The circuit below will be used to driver the directional lights on a motorcycle. The total current for each side will be 4 to 4.25 amps. V2 and V4 will actually be 3.3V digital outputs from a microprocessor.

I have 2 questions:
1. Will these 27 amp mosfets require heatsinks?
2. LTSpice shows the outputs working correctly with a wide variety of resistors. It even works without r5 and r2. The only difference I see is the currents vary with resistance. Any tips on how to go about selecting these values?

 

Resistor values aren't critical.

I'd get rid of R5 and R2 as they aren't necessary.

I'd put R9 and R10 on the bases of Q2 and Q1, respectively, because they're just to ensure that the NPN transistors are off with no voltage applied to R8 or R7. I'd omit them for some other applications. You could omit them if the microprocessor is always driving the base resistors during normal operation.

Since you only need 0.12mA of base drive, R8 and R7 could be larger (up to 20-22k).

Since MOSFET on resistance is 70mΩ max, you don't need to worry about power dissipation (less than half a watt).

EDIT: attach datasheet

 

Hi PB,
Please post your LTSpice asc file.
E

 

Okay ericgibbs, here is the asc file.

dl324, I just hooked up a test circuit with only a single 1157 bulb and the mosfet got really hot really fast (about 30 seconds). I must have something figured wrong.

 

I just hooked up a test circuit with only a single 1157 bulb and the mosfet got really hot really fast (about 30 seconds). I must have something figured wrong.

What was the gate voltage? The 1157 spec I saw said it drew 2.1A. If the Vgs on the MOSFET was 12V, it should have been dissipating 0.15W. Case temperature should have been 10C above ambient.

 

Hi PB,
Got your asc file, post a copy of your FQP27P06.lib
E

 

Made some changes reflected in the attached .asc file. It doesn't get hot as fast, but after about 6 or 7 minutes the mosfet burned the contacts in my breadboard.

Just as a test, I connected a new mosfet directly to the light and power and tied the gate to ground. Same result, it gets hot after a couple minutes. Got a little blister on my test finger. Don't know what I am doing wrong.

attached is the FQP27P06.lib file I use.

 

Perhaps this is the problem???

Brand: YEGAFE

20PCS FQP27P06 27P06 P-Channel Power MOSFET 27A 60V TO-220 Transistor: Amazon.com: Industrial & Scientific

 

Hi PB,
On LTSpice the power dissipation is only approx 0.6Watt.

Do you have the PMOS wired up correctly, perhaps a photo of the PMOS area would help.
E

Update:
It is highly possible that the MOSFETs are fakes????


Also: Recheck the pin out connections.


E

 

I'll wire up a new one like this and it will have real wires, not a breadboard.

 

Hi PB,
Let us know how it goes, good luck.
Did you confirm the MOS pinout?
E

 

I had a wiring error so it's working now. I showed a 10k resistor from the gate to ground in the schematic but in fact it was going from the gate to the drain. So, at the start the mosfet was fully on until the drain came up near the source and then it was only partially on, hence the overheating.

After fixing the wiring, there is .107V across the mosfet and it's drawing 2.15 amps so it is dissipating .230mw. It will be double that on the motorcycle, but the directionals will be on maybe 1% of the time and when on it will be a 50% duty cycle.

The answers to my questions are:
1. No heatsink is needed for the mosfets.
2. 2N3904 collector to Mosfet gate resistors are unnecessary, and the others really have a wide range of value that work just fine.


Thanks all for your patience, recommendations, and help.

 

Hi PB,
Many thanks for the positive feedback on your project, happy biking.
E

 

The answers to my questions are:
1. No heatsink is needed for the mosfets.
2. 2N3904 collector to Mosfet gate resistors are unnecessary, and the others really have a wide range of value that work just fine.

If you have trouble with MOSFET heating, you could make R1 and R4 smaller to shorten turn off time. That would require corresponding reductions in the values for R7 and R10 to support the higher collector current.

In the future, you can make breadboarding changes to be able to handle currents above 1A. Connect to the leads of the components that have high current and route it through wires vs the breadboard contacts.