Current limiting op amp circuit debugging

[Razkid]

I have a current limiting circuit posted here-

http://home.comcast.net/~catnipone/wsb/html/view.cgi-home.html-.html

It has an input voltage between 14 and 30VDC and is supposed to limit the current to 4 amps. Connected to the inverting input is a 200 ohm resistor and a 1 mA current regulating diode which places a Vin - 0.2V signal on that input. When the non-inverting input reaches that same voltage there is a0.2V drop across the current sensing resitor (0.05 ohms) which would be a 4 amp current draw through the FET and 2 ohm load. My problem is the current is not limited to the 4 amps but just keeps rising along with the input voltage to the non-inverting input. This circuit was modeled after the one on my photo's page of my website which is a working circuit. I have simulated the circuit with PSpice and it limits the current to 4 amps. The circuit is already on a circuit board so a complete redesign is out and the original designer has left the company. The circuit in my Photo's page has an 18V input and limits the currrent to 5 amps. Any help would be greatly appreciated.

 

[geekEE]

I have a couple of thoughts on looking at the circuit:

1. Since the op amp is powered by the voltage drop across R2 and R1, it may not be high enough for normal operation. If the voltage between pins 7 and 4 is too low, the output will be indeterminate. In this case, the output may follow the V+ supply and keep the MOSFET on, which in turn would keep the supply voltage low. Perhaps tying pin 4 to the return would work.

2. Of course, make sure that circuit is actually built the way that you think it is. I've been bitten a few times by swapping the drain and source of a MOSFET and having the internal diode cause output to follow the input.

 

[geekEE]

Oops. I meant to say "the op amp is powered by the voltage drop across R2 and M1"

 

[Razkid]

The CA3140 is supposed to work with a supply voltage of 4 to 36V. I am measuring 4.75V between pins 4 and 7. This is the reason the op amp was changed from a HA2640 which could only operate down to 10V between pins 4 and 7. I've checked the FET numerous times and have made the circuit on a breadboard to rule out problems with the circuit board. I have even installed a new op amp and FET with no luck. Thanks for your input geekEE.

 

[geekEE]

If there's 4.75V between pins 4 and 7, then it does not sound to me like the output is the same as the non-inverting input.

For troubleshooting purposes, it would be helpful if you could apply a fixed Vin and then tell us what the voltages are at all the pins of the op amp. It'd probably be best if you reference these voltages to the return path (minus pin of the Vin supply).

Also, when running these tests, is the load a simple 2-ohm resistor, or is it something more complex? If the load were the equivalent of an 8-ohm resistor, then the output should just follow the input.

 

[Razkid]

Using pin 4 (-Vin and source of the FET) as the reference point the output pin 6 is always above 4V when I believe it should be less then 2.5V (according to the IRL540S data sheet for Gate to Source voltage vs Drain current). It is driving on the FET more then it should be thereby allowing more current to flow through the 2 ohm 25W resistor load. I've been on vacation this week so am going by readings I took last week and still remember! Vin is supposed to be between 14 and 30V but I can't go any higher then around 17V before the load curent reaches 6 amps. The Vin input bench supply is set to limit current to 6 amps.

 

[BrianG]

Your problem may be the Input Common Mode voltage limit of the 3140. If you check its data sheet the limit for positive VICR figure is V supply positive, minus 2.5V, i.e. your inputs can only operate up to the positive rail Vin - 2.5V.
Your inputs to the 3140 are within 0.2V of the positive rail and whilst this may not cause damage to the device it may cause unpredictable operation under such conditions.

Perhaps a rail-to-rail op amp would work better. Can't think of one right now, but one with similar characteristics to a section of the (quad) CA324 would work under these conditions.

 

[BrianG]

Just checked my own post - even a 324 type is not fully specified with common mode up to the V+ rail.

I know you have a circuit board and don't want to re-design, however you may be able to fix it with minimal changes. As most op amps have a better range for negative common mode, including "below" the 0V rail, a better solution would be to put the sense resistor R2 in the V return (ground) leg of the circuit (you may be able to do this by placing it within the external wiring). Also, you need to swap the positions of the current source diode and reference resistor R3. Both op-amp inputs are thus around 0.2V above ground, however you would have to swap the inverting/non-inverting inputs to the 3140 so that it changes the drive voltage to the FET in the correct direction.

 

[geekEE]

I think I may not have been clear when I asked for voltage readings from the op amp. Here's what I meant:
1. Apply 14V power supply as your Vin
2. Put the negative lead of meter on the negative lead of the power supply. On your circuit, that is NOT the V- pin of the op amp.
3. With the positive lead of the meter, measure the voltages at pins 2, 3, 4, 6, and 7 of the op amp. I realize that pin 7 should be 14V in this case, but it's always good to make sure.

 

[Razkid]

BrianG, You are correct about the CMVR. I found that problem on Tuesday and have ordered a OP184 which can handle the rail voltage on the inputs. I'm going to put a zener on the output to avoid having the gate voltage go too high. I could have changed the value of R3 to drop the voltage down but then I would have had to add another resistor and current regulator diode to the + input. I'm trying to keep the rework to a minimum! I bought some "Surfboards" to mount the components on and soic op amps to make the test circuit cleaner and easier to test then on perf board. I'll post again with the results. Thanks for all the inputs. I finally got to talk with the EE that had put that circuit on the board and he told me he thought there might be a problem with it!!

 

[Razkid]

The OP184 cured the problem but now I have a problem of a voltage and current spike before the regulator can take affect. I have posted the circuit on my website and the scope pictures are in the photo page of the site. Even if I remove the 1uF cap I still get a current spike above the 5A rating of the relay contacts. I've thought about putting an inductor in series with the load while someone else suggested lowering the value of C74. Any comments or suggestions for this latest problem are appreciated. I should note that the op amp, FET and C74 replaced a three terminal regulator that became an obsolete part. I had the original circuit with the three terminal regulator tested and it also had the >5A spike.

 

[ScottyUK]

The transient exceeds the relays rating for 25[μ]s. Do you think it cares? Relay ratings are generally limited by thermal considerations. The effect of the switching transient is a non-problem.

Curious: what was the original 3-terminal regulator?


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[Razkid]

I'm at home so I don't have the documentation to know what the original 3 terminal regulator was. I sent the scope images to the relay vendor and they were the ones that said the current has to be limited to no more then 5 amps even though the spike is of short duration.

 

[Skogsgurra]

Don't listen to the vendor. It is crap, what they are telling you.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Razkid]

Sorry Skogsgurra but I have to go by what the vendor tells me even though I know it will work. I can be liable if something goes wrong and I haven't followed the vendor's specs.