I attached some 12V leds to the output pins. I also tested them on a 9v battery. They shone way brighter on the 9V battery. Also they don’t turn on when the motor is at HiZ.
I’m in the same camp as @Brett - get a Eval board…
Did you use the above or other ST doco’s for reference?
Yes, the pcb engine used the eval board’s schematic and layout as reference (at least he was told to). From comparing them myself I can’t see any obvious differences.
So next step for me is to get an eval board? I wanted to get the board that came with the STM32 on it so that I can plug in values on the workstation and have it just work, but I could not find any sellers.
To me your mosfet layout looks wrong - drains should be both tied to VS??
**Edit - having looked again at your mosfet fluff my initial concerns my be incorrect…
There is a charge pump that is used to generate the higher voltage needed to switch off the MOSFET’s. This has to be configured so that the voltage is higher than the supply to ensure that the MOSFET’s are switched off.
Is this configured correctly in your design? What is the voltage at VBOOT?
I will have to add a short wire to the VBOOT net (on one of the passives) to find out. While the soldering iron is hot, are there any other voltages I should be probing?
That is more than enough to ensure that the MOSFET’s are off.
I guess I’ll probe the gate voltages next then.
EDIT: Btw, what voltages should I be looking for on the gate? If it’s 18V or more then that means the gate is off? If it’s less than 12 then the gate is on?
Ok the SQJQ910EL-T1_GE3 that I purcahsed says in it’s datasheet:
Gate-source voltage 20V (Limit). Is that a minimum limit or a maximum limit?
Ok, from the application note under the topic 2.4 Wrong setup issues
Starting from VCC the gate voltage is decreased down to the Miller plateau region, but the gate charge is not enough to complete the plateau and the Miller clamp is closed before the Vds commutation is ended. The MOSFET is immediately turned off and as a consequence the output of the power stage is subjected to a strong slew rate which could cause critical issues.
In the second case the gate charge is not enough to complete the Miller plateau region (see
Figure 12). In this case the output commutation is not completed and the overcurrent
protection, which measures the voltage drop on the MOSFET, is triggered even if very low
or even no load current is present.
The document itself is about the setup of the system to suit the selected MOSFETS, however they don’t give an actual example! (and I don’t understand).
I’ve determined that having the eval board isn’t going to help. The problem seems to lay in how the mosfets are being driven. The settings for the eval board most likely won’t work with my board.
After making some changes to the parameters, I am now getting an undervoltage lockout. This implies that VCC is being shorted to GND. I suspect that the low side mosfets are turning back on before the high side ones are fully off.
Whoops, gate is on the wrong pin.
Have you tried doing a bodgey wire hack?
Solder them back on but bend the 2 pins that are wrong upwards first. Also solder the jumper link to the board before you solder the mosfets back on.
Well, that was an epic fail.
?? What happened, magic smoke?
Was trying to build something that had already been done years ago. Might as well use ST’s solution.