LCS708 does not work properly.
Hi all!
I build ac/dc converter based on LCS708. I used the PI Expert to calculate the design. But I have problems with starting-up the LCS708. In accordance with the AN-55 "Powering Up HiperLCS and troubleshooting" chapter, I tried to powering-up the converter from two power supplies: 15 V for VCC and 20 V for VCCH. After start up I see the 20V meander on the HB pin with 150 kHz (Min freq.) but it dissapears after 3 seconds. If I short the optocoupler transistor pins, the switching starts for 3 seconds and stops again. Is it normal? I checked all necessary voltages on the pins- they are OK. If I try to power up the converter from nominal VCCH - 310VDC with zero load, the LCS708 goes to over current protection, but after 10 seconds the output voltage reaches nominal value (output voltage ramp up slowly). Then I connected 25% load and converter appears good working and parameters that are close to calculated by PI Expert. If I connect 25% load to converter without input power and then connect the power, the converter will never reach nominal output voltage.
I checked transformer and capacitor parameters- it's OK.
I attached the PIxls design file and schematic for more information.
Comments
Hi,PI-Neela! Thank you! It was very helpfull!!!
What is the reason to use C109 330uF for this design? - This part of converter include bridge rectifier was used in previous design based on FSFR2100. The load of these converters is the full bridge high voltage class-d power amp and it must have very clear dc bus.
I decreased capacitance of the C109 to 47 uF and now the converter is working very well! But, at the same time, I had to increase capacitance of C soft start to 4.7 uF for stable start-up. With 0.33 uF I got auto-restart by OCP. Now I have 2V p-p ripple at 150W output power. I will be proceed with tuning of the converter.
Another question: I used the C3D02060E Silicon Carbide Schottky Diodes in the output rectifier, in the old converter at 90 kHz they were OK and cold, but in the new converter at 180-200 kHz the diodes get very hot at same output current. Can you give me some suggestions about diodes selection?
Thanks for help!
Hi all!
I tuned up the converter and now it woks fine with 220 uF output capacitor with 10 uF softstart capacitor. Efficiency - 95.2%.
There is one little issue - the actual f min resistor value does not match to calculated one by PI expert. I don't know why. If I use the calculated value the fmin too low.
Another question: why the lcs doesn't regulate output voltage at no load condition? It enters in the burst mode normally, but the output voltage is not stable and always 10 V - 15 V above It's programmed value. The output capacitor is 47 uF, soft start capacitor is 0.33 uF.
If you are getting the high voltage at high frequency for some time then it mean any transient is causing interfernce in your circuit. You must check your all connections. Also check your ground connections. Go for the good isolation method.
As you mentioned, LCS goes into burst mode at no load. It is expected to have output ripple. You can check the switching behavous and Vo ripple together to see if the ripple is coming from bursting activity. Also, I wonder what is the particular reason you concern about VO ripple at no loas ans what is the specs you allow the Voripple on no load?
Hello!
My design has wide load range: from 10-20 milliAmps to 1 Amp. The my converter's load is the special class-d amp, and it's load can vary very widely.
I made little exploration and I noted that the converter behavior at no load\light load depends on feedback parameters. I changed it's parameters from slow to fast response but a got no good results. If I got stabilty at no load\light load conditions then converter didn't work properly at medium and maximum load conditions (has oscillations) and vise versa. Maybe I made mistake in the board layout? Can you take a look at this?
| Attachment | Size |
|---|---|
| LCS_ROUTE.jpg (338.91 KB) | 338.91 KB |
You should set your stability at full load. And you are going to expect bursting at no/light load. It may not necessarity a result of poor stability problem. bursting at no laod is an expected performance. If your application does not allow bursting, then you could consider set the burst mode differently for improved performance.
Datasheet has the layout examples and design guidlines you can follow.
your Vo=25V, you use a very big cap for 250Vo. What is the reason to use C109 330uF for this design? You could reduce to much lower value 10-20x smaller and see if it helps for the start up.
With big cap, you can check the pri current waveforms and see if it is auto restart.