AVS Forum banner

41 - 60 of 430 Posts

·
Registered
Joined
·
875 Posts
Discussion Starter · #41 ·
For fs sake, this again? I was trying to have a constructive conversation here, and you completely ignore all that and go off on this again. I was clarifying since the numbers would be different to those looking. I even said "Just clarifying since this horse has been beaten to death"
Your post read "my numbers are actual watts, not 'peak'." Actual watts, implying the same ol' argument that there's no such thing as peak watts. Why do you have to bring this up when by now everyone should be aware of both your and my testing methodology because we've both been over it too many times now?

sigh.. I actually was, I was making light of the past crap and actually trying to start a meaningful discussion, but you ignored the rest of the post completely. I even said it was nice to see you sharing data. I'm actually trying to get along here.
You'll just have to forgive me then if I'm being a bit defensive but it was under a week ago in another thread and you were being anything but friendly to me. If you want to be decent, I'll play nice but you gotta cut out the snarky quips at my expense routine, I've had about enough of it.

Note the peaks are at almost identical levels on top and bottom.
So what's the difference in using the positive peak unclipped value like I do? Multiply it by .707 if you must but at least it would be a clean unclipped signal.

Furthermore, since there's some flat topping, the RMS values are actually higher than using the p-p numbers. In the end, it comes out to within .1-.2db. The temperature of the amp and unit to unit variances will introduce more error than this method.
So you're admitting that there is error in your method, this was my point for like a year now. You aren't really giving the "actual power" of 1.5 cycles so why are you telling people that when you know you're not?

I wasn't even harping on your method, so just stop. I was clarifying that it wasn't 'peak' watts so the numbers won't line up.
You stop first and I will. Hey, like I said: I don't care what methodology you adopt for 2010 bursts, there's no standard but you started this by telling me I was wrong so you're going to get it back unless you drop it.

fluke 87-v doesn't report watts. Not sure what a peak voltage rating from a basic volt meter has to do with this discussion. It's not a tool to measure complex signals.
Also, their docs say it displays min and max readings.
Maybe you contact them and ask for a user manual ;)
Comon man, you really think I didn't test my tools? This ain't my first day. :rolleyes: Oh, just the same snarky talk down to me BS. Must be more of that trying to get along stuff I keep hearing about.

The Fluke agrees with the Rigol within 2-5% with the positive most peak even in dynamic source material. Meaning: you can find the peak wattage using both tools which is fairly significant seeing as it's a multimeter. Why do you think they went to the trouble to design the functionality into the tool? ;)
 

·
Registered
Joined
·
875 Posts
Discussion Starter · #42 ·
For clarity, are you saying that the phase should be set to 180 in the inuke when connecting from rca(avr) to xlr(inuke)? Or is that only when using vintage equipment that uses the different hot pin? Or did I read into all of that wrong?
I tested a version 3 of the DSP6k. I don't know if all versions with and without have the same problem but I suspect it based on Brian O. mentioning it and in my experience dialing other people's systems. I was thinking it was the minidsp or a mixup in the wiring up to now but evidently not.

The easiest thing to do is test your system set 180 out of phase and then reverse polarity. See which sweep shows better response at the x-over. If you don't have a measurement system, or the DSP version of the amp, reverse the speaker leads for both channels and listen to source material that you know well and listen for the quality of the mid-bass.
 

·
Registered
Joined
·
10,567 Posts
Your post read "my numbers are actual watts, not 'peak'." Actual watts, implying the same ol' argument that there's no such thing as peak watts. Why do you have to bring this up when by now everyone should be aware of both your and my testing methodology because we've both been over it too many times now?



You'll just have to forgive me then if I'm being a bit defensive but it was under a week ago in another thread and you were being anything but friendly to me. If you want to be decent, I'll play nice but you gotta cut out the snarky quips at my expense routine, I've had about enough of it.



So what's the difference in using the positive peak unclipped value like I do? Multiply it by .707 if you must but at least it would be a clean unclipped signal.



So you're admitting that there is error in your method, this was my point for like a year now. You aren't really giving the "actual power" of 1.5 cycles so why are you telling people that when you know you're not?



You stop first and I will. Hey, like I said: I don't care what methodology you adopt for 2010 bursts, there's no standard but you started this by telling me I was wrong so you're going to get it back unless you drop it.



Comon man, you really think I didn't test my tools? This ain't my first day. :rolleyes: Oh, just the same snarky talk down to me BS. Must be more of that trying to get along stuff I keep hearing about.

The Fluke agrees with the Rigol within 2-5% with the positive most peak even in dynamic source material. Meaning: you can find the peak wattage using both tools which is fairly significant seeing as it's a multimeter. Why do you think they went to the trouble to design the functionality into the tool? ;)
So you may have 2-5% error then that way? :)

The reason this whole crap started was because you were using "peak watts" from instantaneous spikes to make it sound like the amps you guys were selling were putting out more power, but again, this is beating a dead horse. It was about peak vs actual watts, not using the single center peak from cea-2010. Again, we've beaten this to death, so let's agree move ahead with some civility.

None of these methods are perfect, we're testing into purely resistive loads after all. They get close enough, though.

Also for ****s, lets compare our numbers. Granted, NU4-6000 and NU6000DSP. Shouldn't be too different.

cea2010
freq, your test, my test
5hz 1389 2227
10hz 2326 2553
20hz 2587 2749
30hz 2769 2953
100hz 3609 3781
1khz 4977 4043 this is voltage limited, lower impedance load explains this difference.

The 5hz results are interesting, the rest line up.

My sustained my numbers were a bit higher. This was on a dedicated circuit with no other load and only for 1-2 seconds. The PS can do about 2400w for this duration without much issue. Again, I saw much more linear results which is interesting. The power supplies on these look to be just about the same.
5hz 1315 1980
10hz 1872 2356
20hz 2147 2422
30hz 2093 2422
100hz 2161 2489
1khz 2100 2489




Taking this from MadHek's thread



There's some decent rolloff on the inputs. I imagine your output device has enough voltage to compensate for this. Did you graph your input along with the output to ensure no clipping there? Still not sure why less power on the 6kdsp.

Anyway, including those previous questions again:

The idle load you reported seemed high. How did you measure this? Was the load connected? Were the input hot/cold shorted?

Did you test output noise with the input hot/cold shorted? What probing method? Any diff probe?
 

·
Registered
Joined
·
1,416 Posts
I tested a version 3 of the DSP6k. I don't know if all versions with and without have the same problem but I suspect it based on Brian O. mentioning it and in my experience dialing other people's systems. I was thinking it was the minidsp or a mixup in the wiring up to now but evidently not.

The easiest thing to do is test your system set 180 out of phase and then reverse polarity. See which sweep shows better response at the x-over. If you don't have a measurement system, or the DSP version of the amp, reverse the speaker leads for both channels and listen to source material that you know well and listen for the quality of the mid-bass.
Do you have other amps you tested that are not showing this same polarity that you are seeing as the Inuke? From everything I have read the Inuke is using the most current standard so if you have other amps showing something different I would think they are using the old standard.

If so what amps are they and I'll try and look up which standard they use.
 

·
Registered
Joined
·
1,416 Posts
For clarity, are you saying that the phase should be set to 180 in the inuke when connecting from rca(avr) to xlr(inuke)? Or is that only when using vintage equipment that uses the different hot pin? Or did I read into all of that wrong?
It depends on if you are using other amplifiers like the integrated amp of an AVR or another amplifier that is connected using RCA to RCA. But it has to be RCA to RCA on the amp to AVR/pre amp for it to be reversed of what the Inuke is or using the integrated amplifier of an AVR.

Like I said on my Marantz 7702 going from my Sunfire amp using XLR it is in phase with the Inuke. It is only when going from the Marantz using RCA to the Sunfire RCA that it will be reversed. Using an RCA to XLR adapter cable it is still in phase with the Inuke because it is the XLR that is the reverse of the RCA's polarity.

I also have a Marantz 5010 that is how I verified that using the integrated amplifier will be 180 out of phase with the Inuke. I used the impulse response in REW to check the phase.

Just add more confusion for people that are mix ID subs with XLR input they will be out of phase if they mix them with Inukes. I take it Speakerpower XLR is usings the old standard/USA standard. Which in a way makes sense as it simplifies it for people adding/mixing subs from an ID company with subs connected via RCA and using the AVRs integrated amps to power their speakers. I verified this with my Submersive that using XLR to XLR and RCA to XLR is out of phase with the Inuke 3000 and 6000 I have.
 

·
Registered
Joined
·
875 Posts
Discussion Starter · #46 ·
NU6000DSP Limiter Measurements...

I fed the input enough signal for max clean output for a sustained sine at 30Hz and only changed the limiter settings.
I calculate the actual change measured from the unlimited output in dB in blue.




Here's an example of the waveform when being limited.
It looked like this on all recorded measurements.


Note: for the top limiter data at 5.2ohms, I forgot to write down the DSP's guessed watts figure so I just calculated it. I no longer have the amp and can't get the software to give me the 4 ohm load figures without connecting to it. Someone can correct me if they feel up to double checking.

The waveform is not too bad looking but I honestly wouldn't really know where to set the limiter unless I had my tools since the voltage peak value in the DSP is just way off. It's like the software guys didn't communicate with the hardware guys. I hope this helps some of you guys who use this amp but I realize that most users are trying to squeeze every drop of clean power out of the amp as it is.
 

Attachments

·
Premium Member
Joined
·
4,384 Posts
Excellent posts !


More about the low end roll off please.


I am using this....well a 1000dsp anyway, amp to power a sub/subs as I imagine most are.


Here is a sweep of what I am getting in room with no mods running the 1000dsp model......I am using a minidsp as the input to the inuke and adding 16db at 10hz, not represented in the sweep picture. Sweep only has negative dsp at around 53hz to tame down a peak.


Based on the input roll off you pictured, 16dd at 10hz with a .5 Q is probably just making it flatter input and not a real world 16db boost.


Just trying to learn more about all this and this looks to be solid info being presented here in this thread.
 

Attachments

·
Registered
Joined
·
10,567 Posts
Excellent posts !


More about the low end roll off please.


I am using this....well a 1000dsp anyway, amp to power a sub/subs as I imagine most are.


Here is a sweep of what I am getting in room with no mods running the 1000dsp model......I am using a minidsp as the input to the inuke and adding 16db at 10hz, not represented in the sweep picture. Sweep only has negative dsp at around 53hz to tame down a peak.


Based on the input roll off you pictured, 16dd at 10hz with a .5 Q is probably just making it flatter input and not a real world 16db boost.


Just trying to learn more about all this and this looks to be solid info being presented here in this thread.
16db at 10hz is a lot of boost. They're only down 3.5db or so at 10hz
 

·
Premium Member
Joined
·
4,384 Posts
16db at 10hz is a lot of boost. They're only down 3.5db or so at 10hz

Here in winisd is the transfer function screen shot of my build, which shows it down 22 db at 10hz. Plus the 3.5db of the amp inputs...25.5db down.


I guess I need to do a new rew sweep to see where the 16db hot at 10hz puts it..........I used 2 dsp peqs at 10hz to bump it to 20db hot but backed it down to 16db at 10hz out of fear of destruction.:)


The transfer function showing down -22db at 10hz is identical to the SPL tab as it shows the same curve with the max at 118db at a little over 50hz, which is why I have -7db or so at 51hz with a .5Q to get a flatish line from as low as possible.
 

Attachments

·
Registered
Joined
·
10,567 Posts
Here in winisd is the transfer function screen shot of my build, which shows it down 22 db at 10hz. Plus the 3.5db of the amp inputs...25.5db down.


I guess I need to do a new rew sweep to see where the 16db hot at 10hz puts it..........I used 2 dsp peqs at 10hz to bump it to 20db hot but backed it down to 16db at 10hz out of fear of destruction.:)


The transfer function showing down -22db at 10hz is identical to the SPL tab as it shows the same curve with the max at 118db at a little over 50hz, which is why I have -7db or so at 51hz with a .5Q to get a flatish line from as low as possible.
Just saying that's a lot of boost for the amp/woofer. 13db is about 20x the power requirement.
 

·
Premium Member
Joined
·
4,384 Posts
Just saying that's a lot of boost for the amp/woofer. 13db is about 20x the power requirement.
Which is why I think I have not damaged the speaker yet as I do not have that kind of power....:D



I designed the box to take full amp power with no HPF, that was the reason for the clipping thread also........I guess I listen low enough to not clip the amp with that kind of boost as it only lights up 3 of the front lights on the inuke sometimes....mostly not even 3 lights and never 4.


I would guess now based on my understanding gained from this thread........I am possibly getting some massive distortion and or signal clipping......it sounds good to me, so what do I know.......just trying to learn/understand it all.


+16db at 10hz makes for some great bass anyway...at this point...until I learn better or something lets go.......;)


Maybe that's part of why I am so impressed with the HT-18.....:eek: A tad bit of unnatural boost makes it really seem like it has some power.

I better shut up as I have deviated way away from the part of lack of understanding that concerns the inuke amp.....


EDIT:...That's not counting running the LFE on the AVR at +12db.......:eek: I am starting to realize I might have a small bass craving I am tryin to satiate here.
 

·
Registered
Joined
·
4,344 Posts
Sounds like you need a bigger amp:)
 

·
Registered
Joined
·
1,416 Posts
Gotcha, so being I go rca from avr to rca on my b&k's, I likely need to go 180 on the inuke phase.

I did get a mic recently but haven't had time to do any measuring. I will have to check that out soon. That's crazy
Correct, when you do get your mic and if you verify it with an impulse response would post back to confirm it is correct. I know it is from what I have tested but maybe other equipment or brands will differ and at times it makes me question my sanity :confused: when everything is different. I use to think it was me or my setup was screwed up until I started digging into it and then seeing others confirm it in this thread and other threads.

In the past, I had posted for people to try their rear subwoofer out of phase because that is what I had found worked best for me for phase and FR. It didn't click until last night why that was. It was because I was using the Inuke on my rear subwoofers and had been running them combined with ID subwoofers connected using RCA to RCA for the other subwoofers placed in the front of my room and that is why they were 180 out of polarity.
 

·
Registered
Joined
·
875 Posts
Discussion Starter · #56 ·
Hulk Punch Scene...

I played the scene in TIH where Hulk punches Abomination into the ground. It has major ULF and is a scene that demands a lot from the amp of a sealed system. Here is the tested scene:



Here is the preamp signal of the sub out feeding the amp:


The major ULF hit in the scene lasts ~1.7 seconds and has strong content from 5Hz to 18Hz. I eased the signal up until it was just nipping the clip lights with stereo 5.2 ohm loads and recorded the waveform and calculated the peak wattage that the amp was able to put out clean:

The total for both channels comes to 6482.32Wpk.

Here is the same test into 10.43 ohms with only 1 channel driven:

 

Attachments

·
Registered
Joined
·
21,343 Posts
paul, this peak watts thing is out of control. you just reported that "The total for both channels comes to 6482.32Wpk."

3240 watts is a good enough showing.
 

·
Registered
Joined
·
21,343 Posts
Please show us how to calculate accurate RMS voltage from the above complex waveform to come to 3240Wa.

we can drill down with one. the same concept applies to them all.


if we are generous, we could allow for the max amp power to be calculated over 1/2 a wavelength. anything less than that is dubious. better would be *at least* one full cycle, but i'll work with 1/2 wavelength here.


what you posted for the right channel into 5.2 ohms. -128v was used to calculate the 3150.77 "peak watts".







zooming in on that one half wave and overlaying a 1/2 sine wave gives this:










the power across the 1/2 cycle is peak volts * 0.707 to get rms volts. then v^2/ohms to get power over the half wave.


in this case, 128v * 0.707 = 90.5v


90.5v^2 / 5.2 = 1575 watts


now, that is the theoretical maximum because the voltage doesn't follow a sine wave. what the actual rms is over that half wave form, i can't say**, but it cannot be any more than 90.5v because the measured half wave is less than or equal to the sine wave at every point.






** with careful geometric analysis of the 25 pixels, the curve could be approximated reasonably well. then just manually calculating the rms should get it quite close.


yes, it would be tedious (hey that's what folks were doing before newton and leibnitz), but this is essentially the idea.







it should be obvious that the rms of the blue bars will be less than the rms of yellow curve, so the watts calculated for the yellow curve would represent the maximum possible value, the actual calculated value would a good bit low just eyeballing how much empty space there is--probably something in the 80-90% of the theoretical power based on the yellow curve, which would take the 1575 watts down to about 1350 or so. of course, that is one channel, the same would be done for the other, so even 3240 watts is overstating things by a fair amount.


just to be clear, i'm not suggesting this level of analysis is necessary. calculating peak power based on the rms for the best 1/2 wave form that you capture is probably good enough, so long as it isn't waaaaaay too far off being a sine wave.








.
 

Attachments

·
Registered
Joined
·
875 Posts
Discussion Starter · #60 · (Edited)
the power across the 1/2 cycle is peak volts * 0.707 to get rms volts. then v^2/ohms to get power over the half wave.
in this case, 128v * 0.707 = 90.5v
90.5v^2 / 5.2 = 1575 watts
Or divide the peak number I give by 2 for a very rough idea, yeah I've been saying this for about a year now.

now, that is the theoretical maximum because the voltage doesn't follow a sine wave.
So me measuring the actual undisputed peak of the waveform is me being "out of control" but "theoretical maximums" of a very complex waveform that doesn't resemble a synthetic tone at all is cool? :confused:

what the actual rms is over that half wave form, i can't say**
You can't even say that, let alone over the 1.7 second passage with content on the whole sub bandwidth. You and I both know that this is my point.

yes, it would be tedious (hey that's what folks were doing before newton and leibnitz), but this is essentially the idea.
They sure weren't doing that to a waveform like the hulk punch scene. :p Look John, I understand your idea and if you would like me to zoom into every nook and cranny of this waveform and email the screen caps to you for your manual napkin calculations I wouldn't mind doing that and the data would be interesting but I simply don't have the time to put forth myself to what you're suggesting and even if I did, I wouldn't do it because peaks tell me more of the story.

it should be obvious that the rms of the blue bars will be less than the rms of yellow curve, so the watts calculated for the yellow curve would represent the maximum possible value, the actual calculated value would a good bit low just eyeballing how much empty space there is--probably something in the 80-90% of the theoretical power based on the yellow curve, which would take the 1575 watts down to about 1350 or so.
So we've got some eyeballing, some probably and some theoretical power. Sounds like some rock solid data to me.

just to be clear, i'm not suggesting this level of analysis is necessary. calculating peak power based on the rms for the best 1/2 wave form that you capture is probably good enough, so long as it isn't waaaaaay too far off being a sine wave.
All real world dynamic content is far off from synthetic tones. The area that you've analyzed is a little over 40ms in duration for the half cycle. The top portion (half cycle) of the 10Hz CEA-2010 I measured from this amp is very close to that but the burst peaks out at 110V while the section that you looked at peaks at 128V. Why the difference? Because an amp being fed real world content doesn't behave the same as with any synthetic test.

Don't get me wrong, synthetic tests have their place and can teach us things but in the past 20 years we've seen theater subwoofer technology grow in leaps and bounds... If you could go back in time and play the guys who came up with these standards the scene from Hulk at reference level, they would run out of your house screaming "Witch! :eek:... Burn him!"

And even today there aren't a lot of people in the world using these high powered amps for our applications and all I'm trying to do is come up with new ideas for testing the limits. I'm the first guy to test a subwoofer amp with bandwidth limited real world content to see where it's limits are and all I get is S*** from you guys. I don't really understand that. Like you guys don't find this interesting in the least, or understand how to interpret the data or what is it? :confused: In other words, if it was up to you we'd still be using graph paper and an SPL meter instead of REW and a USB measurement tool?

With the tools I have at my disposal, all I can do is to keep the amp output clean by comparing it to the input signal and then analyze the peaks as a comparative tool. You all talk about headroom, well peaks show the ceiling. There is no cheating with peaks. You can't raise the level and have the value go up like RMS. Doesn't matter if the amp is voltage or current limited, the peak is the peak and it's either clipped or it's clean.

John, when I post things like this and you feel the need to come in behind me and post your guess as to what the average power is I really don't have a problem with that. Just don't say things like I'm out of control. That's dismissive and disrespectful and I've never had a problem with you that warrants that.
 
41 - 60 of 430 Posts
Top