# Beginners question about using a switch with a mosfet

#### shortbus

Joined Sep 30, 2009
10,049
Reason for not interrupting ground. It is common for an amplifier to have a input that is connected to ground and a output connected to ground. When you cut open "power ground" the input and output are not happy. We try to connect all grounds together verry well. Switching power is normal.
I agree if it is mains powered but going to the link, it is DC powered from probably a wall wart.

#### MrChips

Joined Oct 2, 2009
31,087
If you were switching power to a light bulb or motor then a low-side switch would work.

Since this is an audio amplifier, all signals are usually referenced to COM and EARTH.
Hence you want a high-side switch. Keep COM at EARTH potential.

#### sagor

Joined Mar 10, 2019
927
What happened to Ohms' law?

The dissipation is the current squared times the MOSFET on-resistance, thus the MOSFET ON power would be 5² * 0.32 = 8W, which definitely requires a heat-sink.

If you use a MOSFET with <40mΩ on-resistance to keep the dissipation below 1W, than you shouldn't need a heat-sink.
Doohh... Was thinking voltage drop x current.

#### crutschow

Joined Mar 14, 2008
34,843
Doohh... Was thinking voltage drop x current.
Of course, that also works.
Vdrop = I x R, then P = I x R x I

#### BobTPH

Joined Jun 5, 2013
9,278
Yiipes. Let’s square that at current!

Bob

#### Audioguru again

Joined Oct 21, 2019
6,783
Since you are probably not buying hundreds of Mosfets to select a sensitive one, I calculated the maximum on resistance creating the maximum amount of heat.

Which amplifier part number draws a current as high as 5A from only 15V? A stereo TPA3116 produces 20W per channel into 4 ohms at low distortion and its heating will be about 4W. Then its total power of 44W draws a current of 44W/15V= 2.9A.

I corrected your schematic and discussed the poor Mosfet selection:

#### Attachments

• 63.4 KB Views: 11

#### daeiron

Joined Feb 28, 2022
11
Thanks for all the quick responses so far. I've made some adjustments and came out with these:

I think the high side design is basicly the same as @Audioguru again one. I have only added the R2 resistor as to have it act as a voltage divider, but I don't know if that makes sense. As I said I am a noob . Can someone explain to me why R2 might or might not be necessary or smart? And would the low side version make any practical difference?
Also the mosfets were definitely a bad choice, just dragged a random one in. I have now chosen mosfets which make a bit more sense I think.

#### SgtWookie

Joined Jul 17, 2007
22,230
Note that the P-ch MOSFET in the schematic is rated for 7A, but that is if the junction is held at 25°C. Current rating degrades significantly as the junction temperature increases, which it will certainly do with 5A current flowing through it, even with a heatsink. Better use a MOSFET rated for 3x or more your expected max drain current; that way you can use a smaller heatsink.

I tend towards overkill, because I don't want my circuits to break.

#### AnalogKid

Joined Aug 1, 2013
11,201
I have only added the R2 resistor as to have it act as a voltage divider, but I don't know if that makes sense. As I said I am a noob . Can someone explain to me why R2 might or might not be necessary or smart?
As above, because the voltage from the gate to the source never is greater than the datasheet limit, R2 is not necessary. Also, the value you have for it is not optimal. A better value would be 4.7K, as this makes the gate-source voltage close to 10 V.

For this application, FET selection is pretty straightforward. Four steps . . .

A common rule of thumb for long-term reliability is 2x. If you have a 15 v circuit, use components rated for a minimum of 30 V. If you are switching 5 A, use components rated for a minimum of 10 A. Etc. Those two parameters should reduce greatly the number of possible part numbers. After that, consider mounting. TO-3 package parts are very good for high power applications, but a pain to mount. Something in a TO-220 package should work for you. Now that the pool is much smaller, go for price.

ak

#### Audioguru again

Joined Oct 21, 2019
6,783
The IRF4905 Mosfet is a much better selection because its drops a max of only 0.1V at 5A and gets only warm without a heatsink.
Its max allowed Vgs is 20V then R2 should be 5.1k when R1 is 10k for a Vgs of 9.93V.

Use the P-channel Mosfet to switch the positive supply to the amplifier then the grounds can stay connected together.

Again I am asking, which amplifier part number uses as much as 5A from only 15V?

#### daeiron

Joined Feb 28, 2022
11
Again I am asking, which amplifier part number uses as much as 5A from only 15V?
I won't push it that far, but just to be safe. According to the TI website:
• 2 × 50 W Into a 4-Ω BTL Load at 21 V (TPA3116D2)
And according to this 2x 25W @ 15V 4A (4 Ohm)

#### Audioguru again

Joined Oct 21, 2019
6,783
The TPA3116 datasheet says 2 x 50W into 4 ohms with a 21V supply with horrible-sounding 10% clipping distortion. Their graph shows 57W for each channel.
But you have a 15V supply then the graph shows 20W per channel at low distortion which might actually be 50/57 x 20= 17.5W per channel at low distortion.

You never want the amplifier to produce clipping distortion.