Arrggh, I am going mad. Electronics gurus needed.

I've no info on the LN324, however have you tied the input to zero via ,,,say 10K R and have you got NFB say 100K R (gain of 10) or are you using it in open loop?? Don

Oh, its all properly in a circuit with various gains.

The point is that the chip itself is showing nonsense on the pins. IF the +ve input is higher than the -ve it should drive towards the positive rail. It isn't. Nothing is holding the output down..there's a

All I can think of is I have it in the wrong way round, but I've quadruple checked the data sheet, and definitely pin 4, the positive rail, is 4 anti-clockwise from the notch, looking on top.

Check the pin out. The LM324 is weird and has the +,- and out in strange places.

The + inputs are next to the power supply pins. Next out are the - inputs. In the corners are the outputs.

Christian.

Full circuit on Web please, so we can see exactly what you're doing. Circuit construction? - rat's nest lash-up, Veroboard or similar, or a pukka PCB? Does the actual circuit you've built agree with the circuit diagram you've drawn? Shorts and opens on PCB tracks are not unknown. Any unused devices in the package? - if so, how wired?

'324s are easy, and usually trouble-free. Is anything oscillating? - have you had a 'scope on it? What's the supply voltage between the rails? Is the supply current being drawn normal (around 700 uA per amplifier plus o/p load current), or excessive?

Yep, that's right.

Note that this is different to most packages, where the -,+ and out are in a fixed order, with the amps "pointing" in one direction. This would easily explain why two appear to work and two do not.

Christian.

When you have eliminated the impossible...

Even if both chips are now faulty, probability suggests that they died in the same way. Pull the chip and check that the circuit board corresponds to the diagram, that the output resistor is K and not ohms, etc.

It only needs a single rail. Attach the bare chip to nine or so volts with all outputs taken to inverting inputs, non-inverting to half-rail. If the outputs aren't within a few mV of the NI inputs then the chip is NBG, if it is OK then there is definitely a map-is-not-the-territory type problem, and if both chips are dead this is even more likely.

I am assuming you're using a scope and the beast isn't oscillating?

The message from The Natural Philosopher contains these words:

Are they oscillating like mad? That can sometimes happen and without a scope you'd not spot it.

Have a read of

Circut on the web please...

As others have said, check for oscillation - unloaded op-amps can pick up crap and fly into oscillation, terminate the outputs of the unused op-amps.

Andy.

It was oscillating...bugger. I only WANT DC response..and I gave away my scope years ago,.

Will have to pop over to a mates next week.

Yep. It was hooting away madly. Its quasi stable now to the point where its beginning to do what it should, but readings are affected by leads moving around, which suggests its still unstable.

Yup.

It was.

Circuit severely modified now, and sort of half works. Needs a scope now tho.

Try an 0.01 uf cap down to ground from the amp outputs. It's quick and dirty.

If it works try a smaller cap ( 'till it doesn't :-) )

DG

Often happens with this sort of op-amp. Stick a 510 ohm resistor in the supply with a 47 µF between + connection and ground - both as close to the op-amp as possible.

Essential bit of kit. Dunno what I'll do when my 'ol Tek 465B bites the dirt. It's one piece of test gear that doesn't seem to have come down in price for decent stuff.

This is not recommended. Introducing an RC delay in the closed loop is how oscillators are designed. Also loading opamps with C is a famous cause of oscillation. It might work in this case, but it wont get you a stable circuit. It also increases supply i draw.

Check out National semiconducors application notes series, they have tons of great stuff on opamps, and its all online free.

Nearly forgot, with their slow speed, a free PC software scope should be adequte for working with them.

NT

ASCII art below. View with a fixed font such as courier new. Cf +----||----+ Vout/Vin = 1 + Rf/Rs. | Rf | +---/\/\---+ | __ | +--|- \|+V | Rout | |324>---+---/\/\-----+-->Vout Vin--------|+_/| 20-40ohms | | | | \ | ===Cstray Rs/ | | \ | | | | | 0v------+------+----------------+--0v

Always have a capacitor Cf across the feedback resistor to roll off the frequency response, even just 100pF or so.

Most opamps are sensitive to a capacitive load, even the Cstray of output cabling. So isolate the output pin from the Cstray with a low-value resistor, Rout, close to the output pin.

Have a power supply decoupling capacitor near the package of course.

Wire any unused sections as above so that they are safe and don't flap about.

Ahah. Thats an option I hadn't tried.Ta m8.

Where do you live? Could maybe lend one to you?

Near Newmarket, Suffolk.

In message , The Natural Philosopher writes

Ah, bit too far then. Manchester.