# Replacement of a rotary switch

I have a precision burden box, whose burdens are switched by a rotary switch, and there are 21 positions. I want to make it "digital controllable", that is, using MCU to switch between these burdens. The current flow through the burden may be 10A max. Now I list the possible methods I know. In all method, I need to consider the current capacity and on resistance, and for some methods, I need take more into consideration.

1. Using solid state relays. (The on resistance and current capacity maybe a problem.)
2. Using MOSFETs. (I need work more on the driving circuit and heat-sink design, as well as the space maybe a problem.)
3. Using electro-mechanical relays. (The on resistance maybe a problem. Is there a relay with Ron comparable with a rotary switch, that is, <10 mOhms? )
4. Using a stepper motor to drive the rotary switch. (If using open-loop, I can't know exactly the position of the 'wiper', so it may not be reliable.)

That's what I know for now. Can someone give me some suggestions on the possible methods, or some recommendations on parts selection.

• Is it used exclusively on the output of a current transformer? – Andy aka Feb 2 '15 at 9:15
• Yes, it's used as a burden on the secondary of a current transformer. – diverger Feb 2 '15 at 9:19
• Are you sure the burden resistor will be taking 10A - I've never seen a CT where the burden current is 10A. It seems massively excessive! – Andy aka Feb 2 '15 at 9:48
• It's a power current transformer. Normally, 6A is enough. But occasionally, I need 10A. – diverger Feb 2 '15 at 10:44
• What current is flowing through the measurement wire of the CT - I'm asking this because I have a suspicion you may be talking about the current thru the measurement wire. Convince me you are not!! – Andy aka Feb 2 '15 at 11:20

Is this being used for AC, DC, or both?

What is the highest voltage that will be across an open switch contact?

Do all switch positions need to handle your full current rating?

How fast does the switch position need to change?

My gut tells me to use relays for this. They are inexpensive and have contact resistance comparable to a rotary switch. But I also like your idea of using some form of motor drive (servo, whatever) to drive the existing rotary switch. Position feedback can be arranged.

• It's a AC burden box. The max. voltage on the "burden" when current flowing on them is 400V. All the switch positions must can handle the 10A current. The time is not so important, but shouldn't too long, less than 5s or so is acceptable. – diverger Feb 2 '15 at 8:12
• Since the "precision" problem, and the unknown rotary switch parameter, I'm afraid the contact resistance of the relay. I find the lowest contact resistance with OMRON's relay is less than 30mohms. But in my opinion the rotary switch's contact resistance may as low as 3mohms (and in my burden box, there are actually 3 contact in parallel, so the contact resistance may even less!) – diverger Feb 2 '15 at 8:32
• How large is your burden box? Something doesn't jive: you mention a maximum of 400V across the burden box and a maximum of 10A. Do both of these extremes happen at the same time? That is: do you have 400V applied across the burden box while 10A current is flowing? – Dwayne Reid Feb 2 '15 at 16:07
• Yes, actually, each "channel" of the burden box is a resistor in series with a inductor. The current flow in them is 10A max. So if the maximum current flow in the "channel" with maximum impedance, there will have 400V voltage drop on it. – diverger Feb 3 '15 at 1:10

I understand your set up now. The burden resistor is on the output of a CT and various switches are used to select different burdens. If you had a 1 ohm burden resistor and the internal secondary winding of the CT measured 0.1 ohm, it is likely that the accuracy of the CT (current in primary to voltage across burden) will be still maintained.

I'm not saying definite but once you have chosen the correct CT, it will work with a range of burdens and still give the precise current ratio as implied by the turns ratio. If the burden gets too big, the impedance reflected onto the primary will rise and the current taken through the primary will start to divert thru the magnetization inductance - this then starts to produce an error.

So, if you can live with (maybe) 0.1 ohms of series resistance and leakage inductance in the secondary winding then adding a relay with a closed contact resistance of 10 mill ohms isn't going to make a big deal.

How you manage the monitoring of the correct burden resistor is another issue but I'll presume your monitoring circuit is high impedance and therefore if you use relays on the monitoring side (to connect the chosen burden to the signal amplifier) they won't have to be low on resistance at all i.e. virtually any relay would do if it can handle the open circuit voltages.

• Thanks for your answer. Here the the burden box itself is use as the burden of a CT. Sure there will be different error under different burden. But that's not important. I just want to make the burden box "digital adjustable" and because the burden box's burdens are calibrated with the rotary switch, so I need a method to make it "digital adjustable" but won't degrade it's performance. I think you can understand me well. The lowest burden is 100mohm, and the current flow through it may be as low as 10mA, or even small. So can you give me some suggestion? – diverger Feb 3 '15 at 9:54
• I don't understand why the burden resistor needs to be calibrated with the switch - the current flowing through the burden might be 10mA and this will still flow even if the switch introduced an extra resistance of 10m ohms - you measure across the burden only - not across burden and switch. The current that flows through the burden is fairly independent of extra series resistance introduced by the switch's contact resistance. This is how CTs are meant to work. – Andy aka Feb 3 '15 at 10:41
• The switch's resistance will be in series with the true burden, that will behavior as the burden of the CT. The CT's error with different burden will be different. And the CT's error is the current error, not voltage. So we don't measure "across the burden only". Actually, we measure the current. – diverger Feb 4 '15 at 3:28
• If you measure the current (as you said) then an extra 10 milli ohms won't make a difference. I feel you are not understanding me or I'm not understanding you (or both) so maybe you should draw a circuit of what you think explains this best. – Andy aka Feb 4 '15 at 11:10
• But the extra 10 milli will add to the burden of the CT. I've shown the CT will show different error under different burden. This error will be reflected by the secondary current's amplitude and phase, call current error and phase error. My purpose is to make the extra error caused by my "modification" as lower as possible. – diverger Feb 4 '15 at 11:38