Power supply voltage drop

[soundguy]

I have an old supply that a friend built for me, I should probably trouble shoot it, but need a quick band aid instead. Its a +/- 18v supply, the negative side is putting out little more than -19v. Can I just put a resistor in series with the output to drop the voltage? Seems like a 180K resistor will do the trick. The regulator is a 7918a, its for a pair of api 312's, current draw should be fairly low.

dave

 

[Samuel Groner]

Did you test with load? Most regs sag a bit.

And I don't think there's anything wrong with just using it as it is. What opamp are you using?

180k sounds hellish high. 180 kohm * 1 mA = 180 V well...

180 ohm is better (and you probably wanted to type that). Use a 220 uF or bigger 'lytic after the series resistor to make things low impedance at higher-than-DC frequencies.

Samuel

 

[soundguy]

hi samuel-

I didnt hook the amps up to it, afraid of messing them up. Unloaded, I get +17.9 v and -19.6v, seems like a big enough difference to be concerned about. Running 2520's in the 312's. I started with a 200ohm resistor in series, I didnt get the neeg side down the -18 volts until I stuck a 180K resistor in there.

dave

 

[gyraf]

try loading the supply with a 2k2 resistor and see if that dosen't help. The 78xx - and specially the 79xx's don't regulate well with no current drawn..

Jakob E.

 

[Samuel Groner]

As Jakob said, I guess that everything is OK and that they will drop when you connect the load. Are the 2520 rated for +/- 18 V?

Regarding series resistor: you cannot determin it's value without connecting the load, that's why you get completely strange values for it.

U = R * I - I'm sure you heard that one. It means that the voltage drop is dependent on how much current is flowing. Your volt meter draws around 9 uA, thus 1.6 V = 180 kohm * 9 uA, but the opamps will draw at least 5 mA each (blind guess, might be 3 mA or 30 mA), thus the (theoretical) voltage drop will be some hundred volt, which means your amps don't get more than a few mV at best.

Samuel

 

[bcarso]

[quote author="gyraf"]try loading the supply with a 2k2 resistor and see if that dosen't help. The 78xx - and specially the 79xx's don't regulate well with no current drawn..

Jakob E.[/quote]

Jakob thanks for reminding us about the 79xx minimum load requirements. For a moment there I was inclined to say that the negative regulator must not be working.

Nonetheless, I'd measure that supposed 180k resistor---it's just not plausible that one could put that in series with much of anything low-Z and not have almost zero volts.

(I'm noticing a tendency lately to talk in chains of negations---what does this mean ? :? )

 

[Svart]

what brand 79xx?

the ST 79xx parts need at least 100ma drawn before they regulate well. my own testing figured this out after a bout with them a while ago.

also, 79xx parts ALWAYS need 1uf of bypass on the output, not .1uf like the 78xx parts.

check it and let us know.

:thumb:

 

[Rob Flinn]

Can I just put a resistor in series with the output to drop the voltage?

This is not always a good idea because if the current being drawn fluctuates then the voltage drop across the resistor will also fluctuate. Therefore you will not have very good regulation.

 

[Gus]

As posted by others some 78xx 79xx don't regulate well under light or no load often you need a few ma. The voltage often measures higher with no load.

FWIW an old HiFi trick is to load a 78,79 at a min of 150ma because some of the older regs work well at that current up to about max current. So even if the circuit only used 20 ma you would add a resistor load to make it 40 ma to 150ma.

where did that 40ma come from, well if you load a series reg with a load X2 or more the static drain of the circuit it can work something like a shunt reg.

Find a data sheet for the reg and look at the output Z vs load.

 

[soundguy]

thanks for all the info guys! I tested the power supply under load and its now within half a volt which is close enough for what Im doing with it.

always learning something new.

dave