Hello everyone,
I have a slight problem: I need a high-efficiency (class D style) amplifier that has to be powered from a 12V battery system, so it should be running with 9...15V input voltage.
Efficiency should be in the 80% range.
It has to output 400 or 500 WRMS into an 8 (eight!) ohm load without problems.
It has to be reliable. I can't have the amp break down. I originally wanted to get a premium car audio amplifier for this application, but the only ones that I can consider are the larger digital monoblocks that are totally tailored towards very low impedances.
When I find an amp that can deliver ~1000WRMS into 4 ohms, that amp is a monster that can deliver 2KW into 2Ohms and sometimes even more.
I am afraid that the power supplies are so large that the amp will be draining a lot of battery when I drive it at "only" 400 or 500W. At least more current than an amp that is designed to deliver half a KW.
Is that assumption correct?
I found the ebay-versions of that international rectifier evaluation board (I think [SIZE=+1]"IRAUDAMP7S"[/SIZE]) that seems to do the job, but I am not too confident that this is reliable enough. And a SMPS to supply the bipolar supply voltages from 12V is going to be another thing that I'll have to get from china and I will not know how reliable that is either.
What is your suggestion? Is a "digital monoblock" car amp really as bad as I think for battery life when driven at low output? Would I be better off using a class A/B car audio amp?
Thanks!
I have a slight problem: I need a high-efficiency (class D style) amplifier that has to be powered from a 12V battery system, so it should be running with 9...15V input voltage.
Efficiency should be in the 80% range.
It has to output 400 or 500 WRMS into an 8 (eight!) ohm load without problems.
It has to be reliable. I can't have the amp break down. I originally wanted to get a premium car audio amplifier for this application, but the only ones that I can consider are the larger digital monoblocks that are totally tailored towards very low impedances.
When I find an amp that can deliver ~1000WRMS into 4 ohms, that amp is a monster that can deliver 2KW into 2Ohms and sometimes even more.
I am afraid that the power supplies are so large that the amp will be draining a lot of battery when I drive it at "only" 400 or 500W. At least more current than an amp that is designed to deliver half a KW.
Is that assumption correct?
I found the ebay-versions of that international rectifier evaluation board (I think [SIZE=+1]"IRAUDAMP7S"[/SIZE]) that seems to do the job, but I am not too confident that this is reliable enough. And a SMPS to supply the bipolar supply voltages from 12V is going to be another thing that I'll have to get from china and I will not know how reliable that is either.
An externally hosted image should be here but it was not working when we last tested it.
What is your suggestion? Is a "digital monoblock" car amp really as bad as I think for battery life when driven at low output? Would I be better off using a class A/B car audio amp?
Thanks!
Check out this amp.. Its a class d monoblock and will probably do what you want..
DB1.2 - Class-D Monoblock - Bass Face Car Audio
DB1.2 - Class-D Monoblock - Bass Face Car Audio
The power supply requirements for ClassAB amp's also apply to ClassD amp's.
In order to get 1000watts into 4 ohms will require 89.44v power supply rails, Both positive and negative rails.
The same apply's to 500 watts into 8 ohms.
This not include any output device or power supply diode voltage drops.
FWIW
jer
In order to get 1000watts into 4 ohms will require 89.44v power supply rails, Both positive and negative rails.
The same apply's to 500 watts into 8 ohms.
This not include any output device or power supply diode voltage drops.
FWIW
jer
Last edited:
Thanks for the help so far! A battery array with sufficient current capabilities is of course going to be a part of the system. several hundred amps is no problem at all. However, I want the amp to run at very high efficiency even if I drive the speaker with only 70 watts. How will a class D monoblock that is rated at 1000W @ 4ohm behave in that situation?
I guess I don't trust the IR amp either. I think modifying it is not an option for me right now.
Phaedras, thanks! That looks like a nice amp and it has a very compact form factor. Great. And it is even affordable. Have you got any experience with these?
I guess I don't trust the IR amp either. I think modifying it is not an option for me right now.
Phaedras, thanks! That looks like a nice amp and it has a very compact form factor. Great. And it is even affordable. Have you got any experience with these?
It is not uncommon for a Class D amp to have 90% efficiency or better.
FET's with the lowest Rdson will give you the best results of high efficiency.
Having a very short dead time is also critical to increasing the efficiency as long as you don't have any shoot through current through the output devices or else they will fail with a very loud bang and flying peices.
Take a look at this thread as projects such as this are already on going,
http://www.diyaudio.com/forums/clas...s-1200-watts-using-2-mosfets.html#post2173932
Good Luck!!
jer
P.S. The best thing to do is build a low voltage one and learn how they work before messing with one that has 90v rails at over 5 to 10 amps each.
FET's with the lowest Rdson will give you the best results of high efficiency.
Having a very short dead time is also critical to increasing the efficiency as long as you don't have any shoot through current through the output devices or else they will fail with a very loud bang and flying peices.
Take a look at this thread as projects such as this are already on going,
http://www.diyaudio.com/forums/clas...s-1200-watts-using-2-mosfets.html#post2173932
Good Luck!!
jer
P.S. The best thing to do is build a low voltage one and learn how they work before messing with one that has 90v rails at over 5 to 10 amps each.
Last edited:
the idea is in fact very tempting and I was thinking about doing it like that for some time. the tda8954 seems nice but is hooking it up directly to batteries really an option? when using 6 or 8 12V batteries in series, they will very likely be at slightly different voltages each. so there would be an asymmetry in the bipolar supply to the chip. will that not cause failure? would I need voltage regulators? do you know a class d chip that can handle +-48V and put out 400W @8R at low THD?
Last edited:
No. Batteries in series will not show different voltage potential on the 2 rails unless they are completely separated or subjected to different loads on the two rails but even then the rails will quickly align themselves.
I know the Bassface amp suggested above, and was the first to propose the use of that or similar low cost simple class D amps as they have very good efficiency, probably most in the 80% region or better in their limited frequency range, and they run directly on 12V which makes integration in a portable system much easier as you will probably have a lot of other things that run on 12V. The other point of using them is that the since sound quality in a regular hi-fi sense is basically not needed, or you could even argue that it would be non-sense to even call it that, for a bass frequency only amplifier. The only requirement is that it can reliably output sufficient power into the loads presented to it, and that it has integrated active filters that makes it easy to incorporate. In that sense, I, and indeed most people who have heard it, can testify that the Bassface amp or other similar amps does in fact do that.
I want to attack the logic here as well. 500W is not half as loud 1000W, 100W is (theoretically, in practice it's closer to 85W being half as loud as 1000W). Let's imagine a volume control that has 10 steps. To have any real meaning each of these step would have to be 6dB otherwise the dynamic range of attenuation would be ridiculously low. So that means if your amp output 1000Wrms in the 10 position, then it will output 250W in the 9 position, 62.5W in the 8 position and so on dividing with 4 every time.
When we speak volume it takes 10dB for a perceived doubling or halving of volume. Theoretically, 10dB equals 10 times as much or less power. In reality that's not true either, as we have something called power compression in speakers that lowers both the efficiency and the sensitivity of a speaker (yes, they are different things) the more power you out into them, so in broad general terms it actually takes about 12 times as much power into a speaker to be twice as loud when the power is nearing the limits of that speakers power handling capabilities (which 1000W would be far exceeding for most regular speakers).
Also, music is not RMS test signals. Far from it. The accepted standard is that for a music signal to have sufficient dynamic range the highest average power level has to be 12dB under peak, in most cases realised as -9dB under max RMS -1dB. This 10dB level is confusingly called RMS in the recording industry. That aside, what it means is that the average output of an amplifier at full possible volume without distortion is 10% of the nominal RMS rating of the amplifier.
So in the case of the 1000Wrms mono class D with 80% efficiency it would actually consume about 125W at full volume, or just about 10A. A regular class A/B amp in the same circumstances would have an efficiency of 20% at best and would therefore consume 500W or 40A at exactly the same output on the speaker terminals.
That is so crazy, you know what- I have spent a lot of time trying to figure out what you just said. It seems no one could give me this kind of clear explanation. Thanks a lot! Really appreciated.
What you said about decibels and watts is very clear to me, and I know that we are not listening to test signals. What I said previously was probably not very precise, I am sorry.
What I meant is this:
Let's say I have an amplifier-speaker system that works well together and I can drive it to a maximum volume (before the speakers distort or the amp clips) where I need no more and it is absolutely loud enough. So that is fine. But when I reduce volume and therefore reduce power to the speakers, the PSU in the amplifier of course still provides the full positive and negative rail voltages.
My concern was that some amplifier designs *might* be significantly less efficient at lower set volumes.
I gathered that most designs do reach their peak efficiency as output power rises, I just do not have the knowledge of exactly how class D car amplifiers are designed and was worried that they would be rather inefficient when driven very low. Again, it is just a matter of not knowing, therefore I wanted to ask.
now this is interesting- in my understanding, class A/B amplifiers do reach 60-ish percent efficiency. At least with test signals at rated power. Do you state that high-wattage class A/B amplifiers do in fact become a lot less efficient as power output drops?
Thanks!
Last edited:
Yes. That is a known fact. Take any class AB amp and have it run at 10% of nominal maximum output and the efficiency will typically be around 20% when losses in the dc-dc converter is considered. It is possible it could be a few percents higher, say around 22% at best, but it's more likely to be several percents lower.
The same is true for a class D amplifier, however, the drop in efficiency is much much lower and will likely be in the 70-85% range after dc-dc converter losses under the same conditions.
That also means the even lowest efficiency of any well designed class D amplifier will under real life use with actual music signal will still surpass even the theoretical maximum efficiency of any class AB, G, or H amp.
Class AB amps are typically 60% to 70% efficient not 20%.
Although if it is run at a low level and it is still in a class A mode (determined by it bias current) then the efficiency can be quite low.
Class A amps have a very low efficiency into the 10% to 25% range.
A bridged Class A can be as high as 50% efficient.
Class B amps are about 78.5% efficient max.
jer
Although if it is run at a low level and it is still in a class A mode (determined by it bias current) then the efficiency can be quite low.
Class A amps have a very low efficiency into the 10% to 25% range.
A bridged Class A can be as high as 50% efficient.
Class B amps are about 78.5% efficient max.
jer
Show me one example of a class AB amp that is 60-70% efficient at 10% of it's nominal RMS output! Just one.
Last edited:
I was speaking strictly upon amplifier efficiency not including power supply losses.
As I stated low power efficiency is greatly determined by how much bias current Is running through the output devices.
If there weren't any quiescent current flowing and it was strictly Class B the efficiency would still be in the 70% range regardless of its output power level.
http://www.zedaudiocorp.com/Technical/Amplifier-Efficiency.pdf
Amplifier Efficiency
Efficiency of a Class AB Amplifier - All About Circuits Forum
Amplifier - Wikipedia, the free encyclopedia
This is an interesting read,
Exploding the Efficiency Myth of Class D Amplifiers
Jer
As I stated low power efficiency is greatly determined by how much bias current Is running through the output devices.
If there weren't any quiescent current flowing and it was strictly Class B the efficiency would still be in the 70% range regardless of its output power level.
http://www.zedaudiocorp.com/Technical/Amplifier-Efficiency.pdf
Amplifier Efficiency
Efficiency of a Class AB Amplifier - All About Circuits Forum
Amplifier - Wikipedia, the free encyclopedia
This is an interesting read,
Exploding the Efficiency Myth of Class D Amplifiers
Jer
Last edited:
Nor am I. I am addressing the fallacy of your claim that any class AB can ever get 60-70% efficiency at 10% of it's nominal output.
In the same line of thought can you give any example of an audio power amplifier that is purely class B?
Please have a look at the actual datasheets of any class AB amp of your choosing. And you will find that the argument I am making is a fact.
You can for example take whatever chipamp you choose, let's say the well known TDA7294. First we look at what is the nominal power output before clipping. In my mind that is at 70W or where we exceed 1% distortion. Then you look on the power dissipation vs. power output chart.
At 70W Pout, Pdiss is 38W. Hence the total power consumption must therefore be 108W, and the total efficiency is 65%.
At 7W Pout (10% of nominal, remember), Pdiss is 28W. Hence the total power consumption must be 35W, and the total efficiency is 20%.
This is before any losses in the power supply is considered. It is losses due to the very operation of the amplifier itself.
Since we know that average level in music is roughly around -10dB which equates 10% of maximum before clipping we can therefore conclude that the maximum efficiency of the TDA7294 in actual class AB use is 20%.
The datasheet then goes on to describe a high efficiency mode where you can at best achieve 37.5% efficiency under the same conditions. But this is achieved by switching between a higher and a lower voltage supply rail on the output power supply pins. Or what we also call a bi-voltage class G design.
Last edited:
No. It isn't. It's a fallacy.
Remember. This guy wants to sell his own chips and has made an argument that while it on the surface sounds plausible is actually false.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.