# Current Source Project Ideas

Discussion in 'General Electronics Discussion' started by ggbc11, Apr 23, 2013.

1. ### ggbc11

8
0
Apr 23, 2013
Hi,

I'm doing a project for my final year of BEng. As part of it I need to design a module that can take mains supply,ie 240/110VAC 50Hz, and output 0-100A (AC) at a resolution of 100mA controlled by a PIC.

My current solution is to have a variac actuated by a stepper motor controlled by a feedback loop via a current transformer (CT). So the PIC will recieve a set point value ie. 50A, and move the stepper motor comparing the set point to the input from the CT with some kind of PID control. I think this will work but itll be heavy as this is supposed to be a portable device.

I work with distribution and control so three phase theory and protection is more my thing not power electronics but I have some understanding. I would like to ask if anyone could point me in the right direction of a lighter, electronic solution. Im not asking for blueprints as I will research and learn anything I need to make a good project, just a rough concept or research area.

My last project was a single phase inverter controlled by a PIC, could this be backed by a current amplifying circuit and followed by a transformer?

Would appreciate any advice on this,

GG

2. ### Harald KappModeratorModerator

12,348
2,935
Nov 17, 2011
1)
What output voltage (compliance voltage) is required? It can't be 240 V because at 100A this amounts to 24kW! What output frequency (frequencies) is required.

2)
A standard mains transformer (even in the form of a variac) is a voltage source. I.e., you control the output voltage, not the current by the setting of the tap. If your load is variable, you'd have to continuously re-adjust the setting of the transformer to control the current. This will be a slow process (electromechanical). Also the frequency is fixed as it is the same as the maikns frequency.

3)
I'd tackle this task as follows:
- create a low voltage, (required compliance voltage + some headroom) DC source from the maisn supply. The low voltage source must be capable of supplying the load current (100 A). The next steps will be easier if you create a dual power supply (+ and -)

- create a voltage controlled current source that operates on the low voltage supply and regulates the output current with respect to the control voltage (input).

- create a control voltage e.g. by using a PWM output from the PIC. Since the PWM from a PIC is unipolar (0V...Vcc), you need to shift the signal by 1/2Vcc such that:
- PWM=0V is -1/2Vcc at the input of the current source
- PWM=1/2 Vcc is 0V at the input of the current source
- PWM=Vcc is +1/2Vccat the input of the current source

Ad filters and gain stages as required to accomodate the full range.

Instead of using the PWM from a PIC you may use a DAC chip to generate the control voltage. The design decision depends on the required accuracy and the frequency of the AC signal to be generated.

Last edited: Apr 23, 2013
3. ### ggbc11

8
0
Apr 23, 2013

The input impedance of the device this project will be feeding is only 20mOhms which at full scale would only require 2 Volts. So if I set the low voltage at 5-6VDC I can power the PIC from the same rectifier circuit without extra conversion. The required frequency is fixed at 50Hz but it would be a good idea to keep options open for further development later.

For the voltage controlled current source would you suggest researching op-amp circuits as a start? I think I understand the control voltage shift from a PWM to give accuracy ie Vcc=5, DC=0% > -2.5VDC, DC=50% > 0VDC, DC=100% > 2.5VDC. Would the AC signal then come from the supply and be altered by this PWM controlled voltage into the voltage controlled current source circuit? Or have I mis-understood? What about if I was to generate a sine wave using the PWM function of the PIC and control the current source with a DAC (as you suggest) taking inputs from the PIC? Would this be feasible?

Many Thanks,
GG

4. ### Harald KappModeratorModerator

12,348
2,935
Nov 17, 2011
Right, but at 100A the supply needs to deliver 500W, 300 of which are going to be converted to heat in the current source. Not a good idea. Your source voltage should be as low as possible.

Before I come back to this topic, one general discussion:
You can make the current source an analog one (based e.g. on OpAmps) which would be controlled by the smoothed PWM. This will always cause high losses in the current source because the difference between the load voltage and the source voltage times the load current will be dissipated as heat in the current source. So you will need a big heatsink and your circuit is not very efficient.
You could also make the current source a digital one. This essentially means you amplifiy the PWM so the output of the amplifier will be able to drive the 100A. You will then need an LC filter to smooth the PWM current to a steady current. This is a bit liek using a class D amplifier. Note that the inductor for 100A will be very big. As a pro this technique will be comparatively efficient.

Coming back to the introductory sentences: You can also combine the analog current source with a power supply that tracks the output voltage. The idea is to use a switch mode power supply whose output voltage is controlled such that its output is just enough (milli)volts above the load voltage so the analog current source can actively control the current without loosing too much energy in the current source. This, too, will require a rather big inductor in the switch mode supply to carry the 100A. Or apolyphase supply with multiple smaller inductors (as it is used on PC motherboards, for example).

Whichever way you choose, you will need some kind of feedback to achieve an accuracy of 0.1% (100mA at 100A). You can either trust the design of the current source to accurately output current as instructed by the control voltage, or you can measure the real current by an ADC (at least 10 bit, better 12 bit for 0.1% accuracy) and use the measurement to adjust the control voltage.

All these switched techniques, while efficient in terms of energy, are not easily handled, specially not at 100A and for the faint at heart. So it might be a good idea to go back to start and have a look at the requirements:
- 50 Hz (or 60 Hz) only is o.k. So you could directly use the AC from mains (with some transforming etc., of course)
- 2 V compliance voltage is all that is required. So you could basically use a transformer rated 240V/2V. That's a ratio of 120/1, meaning the secondary current of 100A reduces to a primary current of 100 A/120 = 0.833A. This is easily handled by e.g. a variac.
- Nothing has been said about the stability or repeatability of the output current. Let's assume it suffices to set the current once to a given value and that you can tolerate any changes in current due to e.g. changes in mains voltage.
This leads to the following, much simpler setup:
- A variac (possibly controlled by a servo motor and a PIC, as mentioned in your 1st post) generates an output voltage of AC 0V...240V 50Hz with a max. current of <1A.
- A second transformer, rated 240 V / 2 V at >=100 A converts the variac's output voltage to an output current AC 0 A ... 100 A 50Hz.

You can use the PIC to just set the variac. Or you can measure the true output current with the PIC (you need to make an RMS measurement and smoothing so the servo doesn't start to track the sine waveform of the current) and track the variac accordingly to set the current. This will increase accuracy at the cost of the additional current measuring and feedback loop.
It is also a good idea to include some short circuit protection e.g. such that the variac is turned down to 0% if an overcurrent condition is detected at the output. This would be easy once you have installed the feedback loop anyway.

Hope that gives you some ideas?

5. ### ggbc11

8
0
Apr 23, 2013
Thank you again for so much info, very interesting and a great help!

I have until August to do this so I'm going to give the class D type amplifier method some thought and research, I like the idea of ~90% efficiency's and a lighter solution than variacs although more challenging. So far I need to research MOS gate drivers, LC filters, buck converters, efficient rectification (and voltage regulators), 12 bit ADC devices and current transducers and their conditioning circuits.

So the concept being taking the PWM unfiltered (square wave) sine signal from the PIC feeding it through a MOSFET circuit and then filtering the high frequencies through a LC filter. This therefore producing a smoothed sine wave of varying current dependant on the switching frequency of the MOSFET's. Then a feedback loop back to the PIC via a CT and ADC.

I think I might lower my output specification to make things a little easier, maybe 50A or 20A and increase to 100A for a second project once I have a better understanding.

Thanks,

GG  