I was thinking about making a kit for the BPA-200 design. For those who aren't familiar, it uses the National LM3886 power amplifier IC's in a bridged-paralleled configuration to produce 200 watts into 8 ohms. The plan is to make a very high quality double sided PC board, and to include a suitable heatsink and all components with the kit. You would have to supply the power supply and a suitable case and hardware. I will most likely modify the circuit to include balanced or single-ended inputs, and possibly other opamps for the input buffer/s or servos.
At this point, this is only conceptual. I'm just trying to get a feel as to how many people would be interested in such a kit. Please offer any suggestions or criticism, as this will help me decide if it's worth doing, and how to implement it if it is.
Thanks,
Steve
At this point, this is only conceptual. I'm just trying to get a feel as to how many people would be interested in such a kit. Please offer any suggestions or criticism, as this will help me decide if it's worth doing, and how to implement it if it is.
Thanks,
Steve
You mean something like this:
http://www.diyaudio.com/forums/showthread.php?postid=223348#post223348
http://www.diyaudio.com/forums/showthread.php?postid=223348#post223348
Hi sure, that would be great. Can you share with us your findings if any in regard to this topology since in my oppinion looks like this issue of bridging/paralleling configuration is like black magic....some get it to work great while others seem to get inconsistent results. Will it be with the servos or very basic (which seems to work good for some)? Please let us know. I believe you will get a great response to this project, and me in particular, I will be very interested in the PCB only option. Please keep us posted.
Thanks for the replies...
Yes, I understand that there are other implementations available for this amplifier.
I will be experimenting with different configurations and doing extensive prototyping before designing the pcb for the project. I tend to like the most simple implementation possible, but as you said there are some stability issues with this topology. It will most likely have the servos. The idea is to make the design as easy to build as possible, so that anyone who can solder can assemble it. This will mean complete and detailed instructions, parts lists, suggested power supply, etc.
Yes, I understand that there are other implementations available for this amplifier.
I will be experimenting with different configurations and doing extensive prototyping before designing the pcb for the project. I tend to like the most simple implementation possible, but as you said there are some stability issues with this topology. It will most likely have the servos. The idea is to make the design as easy to build as possible, so that anyone who can solder can assemble it. This will mean complete and detailed instructions, parts lists, suggested power supply, etc.
If you want, please keep us informed about your progress. I'll guess that you already have seen my Ultra universal Gainclone. Feel free to steal anything from me.
BTW: I'm amazed that Fred hasn't complained yet about your "commercial" movements.
BTW: I'm amazed that Fred hasn't complained yet about your "commercial" movements.
I have some questions;
1- What if I remove the servo circuits (op-amps)? In the app notes they say "it reduces offset voltage".. If there are output voltage differences between power op-amps what happens? They get warmer?
2-In the same app notes the power output table; for 4 ohms the BPA200 can drive 450W "Burst Clipping Power". What does is mean? Can we say it is RMS value or PMPO? So we can really get 450Watts / 4ohm?
3-In the reference BPA200 design, they use %0.1 tolerence metal film resistors. I cannot find that resistors. Can I use %1 resistors instead of them? What happens if? And for the serial 0.1ohm resistors I can find only wired resistors and they have %5 tolerence. Can I use they?
Thanx...
1- What if I remove the servo circuits (op-amps)? In the app notes they say "it reduces offset voltage".. If there are output voltage differences between power op-amps what happens? They get warmer?
2-In the same app notes the power output table; for 4 ohms the BPA200 can drive 450W "Burst Clipping Power". What does is mean? Can we say it is RMS value or PMPO? So we can really get 450Watts / 4ohm?
3-In the reference BPA200 design, they use %0.1 tolerence metal film resistors. I cannot find that resistors. Can I use %1 resistors instead of them? What happens if? And for the serial 0.1ohm resistors I can find only wired resistors and they have %5 tolerence. Can I use they?
Thanx...
saletel said:
I am considering offering it 2 ways:
1- PCB with 3886's and heat sink adapter only.
2- PCB with all components supplied.
I don't have the biggest budget to work with, so it will most likely be option 1, with a parts list and possible vendors (for those in the US and Canada) for the other components. If there are any components that are hard to find, they may also be included.
Moderators:
Since everyone seems so concerned that I posted this here, I have no problem with it being moved to the appropriate forum. It wasn't my intention to break the rules.
Thanks.
Steve
Yes.Dxvideo said:
I have also some trouble to figure out the _real_ max output power but I suspect this means power a few seconds until the chip is hot.Dxvideo said:
You can use 1% which you match. Normal 1% is usually around 0.3% according to my experience.Dxvideo said:
For the other resistors I think 5% will do.
takes only time, no money involved.
Thanks for the direction, everyone. I will look into the LM4780. Is this a new product? The only thing that would bother me about a single die is getting the heat out of it.
I believe that the burst clipping power is just that- a very brief "burst" (mS long) that shows the dynamic power of the amp. I think the difference is that it is power output that is possible before the supply sags from the current draw.
I plan on doing a few nice updates to the circuit. These will probably include better opamps for the buffers, true balanced or unbalanced inputs, and the option of using the board as a stereo unit with 4 devices per output, or as a mono unit with up to 8 output devices. Also to be included is a pre-machined heat sink adapter which will be thick enough to have very good heat transfer to the sink, and the devices will be mounted standing up, not laying down on the board.
I am thinking of using all output devices in the inverting mode only, and burr-brown opamps as input buffers/inverters. Any other suggestions would be great! Let me know what you want!
I believe that the burst clipping power is just that- a very brief "burst" (mS long) that shows the dynamic power of the amp. I think the difference is that it is power output that is possible before the supply sags from the current draw.
I plan on doing a few nice updates to the circuit. These will probably include better opamps for the buffers, true balanced or unbalanced inputs, and the option of using the board as a stereo unit with 4 devices per output, or as a mono unit with up to 8 output devices. Also to be included is a pre-machined heat sink adapter which will be thick enough to have very good heat transfer to the sink, and the devices will be mounted standing up, not laying down on the board.
I am thinking of using all output devices in the inverting mode only, and burr-brown opamps as input buffers/inverters. Any other suggestions would be great! Let me know what you want!
US-only and DC-servo
I am rather ”surprised” to see that from some of the replies to this tread it seems like US/Canadian DIY’er prefer to buy “local” PCB’s.
This is almost to imply that only US/Canadian PCB’er are any good. I almost feel offended.........not really
Shipping PCB around the world is very easy, since they can go in an normal envelope are not to heavy and don’t cost much to send. The only problem for US/candians to buy “overseas” could be the payment, but even that is normally easy to solve.
To Dxvideo
1. You MUST take care of the DC-issue when you are paralleling chip amps (or other amplifiers for that matter). Failing to do so will result in huge amount of current flowing between the amps…………..I suspect this is what happens when people try to use the parallel approach and can’t get it to work properly.
Personally I think that use of the DC-servo are the “technically” correct solution. However from a more “esthetical” point of view the DC-servo looks slightly “misplaced” since it almost uses more components than the amplifier it-self. Having a carefull “look” at the gain-resistors and use of a large capacitor in the lower feedback arm will get you a very low DC-offset………………….I am working on a “paralleling basic” that I hope will be usefull for all the “BPA-people”………..I hope.
2. Getting 450 W into 4 ohm is probably to “push it to fare”. However 350W into 4 ohm is possible, but you need more that 2x2 LM3886 or an extreme efficient way off cooling them.
3. One of the reason for using 0.1 % resistors is to get AC-current balance. If your parallel amplifiers don’t have (exact) the same gain, then one will be doing all the work (and get to hot). And we are talking a big difference in distribution here. In teori 1% is not good enough to prevent this. However in my expirence, 1% resistors in real life turn out to be 0.3%-0,5% relatively to each other coming form the same batch. (I have up to now used 1% resistors in my constructions). The 0.1%-issue will also be discussed in my “paralleling basic”. For now, just buy 1% and handpick them to be as close as possible.
About the 0.1% “output” resistor. I was very “suspicious” when I saw this requirement. Up on close inspection/calculation it turns out that using 5% resistors worst-case you will have a 10% “imbalance in AC-current between your amplifiers……..it is up to you if you can live with that….
Again, for now I recommend that you just get the 5%/5W and handpick them to be as close as possible. (they are also normally better that the 5% coming from the same batch).
It is also quite hard to get hold of 0.1 % /5 W resitors. 1% would be fine if you can find them.
Have fun and let us all know how you get one.
Regards
Thomas
I am rather ”surprised” to see that from some of the replies to this tread it seems like US/Canadian DIY’er prefer to buy “local” PCB’s.
This is almost to imply that only US/Canadian PCB’er are any good. I almost feel offended.........not really
Shipping PCB around the world is very easy, since they can go in an normal envelope are not to heavy and don’t cost much to send. The only problem for US/candians to buy “overseas” could be the payment, but even that is normally easy to solve.
To Dxvideo
1. You MUST take care of the DC-issue when you are paralleling chip amps (or other amplifiers for that matter). Failing to do so will result in huge amount of current flowing between the amps…………..I suspect this is what happens when people try to use the parallel approach and can’t get it to work properly.
Personally I think that use of the DC-servo are the “technically” correct solution. However from a more “esthetical” point of view the DC-servo looks slightly “misplaced” since it almost uses more components than the amplifier it-self. Having a carefull “look” at the gain-resistors and use of a large capacitor in the lower feedback arm will get you a very low DC-offset………………….I am working on a “paralleling basic” that I hope will be usefull for all the “BPA-people”………..I hope.
2. Getting 450 W into 4 ohm is probably to “push it to fare”. However 350W into 4 ohm is possible, but you need more that 2x2 LM3886 or an extreme efficient way off cooling them.
3. One of the reason for using 0.1 % resistors is to get AC-current balance. If your parallel amplifiers don’t have (exact) the same gain, then one will be doing all the work (and get to hot). And we are talking a big difference in distribution here. In teori 1% is not good enough to prevent this. However in my expirence, 1% resistors in real life turn out to be 0.3%-0,5% relatively to each other coming form the same batch. (I have up to now used 1% resistors in my constructions). The 0.1%-issue will also be discussed in my “paralleling basic”. For now, just buy 1% and handpick them to be as close as possible.
About the 0.1% “output” resistor. I was very “suspicious” when I saw this requirement. Up on close inspection/calculation it turns out that using 5% resistors worst-case you will have a 10% “imbalance in AC-current between your amplifiers……..it is up to you if you can live with that….
Again, for now I recommend that you just get the 5%/5W and handpick them to be as close as possible. (they are also normally better that the 5% coming from the same batch).
It is also quite hard to get hold of 0.1 % /5 W resitors. 1% would be fine if you can find them.
Have fun and let us all know how you get one.
Regards
Thomas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.