It might be useful if it is a brush motor, so I would leave it. It might damp some arcs and sparks. It should be at the motor.
But your right the voltage is clamped first by the diode (for a very short time) then the FET turns on and shorts the diode.

Yup... it's a brushed DC motor alright... The values I used for R and C are just a shot-in-the dark... I guess calculating them could be quite a hassle, considering that I don't have all the motor's info yet

 

I hate using transformers (I'm funny that way) ... they're bulky, heavy and expensive ... but sometimes using them is unavoidable, I guess ...

Just something to put on your watch list.
If you look real close at the gate of the lower FET you will notice a little positive glitch after it has turned off. It is caused by the fast rise time of the top FET turning on and the gate to drain capacitance of the bottom FET. I first noticed it on one of your sims using a logic level FET and it was actually big enough to turn it back on. Gota love Spice.

 

mmmhhh... I see a glitch, but it's a negative one. It doesn't seem to go below -2.5V. But I guess it's enough to trigger an LL fet... how can it be supressed... using a schottkey diode?

 

Yup... it's a brushed DC motor alright... The values I used for R and C are just a shot-in-the dark... I guess calculating them could be quite a hassle, considering that I don't have all the motor's info yet

I don't think it is critical - just something to slow down the edges. Whenever you are throwing around 170 volts there is going to be some noise.

 

mmmhhh... I see a glitch, but it's a negative one. It doesn't seem to go below -2.5V. But I guess it's enough to trigger an LL fet... how can it be supressed... using a schottkey diode?

View attachment 105773

I don't think the negitives are a problem. Blow up the little bright spots on the positives side.

 

You're right!

​

But they don't reach more than 2.5V (although that might just be the LL nFet's threshold), and don't last more than half a nanosecond. So they probably don't last long enough for there to be a problem, right?

 

View attachment 105775 You're right! View attachment 105776

View attachment 105777​

But they don't reach more than 2.5V (although that might just be the LL nFet's threshold), and don't last more than half a nanosecond. So they probably don't last long enough for there to be a problem, right?

I think it is okay. Just interesting.

 

I see a glitch, but it's a negative one.

I don't know how much reliance you should place on the glitch magnitude. It may be partly an artifact of the sim model. I tried using the 'official' model, but couldn't get that to give proper LO output pulses at all! Kept getting a 'Timestep too small' message and the sim ground to a halt. The feeble LO pulses it did briefly produce also exhibited glitches, albeit about half the amplitude.

 

I don't know how much reliance you should place on the glitch magnitude. It may be partly an artifact of the sim model. I tried using the 'official' model, but couldn't get that to give proper LO output pulses at all! Kept getting a 'Timestep too small' message and the sim ground to a halt. The feeble LO pulses it did briefly produce also exhibited glitches, albeit about half the amplitude.

Strange. It is a little smaller with my model, but not much.
Anyway the threshold for that FET is 3 volts.

 

I don't use LTspice, but how good is it simulating a DC brushed motor (L1)?
A DC motor is quite a bit different from simple inductance?
Max.

 

I don't use LTspice, but how good is it simulating a DC brushed motor (L1)?
A DC motor is quite a bit different from simple inductance?
Max.

Not so good. No back emf, etc. What I usually do is look at the beginning with the armature resistance then change the resistance to represent the run current.
I'm trying to work up the courage to play with @Alec_t 's model. The thing that I really miss is regen.

 

I'm trying to work up the courage to play with @Alec_t 's model. The thing that I really miss is regen.

And I'm trying to work up the knowledge to understand it....
If you decide to experiment with Alec's model, I'd be very interested in seeing the results.

 

I don't use LTspice, but how good is it simulating a DC brushed motor

Feed LTS a good model and it does a good job .

I had a look at the 'official' model for the IR2104 and spotted that the switches used there to provide the HO and LO outputs were modelled as having Ron=10Ω. My posted model uses two switch types, ms1 and ms2, modelled with Ron=71Ω and Ron=42Ω respectively. I played with reducing those resistances to 10Ω and 5Ω respectively and the glitch magnitude is much less. You might want to edit your copy of the IR2104g.sub file accordingly.

Re the DC motor model, the DCMotor-Test.asc file shows the elements of the model and was used to generate a netlist which, with a bit of editing, is the substance of the .sub file.
The somewhat simplified model makes these assumptions:-
The armature size and inertia are proportional to rated power (rp).
Rated power = rated voltage (rv) x rated (no load) current (ri).
Generated torque is proportional to motor current.
This torque is opposed by extenal applied load torque (Lt) and by internal frictional/windage torque.
Internal torque is proportional to angular velocity of the armature.
Acceleration is proportional to nett torque divided by armature inertia (hence rp).
Velocity is proportional to the integral of acceleration.
Motor resistance is a fraction of the value given by rv/ri.
The various k values are proportionality constants empirically derived.

Give us a shout if you want further explanation.

 

You might want to reduce the value of the motor's inductance in the sim. From extensive googling today (practical figures for motor inductance are scarce on the web), DC motors of ~1HP rating generally seem to have an inductance in the 5-50mH range. I've just measured a couple of hobby motors (an R540? and a non-descript smaller one) and their inductances are 11mH and 43mH respectively. I'm surprised that inductance doesn't scale nicely with motor power/size.

 

and a few TVS see how things work out.

@cmartinez Hi, would you be willing to post about the TVS, how you chose them and why you place them where you do? It would be much appreciated.

 

@cmartinez Hi, would you be willing to post about the TVS, how you chose them and why you place them where you do? It would be much appreciated.

Sure thing... as soon as I get back from my trip

 

@cmartinez Hi, would you be willing to post about the TVS, how you chose them and why you place them where you do? It would be much appreciated.

Alright, my friend, I'm back, and ready to rumble again. I've attached an interesting document describing the how and why's of TVSs for a mosfet. I intend to use higher powered TVSs (at least twice as much) than the ones suggested, since the load will not only be of the inductive type, but because abundant transients will also be present due to its being a brushed motor.

The FDA33N25 datasheet provides the maximum following values:

• VDSS Drain to Source Voltage, 250 V
• VGSS Gate to Source Voltage, ±30 V

Funny that no Drain to Gate Voltage is mentioned in the specs. It's probably because the gate is driven relative to the source, and its drain to gate breakdown voltage is higher than the drain to source voltage anyway. But of course I could be wrong.

So, considering the previous values, I've decided to place a TVS between drain and source with a breakdown voltage of about 150V, and 3,000W capability. And for gate to source, anything between 20V to 24V rated at 600W will do just fine.

After doing some searching at Digikey, this is what I've come up with:

• I think that the SMDJ150CA should fit the bill for drain to source, since it has a minimum breakdown voltage of 167V, a maximum clamping one of 243V, and a rating of 3kW
• As for gate to source, the ICTE18-E3/54 should do just fine. Minimum breakdown voltage is 21.2V, maximum calmp at 25.2V, and power rating of 1.5kW

The SMDJ150CA would be replacing the 1N4004 shown in my last schematic.

@ronv, what do you think?

 

Alright, my friend, I'm back, and ready to rumble again. I've attached an interesting document describing the how and why's of TVSs for a mosfet. I intend to use higher powered TVSs (at least twice as much) than the ones suggested, since the load will not only be of the inductive type, but because abundant transients will also be present due to its being a brushed motor.

The FDA33N25 datasheet provides the maximum following values:

• VDSS Drain to Source Voltage, 250 V
• VGSS Gate to Source Voltage, ±30 V

Funny that no Drain to Gate Voltage is mentioned in the specs. It's probably because the gate is driven relative to the source, and its drain to gate breakdown voltage is higher than the drain to source voltage anyway. But of course I could be wrong.

So, considering the previous values, I've decided to place a TVS between drain and source with a breakdown voltage of about 150V, and 3,000W capability. And for gate to source, anything between 20V to 24V rated at 600W will do just fine.

After doing some searching at Digikey, this is what I've come up with:

• I think that the SMDJ150CA should fit the bill for drain to source, since it has a minimum breakdown voltage of 167V, a maximum clamping one of 243V, and a rating of 3kW
• As for gate to source, the ICTE18-E3/54 should do just fine. Minimum breakdown voltage is 21.2V, maximum calmp at 25.2V, and power rating of 1.5kW

The SMDJ150CA would be replacing the 1N4004 shown in my last schematic.

@ronv, what do you think?

Well I'm not a big fan of TVS for this application. Here is my thinking:
On the drain to source the inductive kick will go either to the clamp voltage in one direction and the body diode of the FET in the other direction. So more noise in one direction and slow turn off in the other case. My preference is for fast diodes rated at at least 50% of the maximum motor current. Using the TVS across the supply is maybe a good idea. Also a large cap from the top to the bottom of the FET to reduce the effect of inductance in the power leads.
Again, just my preference, I would use back to back zeners to try and protect the gate. They can be run closer to the supply voltage and have lower capacitance so don't slow down the turn on/off of the FET.
Stay away from any from drain to gate. The capacitance can cause the lower FET to turn on when the upper FET turns on.

 

Stay away from any from drain to gate. The capacitance can cause the lower FET to turn on when the upper FET turns on.

That's very valuable advice. Thanks!

 

I would use back to back zeners to try and protect the gate

How about a bidirectional TVS like this one? would it do the job just as well?