Lm3886 Bridged Problem

[kanifee]

Hoping someone can help, today i made up an lm3886 bridged amp, well atleast one channel for now on some matrix board, ive constructed the psu and tested voltages rigged up the amp and its powering on fine no shorts or anything nasty. i hooked up an old speaker to it with my pc as the source and it has a deep low noise hum, the music is also very very very low, its so low you need to move your head to the speaker to ensure its playing. have you guys got any idea?

[zei]

Hi

What schematic did you use?
Can you provide some pictures.

Thanx
/JZ

[kanifee]

It was nationals schimatic, the only change i had to make was for the 51.1k resistor i had to use 3 in series to atain the value i needed this was the same for the 51.1k ohm.

Like i say there is faint music so its sort of working lol!
this is board, i doubt youll be able to make anything out i struggle and i made it! everything is wired correctly, ive checked double checked and triple checked

[sangram]

How are your grounds routed?

In a bridged amplifier the load is floating, but each amplifier needs to 'talk' to the other's ground to minimise errors. It sounds to me like the two grounds are not connected correctly.

When I run amps bridged, I use one signal ground, one power ground and tie up both amps to the return points. If you have a star ground for each amp and connect the two together, noise will appear on the inputs as a function of power supply loading. Theoretically this should cancel out as the load currents in each amp are the opposite of the other. Theoretically.

Most readymade PCBs are optimised for stereo, but since you have a vero you can configure it freely. I do have a bit of this problem on my readymade PCB (Peter's 4780 kit) where the source is a balanced amp. The grounding is a nightmare to reroute, but I got it to work with minimal hum.

Edit: I usually check both halves of the amp when running them in bridged mode. Connecting a speaker between the output and power supply ground will reveal the status of the amplifier. And other mistakes are usually grounding errors.

[jackinnj]

do you have the mute properly working -- that's the first thing to check.

check the value of the resistors you "series-ed" to come up with
51K -- the bridged amps really need the resistor values matched such that the gains of both halves are equal. There is a slight mismatch with the values shown. (I use 46.4k/4.75k and 51.1k/4.75k)

you also have to make sure that the heat sink is properly sized for bridged amplification -- you can use the calculator on National's website. if the heat sink is too small, load too low or supply voltage too high the thermal protection will kick in.

while you should be able to get the amp to work with Vero-Board the reality is that you have a comparatively high gain power circuit which is very susceptible to oscillation, particularly with leads dangling all over the place. Peter Daniel does a beautiful job of point-to-point wiring with some of his amps and I would look to his website for how it is done. With long, dangling leads the chipamp will pick up RFI, EMI unless you bypass the power pins as suggested by Nat Semi.

Just as a matter of precaution I would suggest that on each output leg you place a zobel consisting of 2.7R in series with 100n.

[kanifee]

this is fromt he top side, i done as you asked and used the ground on the psu as the referance for each chips output, i get audio at a very very low level with no hum from the right (circled in black) but i get nothing but hum from the left (circled in red) the red lines are the earth track i have made from the input.


I also think i should mention im using pin 1 and 5 of the lm3886 for v+ input, wasnt sure if it was just one or the other i should be using so i used both!

[pacificblue]

The right is the inverting or the non-inverting IC?

And yes, you should connect 1 and 5. It doesn't look as if you did in your photos though.

[kanifee]

power lead isnt connected int hat photo but it is at the moment, right is the inverted, have gone back and looked over the data sheet, i have left the muts pin open so mute is on yes? how do i figure on the resistor i need to open the mute pin and what do i connect it to? is it v-?

[jackinnj]

Originally posted by kanifee
the muts pin open so mute is on yes?

The mute pin is to be connected to Vee via a resistor such that current flows FROM the pin to the neg supply. I have used 8.2K + 10k, 22K etc.
the mute takes 0.5 mA to turn the device on.

[kanifee]

terminology has lost me a little, vee? is that signal in or voltage+ in?