1
\$\begingroup\$

I have some concerns in a motor control board I designed. Sometimes the MOSFETs unexpectedly start to burn. I suspect one of the reasons is the thin track (0.2mm) I have in my flexible board (0.2mm), but I think the transient short circuit is also a good explanation. I'm just using 4 transistors per motor (2xPMOS and 2xNMOS) like you can see in the picture

enter image description here

I read the information in this website but I still don’t understand how to choose the NMOS and PMOS \$V_{GS}\$ in a way that the transient short circuits stop to happen.

Could someone advise me on this subject?

\$\endgroup\$
  • \$\begingroup\$ is that a requirement? the control voltage must be = or higher than the supply rail voltage? \$\endgroup\$ – Electronics_Guy Jun 19 '14 at 11:08
  • \$\begingroup\$ The datasheet for your p-FETs will specify the threshold Gate voltage for ensuring 0V@Drain at whatever your Source voltage is. If you need negative vGS, logic-level MOSFETS exist, or you can use an additional n-FET to supply rail voltage to the Gate of your p-FET. \$\endgroup\$ – Emyr Jun 19 '14 at 11:15
4
\$\begingroup\$

Since you have six I/Os, as long as you don't have to run both motors at the same time, you could parallel the like gates of the two bridges and mux the grounds to the bridges with your two remaining I/Os, like this:

enter image description here

The transistors and CMOS inverters are what I have in my SPICE library, and they're conceptual only. You should select the parts which fit your project.

If you want to play with the circuit, the model is here, and if you don't already have it, LTspice is available free.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Aside from the pull-up resistor, what other provisions can be made to ensure sufficiently fast turn-off time? \$\endgroup\$ – sherrellbc Jun 19 '14 at 13:55
  • \$\begingroup\$ @sherrellbc: Basically, the MOSFET gate drivers should be stiff enough voltage sources to either charge or discharge the MOSFET gate capacitances quickly and the MOSFETs themselves should be chosen with the lowest practicable gate capacitance. \$\endgroup\$ – EM Fields Jun 19 '14 at 14:21
  • \$\begingroup\$ @Tut: I agree, in that if the MOSFETs are neither fully turned OFF or ON, they'll be dissipating power in the partially-conducting channel. \$\endgroup\$ – EM Fields Jun 19 '14 at 14:36
  • \$\begingroup\$ @EM Fields: I only have 6 I/O's available to control 2 motors (2 full bridges). that is why i only used the 4 pins instead of 8 to control 2 motors. \$\endgroup\$ – Electronics_Guy Jun 22 '14 at 18:13
  • \$\begingroup\$ @EM Fields: If i used a MUX to expand the number of pins to control the transistor gates individually would it work? \$\endgroup\$ – Electronics_Guy Jun 23 '14 at 9:22
3
\$\begingroup\$

You have to understand few important things in bridges like this.

  1. Mosfet has some turn-on and turn-off time - delay after gate signal is off.**

enter image description here

You have to add some delay between you turning off first pair of transistors and turning on another.

Sorry for bad quality drawing. This is supposed to show you what happends if you turn on 2 mosfet pairs are turned on.

By pairs - I mean

  • first pair: top-left and bottom-right
  • second pair: - bottom-left and top-right

enter image description here


2. Your bridge needs "logic-level" mosfet that can be fully opened with 3.3V (thats very low voltage). This is Vgs (gate-source voltage) transistor parameter. On your schematic - it looks like you have 3.3V microcontroller. If you so your gate voltages is 3.3V. It can be as low as 3.0V when LiPo battery is discharged and under load.

Transistors in applications like this should be fully open or fully closed to work efficient, otherwise they have significant resistance, and there are big power losses on that resistance. That power loss on resistance can damage transistor.

Modern mosfets, when they are fully opened have very low resistance (under 1 ohm), thats why they can be used for switching relatively high currents even without heatsinks.

Edit 2:

You should see this topic: How to determine maximum PWM frequency for transistor (2SK2554)

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks for the answer. The drawing is fine. I miself draw much times in Paint =). Regarding the delay that you talk about, i was hoping to create that delay by choosing properly the PMOS and NMOS transistors. I just don't know what to compare in the datasheet of the transistors to be able to choose them correctly. \$\endgroup\$ – Electronics_Guy Jun 19 '14 at 11:01
  • \$\begingroup\$ @user43113 I have added some basic information that could help you to choose transistors. \$\endgroup\$ – Kamil Jun 19 '14 at 16:42
1
\$\begingroup\$

What about this circuit using NOT's and crossed complementary PMOS/NMOS? Would you say something like this work?

The NOT will give a considerable delay to the NMOS i think.

enter image description here

That'll work, and you'll save 2 I/Os and you'll be able to run both motors at the same time :-), but you'll have to be careful; see the red stuff following...

Also, the inverter (NOT) will only add a slight delay through the conducting pair, but will do nothing to prevent shoot-through while the bridge is switching; that must be done by judicious software design.

Take a look at the MOSFET data sheets for their gate capacitances and their ON and OFF switching times for an idea of what must done in terms of delays.

Also, following are the edited schematics annotated with MOSFET reference designators and truth tables.

enter image description here

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.