0
\$\begingroup\$

I have a BLDC motor controller that uses a high side 0.001 ohm current sense resistor (CSS2H-5930K-1L00F) and a MAX40201FAUA (50 V/V) to measure the current during motor operation via a Teensy 3.6 ADC. I am trying to calibrate the current sense hardware to the Teensy 3.6 microcontroller, but I am having issues with the MAX40201 outputting values way too high for the known current passing through the resistor. I measured voltages on each side of the resistor with my oscilloscope and it showed a voltage drop of 30-80 mV for a 2.5 A known load across the 0.001 ohm resistor!

So I am thinking that the solder joint is bad. During soldering, I used a stencil and a rework gun to solder the SMD components. For the high amperage traces, I later added higher temp silver bearing solder with a hand soldering iron to reinforce them. I thought this would help the traces hold up to the high intermittent current.

Since my process apparently led to either incomplete fusion of the large flat current sense resistor or capturing flux between the resistor and the PCB, what is the best method for hand soldering these types of components so as to minimize joint resistance?

Image of the PCB section housing the 0.001 ohm resistor:

enter image description here

\$\endgroup\$
8
  • \$\begingroup\$ Comment not an aswer as not quite the question and probably too late now anyway: 1mΩ is very small and this was always going to be tricky. You can get four-terminal resistors for current measurement, which make this much easier. \$\endgroup\$
    – Jack B
    Commented May 27, 2020 at 12:45
  • \$\begingroup\$ It's not clear where the current flows in the three branches off pin 2 of the resistor. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented May 27, 2020 at 12:57
  • \$\begingroup\$ "So I am thinking that the solder joint is bad." It would be helpful if you could take a picture of the joint then, zoomed in as much as possible. But looking at the board from the side would likely tell if the joint is bad, you'd likely have a gap between the component and the board. \$\endgroup\$
    – Lundin
    Commented May 27, 2020 at 13:16
  • 1
    \$\begingroup\$ @JDD If it is a high side resistor then you probably used 2 scope probes to measure the voltage on either side of the resistor and took the difference to measure the current.This does not work with voltages as small as you're describing. The difference of 2 probes is not accurate enough to measure a difference this small. There may be nothing wrong with the solder. \$\endgroup\$
    – scorpdaddy
    Commented May 27, 2020 at 13:28
  • 1
    \$\begingroup\$ @scorpdaddy Especially if he was scoping the main terminals and not the kelvin terminals. Oscilloscopes aren't known for their accuracy to begin with. \$\endgroup\$
    – DKNguyen
    Commented May 27, 2020 at 13:30

2 Answers 2

Reset to default
3
\$\begingroup\$

This type of circuit would generally use a 4 lead shunt resistor, these break out 2 pads specifically for measuring current, in your case, you could fit bodge wires to the top of the resistor, this will remove more of the offset, but it still will not remove the issue completely.

enter image description here

Image source is from a supplier, so not sure how to handle that with the size rules

\$\endgroup\$
3
\$\begingroup\$

Flux (lack of) is not your problem. You can get good joints by flooding the pads with fresh liquid flux and then hand-soldering the joints with plenty of solder.

The layout looks a bit dubious to me, I would have avoided the horizontal current flow paths and stuck to the datasheet recommended pad layout. It's still going to be tricky getting to a reasonable tolerance on this kind of shunt setup- a Kelvin sensing 4-terminal resistor would be better.

enter image description here

1mm of a 9mm wide 2oz trace is about 27\$\mu\Omega\$ at room temperature so about 3%- so I think < 5% should be achievable, with care.

\$\endgroup\$
2
  • \$\begingroup\$ I never would have imagined that such a small deviation from the recommended design would have such a drastic result. Lesson learned! \$\endgroup\$
    – JDD
    Commented May 27, 2020 at 13:37
  • \$\begingroup\$ @JDD I'm not 100% sure it did, but I think it would be a lot harder for that to happen with the recommended layout. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented May 27, 2020 at 13:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.