I require a circuit that compensates for a varying active load by applying another load that varies with the actual load.
Basically I have a DC power feed that powers a load which modulates the voltage to communicate with devices on the line. The device comunicate back using current pulses. However some devices could have intermittent load characteristics and could draw enough change in current to look like a data pulses.
Is it possible for a circuit to be made that allows a constant current draw from the line say 40mA with the actual load taking 5mA and when the actual load jumps up to 25mA the current draw circuit would reduce its consumption so the actual load remains at 40mA a kind of reverse current limit.
Do you have control of the receiver design? Do you know any of its details?
I suspect that if the current draw on the line changes slowly enough then it won't be interpreted as a data pulse, ditto for small enough current changes. So the problem is not one of making the current constant, it's one of making the current constant _enough_. Depending on the details of your receiver and load changes this could range from a simple LC filter between your 'real' load and the outside world, to a fancy switching supply that _looks_ like an LC filter to the outside world. Since it's job would be to make sure that its input power varied slowly given a rapidly varying load you could think of it as an un-regulator -- presumably this means that you'd need to follow it with a honking big capacitor, then a re-regulator.
Two linear regulators: one wired as a constant current source (with a dropout voltage of 1.25V for the typical circuit) and the other to give you your regulated voltage (possible a low dropout regulator). As long as you keep the load current under your constant current device, you're golden.
My favorite tiny current regulator was an LM317L in a TO92 package with an
0603 resistor to set the current between Vout and ADJ. See page 9 of
and you'll see that the ADJ is the output of the simple 2-terminal current regulator. I snipped the Vout pin after the 0603 resistor and I had a cute little 0.100" spaced 2-terminal device. I used it for a blue LED driver when blue LEDs were new and I wanted to keep the current to 20 mA, not 21.
(apologies to those who hate top posting)
"Paul" wrote in message news:dvmkd2$89o$1$ email@example.com...
Don't have access to the design of the receiver not in or domain. current pulses are approx 20mA with a duration of 240 to 270uS
our device can go from 6 to 30mA depending on what its doing. It does this in burst of some 30ms although at stages could be for as long as 2 seconds.
There are more complications, we need to keep the circuit powered for some
10Seconds on loss of power while drawing some 5-10mA at 3v3 this means I already have a 100mF capacitor for bulk storage. However I'm not allowed to draw more than 1mA for the first 100mS and thereafter preferably no more than 40mA. I already have 2 current limiting circuits on the main DC power feed and then a third Current limit to the bulk capacitor after a stepdown. For this reason the bulk capacitor is below the regulated DC voltage so any current spikes can't be take up by this capacitor.
I will try and post a binary to the alt.binaries.schematics.electronic goup later this evening or tommorow.
I just realized that a critical element is missing: the constant current regulator will limit the current, it won't provide the extra current; to do that you need a separate load, perhaps a resistor through a Zener to keep the shunt load from adding to the current when your actual load is near your limit. I'd add a resistor just to keep things from being too sharp but you may get by without that resistor. A zener in parallel with your voltage regulator and load combination, all fed by the constant current (limiting) source.
There is much you haven't said, though. For example, what is the nominal Vin? What is the modulation like? Would it be practical to pass this modulation voltage through a capacitor to the device as I have shown? How much Voltage do your devices need? Is the minimum Vout from your DC power feed high enough above the devices Vin min that this scheme will work? Will the current spike caused when the transistor turns on cause a drop large enough to confuse the device vis-a-vis Voltage modulation?
From a theory perspective, the TL431 will pretty much keep the Voltage at the junction of R1 and R2 constant. This means that the current through R1 will be pretty much independent of the load presented by the device, which is what I think you want.
R4 would have to be chosen to provide the desired 20 mA of current when the npn transistor is turned on.
Below is the initial circuit. However it doesn't include the bulk hold up cap. CL1 is current limiting circuit of 1mA CL2 is current limiting for
40mA under control of the micro.
The holdup ca could either go before the switch reg which has the downside that it needs to be a large voltage, or after the switching reg in which case another current limiting circuit would be required as the length of time take to charge such a cap would be too great the switching regulator would then put out a higher voltage for the cap and regulated for the micro. This would enable the cap to provide voltage from 5v5 down to 3V as the micro can operate down to 2.7. This is shown in Circuit 2. Apoligies for lack of experience on ASCII diagrams
In your circuit the absolute maximum of the TL431 is 37. This is the absolute of my input although only by the modulated data the max DC component is 28V. As the data would be smothed by the Caps after the current limiting circuitry I guess the 431 would be suitable.