just built a lm3886 amp

[audioguru]

The supply voltages are way too high for a 4 ohm speaker.

The heatsink is way too small for anything more than about 15W output.

Are the power supplies regulated? Do their output voltages stay up at full power?

 

[Gaston]

im running an 8 ohm speaker,the heatsink has a fan on it and stays cool. the supplies are unregulated. and as soon as i go grab a beer i will put the scope on the supplies and crank it up and see how the voltage does. and i'll post it

 

[Gaston]

cranked up enough for the woofer cone to just about jump out of the magnet each supply drops about two volts every bass hit

 

[audioguru]

So now you traded the original 4 ohm speaker for an 8 ohm speaker.

The 38V supplies drop to 36V when loaded and the 32V peak drops to 30V peak. Then the power in the 8 ohm speaker is 56W.
It is good that the IC stays cool.

 

[Gaston]

i never said i had a four ohm speaker. it is working fine now. i just want to know about the hum and no output that i corrected with the cap to ground.

 

[Gaston]

damn i sure did use 4 ohms in my calculation. sorry. i really don't know what the impedance is of the speaker. its a two way speaker with a horn and a woofer and its not for auto sound so i would have to assume its 8 ohms. i guess i could put my inductance meter on it and see?

 

[audioguru]

Measure its resistance. An 8 ohm speaker measures about 7 ohms resistance.

It makes loud hum or no output if it oscillates at a high frequency. It might happen if its input picks up radiation from its output wires, probably when there is nothing connected to its input. Try connecting a 100pf cap from its input to ground.

 

[Gaston]

i measures about 8 ohms. so what you are saying is that the output is picking up radiation and it is getting to the input through the feedback loop?

 

[audioguru]

No.
The output is a high voltage. The input is very sensitive and is a fairly high impedance so it picks up the signal ratiated from the output if the wires are near each other. This is positive feesdback that causes oscillation.

The amplifier's feedback is negative feedback to reduce gain and to reduce distortion.

 

[Nigel Goodwin]

I agree with AudioGuru, you need MUCH larger heatsinks if you want to run any kind of power, that one is really pitiful!.

 

[Gaston]

i think you hit the nail on the head.it never started humming until i moved the circuit board over to make room for the larger heatsink. now the output leads are directly under the input. the heatsink has a fan on it and i ran it really hard for aabout two hours and it never even got warm. it a cpu heatsink off of a pentium three and has a lot of fins

 

[dr.power]

The LM3886 is rated to deliver 68W RMS to a 4 ohm load when it has a positive 28VDC and a negative 28VDC supply. The total supply voltage is 56VDC and the sine-wave output is 16.5V RMS which is 46.6V peak-to-peak. So it has a loss of 56V- 46.6V= 9.4VDC.

It is rated to deliver 38W RMS to an 8 ohm load when it has a positive 28VDC and a negative 28VDC supply. Again the total supply voltage is 56VDC. The sine-wave output is 17.4V RMS which is 49.3V peak-to-peak.
So its loss is reduced to 56V-49.3V= 6.7VDC because the current is lower.

It gets very hot when its output is 68W. Any higher output would fry it even if it has a huge heatsink with a fan.

What is your "32V"? Is it 32V peak-to-peak across 4 ohms? Then the power is only 32W RMS.

Look in Google for The Gain Clone Amplifier. It is made with the LM3886 and LM3875 power amps. One guy makes and sells a stereo amp for $4000.00US!

Good point audioguru,

I just got a question regarding your above points. Can we increase the supply voltage so that we get 68W output while the load is 8 ohms?
The datasheet says that for +-35V supply the said IC gives 50W of power for a 8 ohms load. So how about the possibility of increasing the supply voltage so that we get 68W out of chip whiteout damaging it?

The other question regarding your above calculations for 4 & 8 ohms load at 56V supply is that what are the suitable maximum of p-p and RMS voltages for input of the said chip while its gain is 21 ( i.e the input voltages which do not cause the chip to go to saturation yet let the chip to work decently)? is the response 2.2 Vp-p and 0.78 Vrms for the 4 ohms case and 2.34 Vp-p and 0.82 Vrms for the 8 ohms case?

 

[audioguru]

Can we increase the supply voltage so that we get 68W output while the load is 8 ohms?
The datasheet says that for +-35V supply the said IC gives 50W of power for a 8 ohms load. So how about the possibility of increasing the supply voltage so that we get 68W out of chip whiteout damaging it?

The datasheet spec's a max operating supply of 84V and graphs show output clipping with a 6V loss. The distortion is rated when the output p-p voltage is 16V less than the total supply voltage. Then the power with an 80V supply is calculated at 68W.
The "spike protection" might be activated making it sound awful. It will also get very hot.

The other question regarding your above calculations for 4 & 8 ohms load at 56V supply is that what are the suitable maximum of p-p and RMS voltages for input of the said chip while its gain is 21 ( i.e the input voltages which do not cause the chip to go to saturation yet let the chip to work decently)? is the response 2.2 Vp-p and 0.78 Vrms for the 4 ohms case and 2.34 Vp-p and 0.82 Vrms for the 8 ohms case?

Gain has nothing to do with the max output voltage at clipping. Its distortion is spec'd when its output p-p voltage is 14V less than ther supply voltage.

With a gain of 21 then an input of 2.2V p-p produces an output of only 46.2V which is 33.4W into 8 ohms. A supply of 60V will do it at low distortion.

 

[dr.power]

The datasheet spec's a max operating supply of 84V and graphs show output clipping with a 6V loss. The distortion is rated when the output p-p voltage is 16V less than the total supply voltage. Then the power with an 80V supply is calculated at 68W.
The "spike protection" might be activated making it sound awful. It will also get very hot.

Thank you very much for your reply,
Actually I just wanted to know if we are on the safe side when using loads larger than 4 ohms yet let the chip's output power to be more than 50W to 68W?

Gain has nothing to do with the max output voltage at clipping. Its distortion is spec'd when its output p-p voltage is 14V less than ther supply voltage.

So how to calculate the maximum peak to peak input voltage for a 4 ohms load at 68W output power? I don't know what was my wrong?! Doesn't the 68W output for a 4 ohms load which indicates 46.6Vp-p across the said load give us a clue for the maximum p-p input voltage when the gain of the amplifier is known?
For instance if the gain of the amplifier happens to be 21 then the input voltage should not exceed 2.2Vp-p to avoid clipping? Or you are saying something which I do not understand?

With a gain of 21 then an input of 2.2V p-p produces an output of only 46.2V which is 33.4W into 8 ohms. A supply of 60V will do it at low distortion.

And of course it delivers a power of 66.7 for a load of 4 ohms.
So are you agreed with me that the 2.2V p-p is the maximum p-p voltage which we can give to the input of the said chip while the gain is 21 and the load is 4 ohms? (this reflects +-28V according to the datasheet).

 

[audioguru]

The datasheet shows a graph of clipping voltage. With a plus and minus 28V supply the output clips at 48.8V p-p which is 74.4W. But its distortion is listed at only 68W which is only 46.6V p-p.

Your 2.2V p-p input and a gain of 21 causes the output power to be 68W into 4 ohms.

 

[dr.power]

Your 2.2V p-p input and a gain of 21 causes the output power to be 68W into 4 ohms.

So is that what I am looking for? Sorry my knowledge is poor in this regard. I just want to calculate the maximum input voltage while the power supply is +-28V and the load is 4 Ohms.

 

[dr.power]

First Are you agreed with me that by a supply voltage of +-28V and a load impedance of 4 ohms, The output power of the LM3886 is 68W?
If so, Please consider that +-28V means 56V total supply,. the peak to peak voltage across the 4 ohms load at 68W output power as I calculated in my previous post is 46.64V (Vp-p = 2 x 1.414(sqrt P x R) = 46.64V).
It reflects that by having a 56V supply voltage the the P-P output voltage which you can expect is 46.64V, So the lose is:
56V - 46.64V = 9.36V.
So by 56V supply voltage you will not get any voltage larger than 46.64V in the output, So that output voltage is the reference voltage for calculating the clipping/ going to saturation state. for instance by a gain of 21 the MAXIMUM P-P input voltage must be 46.64/21 = 2.2. any higher p-p input voltage larger than 2.2V causes the output to go to saturation, Right? Unless you say That the chip is able to provide more power than 68W by a supply volatge of +-28V?!

 

[audioguru]

First Are you agreed with me that by a supply voltage of +-28V and a load impedance of 4 ohms, The output power of the LM3886 is 68W?
If so, Please consider that +-28V means 56V total supply,. the peak to peak voltage across the 4 ohms load at 68W output power as I calculated in my previous post is 46.64V (Vp-p = 2 x 1.414(sqrt P x R) = 46.64V).
It reflects that by having a 56V supply voltage the the P-P output voltage which you can expect is 46.64V, So the lose is:
56V - 46.64V = 9.36V.
So by 56V supply voltage you will not get any voltage larger than 46.64V in the output, So that output voltage is the reference voltage for calculating the clipping/ going to saturation state. for instance by a gain of 21 the MAXIMUM P-P input voltage must be 46.64/21 = 2.2. any higher p-p input voltage larger than 2.2V causes the output to go to saturation, Right? Unless you say That the chip is able to provide more power than 68W by a supply volatge of +-28V?!

The datasheet shows graphs of distortion. With a plus and minus 28V supply and a 4 ohm load the distortion is very low up to an output of typically 68W.
There is another graph showing clipping voltage vs supply voltage. The loss is a total of 7.2V which is not the same as the graph showing distortion (but it is close).

All LM3886 ICs are not identical. Some have an output in these conditions of 60W, not 68W. But you cannot hear the slight difference between 60W and 68W because double the power sounds only a little louder.

 

[Jackdaniel]

I had built a 5.1 amplifier using TDA2030A ic which has excellent sound clarity and enough power output to blast my house, i think u should also Try TDA2030A

 

[audioguru]

I had built a 5.1 amplifier using TDA2030A ic which has excellent sound clarity and enough power output to blast my house, i think u should also Try TDA2030A

The datasheet for the TDA2030A shows a circuit with an output of only 12W or 16W into 8 ohms and its distortion can be as high as a horrible 0.5%. It has enough power output to blast your house only if you have rice-paper walls. Most home stereos in North America have at least 100W per channel at very low distortion.