tag:blogger.com,1999:blog-8979045496387455785.post3678154113385134802..comments2019-03-08T13:00:27.022+01:00Comments on starlight: Dimming a 12V LED strip with a mosfet and PWM, part 2Unknownnoreply@blogger.comBlogger5125tag:blogger.com,1999:blog-8979045496387455785.post-33249860385945000712013-03-01T16:05:04.786+01:002013-03-01T16:05:04.786+01:00KAN: nope, sorry, this was a while ago. There is q...KAN: nope, sorry, this was a while ago. There is quite some other information/video out there though on the subject.Joost Yervante Damadhttps://www.blogger.com/profile/09143890844136884489noreply@blogger.comtag:blogger.com,1999:blog-8979045496387455785.post-60412911337153970642013-02-16T20:41:16.712+01:002013-02-16T20:41:16.712+01:00Joost, do you have any videos of the LED strip dim...Joost, do you have any videos of the LED strip dimming? I'm designing something inspired by your design and I think this would be helpful.KANhttps://www.blogger.com/profile/06887608371786495598noreply@blogger.comtag:blogger.com,1999:blog-8979045496387455785.post-4090447292491861492012-09-28T06:55:31.952+02:002012-09-28T06:55:31.952+02:00You are probably be right about the massively unde...You are probably be right about the massively underdamped behavior. I'll do some further tests today.<br />10uH might seem like alot but this is a 5m long wire so who knows...Joost Yervante Damadhttps://www.blogger.com/profile/09143890844136884489noreply@blogger.comtag:blogger.com,1999:blog-8979045496387455785.post-48858085675652551082012-09-27T20:50:23.845+02:002012-09-27T20:50:23.845+02:00Of course, these formulas yield approximated start...Of course, these formulas yield approximated starting values only. Some experimenting is always needed to find the best values.Khttps://www.blogger.com/profile/13198159378794954349noreply@blogger.comtag:blogger.com,1999:blog-8979045496387455785.post-29948349988785122642012-09-27T20:49:00.977+02:002012-09-27T20:49:00.977+02:00100nF is too much. Something order of magnitude of...100nF is too much. Something order of magnitude of 1nF seems more appropriate. What you are seeing is that the charging/discharging time of your cap is too large.<br />Recall that there are three types of oscillation: underdamped, critically damped and underdamped. You are now massively underdamped, so much so that your switching period is probably larger than your rise time.<br />Another way of looking at it is that your capacitor is effectively an impedance of 1/sC, which corresponds to 1/(2*pi*100k*100n)= ~16 Ohm for 100kHz.<br />So, what we want is a resistor to dampen the ringing, but it should block current when your transistor is off. You achieve that by placing a small capacitor in series, and that's the basic snubber.<br />So what values to choose? Look at the frequency of your undamped ringing. Based on your plots you seem to have ~200ns.<br />C_DS of your transistor will be around 100pF typically, so that means your parasitic inductance can be calculated by T = 2*pi*sqrt(LC).<br />So the R of your snubber will be sqrt(L/C), and your C of the snubber C = T/R.<br />In your case I get L~= 10uH (which is a lot, so I might have miscalculated here), which yields R = 330R and C = 680 pF.Khttps://www.blogger.com/profile/13198159378794954349noreply@blogger.com