What is the bandwidth of conventional current transformers?

[asrock70]

1. Can conventional current transformers, such as this one, measure the current of a switched converter, with a switching frequency above 100kHz?
2 Is it better to measure the current on the primary (mains voltage and lower current) or on the secondary, in front of the diodes (lower voltage about 10 times higher current?

 

[Warpspeed]

Assuming one volt (?) forward voltage across the secondary, the flux in the core works out to about 12 Gauss which is negligible.
The core material is very lossy, but at only 12 Gauss should be fine.
With a 25mm hole, there is plenty of room to space out 20 turns evenly around the toroid.
There will be 5 amps in the secondary and 4 x 0.35mm looks to be about right for both current capacity and skin effect, although in a constant current circuit, any extra series impedance from skin effect is of no consequence.
I might be tempted to try some flexible insulated multistranded (non litz) cable with thick insulation. The thick insulation might help minimize capacitance, and be less work for whoever has to wind the toroid. Anything over 0.7mm should be fine.

Reset voltage should be several volts, more is better, bearing in mind the reverse voltage rating of your diodes.
I would return D2 to ground through a resistor that was several multiples of Rs to produce the required higher voltage drop.

All looks fine to me.

 

[Easy peasy]

A 1:20 CT, for 100A thru the middle for 45% on time - the sec current is 5A, for a 0.2 ohm ohm burden, the volts are 1V across the burden resistor - but 2.25 watts in the 0.2 ohm R

this is why higher step down ratios are used, e.g. 1: 200 say, then 100A = 0.5A, and a 2 ohm burden gives the same 1V, but now now the watts are 112.5mW, easily handled by a 500mW resistor.

--- Updated Mar 9, 2022 ---

Also - it is a lot easier to measure the current on the pri side of the transformer - assuming you have on in the ckt and the 100A is on the sec side ... the CT can provide sufficient isolation to have its signal on the output side...

 

[carpenter]

Just on the sidelines . ST offer Full Bridge LLC resonant power supply eval board. with resonant frequency 175kHz, for current measurement use CT T2 . CT is completely without resetting, only with diode rectifier bridge 4x1N4148 nad 48 Ohm || $.&nF C.
Unfortunately, ST does not offer a boom for this eval board, nor does the source code of the firmware, although it can be deduced from the text on the web that it does. So it is not clear what CT they used.

Does anyone know how it is with fw for this board?
See It can be modified with the IAR Embedded Workbench toolchain, in zip file is only bin.

 

[Warpspeed]

A 1:20 CT, for 100A thru the middle for 45% on time - the sec current is 5A, for a 0.2 ohm ohm burden, the volts are 1V across the burden resistor - but 2.25 watts in the 0.2 ohm R

this is why higher step down ratios are used, e.g. 1: 200 say, then 100A = 0.5A, and a 2 ohm burden gives the same 1V, but now now the watts are 112.5mW, easily handled by a 500mW resistor.

While the low frequency 50/60 Hz type of current transformes often have 1000 turns (1mA / amp), that is not really practical at switching power supply frequencies.

In this example we are using a very high permiability toroid that gives a satisfactory 2.68mH secondary inductance with 20 turns. The secondary must operate well below self resonance, and 2.68mH requires 945pF to resonate at 100 Khz. Parasitic capacitance will not be high, and the below resonance criteria easily met.

If we try to do it with 200 turns on the same toroid, the inductance increases by x100 to 268mH and that only requires 9.45pF to resonate at 100 Khz. Avoiding the resonance demon might be difficult.

If we were switching at only 20Khz, 200 turns would probably be fine.

 

[KlausST]

Hi,

2.68mH and 945pF..resonate

This is true for an "open" CT. Open on the primary side and open (besides capacitor) at the secondary side.
The lower the burden resistance the more the resonance becomes damped.
Theoreticall you could use an opamp to keep the voktage across the secondary zero (TIA) then there is no external voltage, thus the capacitance is meaningless. (Obviously: no voltage, no function)
The lower the burden the less the secondary inductance counts and only the stray inductance counts.

Thus I still recommend to read the datasheets carefully and keep on the manufacturer's recommendations.

I still miss the OP's schematic and usecase of the CT.

Klaus

 

[FvM]

CT is completely without resetting, only with diode rectifier bridge 4x1N4148 nad 48 Ohm || $.&nF C.

It's symmetrical bipolar current measurement, DC current is blocked by LLC series capacitors. No saturation problem involved.

 

[Easy peasy]

thank you Klaus for pointing out that the sec side wdg cap is swamped by the effective loading of the burden resistor, eliminating so called resonance issues.

 

[Warpspeed]

Klaus is obviously quite right about resonance.
But what then limits the upper frequency limit and rise time ?

I am wondering, because I have a drawer full of assorted current transformers that I have accumulated over the years left over from various projects, and some definitely work a lot better than others in certain applications.

 

[FvM]

Basically, transformer Lh is shorted by the secondary shunt. Bandwidth is rather limted by Ls (leakage inductance) and Cp resonance.

 

[dick_freebird]

AC Current Probes

Features and BenefitsHigh Bandwidth Ultra-low Inductance Very Small Form Factor Characterize Current Waveforms up to <200 ps Rise Times Very Low Loading of Circuit Under Test Fits Into Dense, Closely

www.tek.com

is the datasheet for one of the current transformers I use for small signal.
Its BW is pretty high (1GHz or so). The CT-2 has higher mV/mA ratio and lower BW.
The new CT-6 has higher BW and a different form factor.

If you need BW, I might suggest using a "small signal" CT (for speed) on a "side
branch" of the source (or drain) leg. That is, make a Norton current divider and
measure a down-scaled current (like, a 0.9 ohm in parallel with a 0.1 ohm, put
the CT on the 0.9 leg, get a roughly 1/10 scaled current to measure with a more
"lightweight / agile" CT. You could also do this with scaled cut-lengths of wire
bonded at both ends (cheap DMMs often just use a length of iron wire as the
burden resistor for the 10A scale)..

 

[Warpspeed]

Things have certainly come a very long way.
I still have a very old Tektronix P6021 ac current probe, 20Hz to 60Mhz with a home made clone of the optional amplifier box Tektronix offered to boost the low frequency end.
Its served me very well over the years.

 

[Easy peasy]

Certainly the CT secondary is a complex beastie at high frequencies, and cannot be exactly modelled by lumped values for more than 100 turns or so, at mid frequencies the Lleak and R load form a low pass filter

 

[carpenter]

It's symmetrical bipolar current measurement, DC current is blocked by LLC series capacitors. No saturation problem involved.

You're right, I figured it was an LLC resonnttopology.
By the way, shouldn't an LLC resonant tank contain two inductors and one capacitance?
C I see 3x || 18nF, but there are coils?

 

[Easy peasy]

look to the transformer - it may me integrated ...