measuring floating inductor current in a truesine inverter application

[Jamie Morken]

Hi,

I am trying to make a true sine inverter outputting 120VAC 60Hz,
and am using 200kHz for the fet switch. Also using software PWM
control with either of one or the other of these two algorithms:


method1: (uses high speed DAC)

if inductor current > voltage error, pulse off

if inductor current > current_limit, pulse off

if time > limit pulse off (limit is our max dutycycle for this PWM period)


method2: (uses high speed ADC)

if inductor current > current_limit, pulse off

Duty cycle = ((desired voltage - current voltage) * gain) - (Inductor
current * gain2)



For the current sensing I was thinking of using a hall effect sensor,
like the "S22P025S05" from Tamura,

http://parts.digikey.com/1/parts/945709-sensor-current-25a-2-5v-pcb-s22p-s22p025s05.html

which has an operating bandwidth of up to 200kHz, but this is still too
slow to do cycle by cycle current limiting unless a larger inductor is
used than is otherwise necessary for the circuit. Is there a higher
bandwidth hall sensor that could work for this?

Also I was thinking about using a current sensing resistor and then
feeding this directly into an ADC floating at the high voltage of the
120VAC, and using optoisolators on the ADC SPI interface.

Another possible way to interface to a current sensing resistor would
be to use a high common mode voltage differential amplifier, like the
AD629.

http://www.analog.com/en/prod/0,2877,AD629,00.html


What would be other good ways to measure this high frequency
floating current?

cheers,
Jamie

 

[Paul Mathews]

Hi,

I am trying to make a true sine inverter outputting 120VAC 60Hz,
and am using 200kHz for the fet switch.  Also using software PWM
control with either of one or the other of these two algorithms:

method1: (uses high speed DAC)

if inductor current > voltage error, pulse off

if inductor current > current_limit, pulse off

if time > limit pulse off (limit is our max dutycycle for this PWM period)

method2: (uses high speed ADC)

if inductor current > current_limit, pulse off

Duty cycle = ((desired voltage - current voltage) * gain) - (Inductor
current * gain2)

For the current sensing I was thinking of using a hall effect sensor,
like the "S22P025S05" from Tamura,

http://parts.digikey.com/1/parts/945709-sensor-current-25a-2-5v-pcb-s...

which has an operating bandwidth of up to 200kHz, but this is still too
slow to do cycle by cycle current limiting unless a larger inductor is
used than is otherwise necessary for the circuit.  Is there a higher
bandwidth hall sensor that could work for this?

Also I was thinking about using a current sensing resistor and then
feeding this directly into an ADC floating at the high voltage of the
120VAC, and using optoisolators on the ADC SPI interface.

Another possible way to interface to a current sensing resistor would
be to use a high common mode voltage differential amplifier, like the
AD629.

http://www.analog.com/en/prod/0,2877,AD629,00.html

What would be other good ways to measure this high frequency
floating current?

cheers,
Jamie

CT.
Paul Mathews

 

[Jamie Morken]

CT.

current transformer, Ya I guess that gets rid of the high commom mode
voltage, and then it is just a differential measurement that is
required?

Thanks I will look into this more.

cheers,
Jamie

 

[Fred Bartoli]

Paul Mathews a écrit :

CT.
Paul Mathews

Simpler: if that's possible, just add a few turns winding to your
inductor. It'll give you dPhi/dt that you then integrate (can be as
simple as a lowpass RC).

 

[Jamie Morken]

Paul Mathews a écrit :

Simpler: if that's possible, just add a few turns winding to your
inductor. It'll give you dPhi/dt that you then integrate (can be as
simple as a lowpass RC).

That sounds like it would be the cheapest method. Since that is
integrated/filtered, will it be too slow to detect cycle
by cycle current for comparator based current limiting?

I found a high frequency current transformer, would this do the job?

http://shinhom.com.cn/pdf5/142.pdf

cheers,
Jamie

 

[Fred Bartoli]

Jamie Morken a écrit :

That sounds like it would be the cheapest method. Since that is
integrated/filtered, will it be too slow to detect cycle
by cycle current for comparator based current limiting?

Slow?
You just integrate di/dt.

The only missing info is the DC current, but you the same pb with a CT.
and under some conditions you can restore it.

 

[Jamie Morken]

Jamie Morken a écrit :

Slow?
You just integrate di/dt.

The only missing info is the DC current, but you the same pb with a CT.
and under some conditions you can restore it.

Hi,

since current transformers only work for AC, and the only AC here is the
60Hz output, I don't think current transformers can work to detect the
high frequency current.

cheers,
Jamie

 

[Jamie Morken]

Hi,

since current transformers only work for AC, and the only AC here is the
60Hz output, I don't think current transformers can work to detect the
high frequency current.

So are there any other options to measure the ripple of a 200kHz
current?

cheers,
Jamie

 

[Jamie Morken]

Jamie Morken a écrit :

Slow?
You just integrate di/dt.

The only missing info is the DC current, but you the same pb with a CT.
and under some conditions you can restore it.

Here's the circuit I am working with and the 200kHz current waveform for
one of the output inductors:

http://rocketresearch.nekrom.com/ne...inverter/240VAC 60Hz split phase inverter.jpg

http://rocketresearch.nekrom.com/new/split phase 240VAC inverter/current waveform.jpg

The current waveform is DC at this frequency, but is AC at 60Hz.

I don't think a current transformer can work due to the fact that this
is effectively a DC current, haven't found a fast enough hall sensor for
detecting over-current, and a shunt amplifier would have to be floating
at high voltage.

cheers,
Jamie

 

[Jim Thompson]

Here's the circuit I am working with and the 200kHz current waveform for
one of the output inductors:

http://rocketresearch.nekrom.com/ne...inverter/240VAC 60Hz split phase inverter.jpg

http://rocketresearch.nekrom.com/new/split phase 240VAC inverter/current waveform.jpg

The current waveform is DC at this frequency, but is AC at 60Hz.

I don't think a current transformer can work due to the fact that this
is effectively a DC current, haven't found a fast enough hall sensor for
detecting over-current, and a shunt amplifier would have to be floating
at high voltage.

cheers,
Jamie

You can't use a clamp-on probe?

...Jim Thompson

 

[Terry Given]

So are there any other options to measure the ripple of a 200kHz
current?

cheers,
Jamie

LEM DCCTs. hall-effect devices. design an AC CT that wont saturate with
the DC current.

Cheers
Terry

 

[Terry Given]

You can't use a clamp-on probe?

...Jim Thompson

you have some sort of floating gate-drive for the upper FETs. a current
sense resistor in the output lead(s) sits at the Source of (one or both
of) the upper FET, so there is where you can sit some circuitry to then
transfer that signal to the control circuitry. there are some current
sensing optos, but they tend to want large voltage drops (around 0.5V)
which might not be fun, depending on your power levels. of course a
small Rs, and amp and some circuitry can "fool" the optos...

Cheers
Terry

 

[Paul Mathews]

Here's the circuit I am working with and the 200kHz current waveform for
one of the output inductors:

http://rocketresearch.nekrom.com/new/split phase 240VAC inverte...

http://rocketresearch.nekrom.com/new/split phase 240VAC inverte...

The current waveform is DC at this frequency, but is AC at 60Hz.

I don't think a current transformer can work due to the fact that this
is effectively a DC current, haven't found a fast enough hall sensor for
detecting over-current, and a shunt amplifier would have to be floating
at high voltage.

cheers,
Jamie



- Hide quoted text -

- Show quoted text -- Hide quoted text -

- Show quoted text -

Provided that the DC doesn't saturate its core, a CT will couple the
ac component of the current to its loaded secondary. This is done all
the time. The appropriate core size and material will handle the dc
offset. -Paul Mathews

 

[Jamie Morken]

Provided that the DC doesn't saturate its core, a CT will couple the
ac component of the current to its loaded secondary. This is done all
the time. The appropriate core size and material will handle the dc
offset. -Paul Mathews

I need to measure the DC offset though for cycle by cycle current
limiting at 200kHz, the AC current is 60Hz, the DC current is 200kHz.

What about using a shunt resistor and dual rail opamp powered by a
local isolated bipolar powersupply like this one:

http://rocketresearch.nekrom.com/new/split phase 240VAC inverter/isolated bipolar supply.jpg

(thats the same kind of supply I'm using to power the floating fets,
except not bipolar)

Could even use this supply to power a floating ADC and DAC, I need about
12bits of current and voltage accuracy so I will need to get the supply
pretty smooth, which may be a problem, maybe the current transformer is
better.

cheers,
Jamie

 

[Jamie Morken]

You can't use a clamp-on probe?

Maybe the circuit inside would work, I guess a DC clamp on probe must be
hall based? I'm making a PCB!

cheers,
Jamie

 

[Jamie Morken]

I need to measure the DC offset though for cycle by cycle current
limiting at 200kHz, the AC current is 60Hz, the DC current is 200kHz.

What about using a shunt resistor and dual rail opamp powered by a
local isolated bipolar powersupply like this one:

http://rocketresearch.nekrom.com/new/split phase 240VAC inverter/isolated bipolar supply.jpg


(thats the same kind of supply I'm using to power the floating fets,
except not bipolar)

Could even use this supply to power a floating ADC and DAC, I need about
12bits of current and voltage accuracy so I will need to get the supply
pretty smooth, which may be a problem, maybe the current transformer is
better.

(forgot the flux draining coil!)

like here:
http://www.hills2.u-net.com/electron/forward.gif

cheers,
Jamie

 

[Terry Given]

(forgot the flux draining coil!)

like here:
http://www.hills2.u-net.com/electron/forward.gif

cheers,
Jamie

use diagonal half-bridge - switch at either end of primary winding,
diodes to opposit supply rails. voila, no need for a "flux draining
coil" - lovely terminology BTW.

Cheers
Terry

 

[Jamie Morken]

use diagonal half-bridge - switch at either end of primary winding,
diodes to opposit supply rails. voila, no need for a "flux draining
coil" - lovely terminology BTW.

Thanks This seems to do the job for the dual fet/diode version with
no extra coil:

http://rocketresearch.nekrom.com/ne... snubbing forward isolated bipolar supply.jpg

cheers,
Jamie

 

[Terry Given]

Thanks This seems to do the job for the dual fet/diode version with
no extra coil:

http://rocketresearch.nekrom.com/ne... snubbing forward isolated bipolar supply.jpg


cheers,
Jamie

yep. and for added fun, using P and N fets (or BJTs), the lower switch
can provide the gate drive for the upper switch

Cheers
Terry

 

[Jamie Morken]

yep. and for added fun, using P and N fets (or BJTs), the lower switch
can provide the gate drive for the upper switch

That one is best of all, thanks!

http://rocketresearch.nekrom.com/ne...orward isolated bipolar supply-1gatedrive.jpg

cheers,
Jamie