Solving noise problem with shielded cable, or capacitor?

[cupoftea]

Hi,
I once got called into production to fix a 48V, 20W (or so) BLDC fan (on a production unit) that was stopping and starting. It was also strangely , kind of randomly, speeding up and slowing down. I think it was a Sunon fan but can't completely remember. It was at the end of a 2 metre cable that came out of an electrically noisy unit that was full of switching converters. The fan had a smooth 48V at its terminals , but was still malfunctioning as said. The problem was solved by adding a 100nF film capacitor at the fan's input terminals.
Could the problem also have been solved without the capacitor at the fan's input terminals, if shielded cable had been used instead?

 

[stenzer]

Hi,

there are different approaches to control the speed of a BLDC fan. The speed is proportional to the applied voltage, or its average.

As you mentioned the fan is supplied by a switching power supply. So its output voltage might have dropped below the operation voltage of the DC fan. Further, the inductance of the 2 meter cable acts as a high impedance for fast transients, which prevents a steady current flow towards your fan. The capacitor attached to the inputs of your fan acts as energy storage, which provides energy/current in case the supply voltage drops.

The fan had a smooth 48V at its terminals...

How have you determined the applied voltage, with a multimeter or an oscilloscope? A multimeter does not provide information of "high" frequent switching signals.

Does the SMPS also provides energy to an other load, which is wired in parallel to the BLDC motor?

BR

 

[FvM]

I would consider the option that the fan had an unconnected speed control input which was triggered by noise. Seriously, it's a kind of difficult to analyse hear-say information like "I once saw a yeti in Himalaya".

 

[cupoftea]

Thanks, yes i did view the fans vin on a scope, and it was smooth DC.
Its just that we have a project where we have fans and pumps at the end of cables......and we dont really want to haver to mount each fan on a PCB so that we can mount capacitors at the input terminals......if we can avert possible noise issues by just using shielded cable then we would rather do that.
So we're thinking of problems we might get, (noise) and how to reduce the chance of them happening, with as little hassle as possible.

 

[KlausST]

Hi,

The cause if the noise is still unclear - at least to me.

A shielded cable can just reduce the noise that could be coupled into the wires ... but caused by an external noise source.
And it more reduces common mode noise than differential mode noise.

I don't expect that common mode noise is likely to cause fan malfunction.

Now you talk abput pumps. Are they supplied by the same wires, or different wires in the same cable or are threr extra cables.

Is it DC, PWMed or AC?

You talk about smooth DC. Please confirm that the voltage at the fan input is never outside the specified range. Not even for a microsecond.

****
Indeed you don't give much information. All we can do is guessing.
A schematic, photos, hand drawn sketches would help a lot.
Mind: You don't have to provide the informations for us, you do this for yourself...to get rid of the problems.

Klaus

 

[c_mitra]

Indeed you don't give much information. All we can do is guessing.
A schematic, photos, hand drawn sketches would help a lot.
Mind: You don't have to provide the informations for us, you do this for yourself...to get rid of the problems.

Klaus

It is so frustrating, indeed!

It does not matter whether the fan had a smooth 48V at the terminal.

First thing that need to be confirmed that it is indeed the noise that is responsible for the erratic behavior.

In general, noise is nuisance only at high impedance signal inputs. Some noise at the motor terminal may not matter much.

It is possible that pump and fans on the same power supply is cause of the mess. They need to be decoupled at the device end.

As you rightly say, we cannot do more than guesswork.

--- Updated Nov 1, 2021 ---

Thanks, yes i did view the fans vin on a scope, and it was smooth DC.

I though BLDC motors need software (or hardware) generated three phase AC. Under normal conditions, the frequency of this AC will determine the speed of rotation.

It is very strange that a simple capacitor across the Vin solves the problem. Your original design should have liberal numbers of decoupling capacitors wisely scattered all over.

Most likely diagnosis: some noise taking a free ride from the supply (or even just from the air) and then latching on some high impedance dangling input somewhere in the circuit. But how about the pump?