Yes. Its working well. I put a load on the motor, actually stalled it out, no change in the waveform and NO fire.


I believe we have the motor problem solved. What do you think?

 

It's sounding good - are you still running with the 100W motor?

What kind of current draw are you getting when stalling out the motor?

 

Stall current was about 14.5 amps.

 

Rob can you post an updated schematic?

I'm wondering why you didn't use a totem pole driver?

 

Sorry about the quality. I use autocad and don't have a converter that does a good job.

SgtWookie has been kind enough to help me on this, when I nstarted I didn't have a driver at all.

 

Today I will start building the motor assembly prototype to test the power of the motors I have available.
I need to determine what motors to use. For anyone interested, I have started a chart of Electric scooters motors, these are inexpensive and a lot of them are goint to the trash with the scooters because the controllers are failing.
Two matching scooters turned up in the trash yesterday, 135 Watt motors and good batteries. So hese are the ones I start with.

 

I see that the XYD-158 are rated for 12v/24v/36v - is that switch-selectable or what?

Looks like most of those motors are 24v capable - are you planning on changing to 24v? I know you've been testing on 12v.

 

All motors seem to work at 12Voltsmp. No switch, just two wires for power. "Looking at the spec the only thing that changes is current, not sure why the other specs don't change. If I need to go to 24v I suppose the driver could still work at we and the mosfet would handle the 24v.
At this time I am not planning the change, unless I am convinced it needs to be that way. If I have to go to bigger motors it will probally be necessary to keep the current down.

 

I needed to ask that question, because if you were considering going to 24v, the driver circuit would have to be modified.

The maximum limits for Vgs is +/-20v; if you attempted to use it with 24v as-is, you would destroy the MOSFET.

The change would not be difficult at all; it would involve adding a regulator or Zener, one.

 

If needede could we run the driver at 12 and the mosfet at 24?

Sorry about the typing on the last post, I was using the blackberry and stuff got lost, tryed to edit the post but it would not let me.

 

Yes, the driver board could be operated as-is on 12v, with the MOSFET drain switching the ground side of a 24v or 36v motor. The grounds would need to be common between the 12v source and the motor power.

 

I built a driver board with both left and right motor circuits on the same board. The high current portion was removed and put on separate boards so the can be mounted next to the motors reducing the wire losses. The attached schematic shows the 3 separate boards. the attached picture is the high current board under test and not it final mounted configuration.

Put the tank in the ground last night, and it actually moved and seemed to have enough power. Further test are required, once I get everything mounted and cleaned up, again.

P.S. didn't use heat sinks and the test didn’t even warm the up. As you can see there is room for two isolated heat sinks to be mounted in the final configuration.

 

Attached are the readings I am getting with the Motor running.
the yellow trace is measured at the Gate of the mosfet to groung.
the green trace is measured at the Drain to ground.

Does this look correct to you, and does the switching time look ok?

I haven't tried 24 volts yet, to see if there is more torq. If you have time I would like to know what you think.
Robb

 

Utoh.

Hi, Robb.
Somehow, I missed seeing your prior post. There's so much going on in the forums that it's easy to miss some posts.

I understand you wanting to reduce the losses in the wiring, but the MOSFET driver really needs to be located right at the MOSFET. By separating the MOSFET from the driver board, you add inductance in the charge/discharge path to the gate of the MOSFET, at a rate of about 15nH per 10mm (0.4") or 37.5nH per inch. This adds up very quickly.

You appear to have R6 (4R7) and R7 (10k) mounted on the driver board. R6 and R7 must be right at the MOSFET gate; R7 connected from gate to source.

I don't understand the need for separating the driver from the MOSFET. It would help if you explained this need.

If space is a problem, then things might be done to make the circuit more compact.

With 500uS/Div selected, it's difficult to make out the rise/fall times on the gate. It looks pretty "snappy" going from 0v to >10v, but discharge still seems somewhat slow. This is likely still due to the slow turn-off time of the 4N36 optoisolator. (I suffered some hair loss trying to improve that portion.) C2 was added to snub the oscillation that resulted from the slow turn-off.

What would be interesting to see now is the current flow through the motor. You can get an idea of that by using the MOSFET drain as a reference (ground) and measuring the voltage at K1-87A or K2-87.

The wire between those two points has some resistance. If you tell me the AWG of the wire and the length of it, along with posting a shot of your o'scope measuring it, we can see what your average current is. It should look more or less like a sawtooth (triangular) waveform.

The smaller the p-p amplitude of the waveform, the better. To decrease the amplitude, the PRT (frequency) of the PWM is increased. However, higher frequency PWM has proven difficult (eg: crispy MOSFETS) with the existing driver.

Looks like I need to work on that driver some more, and/or come up with an integrated optoisolator/driver that will work with your 12v supply.

Another possibility is to use those HCPL-3120 drivers (did you buy any of them?) with a DC-DC boost converter. The converter could be made fairly efficient. The average current demand would depend on the PRT (frequency) of PWM.

 

I understand you wanting to reduce the losses.
Before I look at going to 24volts, I just want to make sure I am in the upper part of the efficiency curve and not wasting energy I need to run the motors. I don't need to get every ma out of the circuit.

I don't understand the need for separating the driver from the MOSFET.
This was done for a couple of reasons. I wanted to keep the 12 volt small current driver section separated from the (not sure what voltage) high current section.
This made the sections easier to maintain, wire, and test. Picture below.

R6 and R7 must be right at the MOSFET gate.
This can be done easily and the wires can also be shortened, I just like a little room for mounting and prototyping and testing.

If space is a problem, then things might be done to make the circuit more compact.
Space is not a problem at this point. I am just not sure where things will ultimately be.

You can get an idea of that by using the MOSFET drain as a reference (ground) and measuring the voltage at K1-87A or K2-87.
I measured 5.5 mV on 4 inches of #14 wire with a VOM. See picture of scope. This is good, now I can do the same thing with just pure DC and compare the two.(right?) ok with source to drain shorted measured 30 mA at the same place. Motor is running a lot faster with direct current.

I came up with .0032 ohms per foot. or about .6 amps. Not much current.

Again, not looking for the absolute best, Just trying to figure out if it will do the job.

 

4 inches of AWG 14 should measure about 0.842 milliohms.
If you were measuring 5.5mV p-p, that's a ripple current of about 6.535 Amperes.
5.5mV/0.842mOhms = 6.535 Amperes.

However, it looks like your 'scope readout is saying something like Vm=5.698uV over on the right, but to the left of it is stated VP=57.55555mV. I don't know how to decipher your scope.

 

I pulled the scope picture it was wrong. and
ok with source to drain shorted measured 30 mA at the same place. Motor is running a lot faster with direct current.

sorry about the confusion. having trouble with the scope going to try a real scope and let you know

scope shows about 10mV P-P the sawtooth looks to be about 1.5 mS and as I remember we are at about 500 Hz on the PWM

 

I found that the computer was compensating for the left engine not running and I could actually get more speed out of the motor adjusting the stick.
I work on that seperately.

So I am confused as to tell how well the mosfet is working.

 

I found that the computer was compensating for the left engine not running and I could actually get more speed out of the motor adjusting the stick.

Why does this sound like a pod race?

Very interesting thread.

 

ok here is the DSO view 9.6 mV and 1.4 mS