Go Down

### Topic: mosfet: understanding tOn, td(ON), tr, td(OFF), tf, and tOFF (Read 749 times)previous topic - next topic

#### MartynC

##### Jul 13, 2018, 08:51 am

with reference to the n channel mosfet HUFA76429D3. Data Sheet here.

I am using the mosfet as a switch. Is tON the total time required to go from high resistance (switch off) to Rds(on) typical resistance (0.025Ω) (switch on)?

The switching time graph shows:

From this I see that tON is the sum of td(ON) and tr.

But if I look at the data:

It shows tON as 220ns, td(ON) as 13ns, and tr as 134ns which doesn't add up. The same applies to tOFF. What am I missing?

#### Southpark

#1
##### Jul 13, 2018, 09:18 amLast Edit: Jul 13, 2018, 09:26 am by Southpark
Just go to this link here ....

What am I missing?
Up the top of the column in the data sheet that you linked to (the full data sheet) ..... the labels "MIN", "TYP", and "MAX". Stands for min, typical, and max. So, if measurements were specifically done for a particular mosfet, then the values would be as defined..... eg. would add up. Where-as .... min, typ, max are statistics based.

#### MartynC

#2
##### Jul 13, 2018, 09:56 am
I had seen the linked document. This says the same "The  turn-on  time,  ton,  of  a  MOSFET  is  the  sum  of  the  turn-on  delay  time  td(on) and  the  rise  time  tr." which didn't help. It didn't explain the difference I was seeing in the datasheet. It also didn't help that tON is not always given in data sheets.

I now understand the min and max. I don't why would the data sheet uses absolutes for td(on) and tr but a max for tON? I have now seen other data sheets that have min and max for all values and this makes sense.

#### Southpark

#3
##### Jul 13, 2018, 10:08 am
I know what you mean there. Anyway, from the min column..... it could imply that some of these devices could possibly have a minimum turn-on time of 147 ns ..... 134+13..... if we're lucky to get one.

#### Wawa

#4
##### Jul 13, 2018, 10:25 am
What will you be switching with this mosfet,
and are you going to use a driver chip or a 20mA Arduino pin.
Leo..

#### MorganS

#5
##### Jul 13, 2018, 10:53 am
If you actually need those nanosevond turn-on times then you need a driver capable of forcing several amps of current into the gate for those nanoseconds.

Usually turn-on is much slower because you have much less current available to charge the gate capacitance. Plus the capacitance of the MOSFET is highly variable and cannot be modelled with a single ideal capacitor.
"The problem is in the code you didn't post."

#### MartynC

#6
##### Jul 13, 2018, 03:24 pm
What will you be switching with this mosfet,
and are you going to use a driver chip or a 20mA Arduino pin.
Leo..
Solenoid valves from an Arduino pin.

#### Wawa

#7
##### Jul 13, 2018, 11:01 pm
Solenoid valves from an Arduino pin.
Then don't worry about switching speed.
Leo..

#### Paul__B

#8
##### Jul 13, 2018, 11:42 pm
Solenoid valves from an Arduino pin.
OK, so you put a 10k resistor from the Arduino pin to ground to set a default OFF level whilst the Arduino boots, then you put a 220 Ohm resistor from the Arduino pin to the FET gate.

Done!

#### MartynC

#9
##### Jul 17, 2018, 04:44 am
If you actually need those nanosevond turn-on times then you need a driver capable of forcing several amps of current into the gate for those nanoseconds.

I don't need the ultra speed, just trying to understand more.

#### MartynC

#10
##### Jul 17, 2018, 04:47 amLast Edit: Jul 17, 2018, 04:47 am by MartynC
OK, so you put a 10k resistor from the Arduino pin to ground to set a default OFF level whilst the Arduino boots, then you put a 220 Ohm resistor from the Arduino pin to the FET gate.

Done!

I have a 100 Ohm to the gate. Any reason why a 220 would be required?

This is something that has a lot of conflicting information about online. Some say there is no need for a resister, other say you should have one. Some say 50 Ohm is OK, others like larger.

#### MorganS

#11
##### Jul 17, 2018, 05:06 am
I have a 100 Ohm to the gate. Any reason why a 220 would be required?

This is something that has a lot of conflicting information about online. Some say there is no need for a resister, other say you should have one. Some say 50 Ohm is OK, others like larger.
It limits the peak current that the gate capacitance draws from the Arduino pin. Other processors (eg the SAM3X on the Due) have different current limits.
"The problem is in the code you didn't post."

#### Wawa

#12
##### Jul 17, 2018, 05:25 am
The gate of a fet is like a capacitor.
It needs current and time to charge from 0-5volt or discharge from 5-0volt.
More current is less time to change from one state to the other, so faster switching.

A 220ohm resistor limits the initial charge current to 5/220= 22.7mA, and 100ohm to 5/100= 50mA.
Without resistor we rely on the internal 'on' resistance of the Arduino chip.
>=100mA might flow, which is above the absolute max limit of the chip (40mA).

But chips are tough, and that very short overload might not harm it (or it might in time).

Size also matters. Bigger mosfet = more gate charge = more current needed to switch (or longer to switch).
Leo..

#### MartynC

#13
##### Jul 17, 2018, 07:35 am
Thanks.  220 it is then.

#### MarkT

#14
##### Jul 17, 2018, 12:25 pmLast Edit: Jul 17, 2018, 12:37 pm by MarkT
BTW when the datasheet says sum of turn on delay and rise time, it means current rise time, not voltage
rise time, the drain voltage is actually falling!  The graph should really be showing drain current, not Vds,
in the output waveform to be consistent with their naming scheme.

The quality of datasheets can vary significantly - read lots of them, that's how most people learn
the conventions commonly used in them.

Oh, the other thing, try to find a datasheet with oscillograms of real MOSFET switching waveforms, they
are much messier than the idealized diagram, for instance the Vgs has a voltage plateau as the
channel forms and collapses, and the Vds will tend to ring due to stray inductances in the package
pins.  But for your application the thing you need to worry about is inductive kick back - make sure you
have a freewheel diode across the solenoid (otherwise you'll have to analyze the avalanche performance of the MOSFET compared to the solenoid inductance, which is a whole other story)
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

Go Up