Problem with ssltech power supply

[mkoijn]

Hi,

I'm a newbie here and in electronics. I've just built the power supply (+18V,-18V,+48V)designed by Keith Andrews (ssltech) as shown here

http://www.beatbazar.com/guests/ssltech/kps-1/index.htm

I could only get a 21V 0V 21V transformer. Measuring open circuit voltages I get +20V, +49V and -30V. Will these voltages be lower when they are connected to the preamp I'm building? I'm a little concerned to connect to the preamp I'm building as the That1512 chip has maximum 20V rating. Any advice?

Many thanks

Albert

 

[MartyMart]

Your +49v DC is fine for your Phantom, the -30v is a little "off" !
try this for some help :
http://www.electronics-lab.com/articles/LM317/

Check your connections with the Traffo data sheet etc, voltage will drop when
in circuit and you can drop small amounts with pairs of 1N4001's in series.
I did this to get nearer to the +/- 15v required for the TL052/TLE2082 used
for the 9k, was 1N4148's actually, 2 in series from each rail.
You shouldn't see -30v on the output of a -18v reg though !!

 

[marshman]

Mine did the same thing on the negative side. Seems it needs a load to come down. As soon as I hooked it into the circuit it went right to -17.98.

That was my experience anyways.

mm

 

[mkoijn]

Thanks, After connecting a load, I get -19.6V and +19.3V ???? I'm using lm7818 and lm7918. The phantom power is exactly 48V

 

[marshman]

I used a 22v tranny, and my voltages were right around 18v (+/- .2 or so) on both rails and 44v for the phantom?

Not sure why your voltages would be so much higher. Are you measuring with a good meter? I used to have a cheap digital meter that was not accurate so I bought a nice Fluke that I now trust impeccably.

mm

 

[SSLtech]

The most common ways for the ±18V voltages to be off is if the regulators themselves are off-absolute regulation, or if their 0V reference points aren't connected to the center tap/0V with a good, clean low-resistance connection.

If you measure between +18V and -18V directly with your meter (i.e. not measuring relative to ground, but measuing the absolute difference between the positive and negative rails)d do you read 36V or do you get closer to 40V? -If it's closer to 36V it sound like your meter might possibly be a little quirky.

The accuracy of the +48V output is entirely down to the tolerance of the 'Vset' resistor (between the 'adj' terminal and ground.

Other than that, 79xx regulators are famous for needing a slight load on them to 'pull' them into regulation.

Keith

 

[mkoijn]

Thanks. I measured across +18V and -18V rails, I get 39.7V. It could be just that I have a crappy meter?

Regards

Albert

 

[SSLtech]

Hmmm.... it does indeed sound like your rails are a little high... the meter largely seems to agree with itself.

You can shave off a half-volt at a time by putting some 1N4002 diodes in series with the outputs, just to keep the chips a little cooler.

I'm curious as to why your regulator chips are regulating so high... What brand are they and where did they come from?

Keith

 

[mkoijn]

Hi, the regulators were bought from Conrad her in Germany. they have L7918CV GKOBC VW CHN 529 printed on them. I will try the diodes in series with the power rails. Is it because I'm using a 21V 0V 21V transformer which is a couple of volts higher than in your schematic? I'm also having problems with mains hum on the preamp, which I've just started a new thread about.

Thanks

Albert

 

[pyjaman]

Hi, everybody, I'm a newb here, I've built keith's circuit too, and I get +18.7/-18.5 unloaded and when I plug it in my device, I get +18.2/-17.9 falling down slowly to +18.2/...0 volt after 40 seconds (!) The first time it appened I thought I had fried my 7918, but I didn't: switch on, switch of, and I get +/- 18 unloaded as formerly... each time I Load it, The neg rail falls slowly down to 0v...but remains normal afterwards...
????????
perhaps a pretty-dead electro cap somewhere...