Voltage regulators

i have a 7805

i have connected

20v to the input
gnd to gnd
I expect to see 5v on the output

except i see 20v (near enough)

do I need to put capacitors around the regulator input and output ?

like this

Yes. Your drawing says Transformer - is that correct?
If so you need rectifier also to convert the AC into rough DC, the caps and regulator then make smooth DC for the circuit.

Yes, capacitors are required, but their absence doesn't explain to me why the output isn't pretty close to 5V. Technically speaking, a transformer doesn't supply positive and negative, but AC. What exactly is the "transformer" being used here? DC needs to be supplied to the circuit as illustrated.

i lifted the drawing from another web page

the supply to the regulator is 20v DC (not ac)

i am very confused :-S

Add the caps and see if that helps.
Polarized electrolytic 3.3uF, 4.7uF, 10uF along those lines on the input.
1uF on the output.

i am very confused

Why, you need caps.
You also need it to be the right way round, check again it is.

except i see 20v (near enough)

Maybe your 7805 was abused in a previous life. Have you substituted a different one?

Don

i added the caps and still 20v in and 19ish out

i have tried this on two brand new regs

one is 5v the other is 15v

could i have damaged them running without caps ?

I sometimes wires a 7805 on breadboard without caps and they works fine, there is no reason being damage without caps.

What I am wondering is may be:
1: you may have it wrong way round?
2: is it ground or negative voltage from the transformer?
3: One of them is not a 7805 and different pinout?

i added the caps and still 20v in and 19ish out

Then it is either not connected the way you think or it is faulty.

could i have damaged them running without caps

No.

I sometimes wires a 7805 on breadboard without caps and they works fine,

Rubbish. Unless by working fine you mean at just one current and you haven't bothered to check the waveform with a scope.

I sometimes wires a 7805 on breadboard without caps and they works fine,

Rubbish. Unless by working fine you mean at just one current and you haven't bothered to check the waveform with a scope.
[/quote]

i.e. breadboard, quick test circuit, my point is that it should not damage the 7805 without the caps.

my point is that it should not damage the 7805 without the caps.

That was my point too.

The problem is that many people think if they have a circuit that they perceive as working then it is fine, it IS working. This is not the case. A 7805 without capacitors can never be stable. You might not have the tools to see the instability or you might have hit a quasi stable state due to the current drawn and the stray capacitance in the layout. But it is not working fine.

Rubbish. Unless by working fine you mean at just one current and you haven't bothered to check the waveform with a scope.

I have several simple setups where I only use a capacitor on the regulator output and they work ok (well, for 6+ years). The last data sheet I read on the regulator chips only recommended the caps on the input side when the input was over a long wire. As to stability, it is a relative term as to the actual operating conditions and requirements.

I have several simple setups where I only use a capacitor on the regulator output and they work ok (well, for 6+ years).

No you haven't you have several simple setups where you have not detected the instability.

The last data sheet I read on the regulator chips only recommended the caps on the input side when the input was over a long wire.

That was one manufacturers claim on one specific chip, did you use that chip 6 years ago?

As to stability, it is a relative term

No it's quite specific.

as to the actual operating conditions and requirements.

Hey lets redefine words, I can balance an egg on it's point, that is stable as long as nothing disturbs it.

Gadget999:
i added the caps and still 20v in and 19ish out

i have tried this on two brand new regs

one is 5v the other is 15v

could i have damaged them running without caps ?

Time to post a picture if you can, the whole circuit, maybe several pics if that works best.

good idea on the picture !

i will send one in the morning :slight_smile:

here is a close up of the regulator and caps

here is the voltage on pin1 - input pin

here is the voltage on the output pin - should be 5 V !

any suggestions welcome


the power supply is my laptop power supply - i have another 20v supply, today when i plugged in the supply i noticed my wifes radio had a lot of noise that went when i unplugged it (not sure of this is relevant but i decided to use the laptop supply for this test)

The only thing that I can see is that the capacitor on the input side is connected backwards. The minus (-) lead on both capacitors should go to ground, i.e. the middle terminal on the 7805.

Not real familiar with what comes out of those laptop supplies, is there any particular reason you need to use one? I'd much prefer an inexpensive and much smaller and lighter wall wart similar to this one.

While the regulator will work fine on 18-20V input, without knowing how much current will be drawn by whatever the regulator will power, it won't take much current to cause the regulator to really heat up, because it has to drop 13-15V. So I might be prepared to heat sink it.

As for the radio frequency interference, was that something that was absent before, and now you're hearing it? Or was it just the first time a radio happened to be on. The laptop supply is no doubt a switching supply, so it could well generate some RFI. Or did you mean this laptop supply wasn't the one used before?

I noticed that the capacitors you use are not ceramic ones. A 0.1uF ceramic is a must in my book. The other thing is that you have no load on the output. For test purposes put a 1K or so on the output to draw a little current. And yes that cap is the wrong way round which might account for the interference.

Good points, Mike, I'd missed that. Actually I've gotten into the habit of using both ceramics and electrolytics on input and output. Might be overkill, but I figure it can't hurt. I'll use a larger electrolytic on the input especially if I'm not sure of the quality of the input power, e.g. from a wall wart, which is the situation in the attached schematic. Haven't ever had an issue with that approach. Also a clip from the National Semiconductor 78XX datasheet. Pretty sure I've checked these supplies with no load however, and the voltage is right where it's expected to be.

78xxDatasheet.JPG