One of the designs a lot of folks want is the the most powerful amplifier that can be powered from a standard electric outlet.
This would be 20A 120V or 2400W maximum continuous input power for most folks.
The other consideration is the loudspeaker load impedance. As a loudspeaker is allowed to dip to 1/2 the rated impedance an 8 ohm output should be able to handle a 4 ohm dip.
Now at full power most speakers heat up and can almost double their impedance before failure.
So the amplifier should not be damaged by a short circuit. It should not have a power dip or protection kick in on a burst of energy into 1/2 of the optimum output impedance.
The first issue is what is a reasonable design impedance?
This would be 20A 120V or 2400W maximum continuous input power for most folks.
The other consideration is the loudspeaker load impedance. As a loudspeaker is allowed to dip to 1/2 the rated impedance an 8 ohm output should be able to handle a 4 ohm dip.
Now at full power most speakers heat up and can almost double their impedance before failure.
So the amplifier should not be damaged by a short circuit. It should not have a power dip or protection kick in on a burst of energy into 1/2 of the optimum output impedance.
The first issue is what is a reasonable design impedance?
In the US the maximum for AC plugs with parallel prongs is 15A. For continuous current draw, the maximum is 80% of that or 12A. A 20A circuit requires a special outlet that will accept a plug with one prong (not sure which one) turned 90 degrees. It is a code violation to install 15A outlets on a 20A circuit.
Most amps rated more than 1000W will not produce that much power continuously. They are rated to produce pulses of power.
The question of reasonable design impedance depends on the intended use which you have not defined. Some amps use switching power supplys that can change voltage outputs to optimize the amp for different load impedances.
Most amps rated more than 1000W will not produce that much power continuously. They are rated to produce pulses of power.
The question of reasonable design impedance depends on the intended use which you have not defined. Some amps use switching power supplys that can change voltage outputs to optimize the amp for different load impedances.
A NEMA 5-20 will handle the current. The average current will be under 5 amps. But the design is for real maximum output!
There are several real world considerations in building a large power amplifier. The main two are power transformer and power dissipation.
The heat sink can actually be done quite inexpensively! I can get a 3' heatsink 3" wide for under $35.
That's good because the power transformer will run just under $200.00
So if there is any interest I will go on, otherwise not!
The heat sink can actually be done quite inexpensively! I can get a 3' heatsink 3" wide for under $35.
That's good because the power transformer will run just under $200.00
So if there is any interest I will go on, otherwise not!
And the biggest one of those that you can run off standard 15A wall outlet producing full power sine waves would be a stereo version of the Dirty Harry. I've built bigger ones (1500W, 2 ohms) and they will run from a wall socket - barely - with music. Sine wave trips a 20A breaker.
That's a matter of opinion. Class D is getting better all the time. At the power levels that make sense for those speaker distortion dominates and you won't hear what's produced by the amp. A pro sound subwoofer stack is more like a motor drive than a speaker, anyway, right?
And even with class D, you might get it up to 500 or 600W per channel producing sine wave output before a 15A breaker on a wall socket trips.
A speaker is nowhere close to a motor drive. Even a subwoofer needs a clean low thd signal to sound good, something only a class A/AB amp can do.
Class D in my opinion has no future, and no its not getting any better, its still today the same crap it was back in the 70's when it was first tried.
Who remembers those philips or whatever class d modules from the 70's that were notorious for failing.
Class D in my opinion has no future, and no its not getting any better, its still today the same crap it was back in the 70's when it was first tried.
Who remembers those philips or whatever class d modules from the 70's that were notorious for failing.
Absolutely - we are not talking about chip amplifier class D here and any big PA user will see the benefit of reduced power costs. Frankly, I've heard a lot worse class AB amplification over many years than current pro. quality class D systems.
The point though, given the growing piles of give-away priced, redundant class AB equipment around now globally, is the incentive to build something similar at a higher cost might not be too great.
Coming down to a few hundred very affordable watts might be a more popular project for a wider range of folks.
I've had a chance to listen to the MHzPower 2 by huygens audio and to be honest, a generic Yamaha RX330 natural sound stereo receiver sounded way better.
The sound of a class D amp is cold, flat and lifeless, at lower volume it was ok but turn it up and and everything above a few hundre Hz felt like taking a knife and stabbing your ears with it.
Its an experience i wish to not go though again so i will never again touch a class D amplifier, atleast not for driving speakers.
The sound of a class D amp is cold, flat and lifeless, at lower volume it was ok but turn it up and and everything above a few hundre Hz felt like taking a knife and stabbing your ears with it.
Its an experience i wish to not go though again so i will never again touch a class D amplifier, atleast not for driving speakers.
Since the design is for a cost efficient high power amplifier, to start the electronic design we can start with the output transistors. One of the application specific transistors made for audio power amplifiers are the MJL4281A and MJL4302A. These are plastic and rated at 230 watts, 1.84 C/W derating and have a junction to case thermal resistance of .54! They cost $5.125 ea from Digikey in quantities of 25.
The MJL1302 and MJL3281 are rated at 200 watts, derate at 1.43 C/W and have a junction to case thermal resistance of .625! they cost $3.21 each from Digikey in quantities of 25.
At a junction temperature of 75 C the cost per watt is $.03095 for the MJL4281/4302 and $.02141 for the MJL1302/3281.
So are there other contenders for a high linear beta wide bandwidth output transistor?
The MJL1302 and MJL3281 are rated at 200 watts, derate at 1.43 C/W and have a junction to case thermal resistance of .625! they cost $3.21 each from Digikey in quantities of 25.
At a junction temperature of 75 C the cost per watt is $.03095 for the MJL4281/4302 and $.02141 for the MJL1302/3281.
So are there other contenders for a high linear beta wide bandwidth output transistor?
How about something a little different?
NJL0281D (NPN)
NJL0302D (PNP)
If not I'd vote for the MJL1302/3281.
Last edited:
Very close to MJL4281/4302 except they have a built in thermal compensation diode. As an output pair they are great for an amplifier of 100 watts +.
However they do have a place here!
Now if I use the MJL1302/3281 they actually begin to have Ft and Hfe fall off around 5 amps of collector current. The MJL4281/4302 do the same at slightly less current.
So the next question is how many output devices are needed?
The Vceo is 260 volts for the MJL1302/3281. So if the rails were at 1/2 of that and a 4 ohm loudspeaker is allowed to have a dip to 2 ohms that would require 65 amps.
Now that assumes a classic AB output stage. At this stage of design we can use that for the first cut. There are some interesting techniques we can look at later.
So at 5 amps per output device we will need 13 of each type.
What kind of driver should we look at?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.