Hi everyone, didn't know about this forum... excellent!. Good to find you.
I'm an electronics engineer but i work mostly with digital stuff, micros pgas etc. In regards to audio amps i'm just an amature.
I'm about to build a LM4780 stereo bridged amp (using two ICs).
I've seen already in this forum a lot of fuss about the LM4780 and a lot of your designs seem to be excellent.
I have a couple of simple questions.
How can i calculate the power needs of the amp.? How many VA should the tranformer be to suppport max operation. (I'm looking for the correct way to calculate it).
Is it worth using regulation? (LM317 with hi power transistors or similar)?
Any mains AC filter suggestions ?
Also on the datasheet i can't find anywhere what level the input signal is expeted to be? (based on the default gain values).
Thank your
I'm an electronics engineer but i work mostly with digital stuff, micros pgas etc. In regards to audio amps i'm just an amature.
I'm about to build a LM4780 stereo bridged amp (using two ICs).
I've seen already in this forum a lot of fuss about the LM4780 and a lot of your designs seem to be excellent.
I have a couple of simple questions.
How can i calculate the power needs of the amp.? How many VA should the tranformer be to suppport max operation. (I'm looking for the correct way to calculate it).
Is it worth using regulation? (LM317 with hi power transistors or similar)?
Any mains AC filter suggestions ?
Also on the datasheet i can't find anywhere what level the input signal is expeted to be? (based on the default gain values).
Thank your
You didn't mention what impedance you speakers are or what the target is for the amplifier. Since you are going to bridge each LM4780 for a stereo amplifier I would guess they are 8 ohm or greater. National does not recommend running bridge into lower impedance since each amplifier effectively sees 1/2 the load. The LM4780 is rated for 4 ohms minimum so that means 8 ohms minimum for bridge. You can get away with running 4 ohms bridge but you will not get much power and the parts will over heat. It's all thermal issues. Better to be sure the speakers are 8 ohms or if 4 ohms use parallel operation. The LM4780 has graphs for bridge 8 ohm or parallel 4 ohm in the datasheet.
The VA rating you need will depend on how much voltage AND current you need. Is this going to be a stereo 100W amplifier design? More or less? Is that power with both channels on & in phase running a sine wave? One channel power at a time? Because music is so dynamic it can be over kill to design the power supply to run 100W/Ch. using sine waves at the same time. Adding plenty of capacitance to the power rails might be a better use of money than a bigger transformer. I personally am in the no regulation camp unless the regulation is very fast (linear regulator is not fast enough). Also adds complexity to the design.
A good resource is National's OVerture Design Guide. Get it here --> http://www.national.com/appinfo/audio/
Go down about 1/2 way in the middle column to find it.
If you read and search around enough on the forums you can probably find a fair amount of info to answer some questions. Might also just want to buy some kits for these parts instead of a full design yourself.
-SL
The VA rating you need will depend on how much voltage AND current you need. Is this going to be a stereo 100W amplifier design? More or less? Is that power with both channels on & in phase running a sine wave? One channel power at a time? Because music is so dynamic it can be over kill to design the power supply to run 100W/Ch. using sine waves at the same time. Adding plenty of capacitance to the power rails might be a better use of money than a bigger transformer. I personally am in the no regulation camp unless the regulation is very fast (linear regulator is not fast enough). Also adds complexity to the design.
A good resource is National's OVerture Design Guide. Get it here --> http://www.national.com/appinfo/audio/
Go down about 1/2 way in the middle column to find it.
If you read and search around enough on the forums you can probably find a fair amount of info to answer some questions. Might also just want to buy some kits for these parts instead of a full design yourself.
-SL
Thanks a lot for the reply,
One problem is that i'm not sure what the load will be, either 6 or 8 ohm. But here i see the problem with a 6ohm in the bridge configuration. In the parallel mode though if the speakers are not 4ohm i won't get full power right ?
I'm a bit confused.
Say the speakers are 6ohms which is better parallel or bridged ? (is it even possible to be bridged since the seperate load will be 3ohms?)
I understand your point on the power supply, i'm sure about which is more money efficient, large caps 47000uF for example cost around 20-30$ which is more than the difference between a 300VA to 400VA tranformer. Obviously the best is a good balance between capacitance and wattage of the transformer. Even the case of large caps, how do i calculate the power of the tranformer say the minimum regardless of the caps.
One problem about the regulation is that the same amp will be used in 240V (UK) and 220V(Greece) whilst the actual stability can be quite dodgy from the mains, so i think it'll be better to have regulation. You mentioned that linear regulators are not fast enough. Would a switched mode supply do a better job? I was under the impression that switched mode supplies introduce a lot of noise to audio amps. I've designed doesnt of sw mode supplies for other types of circuits and i know quite well how these work, even to maintain regulation the need to have voltage ripple on their output which is "bad" for audio. I'm not sure what you meant by that or other type of supply ?
I'm not interested in getting something ready, all the fun is to build it
.
Thanks again
One problem is that i'm not sure what the load will be, either 6 or 8 ohm. But here i see the problem with a 6ohm in the bridge configuration. In the parallel mode though if the speakers are not 4ohm i won't get full power right ?
I'm a bit confused.
Say the speakers are 6ohms which is better parallel or bridged ? (is it even possible to be bridged since the seperate load will be 3ohms?)
I understand your point on the power supply, i'm sure about which is more money efficient, large caps 47000uF for example cost around 20-30$ which is more than the difference between a 300VA to 400VA tranformer. Obviously the best is a good balance between capacitance and wattage of the transformer. Even the case of large caps, how do i calculate the power of the tranformer say the minimum regardless of the caps.
One problem about the regulation is that the same amp will be used in 240V (UK) and 220V(Greece) whilst the actual stability can be quite dodgy from the mains, so i think it'll be better to have regulation. You mentioned that linear regulators are not fast enough. Would a switched mode supply do a better job? I was under the impression that switched mode supplies introduce a lot of noise to audio amps. I've designed doesnt of sw mode supplies for other types of circuits and i know quite well how these work, even to maintain regulation the need to have voltage ripple on their output which is "bad" for audio. I'm not sure what you meant by that or other type of supply ?
I'm not interested in getting something ready, all the fun is to build it
.Thanks again
Hi,
If you design your amplifier to work from 240V mains with unregulated supply , running it from 220V mains will decrease the maximum output power a bit due to lower supply voltage , but not very much.According to the datasheet , these IC amplifiers don't need a regulated power supply.
Your should download overture design spreadsheet from www.national.com .It is an excel spreadsheet which allows to calculate both bridge and paralell configurations easily.
I'd say that for 6 ohms , its probably better to go with slightly higher supply voltage (around 36 Volts under load) , and paralell configuaration. You would have more headroom , so Spike protection would not kick in so easy.
Regards,
Lukas.
If you design your amplifier to work from 240V mains with unregulated supply , running it from 220V mains will decrease the maximum output power a bit due to lower supply voltage , but not very much.According to the datasheet , these IC amplifiers don't need a regulated power supply.
Your should download overture design spreadsheet from www.national.com .It is an excel spreadsheet which allows to calculate both bridge and paralell configurations easily.
I'd say that for 6 ohms , its probably better to go with slightly higher supply voltage (around 36 Volts under load) , and paralell configuaration. You would have more headroom , so Spike protection would not kick in so easy.
Regards,
Lukas.
hi otherside,
Seeing you are an electronics engineer working with digital stuff go with the SMPS. That will solve all your PS issues very quickly. Looking around the various forums people are reporting good results.
Why bridging or parallel? Try a single chip per channel first. It's surprising what they can do.
Good luck with the designing. I'm sure you'll come up with a good solution because these chipamps seem to work well over a wide range of parameters. I would normally recommend just build one and experiement, but if you enjoy the design process that's what you should do. Search thoroughly this forum and the web.
Have you seen "Bridge/Parallel Amplifier (BPA-200) Documentation"? - BPA-200.pdf
Seeing you are an electronics engineer working with digital stuff go with the SMPS. That will solve all your PS issues very quickly. Looking around the various forums people are reporting good results.
Why bridging or parallel? Try a single chip per channel first. It's surprising what they can do.
Good luck with the designing. I'm sure you'll come up with a good solution because these chipamps seem to work well over a wide range of parameters. I would normally recommend just build one and experiement, but if you enjoy the design process that's what you should do. Search thoroughly this forum and the web.
Have you seen "Bridge/Parallel Amplifier (BPA-200) Documentation"? - BPA-200.pdf
Thanks a lot for the replies.
Unfortunatelly to get more than 5A from SMPS it's a different story, inductors that can carry larger current and ic's to support are not easy to find. A couple of designs i looked for 5-10Amps require a large amount of components and the whole thing is starting to be over-designed.
It's a long story but mainly i need the power. The problem with trying out is that to do even the simplest test i need a serious PSU which i don't have.(Most of my PSUs are 12, 5, 3.3V, and none of them can carry more than 3A). So initially i need to buy the tranformer and that's what this thread is all about!. To buy a TR i need to know what VA is should be assuming i know the Voltage.
No i hadn't but i just found it and reading it it's quite usefull and maybe what i'll end up using. However it's not very clear what is the final supply voltage to the amps. On page 14 notes +-42 Volts which is the Maximum the IC can take?? is that right? Also it mentions some 40.000uF Caps that don't appear on the schematic.
I have and it's quite usefull, but still it doesn't metion the INPUT power requirements.
I'm still a bit confused so let me ask a simple question, assume the standard design:
IC : LM4780 - stereo configuration
Vin: 35V
Rload: 8Ohm
This results in output 60W/channel.
Using a 2x25Vac TR will result in 35Vdc without regulation. What is the Power requirement for the transformer and the rail caps?
Thanks again guys.
Unfortunatelly to get more than 5A from SMPS it's a different story, inductors that can carry larger current and ic's to support are not easy to find. A couple of designs i looked for 5-10Amps require a large amount of components and the whole thing is starting to be over-designed.
It's a long story but mainly i need the power. The problem with trying out is that to do even the simplest test i need a serious PSU which i don't have.(Most of my PSUs are 12, 5, 3.3V, and none of them can carry more than 3A). So initially i need to buy the tranformer and that's what this thread is all about!. To buy a TR i need to know what VA is should be assuming i know the Voltage.
No i hadn't but i just found it and reading it
it's quite usefull and maybe what i'll end up using. However it's not very clear what is the final supply voltage to the amps. On page 14 notes +-42 Volts which is the Maximum the IC can take?? is that right? Also it mentions some 40.000uF Caps that don't appear on the schematic. I have and it's quite usefull, but still it doesn't metion the INPUT power requirements.
I'm still a bit confused
so let me ask a simple question, assume the standard design:IC : LM4780 - stereo configuration
Vin: 35V
Rload: 8Ohm
This results in output 60W/channel.
Using a 2x25Vac TR will result in 35Vdc without regulation. What is the Power requirement for the transformer and the rail caps?
Thanks again guys.
OK Finally i think i got that.
To get the max required power from the transformer just multiply the peak current with the TR voltage. So for example for the LM4780 at full power in stereo mode for an 8ohm load this would be 3,9A/Channel so 7,8A for both channels.
Without regulation a 25V TR will come up as DC at 35. 25x7,8= 195VA. So a 200VA TR would be fine.
With regulation a 28V TR would be perfect, but you can't get them easily so a 30V one will come up with 42.3 DC. Dropping 7V linearly at max current of 7,8A is 55W disipation which is kind of waste but anyway. The TR would have to be 30x7,8 = 234VA. So a 250VA TR will do.
However i still can't figure out the same for bridged, parallel, and parallel/bridged configuration.
Also how can i calculate the max voltage for the bridged/parallel configuration?
Could someone help about these configurations ?
Thanks again
To get the max required power from the transformer just multiply the peak current with the TR voltage. So for example for the LM4780 at full power in stereo mode for an 8ohm load this would be 3,9A/Channel so 7,8A for both channels.
Without regulation a 25V TR will come up as DC at 35. 25x7,8= 195VA. So a 200VA TR would be fine.
With regulation a 28V TR would be perfect, but you can't get them easily so a 30V one will come up with 42.3 DC. Dropping 7V linearly at max current of 7,8A is 55W disipation which is kind of waste but anyway. The TR would have to be 30x7,8 = 234VA. So a 250VA TR will do.
However i still can't figure out the same for bridged, parallel, and parallel/bridged configuration.
Also how can i calculate the max voltage for the bridged/parallel configuration?
Could someone help about these configurations ?
Thanks again
hi otherside,
Sizing the toriods is not really an exact science because of the nature of music, full power is not required for long periods. The filter caps supply the power for these brief peiods. DIYers often use just use a rough rule of thumb of 3x, 4x or 5x (or whatever) the output power.
I've heard rumours that commerical amps often use a 0.7x or 0.8x factor.
The VA rating of toriods is usually the sum of the secondaries. So a 200VA toriod would usually have 2 secondary windings of 100VA each.
If you want to do the figures properly you must also include toriod "regulation" which is the voltage drop, usually expressed as a percentage, when the toriod is connected to a load. This is something like 3% to 6% and usually depends on the size and quality of the toriod.
If you decide to go with voltage regulation (after the diode bridge), I think you need to design around the lowest mains voltage i.e. 220V less 10% (or more??) to prevent the regulators dropping out. With your requirement of different voltage and quality mains, I think an unregulated supply would be much easier.
See http://sound.westhost.com/articles.htm
Out of curosity, I connected my GC up via an autotransformer to hear what happens when I vary the mains voltage from 180 to 270 volts. To my ears, I could detect little or no difference on my test system. This has always surprised to me becuase I always expect a significant change when adjusting the mains voltage.
Another option to parallel / bridging is bi-amping.
regards
Sizing the toriods is not really an exact science because of the nature of music, full power is not required for long periods. The filter caps supply the power for these brief peiods. DIYers often use just use a rough rule of thumb of 3x, 4x or 5x (or whatever) the output power.
I've heard rumours that commerical amps often use a 0.7x or 0.8x factor.
The VA rating of toriods is usually the sum of the secondaries. So a 200VA toriod would usually have 2 secondary windings of 100VA each.
If you want to do the figures properly you must also include toriod "regulation" which is the voltage drop, usually expressed as a percentage, when the toriod is connected to a load. This is something like 3% to 6% and usually depends on the size and quality of the toriod.
If you decide to go with voltage regulation (after the diode bridge), I think you need to design around the lowest mains voltage i.e. 220V less 10% (or more??) to prevent the regulators dropping out. With your requirement of different voltage and quality mains, I think an unregulated supply would be much easier.
See http://sound.westhost.com/articles.htm
Out of curosity, I connected my GC up via an autotransformer to hear what happens when I vary the mains voltage from 180 to 270 volts. To my ears, I could detect little or no difference on my test system. This has always surprised to me becuase I always expect a significant change when adjusting the mains voltage.
Another option to parallel / bridging is bi-amping.
regards
When I talked about a fast regulated supply that means more complicated than a SMPS or a linear regulator. I have done some test where I use a pair of LM3886s to create supply rails and compare performance to normal bench supply. The LM3886s are faster in response than the bench supplies and a THD plot shows this. The difference is not incredible so may not even be audible. For all my amps I have used just a xformer to make an unregulated supply. I think the suggestion to design for 240V and the 220V will have lower peak output power but probably not noticeable is a very good idea.
For 6 ohms you could do bridge as long as the supply voltage is low enough, as in +/-25V or less since this is what is spec'd for a 4 ohm load. You should still be able to get lots of power but, like suggested, if you want to play it really safe then run them with just a single channel first to see if there is enough power already. 6 ohms is a bit tricky on which is the best because in bridge with 3 ohms/channel it might run too hot while in parallel at 12 ohms there might not be enough voltage headroom to really get much more power than just one channel driving 6 ohms. One way to calculate the supply voltage for other loads is to trick National's design guide. Look at the power dissipation where it gives total power dissipation per IC. For the LM4780 if you enter in the 3 datasheet specs for supply voltage and load you'll find that this number is always right around 65W. Then you can enter in 3 ohms and although the output power will not show the thermal numbers will. Then adjust you supply voltage to get the same ~65W. This tells you that ~+/-21.5V supply for bridge 6 ohm load will have the correct power dissipation. The die will still get hot fast and you will be on the edge of Spike protection but at least you will be in the same range national recommends for this chip. Or you can run parallel with 12ohms seen by each channel and then the supply can be run all the way to max and still stay within the Pd specs. Some things to think about.
I need to dig out my info on xformers to get you a VA rating.
-SL
For 6 ohms you could do bridge as long as the supply voltage is low enough, as in +/-25V or less since this is what is spec'd for a 4 ohm load. You should still be able to get lots of power but, like suggested, if you want to play it really safe then run them with just a single channel first to see if there is enough power already. 6 ohms is a bit tricky on which is the best because in bridge with 3 ohms/channel it might run too hot while in parallel at 12 ohms there might not be enough voltage headroom to really get much more power than just one channel driving 6 ohms. One way to calculate the supply voltage for other loads is to trick National's design guide. Look at the power dissipation where it gives total power dissipation per IC. For the LM4780 if you enter in the 3 datasheet specs for supply voltage and load you'll find that this number is always right around 65W. Then you can enter in 3 ohms and although the output power will not show the thermal numbers will. Then adjust you supply voltage to get the same ~65W. This tells you that ~+/-21.5V supply for bridge 6 ohm load will have the correct power dissipation. The die will still get hot fast and you will be on the edge of Spike protection but at least you will be in the same range national recommends for this chip. Or you can run parallel with 12ohms seen by each channel and then the supply can be run all the way to max and still stay within the Pd specs. Some things to think about.
I need to dig out my info on xformers to get you a VA rating.
-SL
Seems most of my amps (about half dozen) did not give the VA rating just the voltage and amps. I did make one amp using a Plitron transformer and it was a 5x50W design that could do 5x40W at the same time running sine waves so a little over designed. The coil in that amplifier has a rating of 500VA. Looks like Plitron is still in business (http://plitron.com/) but they weren't cheap TF. So maybe the suggestion of designed output power *2 to give a VA. So if you are looking at a stereo 100W then you would want a VA of around 400.
-SL
-SL
Excellent, thank you for all your suggestions and valuable time.
I guess around 2.5x is not very bad. That's around my calcs and what i can considering "value for money", maybe close to 3x. Depending on the voltage.
The trs i'm looking at have some figures on the datasheet, which basically says for example a 250VA (dual sec) is 25Vx5A on load, and off load it goes to around 26,5V, so i guess the VA rating includes the regulation.
Yes that's an issue (the drop out). I was looking for something else at work today and found by mistake some regs that have drop out around 0.6V which is excellent because power dissipation will be minor.Also instead of needing a 30V TR that on DC is 42,5 and the drop is 7V which means a lot of dissipation (for 35V), i can go for a 25Vac= 35,35DC and make the rail at 33V dc. This is on 230Vac mains, +-5% (220-240) = (33,7-37.1).
now at 33V on 8ohms is around 55W which sounds good.
Is it a good idea to run the supply to the max with no headroom?
I had a look on the idea and sounds very interesting, not for this project though, maybe in the future. Thanks for the link
'm still looking for an answer on this, if someone knows.
How do i determine the max voltage for a bridged-parallel configuration?
Guys thanks again for the info, i think someone can write a book with all the knowledge in this forum!. Excellent ideas.
I guess around 2.5x is not very bad. That's around my calcs and what i can considering "value for money", maybe close to 3x. Depending on the voltage.
The trs i'm looking at have some figures on the datasheet, which basically says for example a 250VA (dual sec) is 25Vx5A on load, and off load it goes to around 26,5V, so i guess the VA rating includes the regulation.
Yes that's an issue (the drop out). I was looking for something else at work today and found by mistake some regs that have drop out around 0.6V which is excellent because power dissipation will be minor.Also instead of needing a 30V TR that on DC is 42,5 and the drop is 7V which means a lot of dissipation (for 35V), i can go for a 25Vac= 35,35DC and make the rail at 33V dc. This is on 230Vac mains, +-5% (220-240) = (33,7-37.1).
now at 33V on 8ohms is around 55W which sounds good.
Is it a good idea to run the supply to the max with no headroom?
I had a look on the idea and sounds very interesting, not for this project though, maybe in the future. Thanks for the link
'm still looking for an answer on this, if someone knows.
How do i determine the max voltage for a bridged-parallel configuration?
Guys thanks again for the info, i think someone can write a book with all the knowledge in this forum!. Excellent ideas.
No, not a good idea to run at the max rails. I was not clear. My point was that with a 12 ohm load the power dissipation will be below the ~65W that National recommends even at the maximum supply voltages. So since you will most likely be below these voltages the power dissipation will be less.
-SL
-SL
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.