Inductors for output filter

[bernroth]

Hello!
I'm new to diyaudio.
I have build a two channel class-d amp using two TDF8590TH running at 375kHz.
The amp works fine but when idle the amp consumes around 10W and the output inductors are getting warm.
I am using two DO3340P-103 inductors together with an 1µF foil cap as output filter per channel.
I noticed that when I put the ferrite of another coil just above the output coil, the current goes down.
I'm not that experienced in the design of the output filter and would really appreciate to get some help to understand the issue and to lower the idle current of the amp.

At the end of this message I attach a picture of the amp, complete on a 160mm x 100mm euro PCB.
You will note the output coils are missing. I was trying some different ones, different core material, other HF suited wires and so on but no luck...

Any help and hints appreciated! Thank you and have a nice day!

Bernhard

[mag]

Hi Bernroth

you can use toroids of material -2 or material -6 from micrometals.
T90-2, T90-6 or similar sizes can be ok.

An alternative is to use gapped ferrite cores using possibly closed cores
to avoid EMI radiation. A gapped core like RM or P can be ok.

What is your rail voltage? You need to calculate the core losses based
on the rail voltage and the switching frequency. Core losses are the same
at no load or at full power.

The wire can be just a standard magnet wire of appropriate size. No need
to use litz or multifilar winding because the high frequency ripple current is
much lower than the output current.

What is the speaker impeadance you will use? 10uH seems a bit too low.

Do you have also a power supply on your PCB? It is a 12V input power supply or what else?

-marco

[bernroth]

HI Marco,
thanks for your reply.
Input voltage of the amp is 12V for usage in a car.
Output voltage of the smps is +-25V.
Impendance of the speakers will be 4 ohms and the values for C and L are recommended values out of the datasheet of the class d amplifier (which is mounted below the pcb for cooling on the aluminium block.
I have to admit that my knowledge of adequate LC filter design is quite limited.
I think the output coil gets saturated and adding more ferrite does increase the energy which can saved in the coil. I thought already about a core with an air gap but I really don't know the manufactures very well.
Where can I obtain the cores you mentioned?
Usually I do some tests and measure the inductance to get around 10µH and mount the L onto the PCB.

Best regards,

Bernhard

[Sch3mat1c]

What type of coil are your present inductors?

If they are saturating, adding ferrite will not help, because this does not change the flux in the core. It will increase the inductance while keeping saturation flux constant (i.e., saturation current drops). In fact, adding more permeable materials reduces the energy storage of the inductor.

Do you have any way to measure the current and voltage waveforms?

Tim

[bernroth]

Hi Tim,

the last inductors I tried and with which I have the best results for now are
Coilcraft DO3340P-103 (10µH)
The value is recommended for 4ohm loads in the datasheet of TDF8590TH are 10µH and 1µF

Concerning measurement I have a digital storage oscilloscope (HM407), digital RLC meter and a R&S HF millivolt meter available.

Best regards,

Bernhard

[Reactance]

Eva has done a great job !!! explaining inductor calculations, theroy ect.. this thread will help you even if your designs are not smps based.

Class - D in anycase are a controlled power converter in anycase understanding smps(s) will give anyone an edge in class-d designs and understanding

toroid for car amp smps