H-bridge for Nth time

[idiot7.0]

Hi,

This is a mosfet based h-bridge that seems to work (at least in simulation) but even when the mosfets are in saturation mode, I see only a max of 1,23A flowing even if shorted. What is causing this?

The zeners are rated at 15V. Do you think the 600Ω resistor will slow the switching times a lot or not ?

This is a link to the simulated circuit in case you want to take a look.

Thanks.

 

[audioguru]

Why is your schematic a negative image?

Your upper Mosfet is drawn wrong. Its diode is backwards and the entire Mosfet is upside-down.
It is an N-channel type but it should be a P-channel type.

 

[idiot7.0]

The image looks negative because that's how the simulation software draws it.

The P/N channel mosfets are a mistake from the falstadt circuit simulator. They are P chanels at the high side and N channels in the lower side (gnd) but the symbols are incorrect from the simulator itself.

What about the amount of current flowing during saturation?

 

[ronsimpson]

What is the power supply for the (op-amp/comparator/MOSFET driver)? 24V??

 

[audioguru]

When an old Mosfet is turned on then it conducts a modest current.
When a modern Mosfet turns on it conducts a huge current.

Maybe your simulation uses old Mosfets.

 

[idiot7.0]

What is the power supply for the (op-amp/comparator/MOSFET driver)? 24V??

The comparator outputs 0 and 24V.

When an old Mosfet is turned on then it conducts a modest current.
When a modern Mosfet turns on it conducts a huge current.

Maybe your simulation uses old Mosfets.

Good idea, but if I remove the zener the current doubles and if I increase the supply voltage the current increases hugely. Then I doubt there is a problem in my calculations. I mean in my schematic.

Try the link in the first post it simulates the circuit online.

 

[audioguru]

The absolute maximum gate-source voltage for most Mosfets is 20V. Without the zeners you are feeding almost 24V.
A modern Mosfet conducts a huge current when its gate-source voltage is 10V.

 

[idiot7.0]

The absolute maximum gate-source voltage for most Mosfets is 20V. Without the zeners you are feeding almost 24V.
A modern Mosfet conducts a huge current when its gate-source voltage is 10V.

Yes absolutely. But the simulator indicating that a small amount of current is flowing leads to think that there is a bug somewhere.

 

[ronv]

What is your mos fet type?

 

[idiot7.0]

What is your mos fet type?

What do you mean? There is nothing special about them. I have P-channels in the high side and N-ch in the lower side if that's what you mean.

But I havn't tryed it in real yet. I'm still having a problem in the circuit simulator that i don't figure out. I'll probably try it with an IRF840 and IRF9540.

 

[audioguru]

Maybe your simulator is using old Mosfets that have low output current.
An IRF840 has an output current of about 6A in your circuit because it is rated for high voltage (500V) not high current.
An IRF9540 is a little better than an IRF840.

 

[idiot7.0]

Maybe your simulator is using old Mosfets that have low output current.
An IRF840 has an output current of about 6A in your circuit because it is rated for high voltage (500V) not high current.
An IRF9540 is a little better than an IRF840.

Yes, I know, but to take away the supposition of an old mosfet used by the simulator, I removed the zener diode and increased the supply voltage to about 240V just as mentionned in a post above, then I got a current flowing of more than 100A.

I doubt I'm having something related to Vds Vgs.

 

[audioguru]

If you feed more than 20V to the gate-source of most Mosfets then you destroy the Mosfet. Use about 10V.
Most Mosfets cannot pass more than about 75A because their pins will melt.

 

[idiot7.0]

If you feed more than 20V to the gate-source of most Mosfets then you destroy the Mosfet. Use about 10V.
Most Mosfets cannot pass more than about 75A because their pins will melt.

I know, it's mentioned in every datasheet. What I said is that I did that in the simulator just to prove it's not a simulator's limitation.

This means there is a bug somewhere in the schematic, or the simulator doesn't handle mosfets correctly (it begun with the symbols) but I don't think so.

 

[audioguru]

I think you need a new accurate simulation program. Try the free LTspice IV from Linear Technology.

 

[ronv]

Maybe your simulation is using a FET like this one that has Rds on of 10 ohms.
I don't know how your simulator works, but sometimes if you don't tell it what kind of transistor to use it doesn't work good. Simulators don't keep you from burning up parts either.

https://www.electro-tech-online.com/custompdfs/2012/08/MCH6601-DPDF.pdf

When I put your half bridge in LT Spice it worked ok.

 

[idiot7.0]

Maybe your simulation is using a FET like this one that has Rds on of 10 ohms.
I don't know how your simulator works, but sometimes if you don't tell it what kind of transistor to use it doesn't work good. Simulators don't keep you from burning up parts either.

https://www.electro-tech-online.com/custompdfs/2012/08/MCH6601-DPDF.pdf

When I put your half bridge in LT Spice it worked ok.

Thanks mate,

My simulator is the falstadt online circuit simulator. You can change only few parameters of the parts (only threshold voltage for mosfets).

I'll give it a try and see what happens.

 

[ChrisP58]

Thanks mate,

My simulator is the falstadt online circuit simulator. You can change only few parameters of the parts (only threshold voltage for mosfets).

I'll give it a try and see what happens.

If you can't enter part models for real components, then the simulator won't give you results anywhere close to how the actual hardware might work. There are many parameters for all components, not just mosfets, that are important to model accurately if you want a meaningful simulation.

 

[()blivion]

I am *VERY* familiar with the falstad simulator. I use it all the time as most who know me are aware.

It does not do MOSFET's correctly at all, it's an open bug that has been reported to Paul..... Use the model for "Analog Switche (SPST)" under "Active Devices" right click menu. Use a backward Schottky in parallel to simulate reverse conduction. You can make them effectively P-channel by right clicking them, clicking edit, then checking the box that says "normally closed". You can also change the on and off resistance in the same menu.

+1 if you feel this was helpful. Thanks.

Edit: Paul chose to use a symbol for the MOSFET's that shows what direction they conduct current when on. This is of course wrong.

 

[()blivion]

*HERE* is a link to a falstad simulation of mostly accurate models of your two transistors. I made 80% of the values and parameters for each transistor identical to the datasheet specs. Note that the red box around all the parts is not a part of the simulation, it is only there to show you where the transistor ends. Do not alter anything inside this box, it is a fact of the transistor that can not be changed. You may remove the box and labels if you feel you must.

Here is a picture for those that can not see the simulation....
View attachment 66404

Things not simulated
Max current.
Max voltage.
Max power.
Analog/liner operation.

You will need to be sure you're not exceeding these parameters yourself, (you know how to do voltage and current already). You can hover over the switch and look at Watts/milliwatts when on. If the power dissipated exceeds 5 Watts, your transistor will need heatsinking. If power exceeds 50 Watts, your transistor will need excellent heatsinking. If you go over 100 Watts your transistors will need replaced... Pretty much for sure.

These models are only for switching. It is important you realize that it does not model switching times or any kind of half on biasing. This is because the simulated switch in the model can only ever be on or off. For this reason, make CERTAIN that you have your FET's gates at GND or 10~15 volts, and that your driver can supply in excess of 1Amp to the gates. And do not attempt to drive the FET's faster than ~1Mhz. If you do, the model will leave reality behind and you will likely blow up transistors.