How to take time average of high side current sense amplifier output?
Bookmark this page:
 Yahoo!
 Google
 Windows Live
 del.icio.us
 digg
 Netscape
|
|
|||||||||||||
|
Posted by M. Noone on December 20, 2005, 8:05 am
Please log in for more thread options High - I'm planning on using a MAX4372H (http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2104) high side current amplifier to monitor current use in a small motor. I plan on using a .025 ohm 0805 1% resistor (digi-key WSLA-.025CT-ND, see here: http://dkc3.digikey.com/PDF/T053/1168.pdf) for current sensing. I will also have a .1uf ceramic decoupling cap from vcc to gnd. The output from the MAX4372H will be attached to a 10b ADC. My primary question is this: I plan on sampling from this ADC at probabaly about 1khz. When I take each sample I'd like the sample to be the average current since the last sample was taken - so I'd like it to be the average current over the last 1/(1khz) or the last 1ms. Implementation of this is where I get a bit confused. Would this be an application for a low pass filter? Or something else? Can anybody suggest a circuit that would work for this? Secondary question: is there anything wrong with my plan? At that small of a resistance - do I need to worry about things like the resistance of the traces, the solder, or the pads? I plan on having the current sensing resistor directly between the two current sensing pins (RS- and RS+) - so trace length between the pins and the resistor will be maybe ~1-2mm or so. Thanks for your help, -M. Noone | |||||||||||||
|
Posted by Dan Hollands on December 20, 2005, 4:44 pm
Please log in for more thread options An R in the line from maxim to ADC with a cap to ground at the input of the ADC would "average" out the signal i.e. remove high freq components so that that you don't get false readings if the sample falls on a current spike. Only other concern is whether the .025 ohm gives the correct volts needed at the expected current and to make sure resistor can handle the power. You will need some way to calibrate the circuit - either with a pot or in the processing of the digital output. That should take care of the actual resistor value and anyother minor resistances. Dan -- Dan Hollands 1120 S Creek Dr Webster NY 14580 585-872-2606 dhollan3@rochester.rr.com www.QuickScoreRace.com | |||||||||||||
|
Posted by M. Noone on December 20, 2005, 9:00 am
Please log in for more thread options
Dan Hollands wrote: > An R in the line from maxim to ADC with a cap to ground at the input of the
Hi Dan - can you reccomend how to calculate values for the resistor and
> ADC would "average" out the signal i.e. remove high freq components so that > that you don't get false readings if the sample falls on a current spike. > > Only other concern is whether the .025 ohm gives the correct volts needed at > the expected current and to make sure resistor can handle the power. You > will need some way to calibrate the circuit - either with a pot or in the > processing of the digital output. That should take care of the actual > resistor value and anyother minor resistances. > > Dan > -- > Dan Hollands > 1120 S Creek Dr > Webster NY 14580 > 585-872-2606 > dhollan3@rochester.rr.com > www.QuickScoreRace.com capacitor? About the choice of 0.025 ohms - max current should be under 500ma, but I want to be able to sense past that. Vref on the ADC is 2.048V. so to calculate max current: Imax * 0.025 * 100 = 2.048, which gives Imax = 819ma. The current sense resistor is rated for an 0.125W. So 0.819^2*.025 = 0.0168W, so I have plenty of room there. Thanks, -M. Noone | |||||||||||||
|
Posted by Mark on December 20, 2005, 11:24 am
Please log in for more thread options
regarding your question about the low resistance and the traces.... look up "Kelvin connection" you need to understand and use this type of connection for this application.... this type of connection will eliminate the problem of the trace resistance Mark | |||||||||||||
> (http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2104) high side current
> amplifier to monitor current use in a small motor. I plan on using a
> .025 ohm 0805 1% resistor (digi-key WSLA-.025CT-ND, see here:
> http://dkc3.digikey.com/PDF/T053/1168.pdf) for current sensing. I will
> also have a .1uf ceramic decoupling cap from vcc to gnd. The output
> from the MAX4372H will be attached to a 10b ADC.
>
> My primary question is this: I plan on sampling from this ADC at
> probabaly about 1khz. When I take each sample I'd like the sample to be
> the average current since the last sample was taken - so I'd like it to
> be the average current over the last 1/(1khz) or the last 1ms.
> Implementation of this is where I get a bit confused. Would this be an
> application for a low pass filter? Or something else? Can anybody
> suggest a circuit that would work for this?
>
> Secondary question: is there anything wrong with my plan? At that small
> of a resistance - do I need to worry about things like the resistance
> of the traces, the solder, or the pads? I plan on having the current
> sensing resistor directly between the two current sensing pins (RS- and
> RS+) - so trace length between the pins and the resistor will be maybe
> ~1-2mm or so.
>
> Thanks for your help,
>
> -M. Noone
>