LNK30x MDCM / CCM - what do they depend on?

6 posts / 0 new

Hi guys, my question is just topic's subject line. DS says that CCM demands <35 ns recovery of FW diode and bigger inductor. Does it mean if I replace these components by just mentioned, then I get ~ x1,5 of MDCM output current - with the same topology?


Please read AN-37 as it contains useful information about the design process on LNK-TN such as the trade-offs from choosing between CCM and MDCM. Let me know if you need further clarification.

Sure I have tried to learn the subject before I went to conference. Trade-offs enumerated in Table #7 were quite clear, but I couldn't catch exactly in the end, WHICH WAY I could run the same driver into both modes (MDCM or CCM). I mean High-Side Buck-Boost Direct FB topology (constant current driver) depictured a bit further by the text of AN-37.

My case: I wanna get max current in order to run the chain of 1W LEDs at it's nominal 350 mAmps. For this the AN-37 suggests I should use (for CCM) Fw UF diode with <35 ns and customized (handmade) power inductor with larger L and lower R ___INSTEAD OF___ <75 ns UF diode and ready-made (standard value) pwr inductor for MDCM.

As I understand there are no additional circuit modifications to convert MDCM into CCM, am I right?

Can you tell me which section of AN-37 does it say that the inductor has to be customized if operated in CCM? I just couldn't find it in the article.

Have you tried using PIXLS Designer software? If you haven't done so, I suggest you try using this tool on your design as it can help determine whether you can operate in either MDCM or CCM and it also has recommendation already for the diode to be used as well the standard inductance which you can get off-the-shelf.

"For LinkSwitch-TN designs, the mode of operation is not dependent on the inductor value. The mode of operation is a
function of load current and current limit of the chosen device. The inductor value merely sets the average switching frequency." (C) AN-37

Sure. I'm sorry. It was no more than mere interferention in my poor brain between AN-37 and another quite similar pdf. As for PIExpert suite - YES, I started to explore this amazing soft a couple of days ago. And even found an ancient AC-bridge at my workplace, and made 2200 uH pot-inductor of <2 Ohm DC resistance as PIEx suggests. Now I'm waiting LNK-306 from my supplier. ))

ps: By the way, whether 75 mV 50 Hz ripple voltage is OK for almost no-load (3K3) 6 VDC 120 mA (max) LNK-304 typical buck circuit got from DS? I made six of them to run tiny brushless coolers in my little custom project (LED-bulbs). All of LNK-304' speciments demonstrate up to 200 mV low-freq. output ripple voltage at nominal wattage (~ 1,4W). Is it OK?

... whether 0,075 V ripple is okay for my design. Well, it shows 0,06 V ripple (at the bottom of the spreadsheet) which is pretty close to my test result; i.e. my question is thoroughly answered. Thank you (and me) very much! ))