pwm chip

DigiKey only has rev 3 of the datasheet, ON Semi is up to rev 6.

Many other manufacturers have uc3843 info, TI (who bought Unitrode, who introduced the chip 15 years ago), Fairchild, ST, Microsemi, Linfinity, Bay Linear, Unisonic, etc... The TI UCC38Cxx ICs are advanced low-power BiCMOS versions of the original Unitrode parts.

And check out TI's slua143 U-100A Unitrode 14-page app note.

Normally the pin 1 voltage results from an error-correction feedback loop, so the relevant component in figure 32 is the integrating cap. But you can wire the controller for a fixed current level by using the opamp, or by otherwise forcing a voltage on pin 1. Ground pin 2 to make pin 1 high, which is a roughly 1mA current source, and over- ride that with your setpoint voltage. Keep in mind the approximate nature of the page 8 formula, Ipk = (V1 - 1.4) / 3 Rs, where the 1.4 and 3 factors are uncertain, and a further offset error isn't shown.

Putting the question another way -- I still want to use feedback how would I know the peak inductor current at such times as low feedback voltage on pin 2 causes it to drive full-out. I can't force pin 1 without defeating the feedback. I will have to look at some of those other data sheets for an explanation. Perhaps it has to do with the integrating cap you mentioned.

What feedback, and for what purpose? Explain yourself, man!

This is a circuit that charges a car battery using pulses from an inductor. The idea is to feed back a divided-down battery voltage to pin 2 for regulation when the battery approaches full charge and its voltage rises near the selected set point.

But when the battery is low the pwm chip drives the inductor to some peak current. This being the highest current the inductor will see, I am trying to determine its value. I want to drive the inductor to a peak current that is high but not so high as to saturate it. Let's say for my selected inductor that current is 2 amps. So now how do I determine the value for the sense resistor that will cause the chip to drive the inductor to two amps when the battery is low and the chip is driving full out. Hope I gave a comprehensible explanation. Kell

I want to find the peak inductor current. This bears on the selection of the inductor, not on the average current going to the battery. I asked a straight question about the operation of the chip with voltage feedback. So think of it as a boost converter circuit, with a voltage divider from the output going to pin 2. How does the sense resistor value relate to the peak inductor current when the output voltage is low (low signal to pin 2). If the chip does not operate this way tell me. I can take it.

By the way, Fabio et al. That 2 amps I mentioned is not the charging current going into the battery. You jumped to the conclusion that I really know so little as to equate peak inductor current with average current. I don't even have to know the average current; I can just regulate the battery's voltage, kapeesh? That's why talk about VOLTAGE FEEDBACK. When I put the battery on to charge, I can set the voltage to 14.5 volts or so. Then when I think it has taken on most of the bulk charge, I can set the voltage back to about 13.5 volts. I don't want to consider monitoring charging current right now. So.

Page 8 of the data sheet covers your question. I think. Here is a clip from that page:

Current Sense Comparator and PWM Latch The UC3842B, UC3843B operate as a current mode controller, whereby output switch conduction is initiated by the oscillator and terminated when the peak inductor current reaches the threshold level established by the Error Amplifier Output/Compensation (Pin 1). Thus the error signal controls the peak inductor current on a cycle?by?cycle basis. The Current Sense Comparator PWM Latch configuration used ensures that only a single pulse appears at the Output during any given oscillator cycle. The inductor current is converted to a voltage by inserting the ground?referenced sense resistor RS in series with the source of output switch Q1. This voltage is monitored by the Current Sense Input (Pin 3) and compared to a level derived from the Error Amp Output. The peak inductor current under normal operating conditions is controlled by the voltage at pin 1 where: Ipk = (V(Pin 1) ? 1.4 V)/(3 RS)

Abnormal operating conditions occur when the power supply output is overloaded or if output voltage sensing is lost. Under these conditions, the Current Sense Comparator threshold will be internally clamped to 1.0 V. Therefore the maximum peak switch current is: Ipk(max) =1.0 V/RS

Before your battery reaches the voltage set point, you are effectively in overload condition, so the second formula applies to the current limit. RS is the current sense resistor in the source lead of the external mosfet.

Once the voltage approaches the setpoint, the voltage error amplifier lowers the current limit and the first formula takes over. The voltage on pin 1 (which is the output of the voltage error amplifier) depends on the gain and compensation network you connect between pin 1 and pin 2 and the voltage divider network between pin 2 and the output voltage.

"kell" ha scritto:

Batteries are charged with the mean current value, not with the peak current value. In discontinuos mode (which often appears in battery charging) peak current and mean current are very different. You may encounter unaccurate battery charging without a proper compensation.

Average current mode can be obtained with the slope compensation technique.

Bye

-- Per rispondermi via email sostituisci il risultato dell'operazione (in lettere) dall'indirizzo

bingo.

IOW, your voltage feedback circuitry saturates the EA, up until Vbatt reaches some suitable value (perhaps 12V, whatever). During that time, thye 1V zener on the internal CS comparator input sets the reference voltage to 1V (ish - read the datasheet for the tolerance, it is given). So Ipeak = 1V/Rs

you dont *have* to have an integrator for the EA BTW - fixed gain is also possible, as is any other arbitrarily complex feedback network. fixed gain might be just what you want....

Cheers Terry