Solution Finder Get Tech Support

PLC810 PFC not working correctly (?)

Posted by: mklinke on

Hello,

i am designing a psu around a PLC810 and a TNY278 as support psu. I am testing the PFCallone, so the Transformer is not connected till now. The pfc seems to startup, it begins to swith the PFC Mosfet at around 120kHz with one or several cycles, and than continously turning on the gate, driving the choke into saturation. After a continously pulse o approx. 24ms duration the gate is turned followed by a series of ~120kHz pulses. This happens periodically at around 7Hz producing a clearly audible clicking noise from the choke. To protect the Mosfet from the high current following the saturation i connected a power Resistor in series to the input to limit the current. The board is connected to 110Vac from an adjustable isolation transformer. Vbulk reaches ~95V with a 3W load connected to the support psu.Without load Vbulk reaches ~140V. I tried several current sensing resistors. In the moment a 0.1Ω is used, producing a sinusoidal negative voltage on ISP. Interestingly Vcomp is always 0V. I tried several current limiting resistors and no resistor. But the current gets to high, killing the Mosfet. The Choke is an RM14 with an N87 ferrite. TEK00008 shows the input current (channel 1) displayed as 100mV/A and channel 2 is the Drain to Source Voltage showing the long duration pulses at ~7Hz. TEK00009 is a train of pulses when the long-term gate-on period is switched off. TEK00011 is showing Voltage on ISP.

regards,

Martin Klinke

Files

Attachment Size
TEK00008.png 15.38 KB
TEK00009.png 9.42 KB
TEK00011.png 13.22 KB

Comments

Submitted by PI-Tucker on 07/30/2012

What is the calculated saturation current limit of your RM14 choke?  What is the inductance?  Number of turns?

 

 

Submitted by mklinke on 07/31/2012

i am in my holidays for the next 3 weeks. I created a new Pixis Worksheet. Inductance should be 121µH and is 110µH measured at 100kHz. I included the rebuild worksheet (not completed). Thank you for your help in advance...

Submitted by PI-Tucker on 08/01/2012

Pls. take a scope shot of the ISP voltage *and* the choke current at the same time, with about 5-10 us/div.

Submitted by mklinke on 08/14/2012

...i took the shot. The Ipfc Sense Resistor is a 100mΩ at the moment (just for testing).

Submitted by PI-Tucker on 08/20/2012

In reply to by PI-Tucker

The inductor isn't saturating in this screen shot.

Slowly decrease the value of the sens resistor (e.g. in 20% increments), and look for saturation (current becomes a lot steeper at a certain current)

Submitted by mklinke on 08/21/2012

You are right, the choke is not saturating. I was back yesterday, and now the pfc is running. The LLC is still deactivated but i connected a 820Ω resistor network as load (at 387V at the bulk capacitor this means an output of approx. 180W). Efficiency at a 230Vac supply is measured as 94%. Fine! (at least for me :) ). The problem was, i was reluctant to increase the max allowable current. Therefore the pfc could not work.

I do not have experience in high power designs. The PFC allone produces high emi. This seems to be submitted by rf (isolating transformer is no help in that issue). And now a further question: I have seen two different kinds of current waveform at 180W output. One was an specifically fine sine (i saw these 2 hours ago) and now the waveform shows some noise (i can't tell if this is normal). I am sorry, i have no picture of the first! Could this be a temperature issue? Room temperatue climbed to nearly 28°C here :(

In the shot, ch1 is current at 100mV/A, ch2 is Voltage (both at the mains).

Submitted by mklinke on 08/21/2012

thank you in advance...

Submitted by PI-Tucker on 08/21/2012

You'll need to zoom in on those current spikes, and see what the actual choke current is doing.  Use a current probe on the choke, or look at the voltage on the sense resistor.

 

Re: EMI - are you measuring this with an EMI setup? 

It was just good guess seeing those spikes and hearing overwhelming Noise in surrounding Radios. But i think the noise is coupled to the audioamps, not in the uhf receivers?

I took some shots at 90W Output. Because i am not in possession of a small current probe i measured Choke Current at the sense Resistor (50mΩ). The Current Spikes on the Mains seems to correlate to the pfc switching. I wondered why, at a constant  output, the current is sometimes nearly twice as high as average. Or is this just a badly constructed EMI Filter?

Submitted by PI-Sarek on 08/28/2012

Hi, When viewing a current using an oscilloscope across a current sense resistor, it is important to use a properly set-up probe to avoid noise pickup, which can generate all sorts of false signals. This is also an issue when measuring the output ripple of an SMPS. The oscilloscope is in essence a sensitive, high impedance receiver, and the ordinary probe with ground lead makes an excellent loop antenna to pick up noise. The attached files show a couple of probe setups that have greatly reduced pickup, The first picture shows a common setup for measuring output ripple using a Probemaster adapter. Wires are soldered to the adapter, and the other end is slipped over the barrel of the oscilloscope probe. For general measurements or for bipolar signals, the capacitors should be omitted. The next picture shows a rough and ready probe setup that reduces pickup. A length of tinned bus wire is wrapped around the ground collar of the probe, and held in place with a couple of turns of tape. The length of the ground wire can be adjusted for the situation. A power factor corrector in essence shapes the current draw of a power supply to approximate that of a resistive load. This means that the current wave form near the peak of the input AC waveform will have a higher value than that near the zero crossings. It is also important to note that the output of the PFC will be no better than its input. If the AC input waveform is already distorted, this distortion and its accompanying harmonics will not be corrected. The power quality of a standard AC outlet can vary considerably during the course of the day, depending on the loads in the vicinity. This is especially true if large loads using capacitive input filters are present on the same branch of the power line. In order to accurately determine the effectiveness of a power factor corrector, a sine wave AC source is necessary. An ordinary AC source can be used for the initial stages of bringing up the power supply.