Bridgeclone

[carlmart]

After gathering some information from this Forum and from AN1192, I would like to discuss the options we might use for bridging and paralleling amplifier chips.

The central idea is to use a balancing chip, which can be the SSM2142, a DVR134 or any other, or perhaps a transformer, to buffer the amp chips from the pot. We might eliminate the cap and isolate the amplifying chips.

Bridging and paralleling might bring other benefits, like cancelling some distortions, etc.

Which amp chips you use is up to you.

We can also use lower voltages, like +/-15v or +/-18v, which would allow lower loads without the chips protecting themselves. It would also allow using batteries, for those that would like to try them. If we to higher voltages we will need to regulate the power we fit to the balancing chip.

For the most informed ones, you can see were we might be getting to: Rowland territory!!!

The first Rowlands use an input transformer to balance the bridge, though newer versions are eletronically balanced apparently. To simplify this step I'm using two quite known balancing chips, from BB and ADI.

The Rowland also uses non-inverted versions all over, but we can do that too.

My project involves designing pcbs to probably hold two complete bridges on each, which some links ready to parallel the bridges.

Comments and constructive critics, please!


Carlos

[carlmart]

Here's the second option.

I'm jumping one stage, just paralleling two amp chips, to go to the most complete version.

The only difficult parts here are the 0.1% resistors National recommends for the paralleling. In fact they recommend using 0.1% resistors all over and also DC servos on each amp chip.

I think this might be a bit much, but building them will prove me right or wrong.

The second drawing could only go zipped. Sorry.


Carlos

[Hamish]

allright! this is what i was after! only thing i can see, is that with the balanced driver, BB recommends a buffer. there might be an easier way to do this, as there are three internal opamps and an addition buffer. i have no problem with opamps, but if kept to a minimum the sound should (hopefully) be a little better.

[carlmart]

Originally posted by Hamish
only thing i can see, is that with the balanced driver, BB recommends a buffer. there might be an easier way to do this, as there are three internal opamps and an addition buffer. i have no problem with opamps, but if kept to a minimum the sound should (hopefully) be a little better.

In my opinion, the BB or ADI chips can work as a buffers too. Remember that here they are only feeding inverted and non-inverted signals to the amp chips.

The internal opamps are obligatory, because if you want to go balanced you will need two or three chips to accomplish it anyway. The advantage here is that an important part of that differential output, the high precision resistors, are built in.

A transformer might be simpler than this, but a lot more expensive.

The only thing I designed in but I'm not quite sure how it will work is the pot. All designs I have seen with these chips, mostly mixers, put the output amp in between.

But I think it will work as it is now. I will try it, also adding buffers or lineamps before it to see what happens.


Carlos


Carlos

[carlmart]

Sorry folks. Today I found out the bridge/parallel design I zipped and sent to the Forum was not the final one I had made.

Here's the correct one.

For those interested in seeing a different approach to paralleling OP549s, please have a look here:


http://www.diyvideo.com/forums/showt...8&pagenumber=3

Once again I apologize for my mistake above.


Carlos

[tbla]

[ThorstenL]

Hi,

The central idea is to use a balancing chip, which can be the SSM2142, a DVR134 or any other, or perhaps a transformer, to buffer the amp chips from the pot.

We had this discussion off-line before.

My main concerns regarding the using the line driver chips are:

1) Quality of on board Chips and more importantly RESISTORS
2) Loss of the option for balanced inputs

I would champion still a transformer and if the positive measurable and audible effects from a transformer are undesirable or too expensive I would suggest that an electronically balanced input using Op-Amp's would be a better choice. In this case you have an input that behaves more or less like a transformer, allowing you to us ethe circuit with balanced or SE Inputs as you like.

And of course, an electronically balanced input can still be made to use Op-Amp Chips in Inverted mode, which would in effect allow you to simply add the "balancing" Op-Amp to the actual bridge circuit, taking the actual outputs of the Amplifier Chips as the point to balance. Simply use a really fast Op-Amp here (how about LM6181 or AD811) and the whole balancing process will happen at the speed of main Amp Chip. This would work well in the sense of "simplicity".

Also, in this case you have a fully free choice of the quality of resistors and op-amp's which will give you the ability to use for example high speed Video Amp's or the very nice sounding OPA627 (of course, three OPA627 readily cost as much as a decent transformer).

If you JUST want want an inverted drive to the second Amp from an SE source the easiest way is to use a sonically very transparent Op-Amp as simple gain of 1 inverter and to drastically simplify the circuit. In this case you could use a Dual Op-Amp and use the other halve as unity gain buffer.

Sayonara

[carlmart]

Originally posted by Kuei Yang Wang

We had this discussion off-line before.

We certainly did, and I always enjoy most of them...

My main concerns regarding the using the line driver chips are:

1) Quality of on board Chips and more importantly RESISTORS
2) Loss of the option for balanced inputs

What balanced inputs are we talking about? The option I was considering was an unbalanced input stage with a balanced output that would feed the inverting gainclones.

I would champion still a transformer and if the positive measurable and audible effects from a transformer are undesirable or too expensive I would suggest that an electronically balanced input using Op-Amp's would be a better choice. In this case you have an input that behaves more or less like a transformer, allowing you to us ethe circuit with balanced or SE Inputs as you like.

Once again, I think we are discussing a balanced output here, not input.

What a transformer would bring, certainly a good thing, is galvanic isolation. The price to pay is ... the price. Good transformers are quite expensive and not easy to find. On the other side I do think you can get as good and measurable effects with transformerless stages too.

Implementing a balancing stage with two or three chips would be fine, as we now can get 0.1% resistors rather easily. It will just use a lot of space and would not be as straightforward as using the chips I suggested.

In order to make small pcbs for this project, that could be used in different ways and options, I opted for the single balancing chips.

And of course, an electronically balanced input can still be made to use Op-Amp Chips in Inverted mode, which would in effect allow you to simply add the "balancing" Op-Amp to the actual bridge circuit, taking the actual outputs of the Amplifier Chips as the point to balance. Simply use a really fast Op-Amp here (how about LM6181 or AD811) and the whole balancing process will happen at the speed of main Amp Chip. This would work well in the sense of "simplicity".

OK. This would mean using a balancing stage as I suggested above. It seems these BB and ADI chips are not quite to your liking and/or prefer to ignore them.

Also, in this case you have a fully free choice of the quality of resistors and op-amp's which will give you the ability to use for example high speed Video Amp's or the very nice sounding OPA627 (of course, three OPA627 readily cost as much as a decent transformer).

Indeed! Also 0.1 % resistors do not come cheap.

I'd say something like an AD825/AD8610 or AD826/AD8620 might be used for this too. Some comments from this Forum recommended them as better sounding than 627s, but this might be a mere question of taste.

If you JUST want want an inverted drive to the second Amp from an SE source the easiest way is to use a sonically very transparent Op-Amp as simple gain of 1 inverter and to drastically simplify the circuit. In this case you could use a Dual Op-Amp and use the other halve as unity gain buffer.

Well, in fact that's something I was willing to try, using a dual chip as you suggest, and compare it to the SSM and DRV chips.

I think it's time I build these pcbs and see what I get. I was waiting for ideas to improve on my balancing/paralling suggestions.


Carlos

[carlmart]

Originally posted by carlmart
Following some off line suggestions from Kuei Yang Wang, here's a different implementation for a bridged/parallel gainclone.

Carlos

This week I should be building these pcbs, then comparing results with different chips.

Kuei Yang Wang thinks an LM6182 would be a less noisy alternative. Unfortunately this chip was discontinued by National, but I have some to try and compare other chips to it.




Carlos

[miguel2]

Hi Carlos,

Maybe I am missing something here, but why do you prefer the bridge/paralell configuration instead of the first bridge? Is just for the power or there is something else?

From what I read from NS, the output resistors will have to be precisely matched, which hads complexity to the design.

Miguel