LM3876 Sub Amp Guidance

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi guys I'm going to build a powered sub-amp using two LM3876 in BR100 configuration.

Here is my drawing.Any modifications....!!!
 

Attachments

  • LM3876 Sub Amp BR100.JPG
    LM3876 Sub Amp BR100.JPG
    48.4 KB · Views: 218
Yes. Everyplace you have a 100n capacitor, add an electrolytic in parallel with it. Make them as large as you can. I would use at least 1000 to 2200 uF for each one, and possibly much more. And it would be much better to parallel many smaller electrolytics instead of using one big one.

Make sure that the 0.1 uF caps are RIGHT AT the power pins. Actually, every component that connects to a pin should be as close to the pin as possible.
 
Data sheet suggests a gain of 10 or greater or "oscillations may occur". You might want to consider generating two signals 180 apart with a preamp stage and driving two non-invering amps with gain as the data sheet suggests.

Or build it that way but change R5 to a 100K pot to trim output DC.
Put a 0.1uF in parallel with it.
Watch for oscillations...you may hear no output but the amp may get hot.

Data sheet also suggests a min of 20uF for supply bypass. I would go larger than this.

I agree with gootee in that many smaller caps are better than one large cap and they still need to be close to the pins.

And don't forget:

"1) The LM3876T package TA11B is a non-isolated package, setting the tab of the device and the heat sink at V− potential when the
LM3876 is directly mounted to the heat sink using only thermal compound. If a mica washer is used in addition to thermal compound,
θCS (case to sink) is increased, but the heat sink will be isolated from V−."

Have fun.
 
When you copy applications from a datasheet, make sure that you copy everything correctly and completely.
  1. Check the feedback resistors. They should be different for the inverting and the non-inverting configuration to achieve the same gain on both sides.
  2. Add the RF snubbers aka Zobel networks to the outputs.
  3. Add the 10 µF and 1000 µF caps to the power supply on the board.
  4. Don't go to the upper limit for the provided load with the supply voltage. Check the power dissipation graphs in the datasheet and be aware of the current limiter.
 
Ok guys thanks for the points.

I corrected the mistakes.
*Added snubbers
*Added electrolytics to power rails

Regarding R3,App note says to use 51.1K resistor.That is hard to find so I use 47K so the gain is mismatched on two amps :(

Reg snubbers I used a parallel RC snubber.Do I need to use a LR snubber in series as in the datasheet?

Whats the point of increasing value of R5?What does it do?
 

Attachments

  • LM3876 Sub Amp BR100.JPG
    LM3876 Sub Amp BR100.JPG
    65 KB · Views: 136
Fellows I need to make an active subwoofer over 100Watts.

I need to use a single coil woofer 8 ohms.

When in bridged mode if each LM3876 sees as a 4ohms load is it ok to use a LM3886?

I have plenty LM3886 @home :)

If I configure LM3886 in BR100 (bridge) & load into 8 ohms what will be the output wattage?Supply is 30V DC.
 
OP2 has a gain of 11 and OP1 has a gain of -10.

To fix this add a 4K7 in series with C10.

The LR in the output path helps isolate capacitive loads from the amp.

"...change R5 to a 100K pot to trim output DC."

Vos on that device can be as much as 15mV. (one each way is 30mV)

If you add to that an imbalance of resistance to the inputs the input current could give you more offset between the two amplifiers.

With the adjustment you could trim that to a very low value.

After you build it (without speaker) adjust the pot and measure the output VDC.

If you don't need it, take it out.

If you don't want to put it in then make it 47K like the other ones.

:)
 
3886 is a better choice for this application, but you need to reduce your voltage. At +/-30 volts the two 3886's would be able to produce about 150 watts output power into 8 ohms while dissipating 45 peak watts of heat (pdmax) in each 3886, which is above the 40 watt limit National recommends.

At +/- 25 volts would get you about 100 watts into 8 ohms with a peak dissipation of about 30 watts of heat in each chip, which should be o.k as long as your heat sinking is up to the task
 
OP2 has a gain of 11 and OP1 has a gain of -10.

To fix this add a 4K7 in series with C10.

The LR in the output path helps isolate capacitive loads from the amp.

"...change R5 to a 100K pot to trim output DC."

Vos on that device can be as much as 15mV. (one each way is 30mV)

If you add to that an imbalance of resistance to the inputs the input current could give you more offset between the two amplifiers.

With the adjustment you could trim that to a very low value.

After you build it (without speaker) adjust the pot and measure the output VDC.

If you don't need it, take it out.

If you don't want to put it in then make it 47K like the other ones.

:)

Totally understood.Thank a lot.

Even with mismatched gains (11 & -10) I can run the system :)
 
3886 is a better choice for this application, but you need to reduce your voltage. At +/-30 volts the two 3886's would be able to produce about 150 watts output power into 8 ohms while dissipating 45 peak watts of heat (pdmax) in each 3886, which is above the 40 watt limit National recommends.

At +/- 25 volts would get you about 100 watts into 8 ohms with a peak dissipation of about 30 watts of heat in each chip, which should be o.k as long as your heat sinking is up to the task

Very nice.Thats what I wanted to know :)

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