UCC2818: UCC2818

Hi Moha,

The UCC2818 is intended for single-phase (input) PFC control.  Are you using three UCC2818 controllers?  Can you provide more detail (diagram would be great) on how things are connected?  Also, can you provide scope capture of the input voltage and input current on the same plot?

Thank you,

Ray

Hi Ray, 

Thank you for your replay.

I m using one UCC2818 controller between two lines (A & B phases). I replace the neutral line by B Phase. 

In fact, the PFC input voltage is between 166Vac (96Vac*sqrt(3)) and 225Vac (130VAC*sqrt(3)). 

Here below a power supply diagram :

Here below and example of the input voltage and current: 

- RED one : input current on A Line 

- Green one: input voltage between two lines A and B 

Tank you for your support, 

Moha

Hi Moha,

You should be getting a good PFC with the A/B inputs just as you are with the A/Neutral.

I have a few questions/comments on your setup:

Do you have a scope probe ground connected to P_GND of the UCC2818?  If so, this will result in bad PF.

Are you using isolated AC source(s)?  Any non-isolated source will introduce unwanted ground paths.

It seems the line-to-GND cap isn't present in A-B mode.  Is the scope AC-coupled?

When comparing the phase currents, make sure you remove all voltage probes to avoid introducing unintended ground paths.

This is all I can think of right now with what you've shared with me.

Hi Ray, 

thank you for your replay, 

I dont know the reason why I dont have a good THD with A/B inputs just as I am with Neutral/A inputs.

I dont have any scope probe connected to the primary ground GND_P. I made some measurements with differential probe to control the input voltage. But we I remove it, there is no difference.

For my Harmonic measurement, I m using RL circuit between the power source and the board (R=1ohm and L=500uH), this set up is described in the DO160.  When I remove the RL circuit, there is no difference. 

I m not using a transformer to isolate the input power supply source or to isolate the oscilloscope.

The scope is DC coupled, when I put it in AC coupled, no difference !

I have some capacitors (6,2nF) between A/B lines and MM (mechanical ground for EMI), to filter the common current mode. 

The THD get worse when I increase the input voltage as shown below. In this example the input voltage was 122Vacrms on each line (216,5Vac between lines)

It is maybe an instability signs ? 

If you have any ideas or recommandations, it will be very helpful ?

Thank you for your support.

Moha 

If you raise Ph-N Vac to 130Vrms and get 225Vac-rms across the phases, you'll get 225Vac*1.414= 318Vpk which is higher than the 280Vdc that you are regulating to. So you're getting peak charging instead of PFC for any Ph-Ph Vin > ~114Vrms.  You'll either have to limit Vin when in Ph-N input, or raise Vout to > ~330Vavg.

The irregularity may be the current loop recovering from the peak-charge event.  Also, there aren't many ringing cycles between the peak and zero crossing, the current loop BW may be slow and underdamped.  There may also be an interaction with your line filter if it has low frequency resonance(s).

Hi Ray, 

Thank you for your time. 

You are right, the input peak voltage is superior to the output voltage... 

I changed the PFC output voltage to 390Vdc. With Vin=130VAC, It is better but not perfect as shown below: 

I used two board. board 1 is supplied between PH-A and PH-B, board 2 is supplied between PH-A and PH-C

Red : Phase A current 

Green: Phase B current 

Yellow: Phase C current 

At Vin=115V I have the following signal: Vin=115V, Freq=400Hz

At Vin=115Vac and freq=800Hz I have this:

As you said before I have to change the current/voltage loop components for better response espcially at 130Vac . 

I will be on vacation, I will work on this issue when I returned to work in January. 

Thank you so much for your support.