Hello Gert, When I wrote: > slew-rate), so even high impedance common mode choke should make no problem. I thought of choke parasitics as you wrote later.
I don't agree that 3rd ground wire needs not be filtered. When you put the ferrite ring on your shielded cable you filter even the shield. In test (which simulates the real disturbance) you put the CM mode signal to all lines by resistors (assume 50 Ohm generator and 3 resistors 300 Ohm each). Assume 30V after this 50 Ohm. The 2mH is 2kOhm at 150kHz. So the 30V is divided to 26V at Coil and 4V at 300 Ohm. In signal lines there is high RS485 input impedance, assume infinity (it is the worst situation). So there is no current in those lines and no voltage drop at their 300 Ohms resistors. The CM choke works like trafo, and subtracts 26V from 30V input. giving 4V CM signal on filtered side. It is accepted by RS485 transceiver. I understand it this way (may be wrong). It is one of possible explanations of how ferrites on cables work. I don't agree with calculation making from 2.2mH the 4.4mH. Newer mind how many coils the CM choke has it (as whole) has the same L as one of its coils. (originally I thought 2x2mH as two coils 2mH each). It doesn't mean that I don't believe in DM 44nH because I have no knowledge in that question and may be even wrong way (in my opinion) you come to the right result. It depends on how somebody defined this 1%. I didn't know how to estimate the differential effects of CM coils. Now I know, I think. Thanks. Regards Piotr From: "Ing. Gert Gremmen" <[email protected]> To: "Piotr" <[email protected]>; "EMC-PSTC" <[email protected]> Sent: Friday, February 06, 2004 2:43 PM Subject: RE: RS485 and CM choke > hello Piotr > > Your approach is not bad, so i added a just few remarks. > > Coils are not what they specify. Above > 20 Mhz they will probably not behave as you think. > Coils can be enhanced with ferrites on the wire set (CM). > > Capacitors to ground need a decent ground, most suitable is > a dedicated groundplane or metal enclosure. This means that > you will need to use SMD caps on multilayer PCB or feed trough caps > in a enclosure wall (probably a recessed wall) > > Most receiver chips do a good job in CM voltage suppression, this may > reduce your demands on the filters a lower frequencies < 1Mhz > The 3rd ground wire needs not be filtered, as it is connected to > ground at both sides even if a 100 ohm resistor is used. > Possibly it can be implemented as a screen. The 100 ohm > resistor needs to be bridged by a cap of course. > > Further comment in between your text. > > > > Regards, > > Gert Gremmen > Approvals manager > ====================================================== > ce-test, qualified testing > Member of EMC committee CENELEC/IEC > > + Independent Consultancy Services > + Compliance Testing and Design for CE > + Improvement of product quality and reliability > + Testing services according to: > Electro magnetic Compatibility 89/336/EC > Electrical Safety 73/23/EC > Medical Devices 93/42/EC > Radio & Telecommunication Terminal equipment 99/5/EC > > Website: www.cetest.nl (english) > www.ce-test.nl (dutch) > Phone : +31 10 415 24 26 > Fax : +31 10 415 49 53 > ====================================================== > > > > > -----Original Message----- > From: [email protected] > [mailto:[email protected]]On Behalf Of Piotr > Sent: Friday, February 06, 2004 11:40 AM > To: EMC-PSTC > Subject: RS485 and CM choke > > > > Hello, > > I posted my question at sci.engr.electrical.compliance and was suggested to > post it here. > So here it is. > > I'm trying to find if it is possible to do RS485 without optoisolation. > EN61000-6-2 and EN50130-4 needs interface to work with common mode > disturbance in 150kHz to 100MHz range of 10V (80% AM modulation) and 150 Ohm > source impedance. This gives me 18V top signal level. I've read that it is > good practice to add extra 6dB (measurements inaccuracy and element > distortions). > This way I have 36V. > > > This is the theoretical maximum stress, actual voltages on the chips will > be less. > > I wont to communicate with 115kb/s (I plan to use IC-s with limited > slew-rate), so even high impedance common mode choke should make no problem. > > > RS485 is a differential mode signal, a CM choke (coil) will not be noticed > > ( but see below) by the signal (by design). > > I assume up to 32 devices and up to 1km line. > I see two solutions: > Solution 1. > CM choke 2x2mH + 2.2nF capacitor at each line to ground. This reduces 36V to > acceptable values but gives me 1.1nF differential mode. > Can this 1n1 be a problem ? > > > output impedance of 25 ohms for RS 485 is common. 25 ohm and 1NF > > create a cut off frequency of > 5 MHz No problem for the signal. > > At CM a CM coil of 2.2 equals 4.4 mH > > I think 4.4 mH is too large ( at 150 Khz = 4100 ohm) unless > > you buy such a coil using extreme high mu ferrites so the > > number of windings is limited. > > The induction of a CM coil in DM mode is rarely less then > > 1% of the specified value, so 44 uH, this is a series impedance > > of 31 ohm per wire unacceptable for RS 485. > > Look at the specs, or measure, you will need high quality coils. > > > > > > I can increase L and decrease C. But the higher L the lower its resonant > frequency and I'll have more problems with attenuating higher frequencies. > > > not recommended as one should keep coils as small as possible. > > Solution 2. > CM choke 3x5mH with no capacitors. RS485 with third wire connecting grounds > of communicating devices. I've read somewhere that only 3 wire RS485 is the > right solution. In this third wire I can place 100 Ohm to avoid ground > difference currents. This solution seems to be very good. One coil senses > the CM voltage and subtracts it from signal lines. If it is a good solution > and RS485 is popular than CM choke manufacturers should have a large offer > with 3 coil chokes of several mH inductance but they don't. CM chokes with 3 > (and more) wires are widely offered but for higher frequencies. From that I > think that I am doing somewhere a mistake. But where ? > > > I suggest approx 2x1 mH with good coil symmetry and 2n2 to receiver > ground. > > Don't filter the ground wire but connect it directly to ground of the > > driver PCB. 100 Ohm to safety ground. Capacitor to bridge it. > > > Connected with that subject is the question: Can I connect my circuit ground > with Earth ground via 100 Ohm, and than directly connect circuit grounds of > communicating devices with this third wire. ( YES ) May be the choke will be > not > needed at all (probably if I reduce my needs from 36V to 18V). > My devices are powered from external 12V power supply selected by installer > (not me). > I don't know where the problems can arise from in this situation ? > > Thanks in advance for any comments. > > Piotr Galka [email protected] > > > > > ------------------------------------------- > This message is from the IEEE EMC Society Product Safety > Technical Committee emc-pstc discussion list. > > Visit our web site at: http://www.ieee-pses.org/ > > To cancel your subscription, send mail to: > [email protected] > with the single line: > unsubscribe emc-pstc > > For help, send mail to the list administrators: > Ron Pickard: [email protected] > Dave Heald: [email protected] > > For policy questions, send mail to: > Richard Nute: [email protected] > Jim Bacher: [email protected] > > All emc-pstc postings are archived and searchable on the web at: > http://www.ieeecommunities.org/emc-pstc > > > This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. Visit our web site at: http://www.ieee-pses.org/ To cancel your subscription, send mail to: [email protected] with the single line: unsubscribe emc-pstc For help, send mail to the list administrators: Ron Pickard: [email protected] Dave Heald: [email protected] For policy questions, send mail to: Richard Nute: [email protected] Jim Bacher: [email protected] All emc-pstc postings are archived and searchable on the web at: http://www.ieeecommunities.org/emc-pstc
