??? 05/27/05 10:57 Read: times Msg Score: +1 +1 Good Answer/Helpful |
#93971 - Ground routing and filtering Responding to: ???'s previous message |
Dear Farshid,
more important than to have proper filtering is to have proper ground routing, where all signals leaving your board are referenced to the central ground point (star ground routing). If you use a proper ground routing, then a big number of noise sources cannot introduce noise into your circuit any longer. So, first of all, choose a proper ground routing and avoid ground loops. When you start to design filtering, you first must analyze what sorts of noise develops in your application. If you have noticed, that spark plugs make noise, for instance, then look how they introduce noise into your circuit: By radiation? Then only shielding can help. Or is the noise superimposed to car battery voltage? Then only filtering in front of 5V regulator will help. Good filtering prevents noise from entering your circuit at every vulnerable point. So, you need filtering at every cable which enters or leaves your board. Ok, assume you want to filter spark plug noise superimposed on battery voltage. Then you first need to find out the bandwidth of that noise. What is the minimum rpm of your motor? How many sparks are produced per revolution? So, how many sparks you will see per second? This number of sparks per second gives the fundamental of spark plugs related noise. If you want to filter that noise out, then your filter needs to have a corner frequency being lower than the fundamental of this noise. As sparks are very spiky, they will contain lots of harmonics. These might have more negative effects on your circuit than the fundamental. Especially if you take into consideration, that your 5V regulator shows a much better ripple rejction at low frequencies compared to that at higher frequencies. So, it might only be necessary to filter out these harmonics. This can very easily be done by the help of a 100...220µ electrolytic (paralleled by 100...220n X7R), in combination with a series inductance. Choosing a proper series inductance is extremely important, because it's the high frequency performance of this inductance which decides about the dampening at very high frequencies. So, do not use inductors which consist of many windings, bend on each other. These will show dramatically high stray capacitance between input and output terminal (interwinding capacitance), which will make your coil looking like a capacitor at very high frequencies, allowing all the noise again to enter your circuit. It's better to choose an inductor which contains as little windings as possible, and which are not bend on each other. This will minimize interwinding capacitance and give a perfect high frequency performance. Those inductors shown by you in the picture are all not good suited to filter high frequency noise! Only the triac chokes at the right are good. 220µH is much less than 5mH, but keep in mind, that chokes of 5mH will show a resonance frequency of about 1MHz, where at higher frequencies these chokes will no longer be able to suppress noise! Because of neglecting windings bend on each other, the 220µH choke will show a much higher resonance frequency, where the dampening is much better at frequencies above 1MHz. In combination with 220µF you get a corner frequency of fc = 1 / (2 x pi x SQRT(220µF x 220µH)) = 723Hz. This will give a very nice low pass filter. Put a 1Ohm resistance in series to the inductor to prevent this filter from showing a resonance at 732Hz. When you want to buy inductors for filtering of high frequency noise, don't look at the inductance of this part. Have better a look at the impedance curve in the range between 1MHz and 1GHz. And choose a part which has much impedance here. Then you will automatically choose soft ferrite beads, like the BL02RN2, for instance. These parts do not come with a specified inductance stamped on them, but they come with the promise to have a high impedance between 1MHz and 1GHz. And even if the effective inductance might be miuch lower than that of your triac choke, they will provide much better noise rejection at very high frequencies, just because their impoedance is much higher there than that of a triac choke, suffering from interwinding capacitance! There are very nice SMD ferrite beads on the market, so that you don't need to buy these fat and ugly 6-hole ferrite beads, which formerly offered the best noise rejection performace. I use them a lot, from Murata, Würth, etc. And they will allow you to even neglect the 1Ohm resistance from example above. Why? Because equivalent series resistance of electrolytic will be enough to prevent any resonance of this LC low pass filter. Good luck, Kai |
Topic | Author | Date |
Info on Murata Inductors for LC filter ? | 01/01/70 00:00 | |
Ground routing and filtering | 01/01/70 00:00 | |
Examples of soft ferrites | 01/01/70 00:00 | |
good link | 01/01/70 00:00 | |
Ground routing | 01/01/70 00:00 | |
Ground routing techniques | 01/01/70 00:00 | |
Ferrite on cables and filtering | 01/01/70 00:00 | |
Answers... | 01/01/70 00:00 | |
many of my questions are answered ! :) | 01/01/70 00:00 | |
Long lines | 01/01/70 00:00 | |
the other reason | 01/01/70 00:00 | |
Buffers good ! | 01/01/70 00:00 | |
my next PCB would be way much better !![]() | 01/01/70 00:00 | |
BL02RN2 in Tehran | 01/01/70 00:00 | |
thanks a thousand dear payam | 01/01/70 00:00 | |
On-Board Type EMI Suppression Filters | 01/01/70 00:00 | |
Basics of EMI Filters | 01/01/70 00:00 | |
More... | 01/01/70 00:00 | |
Tell me when it's enough | 01/01/70 00:00 | |
you're doing a favor for me! | 01/01/70 00:00 | |
Not all available in Tehran probabely// | 01/01/70 00:00 | |
BL02RN2 Datasheet | 01/01/70 00:00 |