Current sense schematic question

[lk]

Greetings,
I was looking at different ways to do current sensing, using regular opamps, rather than dedicated chips like the MAX4073T/F/H, which work well, they are just rather expensive.

So while poking around i found an article written by Bob Pease, http://electronicdesign.com/power/whats-all-error-budget-stuff-anyhow, which talks about "error budgets"
And the schematic fugure two, caught my eye http://electronicdesign.com/site-files/electronicdesign.com/files/archive/electronicdesign.com/files/29/12629/figure_01.jpg

As far i understood what Bob was talking about that curuit, should generate 2 volt on the output if we have one amp through the shunt resistor, I decided to run the shcemati through an simulation tool or two to see whats up.

In the one tool i tried where i had the PN4250 transistor, i consistently get around 6 volts on Vout, no matter what the current is. I tried in LTspcie, and that just gives me 11.2 volts (see attached model, i dont know if the lm258 and the 3906  suitable for this)

Sooo can anyone tell me what im doing wrong, as thats usual where the problem resides

-lk

 

 

[Rufus]

An LM258 is not specified to operate with input voltages higher than its supply voltage - 1.5v.

If you add a +14v (or more) supply just for the LM258 the circuit will operate correctly, otherwise you need an op-amp which functions with inputs close to its +ve supply which are not common.

[JensAndree]

Sorry for a bit off-topic here but have you tried the  ACS714LLCTR-30A-T hall-effect linear current sensor?
It's cheap by my opinion and it operates between -30 to +30A and has proven rather accurate in my builds.

[awallin]

That circuit does seem to work. I just tried a simulation in NI Multisim:
http://www.anderswallin.net/wp-content/uploads/2013/04/current_sense_circuit.png

the op-amp output needs to swing quite close to +12V, so it's probably a good idea to provide separate +/- 15V supplies for the op-amp.

Anders

[lk]

Hi Rufus,

I retried the simulation with an ideal opamp, and that solved the simlation issue Now i get the output that i expected.

JensAndree, the hall sensor is 3$ in small quantiites, so its a bout twice the price of the max4073m but i guess that the hall sensor might be better over larger current ranges and with AC.

I have attached the schematic for the working setup.

Again the mistake was mine

awallin: you posted the same time as me


-lk

[croberts]

Hello lk

Sorry if this is also off topic but I did a post on current sensing a while back that may interest you. Cheers!

https://www.eevblog.com/forum/projects/bidirectional-current-monitor-fits-anl-fuse-holder/

[lk]

Hi croberts,

looks very interesting, and i can see that you also found the mcp600x series opamp, i have found that that microchip has a lot of interesting analog parts. I am building a related system to yours, to power a server 24/7 using solar panel s and batteries, with a power brick as a backup.

-lk

[croberts]

Hello lk

The current monitor pcb you saw in the post is 1oz copper and is easily handling +/- 15A. I would like to test the board at the maximum ATO fuse rating of 40A when I get the time. I think I will have to up the pcb copper to at least 2oz for that much current. Good luck with your project. I do all of the lighting in my house with LED lights I designed and all 24 lights are 100% powered by the battery bank you saw in the post. I have attached a photo of the solar array that charges the battery bank ([8] 85W panels).

[lk]

Hi Croberts,

Is this a 12 volt system, for the leds?. In my original design i decided to go with 24 volts at the cable losses would be quite significant, or the cables would be unwieldy.  I must say that im pretty impressed with your solar panels. It almost looks like you have solar tracking on them.

And you are using car fuses right? mjlorton from South Africa, has been using those "automotive" fuses, and he has experienced that they melt, and dont blow. Take a look see  


-lk

[croberts]

Hello lk

Yes I do have a 12V system. I distribute the 12V from a distribution box via 10GA wire pairs fused at 10A. Each wire pair powers 4 LEDs lights max. The wire pairs go to circuit boards I call PDACs (Power Distribution and Control) mounted in 4" electrical boxes throughout the house. The PDACs are all networked together and all are slaves controlled by one master PDAC. Touch switches and motion sensors are connected to the slave PDACs along with the LEDs. The master PDAC reads the motion sensors and touch switches via the network and then after executing control logic sends commands to turn the LEDs on and off also via the network. This allows the master controller to turn lights off automatically with no motion after a time delay....a must for a solar powered system.

The racks holding the solar panel are also my own design and allow for the panels to be rotated down for protection against hail or to be pointed at the sun by adjusting both azimuth and elevation. I have to move the panels manually but it only takes about 10 minutes to do all four racks. Thanks for the warning about the ATO fuses. I use only high quality name brand fuses but you can bet I will be contacting the manufacturer to follow up on this one. Sometimes in spite of every effort to do the right thing with a design there can still be problems. Oh well!

[qno]

Like Bob Pease would have said;

If the circuit works on the bench but the SPICE says it does not, who are you going to call??

The easiest way is to build up the circuit and do some measurements.

http://www.edn.com/electronics-blogs/anablog/4311558/Bob-Pease-didn-t-hate-Spice-simulations

[lk]

Hi Qno,

If you have suitable parts


-lk