UCD 25 watts to 1200 watts using 2 mosfets

Jump to Latest
#3,043
Hi, gents, I am not pouring cold water over this hot topic. But, I am struggling to understand how could this Class-D technology be applied to make Hi-Fi amps, given the high level RF noise at the output. From the schematic at post1#, I figure out that the L&C forms a LPF, that would prevent most of the carrier getting into the spk system. However, I think this is only a 12db/oct LPF and there will be significant residual RF pass the LPF and get into your speakers. For a rough calculation, if you make the switching freq. at 400kHz, and you measure how much carrier get though the 12db/oct LPF at 20kHz, then you will get (-12)*log(400000/20000,2) = -51.86db, which is about 0.255%. That is to say the THD+S/N at 20kHz must be greater than 0.255%. Even if you push the SW freq. into 1MHz, very unrealistic though, the residual at 20kHz would still be -67.7db, which is equal to 0.04%. If the above calculation is correct, A 40 year old JLH class-A design can easily beat the class-D in terms of THD+S/N.

Neverthless I think this class-D design is good enough for a sub-woofer amp. If using the same switching freq. at 400k, you will get only -131.59db noise at 200Hz.
 
#3,045
jazzclassics said:
Hi, gents, I am not pouring cold water over this hot topic. But, I am struggling to understand how could this Class-D technology be applied to make Hi-Fi amps, given the high level RF noise at the output. From the schematic at post1#, I figure out that the L&C forms a LPF, that would prevent most of the carrier getting into the spk system. However, I think this is only a 12db/oct LPF and there will be significant residual RF pass the LPF and get into your speakers. For a rough calculation, if you make the switching freq. at 400kHz, and you measure how much carrier get though the 12db/oct LPF at 20kHz, then you will get (-12)*log(400000/20000,2) = -51.86db, which is about 0.255%. That is to say the THD+S/N at 20kHz must be greater than 0.255%. Even if you push the SW freq. into 1MHz, very unrealistic though, the residual at 20kHz would still be -67.7db, which is equal to 0.04%. If the above calculation is correct, A 40 year old JLH class-A design can easily beat the class-D in terms of THD+S/N.

Neverthless I think this class-D design is good enough for a sub-woofer amp. If using the same switching freq. at 400k, you will get only -131.59db noise at 200Hz.
Click to expand...
there's not a lot of tweeter that's able to play high frequency over 20khz. even then, most people won't be able to hear it anyway.

THD alone don't reflect the actual sound quality. my advice is, build it, test it on your own and deduce your own conclusion. i've built a few of the designs posted in this thread and while i won't personally use them as the resident amp, it does sound quite musical.

heck it sounds even more musical than some single ended class a tube amp that i've heard before.

i'm personally running the irs900d for months now as my subwoofer amp.
 
#3,046
I agree with you that THD along does not reflect sound quality. In most of time, I am trying to be objective. That is why I looked at the schematic and theoritical effect first. I am trying to understand why such an advanced technology is yet to be accepted by the Hi-Fi audience. I know that the renowned Jeff Lowance has, after making some successful Class-D amps, reverted to a class-AB approach.

I am using a pair of Fostex FE166En full rangers which has a response well extended to 22kHz, and even that speaker is suggested, by the Fostex in the manual, to accompany with a super tweeter. Nowadays the Hi-Fi is really going to "Hi-Freq". Looking at the spec for some flagship speaks, I have got the idea that a flat response to 40-50kHz is a de-facto standard for the hi-end speakers.
 
#3,047
Thanks you are correct way Jazz..

jazzclassics said:
I agree with you that THD along does not reflect sound quality. In most of time, I am trying to be objective. That is why I looked at the schematic and theoritical effect first. I am trying to understand why such an advanced technology is yet to be accepted by the Hi-Fi audience. I know that the renowned Jeff Lowance has, after making some successful Class-D amps, reverted to a class-AB approach.

I am using a pair of Fostex FE166En full rangers which has a response well extended to 22kHz, and even that speaker is suggested, by the Fostex in the manual, to accompany with a super tweeter. Nowadays the Hi-Fi is really going to "Hi-Freq". Looking at the spec for some flagship speaks, I have got the idea that a flat response to 40-50kHz is a de-facto standard for the hi-end speakers.
Click to expand...
 
#3,049
jazzclassics said:
I agree with you that THD along does not reflect sound quality. In most of time, I am trying to be objective. That is why I looked at the schematic and theoritical effect first. I am trying to understand why such an advanced technology is yet to be accepted by the Hi-Fi audience. I know that the renowned Jeff Lowance has, after making some successful Class-D amps, reverted to a class-AB approach.

I am using a pair of Fostex FE166En full rangers which has a response well extended to 22kHz, and even that speaker is suggested, by the Fostex in the manual, to accompany with a super tweeter. Nowadays the Hi-Fi is really going to "Hi-Freq". Looking at the spec for some flagship speaks, I have got the idea that a flat response to 40-50kHz is a de-facto standard for the hi-end speakers.
Click to expand...
class d have a general character of being bright and clinical. if paired to a loudspeaker that have extended HF tweeter they might sound thin and bright. and they do sound thin and bright on my main speaker, rated to extend to 30khz (iinm).
but when paired to a cheap speaker that i have in my study room, they sounded nice as the tweeter have a natural slope above ~12khz.

that being said, when absolute efficiency and low heat output are critical, hard to beat class d. if the designer absolutely must reduce the carrier freq as much as possible, then it's always possible to use much steeper slope.
 
#3,051
thanks dear friend
i buy the 1% resistor for change your amp FREQ but i dont know how and dont know value.
sorry
Click to expand...

Hi Hadi,

change 1K 1% to 510 Ohm 1% change IRFP250 to IRFB4227 (TO247) OR any <100nc MOSFET change gate resistor 39 Ohm to 47 Ohm and filter;

Filter: T106-2 0.8mm wire @50Turns (@33uH). {you can use IRFP250 fsw=@400KHz IRFP250 and IR2110 will heat like >45 degree}

now AUD600 fsw will be @400KHz.

my test it shows 388KHz @ 8R spk load

i waiting for your email sir manoj
thanks
Click to expand...

it will take more time

Regards
MANOJ
 
Last edited:
#3,052
THanks to sir manoj

manojtm said:
Hi Hadi,

change 1K 1% to 510 Ohm 1% change IRFP250 to IRFB4227 (TO247) OR any <100nc MOSFET change gate resistor 39 Ohm to 47 Ohm and filter;

Filter: T106-2 0.8mm wire @50Turns (@33uH). {you can use IRFP250 fsw=@400KHz IRFP250 and IR2110 will heat like >45 degree}

now AUD600 fsw will be @400KHz.

my test it shows 388KHz @ 8R spk load



it will take more time

Regards
MANOJ
Click to expand...

in my country the tecnologie was 10 years older than other place even in mosfet
i can not find here irfb or irfi mosfets🙁
we have here just some irf and irfp type
i buy 680 and 820 ohm 1%
i dont have 510 ohm
 
#3,057
Hadighorbani said:
hi sir manoj
i find IRF640
i have IRF820---IRF540n---IRFZ44n---IRF630 too.
how much power by IRF640 i have in AUD600 ?
Click to expand...

IRFB31N20D is the best choice use it, forget IRFZ44N and IRF820, IRF540N can be use with +/-35V rail, IRF640 with +/-55V and you will get 50W or 100W on 8 ohms. With IRFB31N20D power will be over 200W on 8 ohms.
Power will be same with AUD600 or IRS900D (IRS1500D) or D200.
Regards
 
Last edited:
You must log in or register to reply here.
Top Bottom