Amplifier measurements may require improvement

[SimpleTheater]

@SimpleTheater, do you have the WAV files of the sweep of the high SPL Fosi amp test? Would you be willing to share them?

Does REW save them?

 

[voodooless]

Does REW save them?

If not, you can surely extract them somehow as long as you still have the data saved.

 

[SimpleTheater]

If not, you can surely extract them somehow as long as you still have the data saved.

I have the measurements saved, but I'm not familiar with extracting the wav files. Are there instructions for this? If so I can definitely do it, just not familiar with that aspect of REW.

 

[wwenze]

It is only a problem if the new data discovers something that cannot be measured and predicted with old data under reasonable use cases.

If it is clipping, then the data is available as "max power into xxx low impedance load". We have 4 ohm, sure 2 ohm data would be nice, but what about 1 ohm then? Where do we draw the line?

And then any complex load behavior would be the sum/combination of the individual behaviors. Unless, again, if data of complex load reveals an emergent property. But with current information, it isn't.

 

[IPunchCholla]

Buy this

Stand Alone – AudioGraph

www.audiograph.se

Send to Amir.
Problem solved (well, to some extent)

I always thought the power cube was Rubic’s Cube sized…. Weird the assumptions I make.

 

[IPunchCholla]

It is only a problem if the new data discovers something that cannot be measured and predicted with old data under reasonable use cases.

If it is clipping, then the data is available as "max power into xxx low impedance load". We have 4 ohm, sure 2 ohm data would be nice, but what about 1 ohm then? Where do we draw the line?

And then any complex load behavior would be the sum/combination of the individual behaviors. Unless, again, if data of complex load reveals an emergent property. But with current information, it isn't.

Power into a short? But where do we stop at the other end? Power into not a circuit?

 

[voodooless]

I have the measurements saved, but I'm not familiar with extracting the wav files. Are there instructions for this? If so I can definitely do it, just not familiar with that aspect of REW.

Hmm, it seems that is not so easy… can REW view the raw waveform?

Or can you capture the audio at the amp terminals?

 

[Frank2]

Let me add that the ONLY reason this became an issue, which I detail in the video, was I was trying to do a simple level match of the FOSI v3 with a Yamaha amp. The fact that it couldn't break 80 dB was nothing short of astonishing to me since Amir tested this at 88 watts with the 5 amp 48v ps. But that's just at 1 khz. That's when I got out the old Yamaha receiver rated at 50 watts and it blasted past 80 to 93 dB (nothing was changed except the rca cable were swapped - no mic changes, nothing).

Using music, I could definitely get the meter to go well over 100 dB, but music isn't white noise, so in the real world it may not matter to you.

Thanks for joining! Amir also tested other frequencies into 4 ohms by the way. Someone already remarked it could be due to a gain difference between the amplifiers. A low voltage level input signal will then never be amplified beyond a certain output voltage, i.e. the output voltage would be gain limited. So maybe the voltage level of your white noise signal was too low for the Fosi to crank it up beyond 80 dB, and the voltage level of the music track was higher? What was your white noise source, and the SACD's level settings for the channel(s) you used?

(The low frequency dip at 100 dB seems a normal current limiting issue to me, caused by low speaker impedance.)

 

[Frank2]

Another remark: The sharper dips in the frequency sweep of the Fosi could be the result of lower harmonic distortion. If there is destructive soundwave interference at a specific frequency, harmonics of that frequency will still reach the mic and increase measured SPL. Hence, a higher THD makes the dips less visible on your sound pressure meter.

 

[Djano]

@voodooless, Sorry to go back on the previous question, but I would like to understand better.
Earlier, you informed me about the importance of impedance: in essence, the lower it is, the more the amplifier potentially struggles.
However, the V3 can deliver 89W @8ohms before clipping, while it can deliver 141W @4ohms. So, the trend seems to be that the lower the impedance, the more power can be sent before clipping.
Is it possible that in reality, things aren't linear like that? For example, while 4ohms might be easier to drive than 8ohms, is 2ohms more challenging than 4 ?

 

[IPunchCholla]

@voodooless, Sorry to go back on the previous question, but I would like to understand better.
Earlier, you informed me about the importance of impedance: in essence, the lower it is, the more the amplifier potentially struggles.
However, the V3 can deliver 89W @8ohms before clipping, while it can deliver 141W @4ohms. So, the trend seems to be that the lower the impedance, the more power can be sent before clipping.
Is it possible that in reality, things aren't linear like that? For example, while 4ohms might be easier to drive than 8ohms, is 2ohms more challenging than 4 ?

Current is the limiting factor. If my math is right 114 watts @ 4 ohms is 23.7 volts. If the load is 4 ohms, current would be 6 amps. At 2 ohms it would require 12 amps.

 

[Ron Texas]

Believe me, I'm not making a living on YouTube!

The oracle has spoken...

 

[Djano]

Current is the limiting factor. If my math is right 114 watts @ 4 ohms is 23.7 volts. If the load is 4 ohms, current would be 6 amps. At 2 ohms it would require 12 amps.

Thank you!

That makes a lot of sense.

So, I guess the limit @8ohm comes from voltage limitation
89W@8ohm. U²=P*R - U²=89*8=712
-> U@8ohm = 26.6V
I=U/R. I=26.6/8 = 3.325 A

141W@4ohm. U²=141*4=564
-> U@4ohm = 23.7V
I=23.7/4= 5.9 A

From this, can I assume that the max intensity is 5.9 or 6 amps ?

My reasoning is that the device had to lower its V to match its max. A.
If it could provide more than 6A, then U@4ohm would be >23.7V (and closer to 26.6V).
If this is true, then yes, it becomes obvious that anything under 4ohm will output less power, and I understand why clipping happends there

Is that reasoning correct ?

 

[Sokel]

I have the measurements saved, but I'm not familiar with extracting the wav files. Are there instructions for this? If so I can definitely do it, just not familiar with that aspect of REW.

You don't have to extrack anything,just find the measurement (.mdat file) in the file that REW saves them,zip it and upload it here.

 

[SimpleTheater]

Here's the FOSI mdat file

 

[Koeitje]

It could, though. Just Ls, Cs, and Rs (and a calculator).
That said...

There's a reasonableness clause to be applied to loudspeaker models, don't you think?
In 2023, would it be reasonable to test amplifier performance into, e.g., a simulated (or real) Infinity IRS Beta or a Quad "ESL-57"?

Infinity IRS Beta loudspeaker Measurements

Sidebar 3: Measurements (from June 1989, Vol.12 No.6)

www.stereophile.com

The "ESL-57" is, in essence, a large and peculiarly designed capacitor -- albeit a fine-sounding one.

SDS Audio Labs - Quad Loudspeaker Refurbishing Main

Both of the above were/are renowned amplifier killers in their time.
Just one anecdote that happened to be easy to find.

QUAD ESL-57's destroyed a Carver TFM-55x amplifier

I sold a Carver TFM-55x to a guy on another forum who tried to use it to power a pair of unrestored QUAD ESL-57 speakers. Amp quit rather quickly. Preliminary testing shows blown outputs, resistors and diodes across both amp sections as well as a blown 12A slo-blo fuse. Any idea what happened...

audiokarma.org

Those are amateur levels, let me introduce the Apogee Scintilla. That speaker had a 1 Ohm version. Very few amplifiers could drive them, most would just simply melt.

 

[restorer-john]

But that's just at 1 khz.

Exactly.

Time is also important in the collapse of power output at LF. Nobody is taking this into account. Both the power supply capability and the systems built into the IC to limit output.

Do a continuous power bandwidth test.

 

[mhardy6647]

Those are amateur levels, let me introduce the Apogee Scintilla. That speaker had a 1 Ohm version. Very few amplifiers could drive them, most would just simply melt.

I did forget that one, yes.

Truth be told, I always think of the ESL-57 first -- since I have a pair. Transformer coupled vacuum tube amplifiers are happy to shoulder the responsibility when direct-coupled ss amps wheeze into dysfunctionality (sometimes dramatically so).

 

[Weeb Labs]

I should just stay out of any discussion based on videos.

I also find most videos (there are exceptions) to be an inefficient source of information but there are means by which the task can be optimized somewhat. Here's a little diagram. It isn't particularly obvious but it works reasonably well.

There's a Chrome extension, too.

 

[Holdt]

What does the "Distortion" graph look like in REW?
@SimpleTheater