Please help with circuit idea for voltage supply with current limiter on impulse load

[kanonka]

Hi,

I'm designing schematic for the following task:
I have a load (unpredictable, i.e. user pluggable) with reasonable values 0 ohm (short circuit) to 10K. Circuit should send impulses to that load (got that part), effectively bringing voltage on the load from 0 to Vx (and back to zero). Vx can change for each impulse. Impulse width is about 0.2ms.

The tricky part is here: Vx should not exceed some arbitrary voltage Vmax, specified by the user input (DAC via SPI), AND current on the load should not exceed some arbitrary current Imax, again specified by the user input (again DAC via SPI).

Now, I've made some progress (DACs are working, as well as current reader and converter to voltage), but what I need, is some circuit that has Vout that:
a) provides Vout with range 0..42V; current should never exceed 80 mA, no matter what load is.
b) takes V1 as an input, and compares Vout with it. If Vout exceeds V1, bring Vout to that level.
c) takes V2 (which is feedback from output current) as an input. If V2 exceeds some arbitrary voltage (can be set to, let's say, 5V), then drive Vout down till V2 goes below that arbitrary level.
d) if load is such that Vout does not reach V1, and V2 is far from a limit - it's ok. Main goal is not to exceed specified voltage and current, and allow for impulse load of arbitrary amplitude within that range.

I can probably design this circuit myself, but right now would welcome any suggestions - may be this was already done by someone, or there is readily available IC, or schematic, or anything? Any idea is welcome!

Thank you!

 

[barry]

A bit confusing. I THINK what you are trying to describe is an adjustable voltage regulator with current limiting. Is that correct?

 

[kanonka]

Sort of. The regular adjustable voltage regulator with current limiting will settle voltage and current after load is attached, but I need to "turn" it on and off with period of 0.2 ms, and it should settle in that peiod.

 

[barry]

I guess what's confusing me is your statement c). You're not really trying to control V2 are you? You are trying to limit the current, right? There's a difference. You can't have a constant voltage AND a constant current unless you've got a fixed load.

The way you've described it, you want to control the output voltage by V1, and the output current by some other "arbitrary input". But the output current depends on the output voltage.

 

[kanonka]

I'm trying to control upper limit of the current. If V1 says that output voltage to be 20V, but V2 says that current is too high at 20V, and should be, for example, 10V instead, then output voltage goes to 10V.

 

[barry]

I'm trying to control upper limit of the current. If V1 says that output voltage to be 20V, but V2 says that current is too high at 20V, and should be, for example, 10V instead, then output voltage goes to 10V.

That's what I'm trying to get you to say: you are not CONTROLLING the current you are LIMITING it.

So, do you actually need a voltage to set the current limit point, or can you do it with a pot? (that would be easier, I think).

 

[kanonka]

No, I really need voltage, as the whole thing is controlled via MCU.

 

[barry]

So you need a voltage-controlled current limiter. I'd have to think about that a while. Have you tried a web search?

A simple current limiter has a small sense resistor in the output line. This resistor is connected between base and emitter of a transistor whose collector goes to the base of your pass transistor. When the voltage across the resistor exceeds Vbe of the transistor, it starts to turn conduct, sucking base current away from the pass transistor and limiting the current. You need to add some circuitry that will change the dependency on Vbe to depend on your input voltage.

 

[kanonka]

Ok, here is my original variant (R2 is the load):


It worked pretty well, except I didn't like capacitors on the load for pulse smoothing - they should be pretty large, and thus not allowing faster pulses. And anyway, whole circuit was kind of not to my liking.

Now I'm working on this variant:


(Sorry for kind of messy schematic - it's work in progress with LOTS of tries).
Basic idea in this one is:
1) U1 + M1 measure current on R1, converting current to voltage on R6, with multiplication dependent on R6 value.
2) V2 is DAC that provides voltage to compare. If voltage on R6 exceeds one provided by V2, then U2 acts like a switch, opening Q2.
3) Q2 in turn decreases effective value of V3 (impulse DAC) on positive input of U3, this decreasing voltage on base of Q1
4) Decreased base voltage on Q1 decreases voltage on the load, thus decreasing current on R1.

Problem with this approach is that it is very unstable, and falls into oscillation. I understand that this is because U2 acts like a very sensitive switch, which is constantly on and off when voltage on R6 is close to V2. But, I have no idea what to fix, or how change the approach.