Mosfet gate discharge resistor problem

[Yasinji]

I am driving two mosfet working in push pull mode(Electronic CFL ballast) directly from microchip microcontroller I/O output with 25mA source current. I have used a gate resistor of 22ohm and gate discharge resistor of 1K. The mosfet I have used is IRFZ44N the datasheet attached.

Qg Total Gate Charge ––– ––– 63 nc

Qgs Gate-to-Source Charge ––– ––– 14 nC

Qgd Gate-to-Drain ("Miller") Charge ––– ––– 23 nc

td(on) Turn-On Delay Time ––– 12ns

tr Rise Time ––– 60 ns

td(off) Turn-Off Delay Time ––– 44ns

tf Fall Time ––– 45 ns.



The problem is the CFL is not turning on with 1K gate discharge resistor.

When I increase the gate discharge resistor to 10K the CFL is working ok.

The switching frequency is 38Khz. Can please anyone tell me why this is happening.http://www.irf.com/product-info/datasheets/data/irfz44n.pdf

 

[aProgrammer]

The IRFZ44N is not specified to work
with a gate drive of 5V, it need 10V.

You need to use a Logic Level Mosfet specified at VGS=5V drive.
examples : IRL540, IRL3803, IRLZ34...

 

[alexan_e]

The IRFZ44N is not specified to work
with a gate drive of 5V, it need 10V.

You need to use a Logic Level Mosfet specified at VGS=5V drive.
examples : IRL540, IRL3803, IRLZ34...

The datasheet shows that with a vgs voltage of 5v you have a drain/source voltage drop of 0.4V for output current of 10A,
Yes higher gate voltage is better but he can still drive the mosfet with 5v.

I am driving two mosfet working in push pull mode(Electronic CFL ballast) directly from microchip microcontroller I/O output with 25mA source current. I have used a gate resistor of 22ohm and gate discharge resistor of 1K. The mosfet I have used is IRFZ44N the datasheet attached.

You have connected a 22ohm resistor between the gate of the mosfet and the pic port pin, this is a very low resistor and you will definitely have a voltage drop in the pic output, the output voltage drops below 5v when you pull current and your resistor is limiting the gate current to more than 200mA, and this is far more than the 25ma that pic can give.
How did you calculate the 25mA current, and how is the discharge resistor connected, is it gate to source?
I think you need a driver stage to drive the mosfets properly especially with a frequency of 38Khz.
When you say it is working with 10K resistor, is it actually switching?

Alex

 

[Yasinji]

The 22ohm is connected betwwen pic port pin and mosfet gate and 1K resistor is connected mosfet gate and source. The 25mA current details is given in datasheet of pic as given below.
12.0 ELECTRICAL CHARACTERISTICS
Absolute Maximum Ratings(†)
Max. output current sunk by any I/O pin ..............................................................................................................25 mA
Max. output current sourced by any I/O pin .........................................................................................................25 mA
Max. output current sourced by I/O port ..............................................................................................................75 mA
Max. output current sunk by I/O port ...................................................................................................................75 mA

---------- Post added at 06:23 ---------- Previous post was at 06:16 ----------

Also I have checked the waveform using CRO with both 1k and 10K between gate and source.
In both cases the waveform is ok and frequency is at 38Khz.

 

[alexan_e]

You have a pic mc that says that the max output current you should sink/source is 25ma for the pin (or 75 for the port) but the limiting resistor you are using will actually let more current flow, you can't let the pic circuit give the max current (which i assume is more than the specified limit), you should use a higher current limiting resistor to limit the maximum current to the gate in lower levels, but you will still have a low voltage while you give the current.

The output voltage with a vcc of 5v will be much lower than 5v even when you sink/source the specified max of 25ma, take a look at the attached picture, it is from PIC16F87x, for an output current of 25ma your output voltage will be lower than 3v and this is not high enough to drive the gate correctly.
This voltage will rise to 5v as the gate capacitance charges and the current lowers but i still think you need a driver to drive the gate with more current and a voltage that doesn't drop when you source current so that the gate charges faster.
It seems strange to me that you can get a clear waveform in the output when you drive the mosfet this way, if the waves are ok in both ways (1K and 10K) then the circuit should work too.

I have never used any CFL circuit but i'm talking about the mosfet driving in general, i assume the same principles apply to this circuit too.



Alex

 

[FvM]

It seems strange to me that you can get a clear waveform in the output when you drive the mosfet this way, if the waves are ok in both ways (1K and 10K) then the circuit should work too.

I guess, the waveform is clear without an output stage supply voltage. But it's normal operation, I think. Connecting the 1 K resistor reduces the gate voltage to about 4.5 V, which is apparently too low to turn the MOSFETs on. Unloaded 5V is just over the edge, still operating by chance. I can't imagine how the circuit should work reliably this way.

As already mentioned, the µP's maximum ratings will be considerably exceeded by the dynamic gate currents. This operation apparently doesn't damage the device at once, but it's far from good engiineering practice. A reduced lifetime can be expected at least.