Buzz noise from DC brushless fan in TDA2822m

[enigmaticEntity.]

Hi all, I'm trying to implement a tiny cooling fan in my TDA2822m project (yes, it's only 1W), but I just want to make sure it doesn't overheat if driven hard, plus it will look more geeky...

The problem is that it creates a buzz (square wave or spike) directly related to rpm. Only in the right channel, and gets a little louder as the fan gets closer to the chip. Still there to be annoying within the circuit box.

Would resistors either side of the fan reduce the noise (it has to have resistors any way to slow it down)? I can't try anything until next weekend, so I thought I may as well ask here.

What about an RC in parallel? Though getting that exact frequency will be difficult enough.

Any other possible solutions?

Thank you in advance for any help,

E.E

[pacificblue]

Does the fan already have a power supply of its own?
RC in parallel is more promising than resistors in series.

[sasmit]

Originally Posted by enigmaticEntity.

Hi all, I'm trying to implement a tiny cooling fan in my TDA2822m project (yes, it's only 1W), but I just want to make sure it doesn't overheat if driven hard, plus it will look more geeky...

The problem is that it creates a buzz (square wave or spike) directly related to rpm. Only in the right channel, and gets a little louder as the fan gets closer to the chip. Still there to be annoying within the circuit box.

Would resistors either side of the fan reduce the noise (it has to have resistors any way to slow it down)? I can't try anything until next weekend, so I thought I may as well ask here.

What about an RC in parallel? Though getting that exact frequency will be difficult enough.

Any other possible solutions?

Thank you in advance for any help,

E.E

Is the fans ground connected to your signal ground ?

The fan should ideally be connected to power ground i.e. assuming you're using separate grounds.

There are two ways the fan creates noise one will be through the power supply, the other is through EMI generated from the switching components inside the fan , this is most probably your problem since you're saying it increases if you take the fan near the amp.
For this I think you need to isolate your circuitry using a ground plane or a copper shield etc.

[pacificblue]

Switching components inside a fan?

[sasmit]

Originally Posted by pacificblue

Switching components inside a fan?

yes ! The brushless dc motor has a hall effect sensor and a pwm + switching components inside. When you look at it, it doesn't seem like much until I read an elektor article about brushless fans

[stratus46]

Originally Posted by sasmit

Is the fans ground connected to your signal ground ?

The fan should ideally be connected to power ground i.e. assuming you're using separate grounds.

There are two ways the fan creates noise one will be through the power supply, the other is through EMI generated from the switching components inside the fan , this is most probably your problem since you're saying it increases if you take the fan near the amp.
For this I think you need to isolate your circuitry using a ground plane or a copper shield etc.

You're right about the grounds. An RC filter on the fan like 2.7 ohms in series and 220 uF in parallel would help also. He shouldn't need ground planes, just a proper power return so it doesn't get into the signal path. If the input loading of he amp is low (25K, or my preference <10K) EMI should not be the issue. It's the ground 'bouncing' from the power surges of the fan at commutation time.

G²

[pacificblue]

Originally Posted by sasmit

yes ! The brushless dc motor has a hall effect sensor and a pwm + switching components inside. When you look at it, it doesn't seem like much until I read an elektor article about brushless fans

The hall sensor is used as feedback signal for PCs so that you get an alarm when the fan is stuck. I don't think it is connected in this application.
The PWM current limiter that may or may not be built into the fan is usually only active during start-up to limit the inrush current. That should also not be an issue.
Brushless motors do however produce significant ripple of their own, that is why they should have a separate power supply.

[johnr66]

I'm trying to understand the need for a fan in the first place. Are you overdriving the chip? epoxy a U shaped heat sink to the top would be a better passive solution. A fan probably won't do much for the tiny surface area without an additional sink.

[DigitalJunkie]

I usually use a RC on fans to filter out the noise. Depending on how slow you want the fan to run 10-100ohms,followed up with a 100-220uf cap usually does it for me.

[enigmaticEntity.]

Sorry I haven't replied for a while The noise is from EMI I think, as I attached the fan via a longer line and the noise was gone when about 15cm away. It also disappears when the fan is orientated so air is blowing at a tangent to the amp (Not very useful!)

The enclosure as it stands is already overcrowded, and the fan can only go above the amp at a distance of about 3-4cm. I'm guessing I need some sort of EMI RF shield that can let air flow through it easily?

There is a heatsink on the chip, scavanged from a CRT circuit board, about 3*1*1 cm dimensions, black Alu. It is tied on with string, glued with superglue and has thermal compund in between.