Hi,I agree with you. This is exactly the problem of D-class amplifier Having tried everything on the market as I thought the answer to the resolution on medium and high frequencies and obtain a crystal clear sound like a professional class AB.
Unfortunately the low resolution at high audible frequencies can not be measured and not depend solely on the carrier frequency on the class D does not depend on the type of feedback if this is pre or post filter.
Regards
I honestly am suspicious of this circuit, it is very easy to build .. was almost buying components to assemble a class AB amplifier of 1300W on 2R (9 pairs in the output) would be very expensive ....
I now see this in class D 1250W 4R .. I wonder if he has a good quality bass, with punch, can someone tell me if it's really good? even being so simple ....
I now see this in class D 1250W 4R .. I wonder if he has a good quality bass, with punch, can someone tell me if it's really good? even being so simple ....
Half Bridge class D Amps have physical Problem with power supply pumping and not good solution for good quality bass... trust me each low cost class AB amp kick Half Bridge class D Amps out of the job < 100 Hz
for good quality bass its important to use full bridge class D Design otherwise you are not satisfied with sound
i have made the circuit for +/-29Volts ,on LM's pin 8 i have 24Volts....the zenner diode is OK....sorry for my english i use google translator to translate from romanian to englishsorry...
You misconnected something for sure. Check your circuit against the schematic, or send a good photo!
so unfortunately I will build the amplifier class ab ..
just to kill the doubt, if I mount these two, and put in the bridge (the bridge as AB amps), this would be a full bridge?
and more an issue in the region between 100hz and 2KHz, he has a good sound? or worse in this region than in the bass?
just to kill the doubt, if I mount these two, and put in the bridge (the bridge as AB amps), this would be a full bridge?
and more an issue in the region between 100hz and 2KHz, he has a good sound? or worse in this region than in the bass?
Hello Pafi,
He had 24Vdc everywhere, stabilizer circuit was ok, zenners were ok, he even used 12V from external trafo. nothing ...
That thing has no dead time implementation before the IR.
It lack's some things... It should have had one 4070 in place of q1,q2,q3,q4
Hope he will start again soon....
I started making a SMPS for him.
Best Regards,
Savu Silviu
Of course I do not know many things! No one knows everything!
At first, the circuit is not mine. Am not the author.
II. Obviously there are losses, but which have the greatest impact is the output MOSFET Rdson. This determines the estimated efficiency of the amplifier, which is of interest ...
In the third ...
Each one provides the level of security you want to build this amplifier. Margins of safety ... Of course! It is obvious that for high power is necessary: Put an extra heatsink size or a little help from the forced-ventilation ...
As mentioned by the sound quality at high frequencies. According to the author, this design is high fidelity. It is known that when the output inductor is not well done, the frequency response changes ... For better or for worse ...
In my case, I found no differences in sound compared to conventional AB amplifiers, quasi-complementary designs and input based on OPAM.
Greetings!
PS: Much more obvious are the differences between real life and the calculation on paper. We do not need to remember them. And no, I measured the output of my amp ... But, get an idea, the power source I used for testing was a Peavey CS800 with an estimated power output of 1200W ... Capacitance but I lacked, that sounded, Uffff, is fierce ...
Sorry, but I have to say this is simply not true. What you have ignored are very strong effects, in some cases they can be much stronger then Rdson. And these cases are especially at high voltage, high power devices.
If you calculate with the nominal current rating (eg. 30 A in case of IRFP250N) of MOSFET, then you can't put extra heatsink, because nominal current is defined with infinite heatsink, with infinite thermal conductivity, without any insulator between them. What could be bigger then infinite? And it is specified without any switching loss.
If you calculate with the specified Rdson, then the situation is even worse! Rdson is specified at 25C, but at high chip-temperature can be 2-2.5 times more. Let's see IRFP250N again: It's "allowed" dissipation is 214 W. You said "PDmos = I * I * Rdson / 2", so I=sqrt(2*214/0.075)=75 A. Do you really think this MOSFET will handle this much current?
To make things worse: switching loss. Current rise time with this high Rg will be about 50 ns. This is not too much, for example at fsw=100 kHz this means only 2*V*I*tr*fsw=1.5 % loss (almost the same amount what you calculated for Rdson loss). But then add reverse recovery diode loss! Qrr at only Id=18A would be 1.3 nC at 25C, and 100A/us, but at elavated temperature and faster current slope (eg. 20A/50ns=400A/us) it will be higher, about 3 nC. This means P=Qrr*2*V*f=54 W, while Rdson dissipates Pcond_loss=0.15*18*18=49W in a bad case. We can add turn off loss of approximately 20 W, (total loss in MOSFET: ~ 140 W) and we've already reached the practical dissipation limit (at only 18 A output current).
If you use higher voltage, higher power MOSFETs, the ratio between your estimated power, and real power will be even worse because of the increasing diode recovery time and the cooling problem.
Only with MOSFETs optimised for ClassD audio can your calculations be more or less correct, but also with these devices you have to calculate with parameters at elevated temperature, and realisable cooling!
11250 W with 2 FETs, in this circuit is a joke.
Yes. But one who doesn't know, he shouldn't give advice in the unknown fields, or at least attach an 'as far as I know', or 'but please correct me if I'm wrong'! These issues are the basics of high power ClassD amps, everybody should be aware of them.
I don't understand any of these sentences.
Well... i was thinking if it would be worthwhile to try to answer or not to all the things that have been told about this amp. I was the designer of it, and, be sure, i know the limitations of the design. Severe compromises had to be done in order to make this design doable by common people.
Please, take into consideration that the design was severely affected by the semiconductors that are/were readily available on the local market, here in Argentina. Of course you can always buy fancy and more appropiate components from an international supplier, such as Digikey or Farnell, but that would make nearly impossible to hobbists to build it, just because it is extremely expensive and difficult to import electronic components in low quantities, at least, here.
The design lacks several features desirable for a commercial design, specially , all the required protections (short circuit, overvoltage, etc). It is also using old (but cheap and easiliy available) mosfets that are not "suitable" for class D audio. It even uses an IR211x mosfet driver, that "lacks" internal deadtime control, and requires a level shift translator (a discrete one).
And, i also agree that the board should be a double layer one with metallized holes, but, again, such board would put the design out of the possibilities of regular hobbists, because metallized holes on boards can't be done at home... And would require boards to be professionally made, rising costs again.
And , of couse, the inductor should be wound on a ferrite/metal powder core, but, again, it would mean amateur hobbists are left out of the design...
So...
Take this amp as a proof of concept, and NOT as a commercial design... It has room for lots of improvements (use more suitable mosfets, for example... add external deadtime control, redesign board as a double sided one, etc... But please, do not complain... I see none of you posting designs, all i see are people trying to get already designed and foolproof schematics that outperform, if possible, commercial designs with special components, and lots of time used to debug those designs...
Eduardo
PD: regarding the deadtime, the IR211x has non simmetric propagation delays (turn on is delayed more than turn off, so... it can be taken as internal delay... And the circuit features a basic R and diode in series with mosfet gates, to control it... So it is not fair to say no deadtime control was provided... AND i know the internal diode of this mosfet is slow... BUT, that's exactly one of the purposes of the R||D in gate network... To make the switching time a little bit slower, so we don't blow the internal diode in the process... Of course this rises the disspation of each mosfet, but allows to use slower and cheaper devices... So , lot of compromiese are here... You are free to change mosfets, if you need, so don't complain!)
Please, take into consideration that the design was severely affected by the semiconductors that are/were readily available on the local market, here in Argentina. Of course you can always buy fancy and more appropiate components from an international supplier, such as Digikey or Farnell, but that would make nearly impossible to hobbists to build it, just because it is extremely expensive and difficult to import electronic components in low quantities, at least, here.
The design lacks several features desirable for a commercial design, specially , all the required protections (short circuit, overvoltage, etc). It is also using old (but cheap and easiliy available) mosfets that are not "suitable" for class D audio. It even uses an IR211x mosfet driver, that "lacks" internal deadtime control, and requires a level shift translator (a discrete one).
And, i also agree that the board should be a double layer one with metallized holes, but, again, such board would put the design out of the possibilities of regular hobbists, because metallized holes on boards can't be done at home... And would require boards to be professionally made, rising costs again.
And , of couse, the inductor should be wound on a ferrite/metal powder core, but, again, it would mean amateur hobbists are left out of the design...
So...
Take this amp as a proof of concept, and NOT as a commercial design... It has room for lots of improvements (use more suitable mosfets, for example... add external deadtime control, redesign board as a double sided one, etc... But please, do not complain... I see none of you posting designs, all i see are people trying to get already designed and foolproof schematics that outperform, if possible, commercial designs with special components, and lots of time used to debug those designs...
Eduardo
PD: regarding the deadtime, the IR211x has non simmetric propagation delays (turn on is delayed more than turn off, so... it can be taken as internal delay... And the circuit features a basic R and diode in series with mosfet gates, to control it... So it is not fair to say no deadtime control was provided... AND i know the internal diode of this mosfet is slow... BUT, that's exactly one of the purposes of the R||D in gate network... To make the switching time a little bit slower, so we don't blow the internal diode in the process... Of course this rises the disspation of each mosfet, but allows to use slower and cheaper devices... So , lot of compromiese are here... You are free to change mosfets, if you need, so don't complain!)
hi,
when amplifier is finished and working properly (or other amp that uses Ir21xx) then you can lower the THD especially in the low power segment, varying the R (high-side gate) to asymmetrical (lower value than at the low side gate).
This is due to an additional delay to the internal series IR21xx. (asymmetric DT)
e.g. High side=6,8R -low side=15R
Regards
when amplifier is finished and working properly (or other amp that uses Ir21xx) then you can lower the THD especially in the low power segment, varying the R (high-side gate) to asymmetrical (lower value than at the low side gate).
This is due to an additional delay to the internal series IR21xx. (asymmetric DT)
e.g. High side=6,8R -low side=15R
Regards
This was my point exactly ! the driver ic already has internal DT which i know eva has already spoken about in previous designs IR2011 ECT..,why put another DT branch in the circuit unless your specification alters the design but i trust the design as a proof of concept it a great (tutorial) !
When i get some time im going to build this and improve it where i see fit there are already many sub system protection circuits on this site that can slot in with a few mods here and there, especially on the power supply forum, tailored to ones needs.(simple)
I personally feel there are some jealous people on the forum that are deliberately posting negative viewpoints for those of you who are considering to post more negative posts..do this give us / newbies your example i mean a living example instead of text descriptions of your pro designs which is meaningless for newbies that learn from living examples.
Last edited:
This was my point exactly ! the driver ic already has internal DT which i know eva has already spoken about in previous designs IR2011 ECT..,why put another DT branch in the circuit unless your specification alters the design but i trust the design as a proof of concept it a great (tutorial) !
When i get some time im going to build this and improve it where i see fit there are already many sub system protection circuits on this site that can slot in with a few mods here and there, especially on the power supply forum, tailored to ones needs.(simple)
I personally feel there are some jealous people on the forum that are deliberately posting negative viewpoints for those of you who are considering to post more negative posts..do this give us / newbies your example i mean a living example instead of text descriptions of your pro designs which is meaningless for newbies that learn from living examples.
Hi,
Because You know internal IR21xx,then I can not explain async DT?
yes some jealus people attak me, ......you instead?
other bla,bla I not understand,sorry.
If you want a complete description,scheme and examples,please chose arguments and I help diy.
Regards
Hi All,
I have just finished this amp. It works. But I have some question. I connected 20KA volume before the input .When it is in min. and max. postion there is some noise as normall but when it is in any position between min. and max. ,there is an oscillation noise. Why?
I replaced D2 UF4004 with MUR120 , increased C2 and C8 to 470pf ,decreas R5 and R13 to 10 ohm .
I have just finished this amp. It works. But I have some question. I connected 20KA volume before the input .When it is in min. and max. postion there is some noise as normall but when it is in any position between min. and max. ,there is an oscillation noise. Why?
I replaced D2 UF4004 with MUR120 , increased C2 and C8 to 470pf ,decreas R5 and R13 to 10 ohm .
