I have balanced signals from my pre-amp and would like use them.
What's the best way of implementing a balanced topology with a LM3886?
There are plenty of kits and boards around, would it work if I just got myself a dual set of boards and hooked the speakers up between the two terminals (+/-)?
It would be similar to the bridged amplifier but with still different since I have two inputs per channel?
What's the best way of implementing a balanced topology with a LM3886?
There are plenty of kits and boards around, would it work if I just got myself a dual set of boards and hooked the speakers up between the two terminals (+/-)?
It would be similar to the bridged amplifier but with still different since I have two inputs per channel?
Markus
You could always configure the LM3886 as a differential amp using one of the standard circuits for op-amps available in text books - after all the LM3886 is just a big op-amp.
However, I'd suggest a separate high CMRR balanced-to-unbalanced coverter using one of the many op-amp circuits freely available or, preferably, a decent transformer, with the single ended output of the converter feeding a standard (i.e. single ended) LM3886 circuit. The transformer setup has the advantage of higher CMRR and galvanic isolation than the op-amp at the expense of somewhat higher (but still more than acceptable) LF distortion.
What you describe using two LM3886 amps is a bridge circuit if you have balanced (hot/cold) inputs. This arrangement is often used for headphone amps and is described as 'balanced' but it isn't since it offers no CMRR. It does offer other advantages over a single ended setup, however, mainly related to isolating power supply ground currents from the speaker return path, but if you arrange your grounding appropriately that becomes a non-issue.
You could always configure the LM3886 as a differential amp using one of the standard circuits for op-amps available in text books - after all the LM3886 is just a big op-amp.
However, I'd suggest a separate high CMRR balanced-to-unbalanced coverter using one of the many op-amp circuits freely available or, preferably, a decent transformer, with the single ended output of the converter feeding a standard (i.e. single ended) LM3886 circuit. The transformer setup has the advantage of higher CMRR and galvanic isolation than the op-amp at the expense of somewhat higher (but still more than acceptable) LF distortion.
What you describe using two LM3886 amps is a bridge circuit if you have balanced (hot/cold) inputs. This arrangement is often used for headphone amps and is described as 'balanced' but it isn't since it offers no CMRR. It does offer other advantages over a single ended setup, however, mainly related to isolating power supply ground currents from the speaker return path, but if you arrange your grounding appropriately that becomes a non-issue.
Last edited:
Here's another idea:
http://www.diyaudio.com/forums/pass-labs/36398-gc-supersymmetry.html
By the way, I'm about to build a few balanced/bridged LM4780s. The chip is unstable with less gain than 10x, but setting gain to 11x - 13x should work. That will give a fairly normal amp gain of 21 - 30 dB overall. Some distortion components and DC offsets may also be cancelled across the bridge, if the two amps are reasonably well matched. I expect the biggest issues will be heat management and current capacity into low-ish load impedances.
See also the datasheet and the application note. Both demonstrate the improved distortion performance of the bridged configuration ("BR100").
http://www.national.com/ds/LM/LM4780.pdf
http://www.national.com/an/AN/AN-1192.pdf
http://www.diyaudio.com/forums/pass-labs/36398-gc-supersymmetry.html
By the way, I'm about to build a few balanced/bridged LM4780s. The chip is unstable with less gain than 10x, but setting gain to 11x - 13x should work. That will give a fairly normal amp gain of 21 - 30 dB overall. Some distortion components and DC offsets may also be cancelled across the bridge, if the two amps are reasonably well matched. I expect the biggest issues will be heat management and current capacity into low-ish load impedances.
See also the datasheet and the application note. Both demonstrate the improved distortion performance of the bridged configuration ("BR100").
http://www.national.com/ds/LM/LM4780.pdf
http://www.national.com/an/AN/AN-1192.pdf
Last edited:
I was surprised to see the LM4780 giving improved distortion performance in bridged mode, given that its seeing half the load impedance. The paralleled mode reduces the HF distortion quite a bit too, over bridged. I can't figure out why paralleled into 4R gives better HF distortion performance than single channel into 8R though. Anyway, these results seem to imply bridged/parallel would cover both bases
The bridged amp shows improved distortion over a single amp mainly when the bridged amp is built from an inverting stage and a non-inverting stage. If you use two similar amp stages with an inverted signal driving one of them you do not get the same benefit. The improvement comes from the fact that certain distortion products appear with the same (or similar) phase for both inverting and non-inverting topologies, so they appear at the speaker as a common mode signal and effectively cancel each other out. You also get the benefit of reduced ground noise because speaker currents go from supply to supply, not into ground.
You will find updated schematics later in the same thread. I haven't built something like that myself, but I found it intriguing. As far as I can tell, the supersymmetric configuration passes noise and distortion from one amp to the other, where it cancels at the loudspeaker, while the current source lifts the crossover point and should get rid of some distortion.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.