H-Bridge - IRF530 Heating Up

[brunobasto]

Hey, guys! My name is Bruno Basto and this is my first post here so I apologize if I'm not doing it properly. I'm a beginner in electronics design and I do it as a hobby because I'm very interested in learning it

So, here is my question:

I'm trying to build an H-Bridge to drive my NEMA 23 Stepper Motor. The H-Bridge was built using 2 IRF530 and 2 IRF9530 MOSFETs. I have 1,5K resistors to limit current on the Gate of the transistors. Attached are the schematic and a picture of my protoboard (I took the picture before wiring the connectors, but now there is one green connector for the inputs and one green connector for the outputs).

When I put +5V or +12V (Continuous. Not PWM) on one side of the H-Bridge and 0V on the other side, the IRF530 on the activated side gets hot quickly. I tried various resistors to simulate different loads in the output and it always gets hot.

I tried measuring the current given by the power supply and it reads 1,2A. From what I understand, these transistor are capable of sourcing a lot more current than 1.2A (you guys probably have the datasheet of this FETs memorized ). I know that with PWM I can dissipate less power, but I don't think it should be heating up that fast with such a low current draw.

I don't know if I'm doing something wrong or if these transistors are just supposed to heat up and I shouldn't worry. But they heat a lot. To the point where I could burn my fingers if in contact for a bit of time.

I know that there are integrated H-Bridges out there and all, but the lesson I'm trying to learn is how to build an H-Bridge and not how to buy one ;p

Thanks in advance!

[Paul Price]

For the  MOSFETs to fully turn on the control voltage (gate to source) must be at least 5 (preferably 10V  to 18V  for best turn on) or else the on resistance will be high and heat  is generated.

Both sides must be oppositely voltage "activated" for an H-Bridge to work.

[BloodyCactus]

if your using a microcontroller to drive that, you need a fet driver like MCP1407 to turn those fets on properly. If you are using a microcontroller, you may be able to switch to IRL530, which is the 'L' for logic level version. IRL530, logic level driven, IRF530, not logic level driven.

[Kjelt]

Not sure but looks like a Wrong design.
If you put 5V on the left gates the Vgs of the p fet is still -7V so it turns on and the n fet could also turn on creating a dead short. Even when designed poroperly you have a short for a small time. The gates should be seperately driven.
Use a fet driver circuit for h bridges with fixed dead times.

[brunobasto]

Thanks for the prompt answers!

@Kjelt, You mean wrong implementation on the protoboard or wrong design on the schematics? The schematics looks standard as I saw a lot of H-Bridges designed that same way during my research.

If you look at the protoboard, the black wires are the gates and the yellow wires are the output. I've tried putting a resistor between the yellow wires and +12V on one of the black wires and measured the current across the resistor. It was about 500mA and the IRF530 got hot.

Tonight I'll try to measure the Vgs on all transistors to give you guys more detailed input. Maybe my power source is not behaving as expected. My schematics is not completely accurate as actually my power source is only capable of outputting 1A.

@BloodyCactus, I have not tried to drive that with a microcontroller yet. I've just applied +5V from an LM7805 with the 12V as input. I''ve also tried using 12V directly from the power source to one side of the gates.

If you guys have any better schematics that would improve mine, It'd be much appreciated. Just keep in mind that my component set is limited and I'm a beginner. So, the simpler the schematics, the better. I just want a simple H-Bridge using these MOSFETs I have (If that is even possible).

Thanks!

[Kjelt]

@Kjelt, You mean wrong implementation on the protoboard or wrong design on the schematics?

Schematics if your intention is to drive the gates with 5V logic that won't work.
Look at the datasheets. For instance the IRF9530 has a Vgs threshold of -2 to -4V, and the IRF530 +2 to +4V
This means that with a 12V supply the logic Input A should have voltages of
1) between 0V and 1,8V max to make the upper fet IRF9530 on and the lower fet IRF530 off
2) between 10,2V and 12V to make the upper fet IRF9530 off and the lower fet IRF530 on
So what happens if you apply 5V to input A?
Vgs of the upper fet IRF9530 will be 5V-12V = -7V so this will be on and the Vgs of the lower fet IRF530 will be 12V
so also will be on. So you created something you do not want in a H bridge a dead short of the powersupply over Fets M1 and M3.

That's why you need a levelshifter or fet driver. I am not an expert on this so maybe some analogue expert might give you a good solution.

[brunobasto]

@Kjelt,

Thank you so much. I think I get it now. If you don't mind me asking, do you think using voltage dividers I could improve my circuit to satisfy the conditions:

1) between 0V and 1,8V max to make the upper fet IRF9530 on and the lower fet IRF530 off
2) between 10,2V and 12V to make the upper fet IRF9530 off and the lower fet IRF530 on

?

Anyways, I think I can work from what I've learned from this post. Thank you all.

[JohnnyBerg]

Have a look at the IR2104. Its very easy to drive high and low side mosfet with this device.

[eneuro]

Probably galvanic insulated gate driver PSU and optocoupler mosfets gate drivers like TLP251 which can be easy driven from 5V MPU logic  levels are easiest way to forget about problems with H bridges etc.
This  is my Made at Home discrete transistors gate driver with PSU on 1inch2 PCB which from 24VAC output nice +/-12VDC for turning on/off mosfets with 0.4A gate current limited by outpt transistors and 27 Ohm resistors 


MCP1407 with 6A is not needed for simple H bridges, while probably you do not need high switching frequency.

Have a look at the IR2104. Its very easy to drive high and low side mosfet with this device.

Yep, everything is easy if you know exactly how to do it-from my expierience the easiest way is to have galvanic insulated gate drivers PSU and optocupler gate drivers controled by MCU 4pins with in series resistors 

[Kjelt]

@Kjelt,
Thank you so much. I think I get it now. If you don't mind me asking, do you think using voltage dividers I could improve my circuit to satisfy the conditions:

Yeah could be such that if you apply 5V,  the upper fets gate  is still more than 10V's it should be ok. But the best way is to go with a driver ic like suggest above, they push the gate hard 12V and ground and make sure there is some time between them so you never ever get a short. For a good solid design I would definitly look at that.

[brunobasto]

Hum, The IR2104 is about the equivalent to 3 dollars each here in Brazil. Seems a little expensive.. But I'll look into a MOSFET driver then. Seems like the way to go.

[JohnnyBerg]

The IR2104 has also a dead time build in. When switching, there is a dead time where neither of the MOSFET's conduct. Besides that, there is also the level shifting until up to 600V.

Perhaps you can get them cheaper of EBay or AliExpress ..

[Kjelt]

There are a lot of those chips, ST also makes them so just see if there is some cheap one out there.

[nuno]

Here's another discreet H-Bridge FET driver for P and N FETs, for you to take some ideas; it's for 24V Vcc, some simplifications are needed for 12V, namely, from the top of my head, removal of the zener diodes. You should be able to get the components easily in Brasil.

https://github.com/vnevoa/DiffTrike/blob/MarkIV_RasPi_NJAY/Electronics/PowerBridge/hw/bridge-fet-drivers-schematics.pdf?raw=true

updt: Inputs are 5V or even less if you adapt the input BJT base resistors.

[eneuro]

Here's another discreet H-Bridge FET driver for P and N FETs, for you to take some ideas;

TLP251 is priced around $1 there, one needs 4 such Optocoupler; THT; Out: IGBT driver; 2.5kV; DIP8; 0.4A
http://www.tme.eu/en/katalog/?idp=1&search=tlp251&cleanParameters=1
TLP251

If someone needs more current in outpt stage can add discrete transistors in similar totem configuration shown in your links and I also used in my shown prototype PSU with discrete gate driver in output stages 0.8A PNP BC327 and NPN BC337
I sugested forget about not optosiolated gate drivers for very simply reason-we do not like have our MPU destroyed in the case of power mosfets failure, so we want to limit those disasters to indyvidual mosfets gates (4 in the case of 2 phase H bridge), so I prefere something like TLP251 or 6N137+TC426 to get 3A quite cheap optoisolated gate driver.

BTW: If you broke a few cheap non optoisolated gate drivers, than buying a little bit more expensive optoisolaed one you are still ahead  and have MPU alive else it might

[JohnnyBerg]

Optocouplers are very uncertain spec. about turning off (very dependent on series resistor in collector or emitter).

If one side of the bridge turns on, while the other is still turning off, both MOSFETs conduct and could be fried. Thats why a dead time is needed. When isolating the bridge from the microcontroller, it should be done before the level shifter and with a decent optocoupler.

[nuno]

Over the development I've had several FET failures on circuit using those drivers. In my experience these power FETs most often fail shorted; on one of those failures I took some time to detect the failure and it took a driver's output transistor down, because the output got connected do GND and was trying to put a one. Other than that nothing else got damaged on the 8 or 9 FETs I've blown.

Up to a certain power level it's not critical that it gets galvanicaly isolated, although if you do you can be more careless on the layout without problems. But nice driver that TPL; I know of a few similar, from Fairchild I think, but they're more expensive. They tend to be a bit slow on response time, though.

[eneuro]

H-Bridge to drive my NEMA 23 Stepper Motor.

Is it this $15 one?
Nema 23 CNC Stepper Motor 0.88A 0.6Nm(85oz.in) 23HS16-0884

Description
This bipolar Nema 23 stepper motor with 1.8°step angle (200 steps/revolution). Each phase draws 880mA, allowing for a holding torque of 0.6Nm(85oz.in).

Probably this IC L293 might be usable in this case... 
www.ti.com/lit/ds/symlink/l293d.pdf
The L293 is designed to provide bidirectional drive currents of up to 1 A at voltages from 4.5 V to 36 V. T


$2 per   ST MICROELECTRONICS L293B Driver; half bridge; 1A; Channels:4; DIP16 
http://www.tme.eu/en/details/l293b/motor-and-pwm-drivers/st-microelectronics/#

[brunobasto]

Is it this $15 one?

This is the one: https://www.robocore.net/modules.php?name=GR_LojaVirtual&prod=605

The web page is in portuguese. But it costs about $30.

The site says the specs are:

Torque: 7,0 kgf.cm
Voltage: 3,5 Vdc
Current: 1,4 A
Nema: 23
Model: AK23/7.0F8FN1.8

And the datasheet I found says:

[eneuro]

Ok, so it looks like that NEMA 23 Stepper Motor has many different models and your is more powerfull than this one I've found 

[Paul Price]

What I usually do is create a 18V bias supply using the mcu to pulse a transistor for a few uSec  once per interrupt, one 390uh or so coil for the inductive boost, a schottky diode and finally a 18V zener and a small filter cap, then use a npn transistor level shifter/switch on each side of the H inputs to switch the +18V to 0 to control the MOSFETs. It is also pretty cheap and easy to use a 555 timer to  shortly pulse a transistor connected to the coil to generate the +18 voltage. Small SOT23 boost converters are a little harder to acquire but would do the same job to easily create +18V but you get a tiny footprint, namely the .LM27313. Despite the complexity, this inexpensive and  easily repairable H-Bridge can make up for the complexity.

[nuno]

In the next revision I'll be testing this one: FAN3278TMX (P&N FET half-bridge driver 24V)

[eneuro]

It is also pretty cheap and easy to use a 555 timer to  shortly pulse a transistor connected to the coil to generate the +18 voltage.

There are many posibilities and final design will depend on expected end user product requirements  and whether it is hobby project or for mass production or custom solution on demand.
While preparing for journey around Europe in Made at Home electric HPEV, the most important thing for me is keep things simply and easy to repair even in the road conditions just by using hot iron and paper schematics 
So, my choices are around solutions like this i 3 phase motor drivers
Isolated IGBT Gate-Drive Push-Pull Power Supply with 4 Outputs
This is based on LT5030 push-pull driver, but full-bridge converter is I'm interested in most while this allows keep things simple 
Three Transformers to Power All Three IGBT Arms of 3-phase Inverter The push-pull  topology allows connection of transformers in parallel, which allows the IGBTs in all three arms (U, V,  and W) to be powered using a single controller. The LM5030 High Voltage PWM controller contains all of the features needed to implement Push-Pull and Bridge topologies

Quite interesting application notes there Switch Mode Power Supply (SMPS) Topologies

I do not want to mess with non optoisolated gate power supply and make those circuits easy to repair, so I believe in that galvanic insulation will help to limit disasters areas to known boundaries and it will be easy recognize and replace damaged part of the circuit 

[nuno]

Well, if you're traveling, have a spare board. If it breaks you want the fix to be quick so you won't have to do soldering on the road . It's not always that easy to find the faulty component(s), especially out of the lab, and on the road...

[eneuro]

Energy harvesting methods are most important when traveling in electric vehicle.
Unfortunatelly, haven't got yet any kind of linear transformer to be able charge batteries in this vehicle... directly from 400kV 50Hz huge power grid lines, but as I said.... yet 
For the moment sunny beach should provide energy needed to find another house mains socket and full charge  EV...

Anyway, just a while ago tested my simple 24VAC -> +/-11V versatile gate driver PSU shown above and there were no problems to drive 230VAC AC mosfets switch at 100kHz using +/-13V GDT output created on its primary simply by H bridge made of totem BC327/BC337 and current limited to 100mA by 2 x 27 ohm resistors with a few lines of code on ATTiny85 @ 8MHz 5V powered from 12V battery, so physics works and hopefully electrons doesn't care if they flow on PCB designed in NASA labs by best in the world EE or are Made at Home using old school thermal transfer methods, because of when I have some idea I hate wait for manufacturer for prototyping and shippment delays and within a few hours my idea is changed into working circuit 

[poorchava]

1.5k in series with gate drive is way high. Try something in 1R-100R range (aside from other problems people have pointed out)