Linear regulator for Pimeta V2

Looks okay to me, as far as it goes, except that the output caps aren't necessary.  The PIMETA v2 itself provides those.  You'd use them anyway if the regulator were on the other end of 6' of wire, but since it should only be about 2", there doesn't seem to be much point.

The easiest way to fail with that design is to miscalculate (or fail to calculate!) the heat produced. Google "thermal resistance" if you don't know how to do it. (It's scarcely harder than the most basic applications of Ohm's Law, and then only because you have to be sure to add up all the thermal resistances.)

I would make the input cap significantly larger, somewhere around 1000 uF. Also a bypass cap on the output is probably unnecessary as the amplifier (I assume) has its own bypass caps on the power rails. You may want to add a protection diode to save the regulator from any reverse voltage. If your unregulated input is 24 volts that is not high enough, you should have 3-4 volts above your desired regulated voltage to avoid any risk of regulation dropouts. Finally, I don't know what kind of load the pimeta presents but make sure you have adequate heatsinking on the regulator. 

Sounds like a nice spot for one of AMB's Sigma 25 boards. :)

http://www.amb.org/audio/sigma25/

Example:

http://i23.photobucket.com/albums/b393/X86inside/Pimeta%202/Three/P1030157.jpg

(molex hell!!)

Sigma 25 can be used without any ac voltage. If you look at the schematic, you would simply leave c1 and d1 unpopulated, then solder your + input to the spot where the bridge would go that is marked with a plus and your ground could be put into the ground pad next to where c1 goes.

If you want to do your own any of the 1n400x diodes are perfectly fine, and you would connect it from the output voltage pin on the 78xx to the input pin.

Heatsinking can be determined empirically. The Pimeta doesn't draw a lot of current, if you pick a 1A regulator, it'll almost certainly survive a few minutes (probably for ever) even without a heatsink. You can touch it (carefully, they can reach boiling point) and see how hot it is getting. If it's uncomfortable to touch you can bolt on (or clip on) a heatsink as long as there is room left for it on the board. Alternatively, you can bolt it to the chassis or case and make the connections with wandering leads.

If I was building a Pimeta-style amp, I'd use a dual-rail supply and leave out the third channel.

w

R1 power rating is not an issue as it is simply a current limiting resistor for the led - it will only see ~20 mA through it. Once again though, you need an input voltage several volts higher than your desired regulated voltage. If you want 24 volts regulated output I would try to find a 28 volt wall wort. Notice on the sigma25 page that AMB recommends a transformer with 24 VAC secondary if you want 24 volt output - this is because post rectification AC voltage increases to ~1.4 times its AC value, or in this case 33 volts. If you are sticking with the 24 volt wall wort because you already have it then you are going to have to go with 18 volts regulated.

It could work but it's pretty hard to do math without any numbers. ;)

What's the unloaded output voltage or "load regulation" of the "unregulated 24vdc 0.3A power supply" you're referring to?

and..

What's the dropout voltage of the regulator you're going to use?

I miss the days when math came with numbers
 

Quote:

Originally Posted by MisterX 

It could work but it's pretty hard to do math without any numbers. ;)

AC line voltage is really not dangerous as long as you are careful. Just make sure that you don't connect or disconnect anything while it is on, make sure all connections are properly heatshrunk or otherwise covered so they are difficult to touch, use clips on your meter so you don't slip with the probes and short something. Wear rubber sole shoes and keep one hand behind your back if you are working while it is powered on and you will be fine.

Quote:

Originally Posted by bcg27 

I would make the input cap significantly larger, somewhere around 1000 uF.

That's a good idea if the input is raw DC-rectified AC.

If you're feeding this regulator from an AC-DC wall wart, though, there doesn't seem to be much point to heroic amounts of capacitance here. You'd probably get more mileage using the scratchpad space it costs by making it one of the PIMETA v2's C2s.

Quote:

Originally Posted by Sublike 

But I dont know what values I would have to add in it.

You get them from the datasheets of the items in question. What that equation calls Rϴ is normally just called ϴ. The letters refer to the various thermal resistances, either being an abbreviation for one item making up the combined system or one letter each for the two devices being joined. So, "HS" means "Heat Sink", while "HA" would be "heatsink-to-air".

I recommend you find a somewhat different equation that puts temperature rise on the left hand side and doesn't consider ambient temp and such. I find temperature rise more useful to start with. The one you posted is about max current draw, which isn't really helpful here, because that is already decided for you by the amp design.

What you want on the right hand side are the regulator's ϴJC, the heat sink's ϴHA, and the heat sink compound/insulator's ϴCH, plus either the wattage you're pulling through the regulator or — equivalently — the voltage drop across the regulator and the current drawn from it. From all that you get the temperature rise at the regulator's junction. If it's too high, use a bigger heat sink.

Just in case the letters in those symbols aren't clear, the chain is JC-CH-HA: junction → case → compound → heat-sink → air.

Higher thermal resistance values mean higher temperatures. The only way to get a lower ϴJC is to use a bigger package; TO-220 instead of TO-92, for example. ϴCH, is usually too small to worry about in power supplies, so all you need to look out for here is not overapplying the heat sink compound or putting bubbles in it, which effectively raise its ϴCH. The biggest lever is ϴHA, bigger heat sinks and better heat sink designs mean lower values.

Quote:

Originally Posted by wakibaki 

Heatsinking can be determined empirically. The Pimeta doesn't draw a lot of current, if you pick a 1A regulator, it'll almost certainly survive a few minutes (probably for ever) even without a heatsink.

With a TO-220, probably true. If our OP DIYer goes and buys a TO-92 regulator to better fit the scratchpad, though, it's a recipe for magic smoke.

Be warned, that will tie the case to V- in the PIMETA v2 circuit, which means all the other panel components with a shell connection to IG or OG will need to be isolated from the case. (And if you're not using a metal case, there's no point to trying to use it as a heat sink.)

Quote:

Originally Posted by Maverickmonk 

I miss the days when math came with numbers

That's not math, that's arithmetic.