That part of the process completed safely, the first board got hooked into the eventual home of whatever amp I decide to stick with - a portable package with a welded square-steel chassis frame and an integrated LED power meter. Low enough voltage/current that problems won't be catastrophic, while allowing for obvious go/no-go testing, but sound quality's atrocious. The two built-up boards survived my preflight testing - powering up with two 12VDC 1A switching supplies. Need more power or a lower operating impedance? Add a module or two! My gameplan is to build little amp "modules" with their own heatsink/fans (tunnel HSFs for Intel's Xeon processors - which are meant for 100-watt thermal loads!) and everything in one nice neat package. This is about as compact a board as is practical while leaving enoug space for components AND allowing for sufficient groundplane action to reduce loops and noise. If built with 0.1% resistors for the gain and feedback control and 1% or better resistors for output balancing, it SHOULD be possible to parallel several of these boards together - without any design or parts changes - to drive signals into almost-dead-short impedances without releasing the magic smoke.īut alas, I won't know that until I get the parts in from Mouser and build up a couple. )Īn externally hosted image should be here but it was not working when we last tested it.Įach board is 1.875" by 2.5", and will parallel both op-amps in a single 4780. (No way in hades I'd want to cook these up by hand - I don't have -that- much time. Of course it's complicated - it's crawling with groundplanes.īut, since I'm going to have a replicator make 'em right off the drawing I can make it complex. copper sheet too - I could make SERIOUS current capacity traces if I wanna go crazy. if one uses solid bus wire and preshapes it properly before soldering it down. I've used that trick before with supercompact power supplies and it works nicely, esp. To make the traces that carry current wider may run me into problems with coupling since I'll have to ditch some shielding (especially the B amp's output around Rf2 and Rsn2), so I debated making the traces thicker in the vertical plane instead by soldering some wire along the traces. If worse comes to worse I can float the caps off the board and lean them over the Ri pair.Īs you've noted, the board's design goal is smallest size possible with proper design and groundplane layout to keep coupling and oscillation and so forth in check. 47uF 50VDC in Tantalum is plenty small enough to fit, but I'm not sure tants are suited to amplifier DC coupling. Personally I'd rather leave caps out of the signal path but if ya gotta have it, well.Īlso thought about the Ci pair. I can move Rin over a little if need be to allow the cap to have some landing area on the board. I can trim the edge lightly if it misses by a few mils.Īlready debating widening the mounting path for Cin, as you mentioned. The backplate of the IC will be right at level with the edge of the board. I've had problems with contact corrosion in connectors before (I live across the street from a large saltwater bay) so for me it's better to solder everything and risk having to clip a wire later. They'll even take care of all the vias in the groundplanes.Ĭ4 shouldn't be an issue as my intention for the interconnects is to solder wires directly. They'll pre professionally-made and all holes will be plated through, so that's not an issue. Yeah, I saw that tenth-miliohm matching requirment and thought "yeah right, like hobbyists are gonna have gear capable of resolving that small a difference." I'm not going to even waste time trying to reach that kind of precise with the equipment at my disposal (although I do have a calibrated Fluke DMM that I think can read that low).Īs for the top and bottom soldering issue, the editor's ExpressPCB and I'm planning to have them produce a few of the boards.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |