AC current Shunt Resistor

Thread Starter

ilginsarican

Joined Jul 13, 2017
137
Hello,
I use mcp39f521 Power Monitoring IC.
According to demo board schematic, there is a resistor R5.
I could not understand the purpose of this resistor.
The AC current is sensed via shunt resistor 2m ohm.
Why R5: 2,49ohm is used?
According to the datasheet "Differential Input Voltage Range" of I1+ and I1- pins is +-600mV/GAIN.
So I want to measure 50A(peak) AC current, what should be the shunt resistor value?
For example: 500mv/50A= 10mOhm , is this calculation correct? (500mV for safety margin)
If this example is correct, why 2,49 ohm is used?

I would be glad to share your ideas.
Thanks.
 

Attachments

Ian0

Joined Aug 7, 2020
3,533
I’ve no idea. Submit it to Microchip’s technical help - they are very helpful.
My guess is that it was originally drawn for a current transformer and that was its burden resistor.
Or maybe the ferrite bead is forming a LR low pass filter with that resistor, but it already has a low pass filter at 4.8kHz, and the ferrite bead won’t do much below 1MHz.
 

Thread Starter

ilginsarican

Joined Jul 13, 2017
137
I’ve no idea. Submit it to Microchip’s technical help - they are very helpful.
My guess is that it was originally drawn for a current transformer and that was its burden resistor.
Or maybe the ferrite bead is forming a LR low pass filter with that resistor, but it already has a low pass filter at 4.8kHz, and the ferrite bead won’t do much below 1MHz.
Hi,
Thanks your reply.
Here is Microchip's reply:


"In fact, that resistor should not exist at all.

It was originally placed on old power meters with non-SMD shunts (for tens of Amps of current), which were connected to the meter board with wires. Sometimes, the wires could brake, leaving the ADC input pins floating and the values read totally incorrect.

That's why there is always a small resistor in parallel, so the inputs are never left floating.

It's value is chosen at least 1000 times higher than the shunt value, not to interfere with the normal functionality."
 

Irving

Joined Jan 30, 2016
2,197
The evaluation board is designed for 220v/15A rms. 15Arms through 2mOhm is 30mVrms, 84mV p-p.

I suspect the 2mOhm is chosen to minimise power dissipation. At 15A its dissipating 0.5W and is spec'd as a 2W part.

15A rms = 42Ap-p so not far off your 50A (18Arms), so assuming you meant 50Ap-p and not 50A peak. So I'd stay with 2mOhm. Dissipation increases to 0.65W

If you really meant 50A peak = 100Ap-p = 35A rms. Its always best when talking about AC to work in rms unless there is a specific reason to use peak or p-p - it avoids potentially dangerous confusion.

At 35A rms it would need to be increased to a 5W part (if available in SMD?)

While you could go to a 10mOhm shunt, at 35A rms the dissipation rises to 12W, needing a 25W part which is a lot of unnecessary heat to shift...
 

Irving

Joined Jan 30, 2016
2,197
Hi,
Thanks your reply.
Here is Microchip's reply:


"In fact, that resistor should not exist at all.

It was originally placed on old power meters with non-SMD shunts (for tens of Amps of current), which were connected to the meter board with wires. Sometimes, the wires could brake, leaving the ADC input pins floating and the values read totally incorrect.

That's why there is always a small resistor in parallel, so the inputs are never left floating.

It's value is chosen at least 1000 times higher than the shunt value, not to interfere with the normal functionality."
I guessed it might be something like that... but why spec such a weird value and a 1% part as well... given the input impedance of the ADC a much larger 5% part would have been fine...
 

Ian0

Joined Aug 7, 2020
3,533
I guessed it might be something like that... but why spec such a weird value and a 1% part as well... given the input impedance of the ADC a much larger 5% part would have been fine...
My pet hate: I see something like a 221Ω 1% resistor in a circuit, so I think it must be that precise for a good reason. I follow the circuit around and find it’s a current limit for an LED.
 

RPLaJeunesse

Joined Jul 29, 2018
184
My pet hate: I see something like a 221Ω 1% resistor in a circuit, so I think it must be that precise for a good reason. I follow the circuit around and find it’s a current limit for an LED.
When looking at values and tolerences consider the source of the design. Often the value/tolerance called out is what was used (as in the post above), not what is needed. If the designer typically works with stocked "production" parts the value and tolerances may seem too tight just because they were what's available. A company may only stock 1% parts, avoiding all the added part numbers needed for 2% and 5% parts that might otherwise clog the inventory system.
 

nsaspook

Joined Aug 27, 2009
8,788
When looking at values and tolerences consider the source of the design. Often the value/tolerance called out is what was used (as in the post above), not what is needed. If the designer typically works with stocked "production" parts the value and tolerances may seem too tight just because they were what's available. A company may only stock 1% parts, avoiding all the added part numbers needed for 2% and 5% parts that might otherwise clog the inventory system.
Exactly. There is only so much space for prototype parts. :eek:
resist.jpg
trim.jpg
 
Top