hello,
I'd like to build my first power-amp and would like to keep things basic and easy.
I wanted to use 2 OPA549 per channel, because the thd figures for 4 ohm suggested this to me. I thought that 2 in parallel would show lower thd.
Should the shown circuit (it's from a National Application Note) work fine for the OPA549?
Well, except I thought about leaving away those 68uF caps, using a lower gain of 2 or 4 (as often suggested) and another op-amp in front like the opa2132 or so.
Cheers,
Dominique
I'd like to build my first power-amp and would like to keep things basic and easy.
I wanted to use 2 OPA549 per channel, because the thd figures for 4 ohm suggested this to me. I thought that 2 in parallel would show lower thd.
Should the shown circuit (it's from a National Application Note) work fine for the OPA549?
Well, except I thought about leaving away those 68uF caps, using a lower gain of 2 or 4 (as often suggested) and another op-amp in front like the opa2132 or so.
Cheers,
Dominique
Attachments
Yes, it will work.
You have to match manually some resistors to ensure equal gain on both amps, otherwise, you'll get more problems than single-chip operation.
To be sure you have equal gain on both amps, you can once they are assembled send them a continuous signal that would bring the output voltage near the clipping point, and measure separately the output voltage across both amps to ensure they do have equal gain, and eventually match resistors better. You can get differences as low as 1mv with this method !
Edit : and you won't need the stupid ballast resistors, which dramatically decrease the damping factor.
If you want a parallel amp, I'd suggest you to use the TDA7293 : it has internal programming for being used as slave in a bridged or parallel configuration, without any other components and without ballast resistors.
You have to match manually some resistors to ensure equal gain on both amps, otherwise, you'll get more problems than single-chip operation.
To be sure you have equal gain on both amps, you can once they are assembled send them a continuous signal that would bring the output voltage near the clipping point, and measure separately the output voltage across both amps to ensure they do have equal gain, and eventually match resistors better. You can get differences as low as 1mv with this method !
Edit : and you won't need the stupid ballast resistors, which dramatically decrease the damping factor.
If you want a parallel amp, I'd suggest you to use the TDA7293 : it has internal programming for being used as slave in a bridged or parallel configuration, without any other components and without ballast resistors.
youyoung21147 said:
Wow, that's great! I didn't think of that trick of closely matching the amp's gain! I think I'll try this method! Are there any drawbacks or risks?
That chip sounds interesting, I'll look for a datasheet!
I just thought I'd use the opa549 because a friend has exactly 4 left which he promised to give me for free, so I wouldn't be too shocked if I blew them up
(it's my first power amp and I just had to work with line level until now).There is no drawback using this gain matching method, except off course that you mustn't connect the speakers when doing this !
You must also remove the inut capacitors if there are some because they'd block the DC you apply on the input.
Maybe you will not need the Ci caps in the feedback if your DC offset is low on the output (<5mV). You also need to remove them to use my gain matching method.
Finally, if you don't need to drive heavy loads, why not use them bridged ?
it is more adapted to "audiophile" operation than parallel, which is rather adapted to subwoofer driving. The bridged operation reduces THD, doubles output swing, giving more power to 8 and 16 ohm loads. Its drawback is the power dissipation, which is doubled because you use one chip per pole contrary to single-ended operation. The total power dissipation of the amp could be quadrupled because the bridged mode allows a doubled output power on the same load.
See National AN-1192, there is a bridged schematic in it.
Mail moi si tu veux + d'infos, je suis justement en train de me faire un LM4780 bridgé
You must also remove the inut capacitors if there are some because they'd block the DC you apply on the input.
Maybe you will not need the Ci caps in the feedback if your DC offset is low on the output (<5mV). You also need to remove them to use my gain matching method.
Finally, if you don't need to drive heavy loads, why not use them bridged ?
it is more adapted to "audiophile" operation than parallel, which is rather adapted to subwoofer driving. The bridged operation reduces THD, doubles output swing, giving more power to 8 and 16 ohm loads. Its drawback is the power dissipation, which is doubled because you use one chip per pole contrary to single-ended operation. The total power dissipation of the amp could be quadrupled because the bridged mode allows a doubled output power on the same load.
See National AN-1192, there is a bridged schematic in it.
Mail moi si tu veux + d'infos, je suis justement en train de me faire un LM4780 bridgé
youyoung21147 said:There is no drawback using this gain matching method, except off course that you mustn't connect the speakers when doing this !
Sorry, I didn't mean drawbacks of your measuring method, but then using the amps in parallel with no ballast resistors at all. I wonder if the amps could behave slightly different (change with temperature e.g.), so that the load would possibly under certain circumstances not be equal on both amps.
Regarding the bridged configuration...
Doesn't running both amps in parallel reduce the thd too, if the load is a bit difficult for the amp? My loudspeakers are 4 ohm.
I just wonder, if yes, if that would also be true for low output power (low volumes).
Mail moi si tu veux + d'infos, je suis justement en train de me faire un LM4780 bridgé
Merci beaucoup!
LM4780: Oui, je viens de voir les photos I will think about it, read some more, and probably come back with some questions or my finished circuit soon!
Thanks again for your help!
Dominique
I have built many GCs using this chip, my favourite, and there is no perceptable effect with using output resistors, and the benefits they have in ensuring stable operation whatever load the amps are connected to make this a no-brainer in my opinion.
Whilst Youngyoungs theory may be ideal in principle, speakers are not a perfect load, and ordinary differences in driver parameters can lead to instability. Bear in mind, production tolerances on drivers mean that their specs. can vary by up to 10%!
Whilst Youngyoungs theory may be ideal in principle, speakers are not a perfect load, and ordinary differences in driver parameters can lead to instability. Bear in mind, production tolerances on drivers mean that their specs. can vary by up to 10%!
I don't have any particular experience with parallel driving, it was just a suggestion but I see others have more experience than me
I'd rather use those chips in bridged mode, it is the simplest and most rational way to drive fullrange speakers with two chips.
Parallel mode, because the same signal goes through two devices, may reduce the clarity of the output signal.
I'd rather use those chips in bridged mode, it is the simplest and most rational way to drive fullrange speakers with two chips.
Parallel mode, because the same signal goes through two devices, may reduce the clarity of the output signal.
pinkmouse said:
Well, now I'm a bit insecure...
It would be nice to test several configurations, but I'm not sure if I'm patient and disiplined enough.
Pinkmouse, could you tell me in which configs you liked the opa549 best, or, if difficult to tell, if some amp you built didn't work well - this kind of info could also be very helpful!
pinkmouse said:
I don't understand that.
I would think, because those two amps in parallel drive one loudspeaker, it just depends on both amps behaving exactly the same way, but not depend on the speaker.
I'll go to bed, I can't think anymore!
Many thanks to all of you who replied!
Dominique
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.