Carlos regulated powersupply problems

[jaste]

Hello! I just finished the regulated powersupply by carlos fm using two lm338. But when started it I first got ok readings (+-27V) and hooked it up to my lm2875 amp. Then I meassured 27 volts on the amps output, and after that one of the regulators seems to have broken down cause now i only read +27V. I´ve been going over the circuit but can´t find something that´s wrong.
So how can I tell that the lm338k regulator is broken or trubleshoot the circuit?

circuit: http://dogbreath.de/Chipamps/LM338PSU/LM338PSU.html

[CarlosT]

Are you not supposed to use one LM338 and an LM138 for each V+ and V-?

I'm pondering the same thing right now after reading a statement that Carlos made regarding his preference now for regulated power supplies. This is ironic from the guy that layed out such a prolific unregulated design widely used in most GC kits

Will the real Carlos please stand up...please stand up?

[Bob03]

Where did you find his statement regarding reg. power supply?

[pwillard]

Clearly, it was a design based on two 338's since these are the high current devices with the widest junction temperature range (IE: Before POOF perhaps?).

BUT... the 138, 238 & 338 are identical devices otherwise.

It is a bit unorthodox to build a supply that way (with regulation on the path otherwise known as GND), but there ARE no high current negative regulator equivalents to the LM338. As a result, you have Carlos' design duplicated but polarity reversed for the negative leg.

[cjd]

Well, Carlos didn't invent the use of the LM338 in this fashion - he simply took an existing circuit and tweaked it, adding his "snubbers" and whatnot. First I saw of it was some experimenting that Pedja did and that's quite a while ago.

However, I'm not sure you've done anything wrong with this portion of the circuit! I suspect it's a short somewhere in the amp that may have destroyed one of your regulators - really hard to say by looking at the pics you have. Just have to hunker down and patiently check everything over with care.

C

[CarlosT]

My point was that the LM138 is a POSITIVE regulator and the LM338 is a NEGATIVE regulator.

[cjd]

Originally posted by CarlosT
My point was that the LM138 is a POSITIVE regulator and the LM338 is a NEGATIVE regulator.

Well, yes... that much I gathered. But, no... 2 LM338's it is, and it works pretty well I must add.

[CarlosT]

Originally posted by Bob03
Where did you find his statement regarding reg. power supply?

According to Carlos, the unregulated GC design is limited to about 1,500 uF. Maybe a coincidence that the AudioSector GC kit comes with 1,500 uF Panasonic FCs? Supposedly going higher than 1,800 uF starts to negatively affect midrange and treble.

With a regulated design, the sky's the limit...or so it would seem with capacitance before the regulators. Obviously, very high end amps com with huge caps. Is this functionally superior or just audiosnobiphile BLING? I dunno...my little IGC with 1,000 uF Panasonic FMs sounds pretty good


Here's the link to Carlos' posts:

http://www.diyaudio.com/forums/showt...526#post407526

[AndrewT]

Hi,
1mF of smoothing before the regulators risks ripple dipping below drop out voltage with the result of ripple dips on the amplifier rails.
This will sound worse than ordinary sawtooth ripple coming from a properly smoothed but unregulated supply.

[Greg Erskine]

Originally posted by CarlosT
My point was that the LM138 is a POSITIVE regulator and the LM338 is a NEGATIVE regulator.

I think both are positive regulators, well according to the Linear Technology Datasheet anyway. AFIK there is no negative version of the LM338, that why 2 x LM338 are used.

Here's some references (which are probably mentioned earlier in this thread)

http://www.tcaas.btinternet.co.uk/jlhnewps.htm
http://www.tcaas.btinternet.co.uk/jlhlm3x7cb.htm

regards