NCP1653 PFC Controller Question

I'm using the same controller in a PFC circuit, in an application where I cut the output voltage down to 270V by omitting one of the voltage feedback resistors. The output voltage is not intended to be precisely regulated. It will vary by about 5% or so, which is OK for our application. It is intended to be used as a preregulator.

There is also the "boost follower" mode described in the data sheet, which we did not implement.

We are using an inductor of about 470 microhenries and an IRFP360LC switching FET witih a 10 ohm resistor in series with the gate terminal, although the data sheet shows less. FETs have a tendency to go into parasitic oscillation if the gate resistor is too low in value.

You may wonder about the large value resistor from the bridge rectifier output into the chip through a lower value resistor, with a .033 uF capacitor to ground. I did a live chat with the manufacturer about it, because the break point calculated to be only 11 Hz. It is intended that way, to get an average DC component into the chip. There is also a capacitor from one pin (maybe pin 3?) to ground which sets the closed loop control bandwidth. We had overshoot problems when the load was suddenly removed, as we only have 10 uF of output capacitace due to space constraints (electrolytic capacitors are not allowed since this is an airborne application, and we chose the chip as a COTS part based upon its temperature range). The referenced capacitor sets the control loop bandwidth. We were able to reduce it by a factor of 6 since our power supply is 400 Hz. With the value given in the data sheet (.033 or .33 uf, I forget which), the manufacturer stated the bandwidth to be only 15 Hz. Cutting the capacitor size down solved our problem, but I doubt if it would work at 60 Hz since you would begin to sacrifice power factor correction capability from an input current waveform standpoint.

Remember that the output rectifier has to be fast recovery, like an MUR type, preferably not more than 50 nanoseconds.

Hope this helps. TI live chat is not bad if you need it and are a reasonably fast typist.

BernieK

Thanks for your reply, I will do your recommendations then I will write you again the results

Erdinc

Hello,

I tried the circuit again with capacitor at pin2 330nF, and the other capacitor which makes RC constant of sampling the bridge output is 33nF ( with R=470kohm). What about the capacitor and resistor of pin5 which provides Vm for pfc. Did you operate in average current mode peak current mode?

When no load, pf is 0.6.

My circuit is ok when the load is only 2-3 watts. Power factor is about 0.96 but when the load is larger than 3 watt ( max. output power is 100W) the switch is going out of control and input current reaches up to 2-3 amper at 100 Vac input. And power factor changes between 0.3 and 0.6 in very short durations. Have you met this problems when you were working on this controller?

Erdinc

Zanika: Are you certain that the inductor you are using is not saturating? If you have a current probe (especially one which will read the DC component), take a look at the inductor current. It should be a sawtooth with a relatively small AC component compared with the DC component since it customarily operates in the continuous conduction mode. If you don't have a current probe, insert a a low value current sensing resistor (maybe .01 ohms) in the FET source. Observe the voltage across it with a piece of coaxial cable connected directly across the resistor going to the oscilloscope. It should be a linear ramp, not a "ski slope" aside from an initial spike due to effects of parasitic FET capacitance during FET turnon. CAUTION: If you do this, be CERTAIN to power the circuit from an isolation transformer, or (ONLY IF YOU ARE VERY CAREFUL), isolate the scope safety ground, and DO NOT TOUCH ANY METAL PARTS OF THE SCOPE. Turn the circuit off to make adjustments, unless you are ABSOLUTELY SURE that you are not grounded. Common leather sole shoes will conduct enough on a slighly damp vinyl tile floor over concrete to provide a slght shock.

The current probe is the best bet. Floating the oscilloscope is inherently dangerous.

One additional consideration: Are you certain that you do not have the PFC circuit output common connected with one leg of the input? The output and what operates from the PFC circuit must be totally floating from the AC line and house ground. That requirement is due to use of the bridge rectifier at the input end of the circuit

BernieK

Hello again,

I almost solve the problems that you told. But a new problem arises.

I have constructed more power PFC circuit with NCP1653 again. I have questions about starting inrush current. Did you have large inrush current at starting? I used NTC but still high inrush current. Did you use the same EMI filter that is given in the NCP1653 datasheets and application notes?

... I get a fancy current curve, measured on input AC line.
The current seems to be of higher frequency. The inductor
is 150uH/16A. I use SiC-Diodes and a 20N60C3 MOSFET.
Has anyone an idea? The (light) load is about 2 A, but
even under higher or less load the curve doesn't look like
it should...

Have you solved the problems meanwhile? I have quite similar
problems and interested in your solution...