 V 10A PS used, or the 48 V 5 A one?

 

[voodooless]

I don’t think the measurements need to nessesarily improve. Rather the understanding of what is already there needs to improve.

 

[peng]

It's trivially easy to do a direct electrical measurement. Why this wasn't done and fifty new uncertainties introduced baffles me. It's not even worth discussing until this basic measurement- fully documented- is performed and presented.

I would start by asking questions first, and I have as of yesterday. If answers are given, then we can discussed further. Otherwise, as you said, not worth discussing any more.
As to direct electrical measurements, I suppose the reason it has not been done, though suggested by others, is that the OP seems to have presumed the issue found that he felt could not be predicted, was due to insufficient "electrical measurements" already done by Amir. In my opinion, it has little to do that, but a lot more to do with the way the comparison were made, using SPL vs frequency, and tried to draw conclusions without the necessary supporting background information, most of such information still have not been provided after 7 pages of suggestions and questions.

 

[Sokel]

Ok,I watched the video in FF.
The FR shown on the first post is the two amps trying to do 110db (!) SPL sweep.

I don't think much more can be said as it was already clear by the music test where the 80db average produced 105db peaks with one amp and 117db with the other.

Not entirely a user error as the advertised specs can make someone believe that it does has the ability.
Nope.

 

[peng]

Not entirely a user error as the advertised specs can make someone believe that it does has the ability.
Nope.

How do you know that the amp "has the ability"? Do you have the information on the speaker's sensitivity, impedance curve, and the mic distance?

The advertisements did say something about 600 W (300 WPC I assume), but if anyone believe it can get anywhere close to 1/2 that without clipping badly, then okay I have nothing else to say, but then why even bother discussing it on a forum like ASR.

Even if it can actually do 300 WPC without clipping (that we already know it cannot, based on Amir's measurements), we don't know if it is capable of 117 dB or even 105 dB peak, without knowing the detailed specs of the speakers, and information about the room, and more importantly mic distance.

 

[Sokel]

How do you know that the amp "has the ability"? Do you have the information on the speaker's sensitivity, impedance curve, and the mic distance?

The advertisements did say something about 600 W (300 WPC I assume), but if anyone believe it can get anywhere close to 1/2 that without clipping badly, then okay I have nothing else to say, but then why even bother discussing it on a forum like ASR.

We do have all the info,here in this thread,in Amir's review thread as in the video.
No way to reach what it needs to go up there if we add the numbers

 

[peng]

We do have all the info,here in this thread,in Amir's review thread as in the video.
No way to reach what it needs to go up there if we add the numbers

Really, you mean if I read all the posts, I would have the info I asked in my post#10:

1) What's the sensitivity of the speaker?
2 What is the mic distance from the speaker.
3) Is the question on the wide and deep dip related to when push to apparently 100 dB average spl, or at much lower spl as well?
4) What's the operating temperature, had it been working hard prior to the REW sweeps?

Without 1) and 2), we won't know if the amp(s) were pushed to their clipping point, just the FR curve is not enough. Amir's tests does tell the story, of course, but we are talking about the REW graphs shown in the video.

Anyway, I will do some fast reading and see if I can find the answers to my questions. Would like to see the speaker's impedance curve too.

 

[Sokel]

How do you know that the amp "has the ability"? Do you have the information on the speaker's sensitivity, impedance curve, and the mic distance?

The advertisements did say something about 600 W (300 WPC I assume), but if anyone believe it can get anywhere close to 1/2 that without clipping badly, then okay I have nothing else to say, but then why even bother discussing it on a forum like ASR.

Even if it can actually do 300 WPC without clipping (that we already know it cannot, based on Amir's measurements), we don't know if it is capable of 117 dB or even 105 dB peak, without knowing the detailed specs of the speakers, and information about the room, and more importantly mic distance.

So,we have speakers rated at 89db sensitivity with the impedance curve and details other folks in this thread posted.
Mic distance is at 2 meters as he says in the video.

Any calc tells us that at least 300 watts needed for 115db peaks and as for the result it seems (by the peaks SPL) that this little thing can't go over 80-100 watts (by any definition of peak/dynamic/burst power) depending the placement,room,etc.
Still a lot judging by it's size and price though!

 

[SSS]

Those wouldn't help you here as such dummy loads are low power so are only good for frequency response measurements.

Reactive loads can be made for high power.

 

[kchap]

Isn't the Scientific Audiophile channel just a piss take? Haven't we been sucked in to a long debate on one of his other videos? Did not the the first few minutes of his video seem odd, even by fringe audiophile standards?

 

[Sokel]

Isn't the Scientific Audiophile channel just a piss take? Haven't we been sucked in to a long debate on one of his other videos? Did not the the first few minutes of his video seem odd, even by fringe audiophile standards?

I personally only watched the video to see what REW does in there.
It's results seem comparable if he didn't mess with them.

 

[Ron Texas]

Those wouldn't help you here as such dummy loads are low power so are only good for frequency response measurements.

Is there anything you might propose which would give a greater insight into stability as opposed to what the practice is now?