I am trying to determine component values to operate a TL494 in a mode that limits the peak current. I am basing this off the datasheet's "Typical Application" (Section 10.2).
My understanding is there are two error amplifiers in the TL494. I can disable one error amplifier by tying the inverting input to the reference voltage on the chip and the non-inverting input to ground.
The other error amplifier needs to limit the duty cycle once the current in the circuit exceeds a certain value. Here is my conception of this circuit
V1 is a 12 volt switchmode power supply.
V2 is the reference produced by the internal regulator of the TL494.
R1 is the load, I have used both resistive and inductive loads.
M2 is any typical N-channel MOSFET with a maximum
V_gs of 20 volts. I have been experimenting with IRFZ34 and IRFZ44.
Not shown is the connections from
VCC of the chip, as well as the necessary capacitors and resistors to form the oscillator of the TL494. We can assume some low operating frequency like 10 kilohertz.
This circuit drives the MOSFET on and off in a common collector configuration. When the output transistor is on, the combination of R6 and
R5 form a voltage divider which charges the gate. When the output transistor is off
R5 discharges the gate. This is not the best circuit, but I mounted
R5 right on the MOSFET itself. The same is true for
C1. Even up to 180 kilohertz operation I don't see any ringing on the gate. Turn on time is about 0.5 microseconds, turn off time is about 1 microsecond. The components
C1 were added to reduce voltage spikes seen on the source of the MOSFET. Even if I use a purely resistive load for
R1 there are still some spikes, these components reduced the spikes significantly when measured with my oscilloscope.
The combination of
R4 should produce a reference voltage of 2.5 volts. Section 7.3 of the datasheet indicates a recommended "Amplifier input voltage" of
V_cc - 2. This would be 10 volts in my application, so I am well below that. Forward voltage drop across
R2 should produce the same voltage when the current in the circuit is equal to 25 amps. My understanding is that the output turns off early in this case, reducing the duty cycle.
The selection of
R7 is not something I fully understand. My limited understanding is that omitting the resistor would give the amplifier "infinite" gain, which isn't possible anyways. I basically just chose
R7 because I had it and because the datasheet uses a 50k resistor.
If I ground pin 16 of the circuit this circuit works as would be expected, with no change in the duty cycle. If I connect pin 16 as shown, the duty cycle is lowered but not by any logical amount. With this connected, I see oscillation of
V2 up to a value of about 7.5 volts. This seems to indicate I've made some fundamental mistake, as I must be somehow affecting the operation of the internal reference regulator.
Why am I seeing oscillation of the reference voltage? How can I properly select
R7? Is my understanding of the error amplifier flawed?