In a home cinema - sized as kgrlee stated - this clipping power levels are easily reachable. LFE!
30 Hz at 1000W may be loud - but mostly for your jeans ...
I have one 4 channel amp (4 x 200W@8R) only for sub bass / LFE.
Don't forget to take into account that even using a top speaker chassis rated at 300W rms can leave the "linear" drive above 50 - 70W which adds extra distortion.
... no one want's to listen 1000W RMS 1kHz in this room!
BR, Toni
Last edited:
I dunno. Kevinkr reported similar power requirements for piano recordings he had done. I think people forget about the 'forte' in pianoforte. A full sized concert grand will go very loud if thumped correctly. With no gain riding or compression I can easily see big peaks. 20dB crest factor is believable for piano, maybe more.
So whilst this is an extreme case and may have been being played a tad louder than my 75dB ears like I don't see any reason to argue. And don't forget that 1000W only sounds twice as loud to us as 100W, at least until bit of drive unit start escaping
My amplifier using Blameless topology with TMC compensation. It did not use clamping.
Attachments
Yes, I assumed most of the requirement is simply from uncompressed source material.
Say 30 dB peaks, with 90 dB/W (nominal) speakers and 1000 W amp = 120 dB peak or 90 dB ave. at 1 m.
~85 dB at 3 m in the room you described sounds believable, despite my initial doubts, is that in the ballpark?
My latest was a performance by a friend who is world class (invited to play in the Chopin Competition) on a Bosendorfer Imperial at the Australian National University.
People who knock Canberra don't know the joys of the national institutions. And front row seats!
Best wishes
David
If you had smuggled a Sound Devices 788 and TetraMic (21st century equivalent of PCM-F1 & Mk4 Soundfield) into that concert, you would clip a 1000W amp if you tried to replicate that experience in your living room with 90dB/W speakers.
No one is knocking Canberra. It's some of the creatures that haunt Parliament House that we should knock .. or preferably hang, draw & quarter
Yes, that's why I don't have wussy 90 dB/W speakers.
115 dB/W JBL 4" voice coil compression drivers and 15" woofers + 18" subs.
Serious Speakers
Best wishes
David
Sims often don’t show rail sticking because the base charge storage mechanism is often not modeled well. I would look at a real world amp to ascertain if you had the issue in practice. I would just fit theclsmp diodes as standard practice.
Yeah, a bit
I do subscribe to the idea that the amp simply shouldn't clip.
I understand Richard's desire to make sure it clips "nicely", sensible for commercial products that he used to test.
But better if it never clips at all.
I have seen claims that the mechanism is not modelled well but in my experience the problem is not the mechanism but rather, incorrect values for the junction capacitances.
Is this what you meant or do you have an example?
Nice to see you return to activity in the forum too.
Best wishes
David
Clipping simulation with and without clamp diode shows extra distortion added by the voltage dependend diode capacitance mostly correct, but not the correct wave form during clipping. Verified with all my blameless derivates. The measured results using a 150MHz scope compared to the simulation are always worse compared to simulation.
So even a simulation shows nice rounded clipping waves - try to use a fast (!) scope - not the cheap 20MHz ones with low resolution most of us have ...
BTW - for dadod and others: diode capacitance variation is very low during low listening levels because the VAS diode is loaded with half of the rail voltage when mostly idle. There is some explanation in my 2stageef thread ...
BR, Toni
So even a simulation shows nice rounded clipping waves - try to use a fast (!) scope - not the cheap 20MHz ones with low resolution most of us have ...
BTW - for dadod and others: diode capacitance variation is very low during low listening levels because the VAS diode is loaded with half of the rail voltage when mostly idle. There is some explanation in my 2stageef thread ...
BR, Toni
It is spec'd in many products. If you look at page 5 of the BAS21J diode, the 0V capacitance is typically <0.5pF at Vr = 0V and at Vr = 40 volts its about 0.33 pF and a bit lower as the curve goes asymptotic. It is specified as < 2pF in the data sheet however.
Clamp diodes with higher capacitance may show up as a 2-3 ppm increase in HF distortion in sims, but you are going to have a very difficult time in pulling out that specific distortion mechanism in a practical amplifier where other mechanisms are likely to swamp it out.
https://assets.nexperia.com/documents/data-sheet/BAS21J.pdf
Last edited:
Been back about 2 months or so.
In most cases BAV21 has about 5pF and it's highly nonlinear at low reverse voltage. That one from nexperia Mouser does not have (not easy to find).
It's important to look for very low capacitance.
It's easy to replicate modern concert sound in your house: Just dial volume to ten and it will be as unintelligible as most concert sound is - WAY TOO LOUD, lots of room reflections to reduce clarity - just total crap - instant hearing damage - won't matter if the amp is clipping or not.
The BAV21 , as the half wave splitter in my class H CFP ... WAS a problem
@ 20Khz. Its capacitance/reverse leakage had to be R/C snubbered to
get the class H tracking right.
On the Honey badger , the BAV adds up to 10ppm in 20khz THD. Nearly
negating the improvements offered by the transitional miller components.
I tried another "trick" on my blameless , a transistor clamp that was activated
by the excess saturation current at the VAS main Re. Leach used this on his amp.
This technique worked better but had even more of a THD "penalty".
So , I stated to the builders that it was a safety vs. performance "tradeoff".
"just don't clip your amp" !!
Or build one that does not have this "weakness".
Each to their own ....
OS
@ 20Khz. Its capacitance/reverse leakage had to be R/C snubbered to
get the class H tracking right.
On the Honey badger , the BAV adds up to 10ppm in 20khz THD. Nearly
negating the improvements offered by the transitional miller components.
I tried another "trick" on my blameless , a transistor clamp that was activated
by the excess saturation current at the VAS main Re. Leach used this on his amp.
This technique worked better but had even more of a THD "penalty".
So , I stated to the builders that it was a safety vs. performance "tradeoff".
"just don't clip your amp" !!
Or build one that does not have this "weakness".
Each to their own ....
OS
Last edited:
BAV21 is not the same as BAS21J - I stopped using BAV21 many years ago because of the reverse capacitance.
RS has the BAS21J - that's who I get mine from usually.
This ??
https://www.rectron.com/public/product_datasheets/bav21j.pdf
Just a couple PF Ct. ( on the plot).
Might buy a few of those - SMD !!
PS - but I am not using it on a "blameless" , still I
want well under 5pf.
OS
https://www.rectron.com/public/product_datasheets/bav21j.pdf
Just a couple PF Ct. ( on the plot).
Might buy a few of those - SMD !!
PS - but I am not using it on a "blameless" , still I
want well under 5pf.
OS
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.