Question regarding LM317 & 100kHz Fsw SMPS passing through
Hello All,
If I were to use 12V buck regulator with a 100kHz switching frequency and then tie its output to an LM317 set for 5V, would I see the 100kHz peaks get through?
Instead of a gyrator circuit, could this 100kHz be removed with a simple R-C on input to the LM317?
Thanks
Hello All,
If I were to use 12V buck regulator with a 100kHz switching frequency and then tie its output to an LM317 set for 5V, would I see the 100kHz peaks get through?
Instead of a gyrator circuit, could this 100kHz be removed with a simple R-C on input to the LM317?
Thanks
It depends how you bypass the adjust pin of the LM317, and some other things.
It's pretty easy, and very effective, to put an LC lowpass filter between the SMPS and the LM317. I've done it for 60 kHz, using something like a JW Miller 10 uH high-current toroidal choke (model 2200HT-100V-RC, 11.2A DCR .014, or 2201V-RC, 12.5A, DCR .007; about 1-inch diameter) and a nice low-ESR 2200 uF electrolytic to gnd, such as Nichicon UHE-series 50V. It turned the nasty-looking spikes into something more like curved-top triangles, with about .0013X the p-p amplitude.
In my case, I had about 280 mV p-p of 60 kHz spikes riding on about 38 vdc, before the LC filter. After the LC, the 60 kHz was down to about 360 uV p-p. Then I had an LD1084 adjustable regulator, with 35v DC output, that had 121 Ohms from out to adj, and 56 uF from adj to gnd, and 150 uF from out to gnd. After the reg, the 60 kHz should be down to something like 20 uV p-p or so (in simulations, at least; I couldn't measure it.)
I've heard that it might be better to use a larger R from out to adj, on the reg, to get better lowpass filtering with the cap from adj to gnd, and then maybe use a smaller, better cap, there, like a film type.
So, recapping: the regulator gave me something like an 18X reduction of the 60 kHz. But it had already been smoothed and reduced by something like 770X by the LC lowpass filter.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
It's pretty easy, and very effective, to put an LC lowpass filter between the SMPS and the LM317. I've done it for 60 kHz, using something like a JW Miller 10 uH high-current toroidal choke (model 2200HT-100V-RC, 11.2A DCR .014, or 2201V-RC, 12.5A, DCR .007; about 1-inch diameter) and a nice low-ESR 2200 uF electrolytic to gnd, such as Nichicon UHE-series 50V. It turned the nasty-looking spikes into something more like curved-top triangles, with about .0013X the p-p amplitude.
In my case, I had about 280 mV p-p of 60 kHz spikes riding on about 38 vdc, before the LC filter. After the LC, the 60 kHz was down to about 360 uV p-p. Then I had an LD1084 adjustable regulator, with 35v DC output, that had 121 Ohms from out to adj, and 56 uF from adj to gnd, and 150 uF from out to gnd. After the reg, the 60 kHz should be down to something like 20 uV p-p or so (in simulations, at least; I couldn't measure it.)
I've heard that it might be better to use a larger R from out to adj, on the reg, to get better lowpass filtering with the cap from adj to gnd, and then maybe use a smaller, better cap, there, like a film type.
So, recapping: the regulator gave me something like an 18X reduction of the 60 kHz. But it had already been smoothed and reduced by something like 770X by the LC lowpass filter.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
A picture being worth a thousand words..:
http://www.acoustica.org.uk/t/3pin_reg_notes2.html
Yes the noise gets through. But a little RC filter will also remove this sort of noise very nicely, if you can stand teh voltage drop. As little as 4R7 - 10R with a small low-ESR/ESL cap, say 10uF, before the 317 works well. For low curent circuits you can use more resistance, but you may run into stability problems with the 317.
http://www.acoustica.org.uk/t/3pin_reg_notes2.html
Yes the noise gets through. But a little RC filter will also remove this sort of noise very nicely, if you can stand teh voltage drop. As little as 4R7 - 10R with a small low-ESR/ESL cap, say 10uF, before the 317 works well. For low curent circuits you can use more resistance, but you may run into stability problems with the 317.
- Status
- Not open for further replies.