1
\$\begingroup\$

I am currently building variable voltage/variable current (lab) power supply out of an ATX switching power supply. I have done voltage regulation tweak of the TL494 so now it gave me variable voltage of approx 2V-15V.

Now, I want to add some current limiting capability as it is in lab power supplies (the other function of this will be short circuit protection because by tweaking the TL494, power supply lost its short circuit protection). Regulation range from 0-5A will be sufficient. (original rating on 12V rail was 15A).

I searched a lot, but all i can find is circuits using bipolar transistors and low value sensing resistors sometimes with opamps.

My idea is to use simple circuit - P-channel MOSFET connected as high side switch and potentiometer to set required current limit. I have chosen high side switch, because i require common ground, e.g. ground will be connected to chassis.

Here is scheme:

schematic

simulate this circuit – Schematic created using CircuitLab

It will be kind of difficult to set desired current by potentiometer (because hreshold opening voltage is approx 4V and then the current will be climbing rapidly till full open transistor. I think multiturn potentiometer will be partly solution.

Is there any cons of this circuit? Is there any better options to do current limiting with MOSFET? Eventually, how can i add current limit signalization?

\$\endgroup\$
  • \$\begingroup\$ I don't have a solution for you just now, but I do want to ask: what does 'current limit signalization' mean? \$\endgroup\$ – Hearth May 5 '17 at 0:58
  • 1
    \$\begingroup\$ @Felthry Presumably s/he wants an LED to come on when the supply is current limiting, like many bench supplies do. \$\endgroup\$ – Spehro Pefhany May 5 '17 at 1:14
  • \$\begingroup\$ You will be able to fry an egg on the P channel. \$\endgroup\$ – Autistic May 5 '17 at 11:04
  • 1
    \$\begingroup\$ @Autistic: I don't see how that's relevant, that'll be an issue with any linear current source with a pass element. Do a proper thermal analysis and provide appropriate heatsinking (maybe with a fan) and there will be no issue. \$\endgroup\$ – Peter May 5 '17 at 16:39
  • \$\begingroup\$ @Peter K .I have not used a linear current source at these powers for a long time because I use a switchmode one instead .Almost all stuff is SMD these days in my neck of the woods .If the proposed linear current source HAD to be SMD it would be a challenge . \$\endgroup\$ – Autistic May 5 '17 at 20:38
1
\$\begingroup\$

Yes, there are a couple of significant drawbacks to this scheme.

  1. The drain-source current is not just dependent on the gate-source voltage, but also the source-drain voltage. For a variable supply, this means that you will have different current outputs at different supply voltages.

  2. The threshold current of a FET is temperature dependent, so temperature will affect the current limit. This is not particularly desirable behavior.

You need feedback in order to produce an accurate current source. There are a lot of high-side current source options available online which use an opamp + BJT, but you have asked specifically for a P-channel FET solution. I think this should do the trick:

enter image description here

Quick explanation of the circuit: Rsense measures the load current, which is amplified and buffered by U3. U3 is just a differential op-amp circuit, and can be replaced with an instrumentation ampllifier if higher input impedance is desired. The output from the current measuring amplifier is fed into the non-inverting pin of U1, which by servo action will adjust its output based on the difference between the control voltage and the sensed voltage until the difference is zero (ie: current is in regulation).

The output is fed through a differential opamp configuration (U2) which performs the operation Vout = V(rail) - V(in). We need this as in order to control the P-channel FET, we want to drive the gate lower and lower from the rail voltage in order to get more current. Without this, as U1 tries to increase the current it would actually decrease it, causing it to runaway (control authority in the wrong direction).

Under feedback conditions, the current will stabilize such that the current through the load is equal to the magnitude of the control voltage. You can set this voltage however you like.

When the load is too large for the compliance of the current source (ie: too high a resistance to operate from the input rail), U2 will drive the gate all the way to ground via feedback action, and the voltage on the load will just be the supply voltage. This is your "voltage source" operation. If you like, you could put a comparator on the gate line to detect when the gate voltage is driven below or above a certain point in order to illuminate something like a "constant voltage" or "constant current" light.

Hope this helps!

\$\endgroup\$
  • \$\begingroup\$ Thank for your answer but the image you postet is not visible. \$\endgroup\$ – x-fox May 5 '17 at 9:14
  • \$\begingroup\$ Interesting, it works for me. Here's a direct link: i.stack.imgur.com/DuUPE.png \$\endgroup\$ – Peter May 5 '17 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.