Does bridge of two LM3886 sound better?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'd go for bridged sounding better - its a bit harder to mess up the grounding when there are no high ground return currents. Layout is much easier. Of course distortion is going to be higher into 4R than 8R but its probably not going to be high enough to be audible whereas poor grounding is very likely to be.
 
Thanks!

Thanks to all for your comments and suggestions. In the present form the
amp (parallel) has little noise. Hoping I will be able to solve the
RF problem as well. My only concern was at the same volume level whether
the bridge will sound better than the parallel which will justify the hard
work. Will have to cut two tracks and flip the connection to convert one
of the inverted to non-inverted. I think there is another solution without
making this change, that is using DRV134, but do not know if attaching
another IC on the signal path is OK.

Regards
Roushon.
 
Thanks Andrew for the caution and thanks Kuldeep for
the nice reference.

I tried to make the amp PS as future proof as possible.
Each channel has its own PS. There are two toroidal
transformers. Each has 5Amp per rail and 2 x (0-33v).
There is filtering with enough electrolytic caps (around
14500uF per rail) and then regulated by LM338 to get +-
35v dc (which I can increase but this is optimal I think for
LM3886). So I guess bridge can be driven easily.

Regards
Roushon.
 
A (possibly serious) question!

As I look at the application sheet AN-1192 I find that
in parallel of two LM3886 the feedback resistors are 20K,
on the other hand the same is about 47 and 51K in the
bridge of two LM3886. Why is this? My circuit is attached
in the parallel version. I thought I just have to exchange
the Amp input and the output of the AD8620 connection
to one of the LM3886. Do I have to change the feedback
resistors also.

Please let me know.

Thanks
Roushon.
 

Attachments

  • MyGainClone.gif
    MyGainClone.gif
    18.5 KB · Views: 579
The 500R aggregate load on the OPA627 might well result in somewhat higher distortion than necessary. Also did you really intend to have gain > 100 ? If you cut out the T-network 1k resistor it'll be back down to X20 but I'd suggest raising the feedback network impedances by at least a factor of 2 to give the 627 an easier time.
 
Well, this thread is somewhat continuation of the earlier thread. Forgot to
mention that there is the stepped (10K) attenuator in between OPA627 and
the amp.

Abraxalito, you may (not) remember that that 1K T-network resistor was
suggested in the previous thread to match impedance as I was having
distortion.

Kuldeep, I did not mean why the inverted and non-inverted has different
feedback resistors, rather I wanted to know why the feedback resistors were
less than half value in the parallel case.

Please tell me to get a decent bridge what are the changes I should make
in the above circuit. I am planing to remove the buffer stage around OPA627,
but the stepped attenuator will be there.

Regards
Roushon
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.