Author Topic: Really low value current sense resistor (50mV peak) will it have noise issues ?  (Read 1471 times)

0 Members and 1 Guest are viewing this topic.

Offline MiyukiTopic starter

  • Frequent Contributor
  • **
  • Posts: 908
  • Country: cz
    • Me on youtube
Hi,
I have a question if I can use this low value sense resistor to get reliable current mode boost without sense amplifier just next to it
I want to keep sense resistor as small as possible about 50mV at peak current 5A, fsw is 200kHz so will also make some mess
Question is, can I then trigger by it with basically 1mV or even less resolution, or will noise just mess this low signal and Ill need to get some amplifier ?
Also I am little scared of noise in chip itself as it is digitally controlled and some noise can get to analog part of chip
 

Offline TimNJ

  • Super Contributor
  • ***
  • Posts: 1720
  • Country: us
What chip are you using? Anecdotal evidence, but...I'm using LM5122 (75mV over-current threshold) in an application with a peak current of about 25A. I'm using a normal 2512 sense resistor, and an 2R-C filter, as per the datasheet recommendation. The distance between the sense resistor and the chip is around 10mm. I've done pretty extensive testing, and no weird glitching, unexpected shut-downs, instability, etc. If not using a true kelvin package sense resistor, you can still get adequate sensing by routing the traces direct to the pads. Assuming normal high-side current sense method, make sure to route the + and - current-sense traces as close to each other as possible. Route them over a ground-plane, preferably part of the ground plane not carrying heavy current.
 

Offline MiyukiTopic starter

  • Frequent Contributor
  • **
  • Posts: 908
  • Country: cz
    • Me on youtube
It is controlled by Atmel/Microchip SAM MCU and its integrated comparator/current mode PWM
That is why I am little nervous about chip noise itself
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 7044
  • Country: nl
Offset for the comparator looks a wee bit high relative to 50 mV.
 

Offline Weston

  • Regular Contributor
  • *
  • Posts: 221
  • Country: us
What noise are you worried about? Noise has a source. Due to the low impedance of the current shunt you are mostly going to get voltages induced through other return currents flowing through the ground plane.

The most common source of noise is generate by the switching node, converters typically blank the current sense comparator during this period to prevent false triggering.

If you have more than one phase its possible to get switching noise coupled from other phases. You can also get voltage offsets due to DC currents flowing through the ground plane.

Differential connections to the current shunt resistor would reject most of these nose sources, but I doubt the SAM can be set up like that. You could probably try some star grounding with the central ground connection being the low side of the current shunt. A differential amplifier would make this easier but add to your BOM cost.

You might have noise internal to the comparator, but the SAM datasheet possibly has info on that.

Also, with low value shunt resistors the series inductance of the shunt can be an issue, leading to higher current readings for higher frequency components of the current waveform. Often its not that much on an issue if you have a blanking period though.
 

Offline MiyukiTopic starter

  • Frequent Contributor
  • **
  • Posts: 908
  • Country: cz
    • Me on youtube
I have two phases
Blanking is supported
But datasheet is silent about noise transfered from CPU to comparators, or even detailed specs of comparators  :-//
Layout can be tricky to avoid inductive coupled on sense line or ground issue as it will be single ended measure
 

Offline fourtytwo42

  • Super Contributor
  • ***
  • Posts: 1218
  • Country: gb
  • Interested in all things green/ECO NOT political
I use PIC's in similar situations but preceded the internal comparitor with an external differential op-amp r-r io kelvin connected directly across the sense resistor. Even so I use LEB as the noise in an hard switched SMPS of any power (mine are 1-3Kw) is substantial never mind how much care you take IMOP.
 

Offline MiyukiTopic starter

  • Frequent Contributor
  • **
  • Posts: 908
  • Country: cz
    • Me on youtube
I also considering use of high side sense like HV7802 it is reasonably cheap and can be placed before chokes to give much cleaner signal and have configurable gain
 

Offline MiyukiTopic starter

  • Frequent Contributor
  • **
  • Posts: 908
  • Country: cz
    • Me on youtube
I agree with amplifier right next to sense can voltage go much lower
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
will noise just mess this low signal and Ill need to get some amplifier ?
The problem will be wires to that shunt (their lengths) but not the shunt DC resistance because they can be longer than the shunt itself.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf