really dumb 78xx/79xx regulator question

[bluesmoke]

I'm building a small test power supply for playing with op-amps, and had a small question...

I'm using a (2x)15v transformer (measured output of 15.2~15.6v).

After passing through the bridge rectifiers and the initial 680mf capacitors (pre-regulator), I measure a (+-) 20v; no load.

Would I be dumb (please just don't quote me as dumb...^o^ to use 7815/7915 regulators... Or would I be better off using 7812/7912...? (I've been reading the 78xx series need at least 3v over the rating to work properly)

Also, would 12v vs 15v be in any way noticeable when building something like a headphone amplifier...?

Thanks,
bluesmoke

[bluesmoke]

Doesn't matter much now. decided to be safe with the 7812/7912...

[AndrewT]

Hi,
a 15Vac transformer is ideal for a 15Vdc regulated supply.

I am surprised that you have only 20Vdc after rectifying and smoothing on no load. I would expect 22Vdc to 24Vdc when on no load. Vdc=[ Vac * sqrt(2) * (1+regulation) ] - 1.4

How did you find out you had 15Vac? Did you buy it that way or measure it into no load? If measured it may be that the transformer is actually a 12Vac unit.

You are right about leaving an adequate overhead for the "dropout" from the regulator.
The spare voltage should allow for the lower voltage from the transformer when under load and also for the ripple coming from the smoothing. The DMM will measure the average level coming out of the smoothing caps. When on load the ripple will be plus and minus half the ripple voltage either side of the measured voltage. You can see this on an oscilloscope. When on no load the ripple is almost zero (<=10mVpp)

[bluesmoke]

Thanks for the reply.

I had bought the transformer thinking it was 15v AC (as the sticker on the transformer detailed the leads and values). Tested it no load as outputting 15.5~v... (came from a creepy downtown electronics store that I always hate going to... will stop)

Next time will go for an 18v just to be safe.

Not sure if it makes any difference, but I used some heavy-duty bridges just because I tend to over-spec. Not sure if that would lessen the pre-regulated voltage any...

Would love to have a scope for my projects... a luxury for this lifetime...


Cheers,
bluesmoke

[AndrewT]

Hi,
+-12Vdc should be enough to run a headphone amp. It depends very much on the design of the amp.

Potentially, you can get 7 or 8Vac out of that and that is enough for the headphones to blow your brains out.

Try wiring it all up and remeasure the pre-reg voltage when signal is present.

Since phones are very sensitive, you require a VERY quiet output and this usually requires a VERY quiet supply.

[Nordic]

I would go for LM317 and LM337 instead... more current handleing and variable voltage would be the main dif, as well as 10db better noise rejection.

[bluesmoke]

Hi again!

AndrewT:

Will re-measure shortly...

Nordic:

Will give the 317/337 a try on the next attempt. I'm still a NooB with circuits (and calculating values), and had used the 78xx series successfully before. Definitely would prefer the better current/noise rejection.

Cheers!

[paulb]

With the 317/337, you need to bypass the adjustment terminal to get the improved noise rejection. Or you get more by doing it. See the datasheet.

[bluesmoke]

Will definitely check that out!

Cheers!

[AndrewT]

Hi,
read Walt Jung and Jan Didden on their efforts to produce a really quiet power supply.
They assessed 78xx and 317 in their papers, as well as many others, before embarking on refining the opamp controlled regulator.