Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

[SOLVED] need help with mosfet in H-bridge chopper

Status
Not open for further replies.

alexan_e

Administrator
Joined
Mar 16, 2008
Messages
11,888
Helped
2,021
Reputation
4,158
Reaction score
2,031
Trophy points
1,393
Location
Greece
Activity points
64,371
Hi

I'm using the attached schematic in a design I made which consists of 2 H bridges to drive a bipolar motor (2ohm , 1Amp).
I'm using a 50KHz switching frequency with a chopper to drive the motor with constant current.
The schematic is actually 4 times the circuit I'm showing here and I need it to be able to switch the output fast.
I know that there is a gate capacitance that I have to charge so that the mosfet becomes active so I have used a driver with a 50ohm gate resistor which shows a low on delay.
My problem is switching off thew mosfet, I tried many things and I saw that to switch the mosfet off quickly I have to discharge the mosfet gate capacitance quickly and the only way I was able to do that was by using a 100 ohm pull up resistor to the gate (graph1).
When I tried with a 1K resistor the the off delay is so big that the square wave is very distorted (graph2) , and if I change to 10k I don't get a square wave at all (in the output)
IS there a better way to switch fast a mosfet because this 100 ohm resistor also pulls current that delays the on time.

In the graphs green trace (left axis) is input voltage to the base of the transistor
and the red trace (right axis) is the voltage at the drain of the mosfet (output)

Could you please tell me if there is a better way to drive the mosfets using discrete components?
Can i use some other way to discharge the gate fast?

Alex
 

Attachments

  • mosfet.jpg
    mosfet.jpg
    103.9 KB · Views: 216
  • mosfet_graph1.jpg
    mosfet_graph1.jpg
    138 KB · Views: 163
  • mosfet_graph2.jpg
    mosfet_graph2.jpg
    134.2 KB · Views: 174

Re: medd help with mosfet in H-bridge chopper

I consider your question only regarding the high-side P fet, the low side fet seems to be off all the time.

For doing this with discrete components, you need a level shifter, and possibly a limitation for the gate voltage. Both can be done with following circuit:



Don't look too much on component values.

Q1/R1 form a constant current sink only when Q4 is turned on by your digitial command ('0' = on, '1' = off), this current creates a voltage on R2 with respect to the source of the P-fet. THis voltage must be kept within the limits of the maximum Vgs for the fet, and normally 10 to 15 volt is used. Q2/Q3 is an emitter follower to provide enough current to fast charge and discharge the gate. With high input voltages, be careful with the power dissipation of Q2.

Hope this helps.

Stefaan
 
Re: medd help with mosfet in H-bridge chopper

Thank you for your reply

I have also seen a similar push-pull driver in another forum, i assume the resistor/zener is to limit the gate voltage as in your circuit.
My power supply is 17 volts so maybe a zener of about 6v would be useful.

How does this driver compare to your design, is it worse in some way?
I have added the gate resistor for testing, it wasn't there originally.

I also have removed the N mosfet from the schematic, my main problem is the p mosfet.

Alex
 

Attachments

  • push-pull.jpg
    push-pull.jpg
    139.6 KB · Views: 170
Last edited:

Re: medd help with mosfet in H-bridge chopper

This does not look correct to me. The gate voltage must be maximum -20V with respect to the source (power supply in your case) for a driving a P-FET. The zener diode limits the voltage to the bases of the emittor follower referenced to ground.

Also for switching off there can be a probkem, because when the input voltage is higher than the zener voltage, the mosfet will stay on, because the voltage on the gate will be less than the source (the vgs in this case will be -(VCC-15)).

If you want reliable high side switching, you need to add the level shifter like I proposed. The purpose of the level shifter is to mirror the gate drive signal from ground reference to the positive supply rail. If you create this circuit in for example the free LTSpice simulator, all will become clear.

Stefaan
 
Re: medd help with mosfet in H-bridge chopper

You are right, i have simulated the circuit and i have removed the zener.

The simulation result of schematic 3 (graph3) seems to be giving faster switching time but there is no level shifting, this is not a problem with my circuit because i have a 15-17v power supply and the mosfet I'm using has a max Vgs 20v (IRFR9020A).

The circuit you propose (schematic4 graph4) has the benefit of level shifting but the switching seems to be slower, if this because of the values i have used of is it the typical performance of that circuit.

traces :
green color left axis input voltage from control device (transistor base),
red trace right axis mosfet gate

Thank you for your help
Alex
 

Attachments

  • schematic3.jpg
    schematic3.jpg
    64.1 KB · Views: 159
  • graph3.jpg
    graph3.jpg
    134.4 KB · Views: 148
  • schematic4.jpg
    schematic4.jpg
    77.7 KB · Views: 164
  • graph4.jpg
    graph4.jpg
    137.9 KB · Views: 161

Re: medd help with mosfet in H-bridge chopper

Since you are already close to the maximum Vgs and maybe you will try later on with a somewhat higher power supply voltage, the level shifter can be interesting anyway.

Also I don't see the point of using the gate resistor if you want to compare the switching speed. It will switch faster without gate resistor. The current is normally limited anyway by the Hfe of the bipolar driver stage. I personally use the same circuit with the level shifter and no gate resistor.

The speed difference comes also from the different base current you supply to the emittor follower. In the case without level shifter you have roughly 17V/1k = ~17mA max. In the case with the level shifter, you have only (3.3 - 0.7)/560 = ~4.5mA.

Have a look on the current in several componenets in the simulator to understand better... (especially the gate current of the mosfet).

Also consider that the losses in the mosfet happen at the 'platau' in the gate voltage, thats when neither the current or voltage on the mosfet is approx zero. Also consider the reverse recovery of the diode in the low side mosfet when you complete the leg for the bridge. To have an idea of that, you need to simulate with an inductive load.

Stefaan
 
Re: medd help with mosfet in H-bridge chopper

Thank you

I have just one more question, I'm currently driving the N mosfet of my bridge directly using a resistor.
If i were to use a driver circuit what kind of driver would you suggest?
I'm using but i'm asking in general for a N mosfets.

Alex
 

Re: medd help with mosfet in H-bridge chopper

If you want the same performance ballpark for the driver, you should go to a push-pull driver on the low-side too. Again there are a lot of mosfet driver chips available, but if you want also a dirscrete solution, you will end up somewhere with an emittor follower again. Also I don't think the AP4800agm can be used with 3.3V gate drive, and you need to amplify the logic signal to 10 or 15V pulses.

You can always post some schematics for a small review.

Stefaan
 

Re: medd help with mosfet in H-bridge chopper

There are two reasons for the slow falling edge of the gate voltage in the cascode driver circuit, and the somewhat faster, but still slow rising edge. One reason has been told, low base current. The other, not yet adressed, is the rather poor current gain of 2N2222/2907 in the current range of interest. I suggest to use transistors for the complementary follower, that have high current gain (e.g. > 200) up to 1 or 2 A. You can get good ones in a small packages from Zetex or NXP, or use conventional medium power transistors.
 
Re: medd help with mosfet in H-bridge chopper

My design looks like the schematic, but i have 2 H bridge and op-amp current sensing circuits, this is half the circuit except for the cpld which is one.

The controller is a cpld (XC9536) which has a direction and step input from the parallel port , also has two inputs from the opamps that set the input to logic 1 when the motor current goes over the current limit that is preset with the pot , it also drives the 8 mosfets.

I have a frequency of 50KHz for the chopper, between the steps all the mosfets are turned off for 20us, also the duty cycle of the 50KHz is 90%

My design works with a 5v cpld but maybe it will change to a 3.3v in the future.

If i am reading the datasheet graph correctly AP4800agm can give 10amp with a gate voltage of 3v and i only need 1 or 2 amps.

This circuit works fine in every aspect except that the mosfets become warm , you can touch them but they are pretty warm and i don't want to use any heat-sink so I'm trying to make the switching faster to reduce the thermal losses.
The cpld gets warm too depending on the driver changes and the driving needs i might use a buffer like 74LVC244 that can give 50ma.
The motor is 2ohm/1.1A

I know there is an easier way to do what i want with an ic but i like to experiment and i also have the components.

Alex
 

Attachments

  • Hbridge.jpg
    Hbridge.jpg
    217.5 KB · Views: 169
Last edited:

Re: medd help with mosfet in H-bridge chopper

If i am reading the datasheet graph correctly AP4800agm can give 10amp with a gate voltage of 3v and i only need 1 or 2 amps.

In the datasheet I opened, I see with a Vgs of 3V and 10 amps of drain current, there is 0.6V between drain and source (at 25°C). This makes 6Watt, and a very hot FET.

You can drive some more current with using some 74xxx244 buffers in parallel, but even then at least 5V (4) should be used to drive the gate. See also the specification for an Rdson of 40mOhm at Vgs=4V. If the fet was good enough to be driven with 3.3V, I'm sure they will not forget to spec it at this voltage.

Like FvM said, using the bipolar driver stages, the gate current depends on the type of transistors, but anyway, it should be much better than with the resistors.

Another remark, can you reduce the chopping frequency to for example 25kHz? This will halve the switching loss already. I think stepper motors have a fairly high inductance?


Stefaan
 
Re: medd help with mosfet in H-bridge chopper

Oops, i didn't know that the x axis was the voltage drop, but on the other hand i am only using 1-2Amps (and also 5v with the current cpld) so this should be about 0.3W for 3.3v and much better for 5v
.
I have tested 25KHz in simulation before but as i remember the current regulation wasn't as good but i will try it again.
I don't know the exact inductance i bought the motors used from ebay, but i have seen a similar datasheet and i am simulating with 5mH.

Sanyo Denki 103H5208-0842 or -1241
NEMA 17
1.8 deg 200 steps/rev when full-stepping
2.0 ohm motors
Current per Phase: 1.2A unipolar 1A bipolar

Thank you for your suggestions, i keep learning
Alex
 

Re: medd help with mosfet in H-bridge chopper

You should also consider, that the actual Vgs charateristic shows a type variation. AP4800agm is definitely not intended to be driven by 3 or 3.3V Vgs. 5V can be O.K.if you don't need lowest rds,on values. You can get however logic-level FETs, that are good for 3.3V Vgs.
 
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top