Amplifier measurements may require improvement

[peng]

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!



View attachment 312608

Thanks, I started reading through all the pages anyway, and found the following:

Djano post34........................he uses the 48V 5A power supply (which is rated conservatively accorded to Amir)
Djano post44........................S.A. did his tests with a Tannoy DC8 Ti. I can't find the impedance curve
Frank2 post50......................but it is for the DC8, Djano questioned if it could be assume to have the same curve as the DC8 Ti

Sokel post136....................................distance was 2 meters, but did he confirm the graph was for two speakers, or just one?
Sokel post136....................................speaker sensitivity 89 dB, but do we know for sure it is 89 dB/2.83V/m, or 89 dB/W/m. It makes a difference depending on how Tannoy measured sensitivity at the time, for example, was it based on a specific frequency range, and how they based the "nominal" impedance on.

So yes, I missed a few things, but still missing:

- Did S.T. confirmed the speakers were DC8, DC8Ti or something else, I am not sure if Djano got the DC8 from S.T., or from the video?
- Are we sure the REW plots were for two speakers making noise or just one? I guess it is, just don't want to assume.
- Operating temperature at the time of measurements, if he had been playing around for a while before doing those sweeps, the amp might have already been overheating.
- Speaker sensitivity basis unclear, though this one is hard to know without taking more measurements.

Based on Amir's reasonably comprehensive measurements, we can ignore all of the above questions, but if we have all the accurate answers then we can assess better as to how severly the amp was pushed during S.P.'s test. We do know that THD was very high, based on the REW file, >50% at the lowest dip at around 70 Hz.

 

[Sokel]

Based on Amir's reasonably comprehensive measurements, we can ignore all of the above questions, but if we have all the accurate answers then we can assess better as to how severly the amp was pushed during S.P.'s test. We do know that THD was very high, based on the REW file, >50% at the lowest dip at around 70 Hz.

There are more measurements in the video with more reasonable SPL where the differences between the two amps are minor,nothing really.
Is when it takes it to the extreme of trying to sweep it with 110db SPL that collapses (no surprise really) .

 

[Deleted member 48726]

Thanks, I started reading through all the pages anyway, and found the following:

Djano post34........................he uses the 48V 5A power supply (which is rated conservatively accorded to Amir)
Djano post44........................S.A. did his tests with a Tannoy DC8 Ti. I can't find the impedance curve
Frank2 post50......................but it is for the DC8, Djano questioned if it could be assume to have the same curve as the DC8 Ti

Sokel post136....................................distance was 2 meters, but did he confirm the graph was for two speakers, or just one?
Sokel post136....................................speaker sensitivity 89 dB, but do we know for sure it is 89 dB/2.83V/m, or 89 dB/W/m. It makes a difference depending on how Tannoy measured sensitivity at the time, for example, was it based on a specific frequency range, and how they based the "nominal" impedance on.

So yes, I missed a few things, but still missing:

- Did S.T. confirmed the speakers were DC8, DC8Ti or something else, I am not sure if Djano got the DC8 from S.T., or from the video?
- Are we sure the REW plots were for two speakers making noise or just one? I guess it is, just don't want to assume.
- Operating temperature at the time of measurements, if he had been playing around for a while before doing those sweeps, the amp might have already been overheating.
- Speaker sensitivity basis unclear, though this one is hard to know without taking more measurements.

Based on Amir's reasonably comprehensive measurements, we can ignore all of the above questions, but if we have all the accurate answers then we can assess better as to how severly the amp was pushed during S.P.'s test. We do know that THD was very high, based on the REW file, >50% at the lowest dip at around 70 Hz.

It's DC8Ti. Cut from video-->

Impedance graph has been posted (from the DC8T) not sure what the difference between DC8T vs. DC8Ti is though. And the phase graph is missing as well..-->

Amplifier measurements may require improvement

The "right" way to do it, of course, is to measure the amp's electrical performance into a reactive load that simulates a loudspeaker. It is, of course, not a perfect test (like Camelot*, it's only a model ;)) but it is controlled and reproducible. Stereophile has done this for millennia...

www.audiosciencereview.com

 

[peng]

There are more measurements in the video with more reasonable SPL where the differences between the two amps are minor,nothing really.
Is when it takes it to the extreme of trying to sweep it with 110db SPL that collapses (no surprise really) .


View attachment 312620

View attachment 312622

Thanks again, yeah now that you posted those, save me time to watch that long video, it is becoming too obvious what the issue is. I think even the OP may agree now that it isn't about lack of electrical measurements, but about the amp's actual output capability, 110 dB spl, 2 meters, is just way too much. I asked about the 89 dB sensitivity for my ocd, but also the fact that 3 dB would make a big difference, as - 3 dB in sensitivity would make the amp's clipping point 3 dB (half the power output) earlier.

 

[kchap]

It's DC8Ti. Cut from video-->
View attachment 312621

Impedance graph has been posted (from the DC8T) not sure what the difference between DC8T vs. DC8Ti is though. And the phase graph is missing as well..-->

Amplifier measurements may require improvement

The "right" way to do it, of course, is to measure the amp's electrical performance into a reactive load that simulates a loudspeaker. It is, of course, not a perfect test (like Camelot*, it's only a model ;)) but it is controlled and reproducible. Stereophile has done this for millennia...

www.audiosciencereview.com

I check your profile and the links in your signature. Are you sure we not being punk'd? Maybe punked. Off to bed not; I will read all about it tomorrow

 

[Deleted member 48726]

I check your profile and the links in your signature. Are you sure we not being punk'd? Maybe punked. Off to bed not; I will read all about it tomorrow

What do you mean?

 

[jhaider]

The problem as I see it is arriving at a universally accepted 'typical' loudspeaker load and what it presents to an amplifier.

Dummy loads may be boring, but they are at least standardized.

Does the actual impedance curve of the “simulated speaker” matter though, as long as it varies with frequency similar to the way speakers do (peaks at driver resonances, wiggles in between? No matter what it is, the analysis is the same - deviations in FR beyond measurement error tolerance = bad.

 

[IPunchCholla]

Does the actual impedance curve of the “simulated speaker” matter though, as long as it varies with frequency similar to the way speakers do (peaks at driver resonances, wiggles in between? No matter what it is, the analysis is the same - deviations = bad.

Interesting idea. I'll have to think about it. But if the actual shapes of the curves don't matter, wouldn't that imply we only need to test fixed values that represent the highest peaks and lowest valleys typically seen? And maybe a nominal average?

 

[jhaider]

Interesting idea. I'll have to think about it. But if the actual shapes of the curves don't matter, wouldn't that imply we only need to test fixed values that represent the highest peaks and lowest valleys typically seen? And maybe a nominal average?

Intuitively, the rate of change as well as the actual impedance could matter. I can only guess from intuition as I can’t claim relevant expertise though.

 

[peng]

I think it matters for people who want to drive 4 ohm nominal rated or 8 ohm with major dips with amps that barely meet the power output requirement for their applications.

For those who use amps with output 3x (for me I would go for 5 times) or more than their need, phase angle will unlikely be an issue, with rare exceptions. Still, I think manufacturers should at least provide the impedance and phase angle curves on request, by owners, dealers and reviewers.

 

[pma]

This thread is a storm in a tea cup.

 

[IPunchCholla]

This thread is a storm in a tea cup.

I mean, sure. He isn't running ultrasonics at crazy high power in to a load that at points approximates a short circuit, something that 1. could never happen in the real world, and 2. is specifically cited by the amp manufacturer as out of spec. But what are we going to do?

 

[Deleted member 48726]

I mean, sure. He isn't running ultrasonics at crazy high power in to a load that at points approximates a short circuit, something that 1. could never happen in the real world, and 2. is specifically cited by the amp manufacturer as out of spec. But what are we going to do?

I see what you did there.

 

[Sokel]

I mean, sure. He isn't running ultrasonics at crazy high power in to a load that at points approximates a short circuit, something that 1. could never happen in the real world, and 2. is specifically cited by the amp manufacturer as out of spec. But what are we going to do?

I don't see the reason why not run it to extremes (as this video also does),I can think of lots of useful data by it.
Limits,protection schemes,careless user practices,cannibals (yep,that's me),etc.

At this thread one debate over the years is nicely short of demonstrated and it's not the inability of the poor thing to do 110db SPL on the sweep.

It's the inability to play 85db SPL average without killing all the peaks with music that contains it.
Useful result in my book.

 

[Djano]

Thanks again, yeah now that you posted those, save me time to watch that long video, it is becoming too obvious what the issue is. I think even the OP may agree now that it isn't about lack of electrical measurements, but about the amp's actual output capability, 110 dB spl, 2 meters, is just way too much. I asked about the 89 dB sensitivity for my ocd, but also the fact that 3 dB would make a big difference, as - 3 dB in sensitivity would make the amp's clipping point 3 dB (half the power output) earlier.

Yes, I now think that the problem is nothing that the measurments could not predict. In my understanding, we can explain it as follows:

The SPL limit in sub-bass is due to clipping. Clipping makes lower SPL because the signal shape and amplitude does not allow for optimal and maximal speaker excursion.

There is problematic clipping that has big consequences in bass for three reasons. 1) Bass is more power hungry for same SPL, so too much amps were asked. 2) The amp has more difficulty with bass, as depicted in the 4ohm Amir's graph that I did not focused on, because A.S. (=S.T.) speakers were 8ohm. 3) DC8Ti probably has a lowering of impedance in the problematic area, which would require more amps than the amplifier is capable of.

---

It remains to be verified if the DC8Ti indeed have lowering of impedance as said in 3)
It also remains to be clearly explained why the amp struggles more at 45-140Hz than 20Hz. This is not coherent with explanation 1) and 2). Maybe it is because the peak of speaker impedance is at 20Hz, as @voodooless suggested if I remember right. A.S./S.T. also suggested there is a peak power reserve, which became depleted at the start of problematic area.

At this point, it seems reasonable to assume that, if I thought that measurments could not predict the most important aspects of what we saw, it was only because:
- I did not focus enough on a critical measurment (see explanation 2)
- I lacked knowledge about current as an important limiting factor, so I did not understand the importance of the impedance curve of the speaker.

 

[Sokel]

Maybe it is because the peak of speaker impedance is at 20Hz, as @voodooless suggested if I remember right. A.S./S.T. also suggested there is a peak power reserve, which became depleted at the start of problematic area.

One can always do a reverse sweep,if the results are the same it's not only about the power needed for lows.

 

[Djano]

One can always do a reverse sweep,if the results are the same it's not only about the power needed for lows.

Or a continuous xHz (in the problematic area) test, as I suggested to S.A./S.T
If his theory is right, SPL should start high and go lower.

 

[IPunchCholla]

I don't see the reason why not run it to extremes (as this video also does),I can think of lots of useful data by it.
Limits,protection schemes,careless user practices,cannibals (yep,that's me),etc.

At this thread one debate over the years is nicely short of demonstrated and it's not the inability of the poor thing to do 110db SPL on the sweep.

It's the inability to play 85db SPL average without killing all the peaks with music that contains it.
Useful result in my book.

If one wants to test the limits of amplifiers, I have no issue with that. I was more poking a little at the idea that ultrasonics at high energy with a load that approaches short causing issues in an amp makes it unfit for purpose (which was stated in that thread). I find this thread more useful and realistic, even if the conclusion in the video the OP posted is off. The tested amp did struggle significantly under full power into this speaker. That is a good piece of info.

I missed the peak compression at 85dB since I didn't watch the video. That is also a good thing to demo.

 

[Sokel]

I missed the peak compression at 85dB since I didn't watch the video. That is also a good thing to demo.

You can see it at my #132 post.

 

[peng]

Yes, I now think that the problem is nothing that the measurments could not predict. In my understanding, we can explain it as follows:

The SPL limit in sub-bass is due to clipping. Clipping makes lower SPL because the signal shape and amplitude does not allow for optimal and maximal speaker excursion.

There is problematic clipping that has big consequences in bass for three reasons. 1) Bass is more power hungry for same SPL, so too much amps were asked. 2) The amp has more difficulty with bass, as depicted in the 4ohm Amir's graph that I did not focused on, because A.S. (=S.T.) speakers were 8ohm. 3) DC8Ti probably has a lowering of impedance in the problematic area, which would require more amps than the amplifier is capable of.

---

It remains to be verified if the DC8Ti indeed have lowering of impedance as said in 3)
It also remains to be clearly explained why the amp struggles more at 45-140Hz than 20Hz. This is not coherent with explanation 1) and 2). Maybe it is because the peak of speaker impedance is at 20Hz, as @voodooless suggested if I remember right. A.S./S.T. also suggested there is a peak power reserve, which became depleted at the start of problematic area.

At this point, it seems reasonable to assume that, if I thought that measurments could not predict the most important aspects of what we saw, it was only because:
- I did not focus enough on a critical measurment (see explanation 2)
- I lacked knowledge about current as an important limiting factor, so I did not understand the importance of the impedance curve of the speaker.

Agreed, you summed it up really good and sound logical. I would add that once the amp got pushed to that point, it might become more difficult to explain much more details than you already have, as the "rules" we normally tend to apply in troubling it may not apply that much any more.

Also, the better response at the deep bass end you cited earlier may be of much shorter duration, aside from perhaps higher impedance, due to the way REW sweeps, resulting in a little lower distortions. On that possibility, I have to think about it more, just thought it may be a factor too.

For C$120, I may buy one to replace my 50 W NAD I have been using for my desktop speakers (KEF LS50). Then I can sell the NAD for C$350.