[ashmedai]

Yeah, rectifier bridges are just cute. Awesome little trick. The big cap after it too, since that one is there for smoothing using the exponential charging/discharging property of caps.

The Elpac wart looks great, the only problem with it is that it would cost more than the whole project. Half the point of a CMoy is to do it cheap. But I might end up just getting one and calling it done if I can't sort out the regulator bit...unfortunately that's an application that Horowitz seems to acknowledge but not cover in much depth (unless I'm missing it somewhere). The LM317 is the key, so I'm looking into how to use it so I can understand that section of the TREAD circuit. But I'm pretty sure that's also what the zener diode in the PIMETA circuit is for.

 

[Pars]

D1 in the Pimeta circuit is a crowbar diode to protect the amp in case the power is connected backwards. The ZNR/FET section is intended to turn the LED off when the battery is approximately drained, IIRC, and was optional and rarely used IMO. The Pimeta circuit is not regulated, BTW. It was assumed that a linear regulated wallwart would be used, or battery (NiMH preferred).

 

[ashmedai]

Ah, okay. I was inferring it from this:

Zener Regulators

And probably leaped to a faulty conclusion. Thanks. I noticed the LM317 has a heatsink attachment...I'm assuming it just burns the energy corresponding to the ripple voltage, so I was wondering why the diode didn't similarly need thermal protection.

Batteries make life so much easier...

From page 7 of TI's doc on the LM317, it looks like that section of the TREAD is lifted more or less straight out of the first recommended application with the values tweaked. I'm guessing most of it is there to protect the LM317 while it does its thing? Well, also to "set" it to the desired voltages.

http://focus.ti.com/lit/ds/symlink/lm317.pdf

Also if it comes to that, I can wait a couple weeks and fool around with the LM317 in the lab where I have proper equipment to check that I'm using it right.

 

[synaesthetic]

I used to cheaply regulate CMoy amps with a simple 78xx regulator IC. It works. My first PIMETA build actually used this method as well, using a 7815 regulator dropping an 18VDC unregulated wallwart (which, being unregulated, was passing a lot more voltage) to 15V regulated.

The cost was something around 0.85 USD plus another 1.00 USD for a heatsink, bolt and nut.

You can use this with any old wall wart and it'll work. Of course, the Elpac or the TREAD will work better (the TREAD will work a lot better because you can adjust the output).

 

[ashmedai]

Sounds like a linear regulator IC is just the ticket for a wall-powered CMoy...does the job well, without really compromising the basic design goals.

 

[ashmedai]

Update.

LM78xx-based regulator added. CREG 1&2 added to stabilize the regulator IC per spec (as I understand it, these are strictly to keep it stable, and with newer versions of the IC there is no point in making them larger). Switched back to a resistor-based splitter, which should leave the output current mostly governed by the wall-wart used. Whole thing is ~$30ish, and 1/3rd of that is the wart.

Additionally, ordering parts for a Mini^3, and for some random LED shenanigans. I figure a little fun with random LED light projects for the home will make good soldering practice.

I saw *one* reference to the effect that failure to discharge the caps is the main way to kill a Mini^3. Is this right, and what do I do about it - just short the leads before I solder it in? I didn't see anything about it on AMB's site, but that could just be my oversight. >_<

 

[MisterX]

Just turn the amp on after the power source is removed to drain the caps.

 

[ashmedai]

Oooh, it's during the tests! Okay, thanks.

I used to do a bunch of welding, I figure this should be similar but more fun and more useful. Kinda tempted to see what I can raid from Radio Shack again to play with while I'm waiting on Mouser...but my spool of solder from PE is still in the mail, so, waiting...