2
\$\begingroup\$

For part of a project, I am needing to monitor the current draw of 4 low power loads simultaneously. I like experimenting with circuits and I found one online in a TI application note (AN-31 page 35) which looked good.

I had 4 loads to monitor, and I wanted an LED to light up when the current got higher than a set amount (undetermined as of yet). My supply voltage to the loads were 2.5V, and I didn't want a large voltage drop over the sense resistor, so I made that 0.1 Ohms. I then used the formula provided in the App note to calculate the other resistors to give me an output in the mV range.

I simulated the circuit on Proteus, and decided on a current limit of around 125mA, giving my circuit a Vout of 125mV.

enter image description here

I added in a couple of extra ammeters and voltmeters just so I could understand it a bit better. I believe I have an understanding of the circuit, in that there is a voltage drop over the 0.1R resistor, the op-amp will then turn on the FET to equalise the 2 inputs, which puts a current through the bottom resistor (R8 in this case), causing a voltage drop which is then the Vout. This is just a very quick explanation as it shouldn't need much more detail than that.

So, I then duplicated this circuit 4 times, and added a comparator circuit with a reference voltage set to 125mV. This should then turn on an LED when the current passes the 125mA threshold. I simulated this again:

enter image description here

So, I am happy with that, and I decided before putting this in my main circuit, I would build this current monitoring part on a separate PCB, just to double check it all works. It did not.

The current monitoring part of the circuit is here:

enter image description here

I set up a few test points so I can measure different things. I used an OPA4251 as my opamp, and an LM339 as the comparator, mainly because I had some of those around. The 0.1, 10k and 1k resistor are all 0.1% tolerance resistors. The MOSFET used is the IRLML6402, again because there are lots of them here.

Once setting the load current to 125(ish)mA :

enter image description here

The LEDs were not turning on. I took some measurements:

  • Vdrop over 0.1R resistor - 128.2mV

  • Voltage supply to load - 2.5014V

  • Voltage at I1 (Vout of current monitor circuit) - 0.0003V ?????

  • Voltage at output of opamp (to FET) - 2.4591V

  • Voltage between '-' and '+' input of op-amp - 129.3mV

It seemed to me that the FET wasn't turning on, so I tried changing it, and still nothing. When looking at the simulation, I put a few different FETs in the design, and sometimes it wouldn't turn on, so I have managed to convince myself it is to do with this, however, when comparing datasheets online, I am not quite sure what I should be looking for in this. I also don't have any of the FETs that I used in the simulation, and the ones I do have don't have SPICE models on Proteus (always the way eh?!).

So my question is am I correct in thinking that it is likely the MOSFET selection that is causing my issues? If not, could it be the Op-Amp? Rather than asking for components I could use instead, what are the parameters I should be looking out for? And finally, does anyone have any ideas on how to get this working?

\$\endgroup\$
  • 3
    \$\begingroup\$ When asking troubleshooting questions, this is a good example of what to do! It's nice to see clear explanations and screenshots of steps, an explanation of how OP understands the circuit their reasoning, and a range of measurements taken and steps they have tried. Makes it much easier to answer \$\endgroup\$ – Curious Jul 5 '19 at 10:30
4
\$\begingroup\$

When I see a "high side current sensing" circuit like this, I always first check that the opamp can handle the input voltages. It needs to compare a small voltage very close to the positive supply rail so the "input voltage range" needs to support that.

Your OPA4251 has an input voltage range of:

enter image description here

So it can support input voltages up to V+ -0.8 V. In your PCB schematic that V=V+ = 2.5 V, that means the inputs of the opamp will work up to 2.5 V - 0.8 V = 1.7 V. You're using them close to 2.5 V and that's outside that range!

You can either power the opamps with a higher supply voltage (like 5 V, this is what you did in the first schematics but not in your PCB schematics!)

Or use an opamp which can support input voltages up to it's own positive supply rail. Some opamps with a rail-to-rail input support this.

Do note that in your original schematic you fed the opamps with + 5 V which is enough to drive the N-channel MOSFET. At the 2.5 V you're using in the PCB schematic, this might not be enough.

I would suggest: feed the opamps with +5 V (instead of +2.5 V) like you did in your first schematics.

\$\endgroup\$
  • 2
    \$\begingroup\$ D'OH!!!!! That was a mistake that I didn't notice! The op-amp was always intended to be powered by 5V, but it seems I made a mistake on the schematic, which transferred to the PCB! I suppose this is why I make test PCBs before designing in to the real product! I got that tunnel vision where I had convinced myself it was one thing and didn't look at other bits, as I assumed I had not made a mistake like that. Sometimes it just needs a fresh pair of eyes. I'll hack the PCB quick and get back to you with the result! \$\endgroup\$ – MCG Jul 5 '19 at 10:09
  • \$\begingroup\$ Yep, it seems that was the case. Very silly mistake to make, but oh well! That's another thing I'll add to my (double) check list! Now I just have the comparator not swinging fully high for some reason, but that's another issue. Thanks very much for that, it would have taken me quite a while to notice that as I was too focused on the MOSFET side of things! \$\endgroup\$ – MCG Jul 5 '19 at 10:23
  • \$\begingroup\$ My pleasure :-) Regarding the comparator, they're LM339 and these have an open collector output meaning they can only pull the output down. If you add a couple of pull up resistors (I suggest 10 k ohm) that would solve the issue. \$\endgroup\$ – Bimpelrekkie Jul 5 '19 at 11:57
  • \$\begingroup\$ Yeah, I got that up and running about 2 minutes after sorting out the supply voltage issue. I came back to my desk to write the reply then went back to it and realised I had no pullups. Again, it was a case of using a different model to simulate and using what I had lying around! At least that was a simple one for me. All up and running now. Thanks again \$\endgroup\$ – MCG Jul 5 '19 at 12:00
  • \$\begingroup\$ current shunts are typically 50mV to a comparator. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jul 5 '19 at 12:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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