High Side Switch Protection for Inductive Load

MisterBill2

Joined Jan 23, 2018
18,175
OK, now it is clarified a bit, in that the inductive spikes being protected against are not coming from the inductance controlled by the switch. THAT is an unusual situation.
 

StefanZe

Joined Nov 6, 2019
191
don't you think load dump (i.e. high voltage at VCC) can happen if the inductive load is switched off too fast?
Can you list out the reasons causing the load dump?
Yes, but first the MOSFET T1 will be damaged. The current can only flow through T1, there is no other way.
 

Alec_t

Joined Sep 17, 2013
14,280
THAT is an unusual situation.
I think it's common in vehicle electrical systems. If a high current load, anywhere in the system, is suddenly switched off (i.e. a load dump) then the alternator's output voltage can rise dramatically because of the finite response time of the regulator.
 

MisterBill2

Joined Jan 23, 2018
18,175
What I meant was that it is an automotive electrical system. The initial post did not seem to imply that the spikes were on the supply side of the switch.
But still, by using components with adequate voltage ratings, problems are avoided. And I am aware that in most mass produced consumer goods Quality takes a third priority behind profit and cheap production costs. In all of my designs, Quality, as defined by the ability to constantly perform withing specifications, was always the primary target, followed by stability and durability. Max profit was about number 5 or six on the list. Thus the company prospered.
 

richbrune

Joined Oct 28, 2005
126
This is from this application note (https://www.st.com/resource/en/appl...drivers-for-automotive-stmicroelectronics.pdf) associated with the high side driver VNQ7E100AJTR ( which has 4 internal high side switches used to control 4 different automotive loads e.g. Headlights, cooling pump, etc.). I am reviewing protective features that this switch comes with and one of them is Protection against low energy spikes and load dump .
Thanks for the link MisterBill2. There's a nice set of circuit diagrams there that give a sense of the problem and how it is typically addressed. I know from some bad experiences that the 100 volt spike drawn by the tread starter Surya1234 is docile compared to the 30,000 volts developed numerous times a second across the spark plugs by disconnecting the inductor (ignition coil) of an automobile. Keeping electronic equipment isolated from this and other large inductors in the system is a major challenge for auto designers. I've always wanted to model at least the basics of something like what Surya1234 drew at the start of this thread, in LTspice. I don't know what inductor values to use to get the 100v surge from the 12v when the switch is opened. I'm thinking of using a mosfet and putting a pulse of voltage at the gate to turn the voltage to the coil on and off, in lieu of a mechanical switch, if you think that would work. I wouldn't mind any additional advice on how to make the model--thanks.
 

MisterBill2

Joined Jan 23, 2018
18,175
Glad I could provide a bit of help.
To make really high transient spikes, use a really fast switch-off. The faster the current stops the bigger the spike. E= L dI/dt
 

richbrune

Joined Oct 28, 2005
126
Glad I could provide a bit of help.
To make really high transient spikes, use a really fast switch-off. The faster the current stops the bigger the spike. E= L dI/dt
I used a huge inductor, and shut off the base of the MOSFET from 12v to 0 in less than a microsecond, still can't seem to get the coil to get over 12volts. Any thoughts?1685144470355.png1685144470355.png
 

Danko

Joined Nov 22, 2017
1,829
I used a huge inductor, and shut off the base of the MOSFET from 12v to 0 in less than a microsecond, still can't seem to get the coil to get over 12volts. Any thoughts?
Negative voltage of spike goes to source and turns ON transistor (common gate circuit).
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,175
Try a mechanical switch. And also, I think Danko is correct. And just because the gate voltage changes does not mean that the device is actually switching off.
So once again: Try a mechanical switch. A snap switch would be best. Then you can be certain that it switched completely off. Keep in mind that it is the rate of change of the current that drives the spike. You might also get better results with a PNP transistor, so the spike would not be affecting the Vbe.
 

Alec_t

Joined Sep 17, 2013
14,280
I used a huge inductor, and shut off the base of the MOSFET from 12v to 0 in less than a microsecond, still can't seem to get the coil to get over 12volts. Any thoughts?
1) The generic MOSFET rarely does what you want. Select a 'proper' one.
2) Your sim times don't allow the inductor current to build up to its maximum (~12A?) before switching off the MOSFET. If I'm reading your pic correctly the current is only ~1mA.
 

Danko

Joined Nov 22, 2017
1,829
1) The generic MOSFET rarely does what you want. Select a 'proper' one.
2) Your sim times don't allow the inductor current to build up to its maximum (~12A?) before switching off the MOSFET. If I'm reading your pic correctly the current is only ~1mA.
Something like this:
1685185228381.png
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,175
It is obvious that the simulation is not showing a quick switch off. Or else the simulator is confused.

I suggest moving the inductor to the section between the supply and the drain terminal Then it will have no effect on the gate to source voltage.
 

Danko

Joined Nov 22, 2017
1,829
It is obvious that the simulation is not showing a quick switch off. Or else the simulator is confused.
I suggest moving the inductor to the section between the supply and the drain terminal Then it will have no effect on the gate to source voltage.
In this case simulator not lying.
Everybody knows result of using inductive load in drain chain.
But TS asked about common drain circuit, which is self-profected
from HV inductive spikes:
1685199153731.png
 

richbrune

Joined Oct 28, 2005
126
1) The generic MOSFET rarely does what you want. Select a 'proper' one.
2) Your sim times don't allow the inductor current to build up to its maximum (~12A?) before switching off the MOSFET. If I'm reading your pic correctly the current is only ~1mA.
1)True, apparently not, but from the plots I'm not sure why, since the coil seemed to finish charging. 2)I'm not sure I got the plots to render, but no, the current didn't build up adequately, it rolled of level at about ~1.4mA, well before the end of the sim, and when I used a voltage controlled switch with the same sim times, it worked perfectly and built a 30kV charge (spike). So yes I suspect the MOSFET model had something to do with it.
 

richbrune

Joined Oct 28, 2005
126
It is obvious that the simulation is not showing a quick switch off. Or else the simulator is confused.

I suggest moving the inductor to the section between the supply and the drain terminal Then it will have no effect on the gate to source voltage.
Yes, the "snap switch" you recommended gives the right answer (LTSpice image below), I'll go back later and try to figure out why I couldn't get the MOSFET to work.
snap_switch_and_inductor.png
 

richbrune

Joined Oct 28, 2005
126
Negative voltage of spike goes to source and turns ON transistor (common gate circuit).
Yes, I think you're right, but I haven't figured out why I don't see it in the plots. It's a generic N-channel MOSFET. I think maybe you're referring to the circuit that corrects (absorbs) the spike, which I haven't tried yet since I couldn't get the spike to generate. Just now I used a voltage controlled switch and it indeed generates ~30KV. I'll try the common get circuit soon--thanks.
 

crutschow

Joined Mar 14, 2008
34,283
I think you're right, but I haven't figured out why
Simple.
The inductor inductance will keep the current going through it when the MOSFET tries to turn off, which generates a negative voltage at the MOSFET source.
This generates a positive MOSFET Vgs voltage, and keeps it turned on until the inductor energy is dissipated and the current drops to zero.
 

richbrune

Joined Oct 28, 2005
126
Oh, I think I get it now. So the plots of absolute voltage wouldn't obviously show the relative Vgs. Sure enough, when I plot the current through the inductor (and R1) I see it takes a full 25mS for the current to ramp down to zero. Thanks for your concise description.
 
Top