How does this chip perform at low frequencies, any issues with using it for a small powered sub (i.e. with computer speakers)?
The driver wont be capable of incredible ultra-low frequency performance due to its compact size so I don't need something outstanding, but still want it acceptable etc.
The driver wont be capable of incredible ultra-low frequency performance due to its compact size so I don't need something outstanding, but still want it acceptable etc.
I have it to. LM3886
http://www.diyaudio.com/forums/showthread.php?s=&threadid=63461&perpage=10&pagenumber=4
I'm using it for 12" sub for car, with this smps, and I am very happy with LM
Post no.190
http://www.diyaudio.com/forums/showthread.php?s=&threadid=31614&perpage=10&pagenumber=19
http://www.diyaudio.com/forums/showthread.php?s=&threadid=63461&perpage=10&pagenumber=4
I'm using it for 12" sub for car, with this smps, and I am very happy with LM
Post no.190
http://www.diyaudio.com/forums/showthread.php?s=&threadid=31614&perpage=10&pagenumber=19
LM3876, 1 possible setup
Suggested circuit: http://www.national.com/images/pf/LM3876/01183201.pdf ... pretty straight forward. Pin out is at: http://www.national.com/images/pf/LM3876/01183202.pdf . Physical plan at: http://www.national.com/packaging/mkt/tf11b.pdf
" 56W continuous average output power into 8 ohms
¥ 100W instantaneous peak output power capability
¥ Signal-to-Noise Ratio >= 95 dB(min)
¥ An input mute function
¥ Output protection from a short to ground or to the supplies via internal current limiting circuitry
¥ Output over-voltage protection against transients from inductive loads
¥ Supply under-voltage protection, not allowing internal biasing to occur when |VEE| + |VCC| <= 12V, thus eliminating turn-on and turn-off transients
¥ 11-lead TO-220 package
¥ Wide supply range 20V - 94V ..."
From the National specs at: http://www.national.com/pf/LM/LM3876.html
So, one could drive the bass low band pass filter / crossover directly from the output. (I would also suggest putting a 3 db/octive low pass filter with, say, a rolloff around 200 to 500 Hertz in or very close to the input on the circuit board (pin 10) ... to insure that nothing above 500 or so got near the input, thus reducing a lot of headaches with the sub woofer box and/or crossover and cables trying to "resonate" or pass signals above about 300 to 500 Hertz.
National indicates that there may be greater distortion when driving this chip above 40 watts (note the distortion graph with test parameters @ 40 watts into 8 ohms. Chip makers know when their chips look good and when they don't ... so the best graph was done at 40 watts ... you figure it out.) ... and I would make sure that my sub woofer had at least 8 ohms from DC to 500 or more Hertz (installed in the box with crossover) as any less might result in some interesting distortion not otherwise expected.
I would also be very, very generous with the heat sink, say, at least 2 pounds of aluminum. I would also want that 10K variable resistor at input to be the best quality / reliability I could fine. (Cheap heat sink for inside a sub woofer box: 1/4" thick x 1 foot x 1 foot or larger aluminum plate ... mounting hole / connection hole right in the middle. Mount this aluminum plate inside the sub woofer box on the roof ... install your favorite conductive stuffing for static removal as well as dampining ... )
Personally, I would have the power supply outside of the sub woofer box. That way I could add extra power supply filtering as required and be assured that there was little fire danger in side of the sub woofer box. Power supply should have lots of "headroom" too, designing for 100 watts power in and 75+ watts power out. Upper voltage output limit is that 35 VDC (VCC) indicated in the graph ... 35 VDC at VCC (pins 1 & 5) is what results 40 watts audio power into 8 ohms (nominal). I would not include the "mute" feature, leaving that part of the circuit out entirely. I would also make a provision to connect the (internal) 3876 (3886) amp to some external binding posts ... so in the case of a chip or circuit failure I could still drive the sub woofer box from another amp.
..........
FYI: http://www.national.com/pf/LM/LM3886.html is also a good candidate for this = a little more power rating, but I would not loosen up the PS voltage or anything else above as this chip would simply give you a little more "headroom" and possibly a little less distortion than the LM3876.
" ¥ 68W cont. avg. output power into 4 at VCC = ±28V
¥ 38W cont. avg. output power into 8 at VCC = ±28V
¥ 50W cont. avg. output power into 8 at VCC = ±35V
¥ 135W instantaneous peak output power capability
¥ Signal-to-Noise Ratio >= 92dB
¥ An input mute function
¥ Output protection from a short to ground or to the supplies via internal current limiting circuitry
¥ Output over-voltage protection against transients from inductive loads
¥ Supply under-voltage protection, not allowing internal biasing to occur when |VEE| + |VCC| <= 12V, thus eliminating turn-on and turn-off transients
¥ 11-lead TO-220 package
¥ Wide supply range 20V - 94V ..."
Suggested circuit: http://www.national.com/images/pf/LM3876/01183201.pdf ... pretty straight forward. Pin out is at: http://www.national.com/images/pf/LM3876/01183202.pdf . Physical plan at: http://www.national.com/packaging/mkt/tf11b.pdf
" 56W continuous average output power into 8 ohms
¥ 100W instantaneous peak output power capability
¥ Signal-to-Noise Ratio >= 95 dB(min)
¥ An input mute function
¥ Output protection from a short to ground or to the supplies via internal current limiting circuitry
¥ Output over-voltage protection against transients from inductive loads
¥ Supply under-voltage protection, not allowing internal biasing to occur when |VEE| + |VCC| <= 12V, thus eliminating turn-on and turn-off transients
¥ 11-lead TO-220 package
¥ Wide supply range 20V - 94V ..."
From the National specs at: http://www.national.com/pf/LM/LM3876.html
So, one could drive the bass low band pass filter / crossover directly from the output. (I would also suggest putting a 3 db/octive low pass filter with, say, a rolloff around 200 to 500 Hertz in or very close to the input on the circuit board (pin 10) ... to insure that nothing above 500 or so got near the input, thus reducing a lot of headaches with the sub woofer box and/or crossover and cables trying to "resonate" or pass signals above about 300 to 500 Hertz.
National indicates that there may be greater distortion when driving this chip above 40 watts (note the distortion graph with test parameters @ 40 watts into 8 ohms. Chip makers know when their chips look good and when they don't ... so the best graph was done at 40 watts ... you figure it out.) ... and I would make sure that my sub woofer had at least 8 ohms from DC to 500 or more Hertz (installed in the box with crossover) as any less might result in some interesting distortion not otherwise expected.
I would also be very, very generous with the heat sink, say, at least 2 pounds of aluminum. I would also want that 10K variable resistor at input to be the best quality / reliability I could fine. (Cheap heat sink for inside a sub woofer box: 1/4" thick x 1 foot x 1 foot or larger aluminum plate ... mounting hole / connection hole right in the middle. Mount this aluminum plate inside the sub woofer box on the roof ... install your favorite conductive stuffing for static removal as well as dampining ... )
Personally, I would have the power supply outside of the sub woofer box. That way I could add extra power supply filtering as required and be assured that there was little fire danger in side of the sub woofer box. Power supply should have lots of "headroom" too, designing for 100 watts power in and 75+ watts power out. Upper voltage output limit is that 35 VDC (VCC) indicated in the graph ... 35 VDC at VCC (pins 1 & 5) is what results 40 watts audio power into 8 ohms (nominal). I would not include the "mute" feature, leaving that part of the circuit out entirely. I would also make a provision to connect the (internal) 3876 (3886) amp to some external binding posts ... so in the case of a chip or circuit failure I could still drive the sub woofer box from another amp.
..........
FYI: http://www.national.com/pf/LM/LM3886.html is also a good candidate for this = a little more power rating, but I would not loosen up the PS voltage or anything else above as this chip would simply give you a little more "headroom" and possibly a little less distortion than the LM3876.
" ¥ 68W cont. avg. output power into 4 at VCC = ±28V
¥ 38W cont. avg. output power into 8 at VCC = ±28V
¥ 50W cont. avg. output power into 8 at VCC = ±35V
¥ 135W instantaneous peak output power capability
¥ Signal-to-Noise Ratio >= 92dB
¥ An input mute function
¥ Output protection from a short to ground or to the supplies via internal current limiting circuitry
¥ Output over-voltage protection against transients from inductive loads
¥ Supply under-voltage protection, not allowing internal biasing to occur when |VEE| + |VCC| <= 12V, thus eliminating turn-on and turn-off transients
¥ 11-lead TO-220 package
¥ Wide supply range 20V - 94V ..."
wow awesome replies, thanks
so the 3886 can drop in in place of the 3876?
to be honest it's going to be under-used to some extent, for a fairly hardcore computer speaker setup. not under-used in that it wont be playing good quality music at good volume levels (of course it will!
) but i will be building the amp along with two other chip amps inside the sub box, not yet decided on a built-in or separate power supply but would prefer built-in.
if i go that way the amps + power supply will be enclosed in a metal case with the heatsinks protruding outside of the sub box, so no fire hazard or anything really. i will also go fan forced cooling simply due to the fact that generally my computing environment is not silent as it is (computer fans audible anyway).
i'm tempted to simply buy kits for the amp modules, mainly due to the neatness of properly designed PCBs.
so the 3886 can drop in in place of the 3876?
to be honest it's going to be under-used to some extent, for a fairly hardcore computer speaker setup. not under-used in that it wont be playing good quality music at good volume levels (of course it will!
) but i will be building the amp along with two other chip amps inside the sub box, not yet decided on a built-in or separate power supply but would prefer built-in.if i go that way the amps + power supply will be enclosed in a metal case with the heatsinks protruding outside of the sub box, so no fire hazard or anything really. i will also go fan forced cooling simply due to the fact that generally my computing environment is not silent as it is (computer fans audible anyway).
i'm tempted to simply buy kits for the amp modules, mainly due to the neatness of properly designed PCBs.
LM3876 v. LM3886
" ... so the 3886 can drop in in place of the 3876? ..."
It appears so. The pinout is the same.
As long as the power supply is not driven too hard, either has pretty much the same specs too. It looks like the only difference is the output ... the 3886 is able to drive either 4 ohms or 8 ohms, while the 3876 is speced for 8 ohms only. 4 ohm loads would probably bullox up the distortion ratings of the 3876.
" ... so the 3886 can drop in in place of the 3876? ..."
It appears so. The pinout is the same.
As long as the power supply is not driven too hard, either has pretty much the same specs too. It looks like the only difference is the output ... the 3886 is able to drive either 4 ohms or 8 ohms, while the 3876 is speced for 8 ohms only. 4 ohm loads would probably bullox up the distortion ratings of the 3876.
Just be sure you connect pin 5 correctly for complete compatibility with the LM3886, as noted on the datasheet. The LM2876, LM3876 and LM3886 are all fully drop in replacements as long as pin 5 is connected for the LM3886. Pin 5 is a NC for the others.
All of these amplifiers will drive 4 ohms fine but at reduced output power. Notice the spec. no Output Current Limit in each datasheet, LM2876 = 3A, LM3876 = 4A, LM3886 = 7A. Two things are going to limit output power at 4 ohms, current and heat. For the LM2876 and LM3876 there is not as much output current aviable so output power is going to be lower. This is not a problem with the LM3886. Output power is directly related to die size and the larger the die size the better thermals are in the package so larger die can handle higher power dissipation. If you are planning on driving 4 ohms use the LM3886. If doing 8 ohms either of the others will be fine. National test with sine waves so the recomended voltages and output powers in the tables are pretty much worse case. Running music (although a sub it more demanding) is less stressful than sine waves. Check out the Overture Design Guide National has on there web site here--> http://www.national.com/appinfo/audio/
Gives you some good info on designs. Helpful in determining what you need for voltages, heat sink, etc. Keep in mind that the thermals section is calculated using worst case sine waves. I have run amps higher than they say but only with music.
-SL
All of these amplifiers will drive 4 ohms fine but at reduced output power. Notice the spec. no Output Current Limit in each datasheet, LM2876 = 3A, LM3876 = 4A, LM3886 = 7A. Two things are going to limit output power at 4 ohms, current and heat. For the LM2876 and LM3876 there is not as much output current aviable so output power is going to be lower. This is not a problem with the LM3886. Output power is directly related to die size and the larger the die size the better thermals are in the package so larger die can handle higher power dissipation. If you are planning on driving 4 ohms use the LM3886. If doing 8 ohms either of the others will be fine. National test with sine waves so the recomended voltages and output powers in the tables are pretty much worse case. Running music (although a sub it more demanding) is less stressful than sine waves. Check out the Overture Design Guide National has on there web site here--> http://www.national.com/appinfo/audio/
Gives you some good info on designs. Helpful in determining what you need for voltages, heat sink, etc. Keep in mind that the thermals section is calculated using worst case sine waves. I have run amps higher than they say but only with music.
-SL
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.