Current transformer termination best practice - a theoretical situation - UK Vintage Radio Repair and Restoration Discussion Forum

Al (astral highway) · 8th Apr 2021, 4:11 pm

I'm well practiced at using current transformers (CT's), both home-wound and commercial, at high frequencies, so no explanations needed.

But I have a situation that is pre-occupying me and I wonder about best practice in this specific situation.

So we have a CT that is connected to logic on a board I made with SMD components. It's designed to detect overcurrent on my home-built inverter, via a pre-set limit. The CT is home-wound on two ferrites, and 1:1000 ratio.

Previously, I was very cautious and terminated the CT with a burden resistor soldered to the proximal ends of the leads, right where the windings finish. This was connected via twisted pair to another -- this time SMD resistor -- on the board, about 20cm distant.

So we have a CT correctly terminated, a twisted pair about 20 cm long, and then another resistor, right on the board.

Ideally we could ditch the burden resistor attached to the CT and rely on the SMD component on the board. Obviously the calculations are different but this is trivial for me and off-topic.

My sense is that this would be bad practice, since if the SMD resistor (nominally 1W rated) fails OC, then kilovolts would be present on the suddenly unterminated CT - it's 1:1000, remember, and there are pulses of hundreds of A being sampled by the CT.

I've spoken to folk who are happy doing this, but it doesn't sit right with me.

What would you do if you were designing a commercial product and it wasn't possible to eliminate the 20cm of twisted pair?

Also, there at HF (say 250KHz) there must be subtle but noteworthy differences between the CT terminated with a burden resistor right on its windings, and one terminated there followed by twisted pair 20cm long and then another burden resistor. After all, this is a kind of transmission line, and I know transmission lines embody complex theory. I can't assume what these differences are but I would be curious what the theory says.

Thanks folks

Ed_Dinning · 8th Apr 2021, 8:57 pm

Hi Al, in power system practice it is nor unknown for a CT sec to become disconnected on a metering application. The usual trick was to fit a voltage dependent resistor (varistor, etc across the terminals and heatsink this device. Quite a bit of power can be dissipated in this device and it was normal to fit a latching thermal switch onto the heatsink with spare contact to operate an alarm.
Note that a zener will not work well in this position

Ed

Al (astral highway) · 8th Apr 2021, 9:54 pm

Hi Ed, thanks for the confirmation and solution! I'll treat as if this is a possibility.

GMB · 9th Apr 2021, 9:03 am

I would always say that if a single point of failure leads to a dangerous situation then that is bad.

I can't see a problem with the 20cm twisted pair link unless this is working at GHz frequencies.

8th Apr 2021, 4:11 pm	#1
Al (astral highway) Dekatron Join Date: Nov 2006 Location: London, UK. Posts: 3,496	Current transformer termination best practice - a theoretical situation I'm well practiced at using current transformers (CT's), both home-wound and commercial, at high frequencies, so no explanations needed. But I have a situation that is pre-occupying me and I wonder about best practice in this specific situation. So we have a CT that is connected to logic on a board I made with SMD components. It's designed to detect overcurrent on my home-built inverter, via a pre-set limit. The CT is home-wound on two ferrites, and 1:1000 ratio. Previously, I was very cautious and terminated the CT with a burden resistor soldered to the proximal ends of the leads, right where the windings finish. This was connected via twisted pair to another -- this time SMD resistor -- on the board, about 20cm distant. So we have a CT correctly terminated, a twisted pair about 20 cm long, and then another resistor, right on the board. Ideally we could ditch the burden resistor attached to the CT and rely on the SMD component on the board. Obviously the calculations are different but this is trivial for me and off-topic. My sense is that this would be bad practice, since if the SMD resistor (nominally 1W rated) fails OC, then kilovolts would be present on the suddenly unterminated CT - it's 1:1000, remember, and there are pulses of hundreds of A being sampled by the CT. I've spoken to folk who are happy doing this, but it doesn't sit right with me. What would you do if you were designing a commercial product and it wasn't possible to eliminate the 20cm of twisted pair? Also, there at HF (say 250KHz) there must be subtle but noteworthy differences between the CT terminated with a burden resistor right on its windings, and one terminated there followed by twisted pair 20cm long and then another burden resistor. After all, this is a kind of transmission line, and I know transmission lines embody complex theory. I can't assume what these differences are but I would be curious what the theory says. Thanks folks __________________ Al

8th Apr 2021, 8:57 pm	#2
Ed_Dinning Dekatron Join Date: Jan 2004 Location: Newcastle upon Tyne, Tyne & Wear, UK. Posts: 8,194	Re: Current transformer termination best practice - a theoretical situation Hi Al, in power system practice it is nor unknown for a CT sec to become disconnected on a metering application. The usual trick was to fit a voltage dependent resistor (varistor, etc across the terminals and heatsink this device. Quite a bit of power can be dissipated in this device and it was normal to fit a latching thermal switch onto the heatsink with spare contact to operate an alarm. Note that a zener will not work well in this position Ed

8th Apr 2021, 9:54 pm	#3
Al (astral highway) Dekatron Join Date: Nov 2006 Location: London, UK. Posts: 3,496	Re: Current transformer termination best practice - a theoretical situation Hi Ed, thanks for the confirmation and solution! I'll treat as if this is a possibility. __________________ Al

9th Apr 2021, 9:03 am	#4
GMB Dekatron Join Date: Aug 2003 Location: near Reading (and sometimes Torquay) Posts: 3,094	Re: Current transformer termination best practice - a theoretical situation I would always say that if a single point of failure leads to a dangerous situation then that is bad. I can't see a problem with the 20cm twisted pair link unless this is working at GHz frequencies.