Ridiculously easy opamp circuit not working 1

[monkeysolder]

At my wits end here. I must be missing something incredibly basic. Thanks in advance.

I have wired up a simple non-inverting opamp circuit to a LM324A. I have a 10k potentiometer connected across the inverting input and output of the opamp, and a 2k resistor between the inverting input and ground.

Next I have powered it with +9v and ground.

With only the two resistors and power connected, the output is pegged high (7.8V) no matter what I change the potentiometer value to (verified with a multimeter). What could be causing this?

When I feed it an input signal, nothing happens because it is pegged.

spec sheet:

http://www.fairchildsemi.com/ds/LM/LM324A.pdf

 

[OperaHouse]

While I believe there are voids in the universe that pass over my workbench (and others)on occasion that cause the rules of physics to violate themselves, this state doesn't last very long. The first thing to rule out is the belief that "I've done nothing wrong." Maybe a bad chip, wires are on the wrong pins, started with pin 1 as the wrong pin. First thing to try is to set one input at mid voltage with a simple voltage divider (power to ground). Then use the pot as a second voltage divider on the other input. About the middle of the pots rotatation you should see the output flip as the op amp acts as a simple comparator. This transition occurs when the difference is less than a milivolt. Something to remember! In most circuits, seeing a measurable difference between the two inputs indicates something is wrong.

 

[Skogsgurra]

You say nothing about your non-inverting input. Where is it connected before you apply the input signal? Try connecting it to V+/2 if you haven't already. There was a problem with latch up in the original (NS) LM324 if you left the inputs floating - may still be there, the schematics look very much the same now as back in the seventies.

Another thing that you should check is power. Remember that Gnd goes to the "top" pin and Vcc to the "bottom" pin. That is opposite to intuition and often connected wrong.



Gunnar Englund

http://www.gke.org

 

[itsmoked]

Good idea opera.

My thoughts:

If the pot is really a 100K pot then the gain would be so high that any signal will peg the output.

The pot may be open/faulty. This makes your amp a comparator with the reference voltage set at zero. Hence any signal will trip the comparator pegging the output.

 

[MikeHalloran]

The inverting input should be connected to a regulated +4.5V virtual ground, not ground, or the (-) supply should be connected to a -9V supply.





Mike Halloran
Pembroke Pines, FL, USA

 

[Warpspeed]

Yes, all of the above.

If you are using a single 9v supply referenced to ground, then both the input signal and the output signal plus feedback should all be referenced to +4.5v somehow.

Another way to do it is to leave the circuit as it is, and supply the +ve pin from one battery (-ve connected to ground), and the -ve supply pin from a second battery (+ve connected to ground).

The LM324 then operates from +ve and -ve supplies and both the input signal, and the output signal will be with respect to ground.

 

[itsmoked]

GUYS, Guys, guys, you need the +V/2 ONLY IF you want to amplifier an AC signal it is most definitely NOT needed for a DC non-Inverting amplifier application.

Thomas Frederiksen would be shocked to find out you needed a split rail.. (LM324 designer)

 

[Skogsgurra]

Smoked, smoked, smoked...

It is a 10 k pot. The gain will not be more than 1 + 10/2 = 6 times if the pot is at 100 %. (And no more than 51 times if it were a 100 k pot).

Connecting V+/2 is a good test. You can forget about (possible) latch-up and also offset voltage playing tricks. The latch up thing is very real - the amplification changed sig if you went above a certain common mode voltage. Haven't used the 324 for decades so I do not remember the details. They are hopefully better nowadays.

Still think that power is connected wrong way. Most guys that aren't used to this particular pin out do it wrong first time.

Gunnar Englund

http://www.gke.org

 

[itsmoked]

Yeah skogs I've reversed those on occasion. Its right up there with putting in EPROMs backwards and the pretty light that comes out of the window signaling total annihilation.

And I agree AS A TEST V+/2 might be entertaining...

 

[monkeysolder]

Thanks for all the help. I've done some more testing using your suggestions and am still not able to get the thing to work. I've tried setting it up as a simple voltage follower, with +/- supplies, and with V/2 into the non-inverting input. Still pegged.

I tested the same circuit on a breadboard with another opamp and it works fine, so I think I may be having a problem with the circuit layout on the pcb or with the way I am powering it.

 

[itsmoked]

Well heck look at the circuit board and check it. Tiz not hard...

But truly the most likely problem then is the part is blown! That "old" part blows very easily!! Look at the data sheet, there are warnings all over it..... This means while messing with them the likelihood of killing them is pretty high.. A slipped scope probe, etc., is all it takes. Replace the IC and proceed.

Remember cut all the leads off the part then remove each pin out of the board independently.

 

[monkeysolder]

Well I resoldered a chip on there and used a wrist strap this time. It's working fine now. Argh.

Thanks for the help!

 

[itsmoked]

heheheh BOO-YAH!

I have a product with 9 of those suckas on it. Calibration and set up require I insert and remove several of them with the power on. I have a problem with one about every 200 I go through.

Carry on.

 

[Warpspeed]

Hint for the day:

For a first attempt at any new design on a printed circuit board, always fit the IC's into sockets. It can save a LOT of frustration.

 

[buzzp]

Warpspeed has an invaluable point about using sockets on prototypes. I never go without one. An 8-pin device is desolderable but try doing desoldering something with more pins - no fun if not impossible for >- 32 pins.

 

[itsmoked]

Ah yes the good old dips... I don't think I've been able to use DIPs in the last 3 years.

 

[MacGyverS2000]

See? Surface mount devices have their advantages If I need to remove one, I slip a thin piece of magnet wire under all of the pins, heat them all at once with a bar, then gently tug on the wire... instant separation.


Dan
Owner

http://www.Hi-TecDesigns.com

 

[Warpspeed]

An interesting point. How on earth do you prototype something that uses surface mount components located on both sides, and a 99 layer circuit board inbetween ?

It is one of many reasons why I am now very glad to be retired.

 

[itsmoked]

Nice tip mac... I will give that a try!

Warp... Often it ain't pretty.. Trace cutting is a standard.

Actually you try extra hard to not botch things up in the original design... None-the-less I just left a bar off over an enable for a 4051.. I jumpered the enable pin to VCC with a trace completely under the part.

In this case I just had to unsolder the pin, bend it up then bridge it to the pin next to it to tie it to ground. Bridging is simple with smt parts. Just solder the pin without being extra careful.. It'll be bridged.

 

[buzzp]

You proto-type using a PCB, the somewhat finished design. To prove a hardened design requires the actual PCB be used because PCB layout is very critical to proving noise immunity and such. When I proto a SMT device, we always use the actual layout and populate using the SMT equipment. Of course, for a real rough proto, you can always breadboard with through hole parts just to prove the design works without the concerns for environment and equipment near by (immunity).
I have never done a board with more than three layers so not sure if there is something special for the guys who build multilayer boards. I am betting its done with the SMT equipment as well, the only real way to do it.