Are we all on the same page, that the freewheeling diode is the one across the motor (the diode that praondevou hand-drew)?
This diode is absolutely essential, and should be rated according how many amps you intend to run, as full motor current will be flowing through it when the mosfet is off.

Yes and yes and yes.

The diode from source to drain doesn't do anything in the OPs design since it will not conduct as I showed a few pages ago. Even if it was, not necessarily one would need to add a diode because these MOSFETs have an intrinsic diode which however is not one of the fastest....

 

me too, in post #42

The 6a rectifier diode serves no purpose in this instance and would actually be better if you use that 6a diode as your freewheeling diode if thats all you have

 

OMG, are you telling me you didn't have a diode on the motor terminals?

i didn't sorry i did arrive at the uni late so i didn't have enough so i forgot about it

Doesn't make sense. If at point A the MOSFET starts conducting then why the voltage rises at this point??? Either because the waveform displays the gate signal, or because at point A the MOSFET turns OFF which doesn't make sense neither because your pulse is getting wider with increasing motor speed.

ok to make this point clear after i take all pictures i switch the o-scope terminals then i strat recording the video because for me it was and still somehow confused which is ground and which is positive so in the video as shown there are signal going down having a very high voltage (either in +ve or -ve direction) ,

what i understand is when the mosfet is conducting the voltage will rise to form the square wave and by increasing the speed the time of this square will increase just the same as the gate signal and when it is off the voltage will go the "ground value" (again either square is up or down

Did you do this on purpose?
Now we are getting somewhere, have a look at the picture, how big was the peak? If this goes higher than 500V with the IRFP460, the device is dead.

Btw: The first picture from my post #94 is exactly what you have, only much worse. Have you seen it?

i did this on purpose but i am still confused which is right one square up or down!
the peak value did exceed 45 volt from a supply of 12.4 volt

i did see this picture in post #94 and yes it si very similar next time i will try the capacitor
maybe today

btw: the 12 dc motor i am working with is permanent magnet motor i dont know if this will make difference or not

 

That's what I've been thinking. Are we all on the same page, that the freewheeling diode is the one across the motor (the diode that praondevou hand-drew)?
This diode is absolutely essential, and should be rated according how many amps you intend to run, as full motor current will be flowing through it when the mosfet is off.

i forgot to connect the freewheeling diode , maybe today i will do it , i hope

 

what i understand is when the mosfet is conducting the voltage will rise to form the square wave and by increasing the speed the time of this square will increase just the same as the gate signal and when it is off the voltage will go the "ground value" (again either square is up or down

Ok, a misunderstanding.

1. If somewhere is written , measure Vds of a MOSFET, then you connect the voltage probe from "d" to "s" meaning the tip to "drain" and the probe's ground cable to "source". If you do it the other way around than you are not measuring Vds anymore but Vsd. The first letter is where the tip goes.

You can connect the oscilloscope probe anywhere and anyhow you want, always keeping in mind that the ground cable of the probe is connected to buildings ground inside the o-scope. Clear?
Not wanting to confuse you but sometimes you are forced to put the probe's ground cable to a specific point in the circuit, if this would cause an inverted signal, use the invert channel function of the o-scope.

2.
Although the term ground could be any common reference voltage level in your circuit when referring to o-scope measurements, in general people mean when using "ground" either buildings ground, virtual ground or as in your case the negative voltage level of your battery. You are right it's not exactly clear , but in general if someone tells me "ground" I understand the zero voltage of my power supply.

3.

what i understand is when the mosfet is conducting the voltage will rise to form the square wave

When you are measuring Vds (tip on drain, ground on source as in my picture) the voltage will go to near zero when the MOSFET conducts.

Mosfet conducting --> low impedance --> voltage drop very low
Mosfet not conducting --> Open circuit --> voltage drop high

That means: When turning the gate signal off, Vds of the MOSFET will rise, too high in your case. This is where we need to act.


Next steps:
- put the freewheeling diode in there and leave it there
- measure again the voltage peak with the small motor
- try to put the capacitor "C" from my picture in your circuit, you can play around with some values/capacitor types currently available to you, to get a feeling of how they affect the voltage peaks
- Finally you will probably need a bigger electrolytic capacitor and a smaller, low ESR capacitor in parallel for "C"

- when you understood the effects of the capacitors, then we start with the bigger load, however it's still not a good idea to just connect the bigger motor and connect right away the 260VDC. If you don't want to carry your batteries to the university try to source a power supply , they should have one. If you start with 30 to 100V that will be ok too.

 

i am testing now and so far without the capacitor there is no voltage spikes

 

Your MOSFET Vds doesn't go to zero and stays there while OFF. Why? In post #93 , is this the voltage from drain to source? If so, then you need to measure the height of this horrible positive voltage peak. (you had your freewheeling diode of the motor connected, right?)


Just to make sure, you don't need an additional cap on the motor if that's what you meant.
First and more important you need an electrolytic capacitor from the + side of your load to GND, as near as possible to the load and MOSFET.

I attached a picture to show what I mean.

1. picture: no snubber and bypass capacitor, the peak on the dark blue line is what you want to eleiminate. It's caused by 40cm of wire and 80mA in my test, no inductive load whatsoever. The faster your gate turns off and the bigger your wire inductance from power supply / bypass capacitor to load/MOSFET the higher will be this peak.

2. picture: a small capacitor is placed from drain to source. It decreases the peak but will increase power dissipation of the MOSFET

3. picture : an electrolytic capacitor of a few hundred uF has been placed from + side of the load (a resistor) to the source of the MOSFET. The voltage peak is gone

The absolute values in this waveform are irrelevant because I used a much lower load/voltage/current. The principle is always the same.

well when connected the capacitor i didnt notice the same difference maybe the wires i use is too long

 

i am testing now and so far without the capacitor there is no voltage spikes

what frequency is this? zoom in on the rising edge of the Vds waveform.

 

here are some pics
the frequency is 14.4kz

 

Assuming we're looking at the 12V mosfet output (Vds) and 5V/div, you have a 35+V spike every time it turns off. Still not good. what diode are you using?

 

You should also look at Vgs (voltage on the gate, referenced to the source terminal). The gate may very well be "ringing". This will occur if you have any significant distance between the gate driver and the MOSFET gate, with no resistance between the driver and the gate.

The gate driver's ground needs to be connected to the MOSFET's source terminal, of course - and it needs to be a very low-inductance connection.

Posting a photo of your uC, gate driver, MOSFET and load would be helpful.

 

So SGT you think we're looking at the gate? I thought we were past that and posting pictures of Vds traces now. Kandlix are we back to posting pictures of Vgs? I sure hope you don't have 35V spikes on your gate.

 

So SGT you think we're looking at the gate? I thought we were past that and posting pictures of Vds traces now. Kandlix are we back to posting pictures of Vgs? I sure hope you don't have 35V spikes on your gate.

No, he said we should look at the gate. This must be Vds, after the OP measured it with the right polarity.

It doesn't look good.

 

No, he said we should look at the gate. This must be Vds, after the OP measured it with the right polarity.

It doesn't look good.

DOH! I misread, sorry.

 

Yeah, if it WERE Vgs, I'd be very, very surprised.

If the gate is ringing, you'll see Vds oscillating like that. It also means that the MOSFET would be turning into a room heater, as it would be spending lots of time in the partially conducting state.

 

You should also look at Vgs (voltage on the gate, referenced to the source terminal). The gate may very well be "ringing". This will occur if you have any significant distance between the gate driver and the MOSFET gate, with no resistance between the driver and the gate.

The gate driver's ground needs to be connected to the MOSFET's source terminal, of course - and it needs to be a very low-inductance connection.

Posting a photo of your uC, gate driver, MOSFET and load would be helpful.

well first ,i am afraid i dont get the term "ringing " what do you meen with it ?!
second what to be considered a significant if this consider the wire it is not longer the 10cm
and btw i have a 5 ohm 2W gate resistance

 

well first ,i am afraid i dont get the term "ringing " what do you meen with it ?!
second what to be considered a significant if this consider the wire it is not longer the 10cm
and btw i have a 5 ohm 2W gate resistance

ringing happens on turn-off when all the stored charge exits the gate. it can get in a hurry and bounce all around on its way out, bouncing back up into the gate and turning the mosfet back on. this bouncing can happen several times, so your mosfet turns on and off several times instead of just once. that's what the gate resistor is for, to slow the electrons down and force them to leave the gate in an orderly fashion. you can also aid this problem by putting a ferrite bead on the mosfet leg of the gate.

 

If you have nothing else to do and plan to work on other projects involving gate drivers and MOSFETs have a read through this app note:

App note

Don't give up on this one. We get this to work.

 

well first ,i am afraid i dont get the term "ringing " what do you meen with it ?!
second what to be considered a significant if this consider the wire it is not longer the 10cm
and btw i have a 5 ohm 2W gate resistance

In post #60 I attached a picture indicating the wires/traces that need to have a short length. You say 10cm. Does that mean 2x10cm =20cm (gate + gnd return wire).? Although I don't think that's the problem 20cm is not short.

 

those wire should be twisted as well, to reduce inductance