How to amplify trapezoidal current signal on an inductor

Thread Starter

msamet

Joined Oct 20, 2017
12
Hi,
I'm trying to obtain the following waveform through an inductor having an inductance of 300uH and internal resistance of 0.4 ohms.
upyt-sekil2.png

For visualization, the equivalent circuit of the inductor is actually like this
4.png

My method is to using a square wave signal then integrating it through an integrator to obtain a triangle wave and then clipping the triangle wave to obtain the trapezoidal voltage waveform(If there is an effective or better way to do it please feel free to share).

Then i used voltage to current converter circuit and here is the problem: The current level isn't as high as in the figure. The desired current level is 1.5A and i achieved only mA's level:2.png
What should i do the increase the current level ? (I used power amplifier circuit but i couldn't get to the Amp level)
 

kubeek

Joined Sep 20, 2005
5,794
To get current in a 300uH inductor change by 3A in just 5us you need to apply about 180V to the coil. That is apply +180V for 5ms 5us to get the transition from -1.5A to 1.5A, then apply 0.6V for 400us to keep the current steady, then apply -180V for the transition back to -1.5A.
 
Last edited:

Thread Starter

msamet

Joined Oct 20, 2017
12
I appreciate @kubeek @crutschow thanks. Yes that is the corresponding voltage waveform which i also calculated :

IMG_20191121_212540.jpg





But it must be accomplished by real components and real component values(even with tolerances). Thats why i used a netlist of an op-omp in the picture showing the simulation result.
And thats why i go through obtaining the current waveform from voltage to current converter instead of generating the correspoding voltage and passing it through the coil.

Which one is a much more elegant solution? (going through voltage or current)

Could you help with how to accomplish the result by using real components?

Thanks

IMG_20191121_212540.jpg
 

crutschow

Joined Mar 14, 2008
34,284
And thats why i go through obtaining the current waveform from voltage to current converter instead of generating the correspoding voltage and passing it through the coil.
No matter how you do it, you need to apply that voltage waveform to the inductor to get that inductor current waveform.
Using a voltage to current converter will not help to do that.

So you need a high voltage circuit that can generate that voltage waveform, which is not a trivial design.

What is the purpose of this?
 

Thread Starter

msamet

Joined Oct 20, 2017
12
No matter how you do it, you need to apply that voltage waveform to the inductor to get that inductor current waveform.
Using a voltage to current converter will not help to do that.

So you need a high voltage circuit that can generate that voltage waveform, which is not a trivial design.

What is the purpose of this?
The purpose is to use it as a search coil. You might notice that from these 5us series of pulses.

"Using a voltage to current converter will not help to do that. " @crutschow knew its not gonna work:)Thank you.

My first thoughts when i sketched the voltage waveform was using a switching circuit but i couldn't proceed. Then i used the voltage to current converter. High voltage circuit with that fast switching... Could you give any clue ? (I'm not the type that says: "Can you give me the circuit diagram?" or "Please give me the code of this?" Thats why i asked a clue:)
 

kubeek

Joined Sep 20, 2005
5,794
I would go with a standard high voltage H-bridge and a power supply that can deliver 180V, and the current is only about 20mA on average so that should not be that hard to pull off, you just need large enough filtration caps so that the voltage doesn´t sag much during the pulse. The power supply could be a boost converter, but that depends on what is the power source for the complete thing.
 

Thread Starter

msamet

Joined Oct 20, 2017
12
I would go with a standard high voltage H-bridge and a power supply that can deliver 180V, and the current is only about 20mA on average so that should not be that hard to pull off, you just need large enough filtration caps so that the voltage doesn´t sag much during the pulse. The power supply could be a boost converter, but that depends on what is the power source for the complete thing.
Thanks @kubeek "high voltage H-bridge" would do hopefully. As i said i couldn't carry on with voltage waveform needed someone to guide me:)
 

Thread Starter

msamet

Joined Oct 20, 2017
12
Hi again,
Dear @kubeek, as your suggestion i tried the high voltage H-bridge. Even if i am applying that required voltage the current is only at uA levels.
 

Thread Starter

msamet

Joined Oct 20, 2017
12
Can you show your circuit diagram? Is it in simulation or in real world circuit?
Its in simulation. I started with a standart H-bridge circuit lying around everywhere the one with 4 MOSFETs. To keep it simple in the first place, before implementing with real components in the simulation, i used ideal components and evenif they're ideal unfortunately it didn't work :/IMG_20191214_225235.jpg
 

crutschow

Joined Mar 14, 2008
34,284
How can we help when you don't show what's driving the FET's gates?

Note that the top N-MOSFETs require a gate voltage 10V above the supply voltage to fully turn on.
 

Thread Starter

msamet

Joined Oct 20, 2017
12
How can we help when you don't show what's driving the FET's gates?

Note that the top N-MOSFETs require a gate voltage 10V above the supply voltage to fully turn on.
Yes, sorry i just wanted to show a general thing. The pulse waveform we discussed here before is in the gate.
 

kubeek

Joined Sep 20, 2005
5,794
Here is a corrected version - added more suitable timing and actual mosfets instead of the dummy model. Sadly the 0.4ohm coil resistance is making the current droop quite fast, and to compensate for that you would have to be driving it with a linear amplifier instead of a h-bridge, which would lead to significant power lost in the transitors.
 

Attachments

crutschow

Joined Mar 14, 2008
34,284
the 0.4ohm coil resistance is making the current droop quite fast
You could conceivable compensate that efficiently by using a very low duty-cycle PWM signal to maintain the constant-current phase of the pulse, but that would involve some interesting timing.
Perhaps a microcontroller could do that.

To minimize having to switch two MOSFETs at the same time for the PWM, you could pulse the appropriate bottom-ones, and leave the opposite top-one on, while letting the diode of the othe top-one carry the inductor current when free-wheeling.
The diode drop will increase the rate of inductor current decay, but that's easily compensated by a slight increase in the PWM duty-cycle.
 
Top