H-Bridge: Current Sensor Placement

[dknguyen]

Hi. I was just wanting to run this by someone else. WHen you place a current sensor in series with the H-bridge you measure the battery current going to the H-bridge, while when you place the sensor in series with the motor you measure the current going through the motor right? So the motor way lets you measure the current in the motor at all times (including flyback when zero battery current is flowing so you don't need a lowpass filter), but you do need a low pass filter when using the H-bridge way right? Since there will be times (when the transistors are off) that the current will read zero?

I haven't seen too many schematics of bridges like this, so I'm just wondering if they always stick a low pass filter on the current sensor if it's in series with the H-bridge.

 

[picbits]

All the H bridge motor control circuits I've used have had the current sensing resistor in the ground feed to the lower transistors.

 

[dknguyen]

All the H bridge motor control circuits I've used have had the current sensing resistor in the ground feed to the lower transistors.

Do they all have low pass filters on them?

 

[picbits]

The ones I use do as I use PWM to modify the motor current. Without a filter all you'd see would be the PWM rather than the average current (which is what I monitor).

 

[Hero999]

I'd put a low pass filter on all current sensors used with DC motors because the current will always be periodically interrupted by the commutator in the motor which will generate noise.

 

[dknguyen]

I'd put a low pass filter on all current sensors used with DC motors because the current will always be periodically interrupted by the commutator in the motor which will generate noise.

I'm trying to decide whether or not to use 1 current sensor at the H-bridge with a LP filterpower terminals to measure battery current since it's simpler but I lose the ability to measure motor current during flyback.

Or to use 3 current sensors with no LP filter on each phase of the BLDC so I can monitor all currents all the time. It's a brushless motor- so there's no sparking going on. I don't think that would still interrupt current readings like with brushes would it?

 

[Nigel Goodwin]

I don't think that would still interrupt current readings like with brushes would it?

I don't see why not?, they work in exactly the same way as a normal brushed motor - by switching between different windings.

 

[dknguyen]

Wait a second, is Hero999 referring to commutation distorting the measurement of the current waveform, or is he referring to how current will continuously jump between windings due to commutation. (If it's the second one, that's fine because the uC is controlling which windings have current in them and it will know which one to monitor). The way it's set up right now, each of the 3 phases of the motor has an independent current monitor.

 

[Pommie]

Commutators don't interrupt the current unless you use the 2 pole demonstration type. If you think about the simplest self starting motor with three windings, when one brush is in the middle of a commutator segment, the other brush is straggling the other two. On some multipole motors the brushes are bigger than the individual commutator segments.

Mike.

 

[Pommie]

dknguyen,

What advantage is there to having three sense resistors rather than one. Surely the current will be the same in all three phases. Do you currently see any difference between phases?

Mike.

 

[dknguyen]

No operaitonal purpose. BUt I thought it might help me to characterize the particular motor if I Could record and graph the current through all the phases so I would know how far the current peaked and dipped during a PWM cycle. It wouldn't matter if the current was different between the phases unless it was more than 10% (the error between current sensors is 10%). I'm mainly trying to be as thorough as possible with this driver since it's costing me a buttload to make anyways.

Since I haven't actually built the circuit yet or have the motor, I dont know if the current is different in different phases (I sure hope not! because then my sensorless scheme gets thrown out the window)

 

[Hero999]

Commutators don't interrupt the current unless you use the 2 pole demonstration type. If you think about the simplest self starting motor with three windings, when one brush is in the middle of a commutator segment, the other brush is straggling the other two. On some multipole motors the brushes are bigger than the individual commutator segments.

Mike.

Try connecting a small motor in series with an 8hm: speaker to a 3V battery and it'll buzz at the same frequency as the motor.

Wait a second, is Hero999 referring to commutation distorting the measurement of the current waveform,

Yes, and it doesn't matter whether it's electronically or mechanically commutated, the result wiil be the same.

 

[dknguyen]

*whistles* I just got around to pricing the parts for two of the controllers and a ridiculous amount of the costs are in the sensors. So I guess I'm going to use only one current sensor to measure battery source current and then use just one thermocouple for motor temperature (maybe a leaded thermistor if I can find one) and replace the heatsink temperature sensors with some cheapy thermistors. It's more than $330 worth of parts! Jeez. That's not including the PCB.

8 AD597AR-ND IC THERMOCOUPLE COND 8-SOIC $6.28000 $50.24
6 620-1106-ND IC CURRENT SENSOR 50A BI CB-5 $5.86000 $35.16
6 478-4528-1-ND CAP CER 10UF 50V X7R 10% 2220 $6.61 $39.66

I'm pretty sure I can reduce the size of the bootstrap caps by a lot since I seriously overrated them (even for if they were recharged once per commutation). I'm using a complimentary switching scheme and the bootstrap caps get recharged after every PWM period (as opposed to requiring a commutation to occur if I didn't use complimentary switching).

 

[Pommie]

Try connecting a small motor in series with an 8hm: speaker to a 3V battery and it'll buzz at the same frequency as the motor.

I didn't say the current didn't vary. I said it wasn't interrupted.

Mike.

 

[dknguyen]

Happy news, I found a dirt cheap RTD (as far as RTDs go) that's like $4 from Digikey- the TD5A. It also has an insulated back (rather than connected to ground), so I can place it right against the MOSFET heatsinks or pads to measure the temperature without having it get all screwey because the voltage at those points. It's also quite high resistance (1.5K minimum). Since it's just a local PCB temperature measurement and the resistance is so high, I can just use a resistive divider or something right? No need for fancy conditioning circuits or whatever?

I am just going to go with the single current sensor to measure bridge current rather than motor phase current- frees up a lot of inputs for measuring and limiting of voltages and temperatures...and it's way cheaper...$50 cheaper. I only had one fault limiting input and one ADC input to measure the temperature of 3 phases though. I am considering just sticking them in parallel and assuming they are always at equal temperature (I doubt one driver phase will be 25C and the other is 150C to skew readings that much).

 

[3v0]

One of the reasons I liike this forum is that people bring up things and parts I have never heard of.

Thanks for posting the number and source

 

[dknguyen]

Yeah, I looked at RTDs a long time ago (reflow oven and general temperature sensing) and they were so expensive and required so much signal conditioning so I moved to thermocouples for the most part and silicon sensors. BUt i needed to measure past 100C on a PCB now. I was so happy when I saw some for so cheap I almost cried (well, in my mind anyways...figuratively speaking...weird huh?)