BR, PA or BPA PCB's?

[Scribble]

Ive been waiting for a group buy to spring up with either bridged, parallel or bridge/parallel PCB's for the LM3886 for quite some time. Doesnt look like theres much interest.

Would anyone be happening to make either of these boards soon i could jump in on? just out of curiosity.





Thanks for your time,
Scribble

[peranders]

You have already quite many pcb's you right away can buy for decent money. Have you checked the market?

[Scribble]

I have checked in the both the GB section and the trading post. Is it just me? is there something really obvious im missing?

[peranders]

www.google.com and the keywords LM3886 LM3875 pcb kit. Tryt it!

[tlmadsen]

Hi Scribble

You have a good point. There seems to be a number of LM3886-PCB/kits avaible, but most of them seems to be focused on one or max. two LM3886's (................makes it very hard to go bridged/parallel .

(I hope this is not getting to commercial):

I was thinking an "universal" PCB for 4x LM3886 in parallel.
By "universal I mean several things.

First of all you could put from one up to 4 LM3886 on it (and they woudl run in parallel). If you need the bridged function you would go for two board and stuff them accordingly. When bridging there should be a small driver board, that whould take an unbalanced signal and make it balanced (using something like the DRV134 or sim).

( I have seen a guy on the net using this approach, but can't remember where...........any credit should go to him, but I am not sure he can get IP on it).

If the PCB is made in high enough numbers I thinkt that this approch would be cheaper and easier for everybody, even if you only put two LM3886 on each board.

Some of the other "univeral" part of the boards should be different options to fight DC-offset; i.e. DC-servo, Trimmer adjust or just bigger outout resistors/small value gain resistors.

Going into high numbers, the universal board could also have 4-layer and 70 um kobber (=x2 normal).................nice



The board would be around 14 x 8 cm (=5½" x7½") and if we get past the +100 pices, I take it that one "universal board" as described could be sold for around 12-15 US$

I will keep on dreaming about getting time for this, but I hope somebody is going to do it.


Have fun

Thomas

[traw]

i'd thought about sending off a layout of the an1192 implementation of bpa200 to get fabbed... was going to be reasonable cost wise. still may do...

attaching photo of eagle layout i slapped together, cannot remember where i left it so may have some errors in it. let me know if should attach eagle file as well.

on another note... i know thomas is versed on the bpa200 impl woes (das module?). app notes in an1192 mention if 0.1% match on resistors, could leave servos off. really possible?

would like explanation as to why 2 parallel halfs ( the 2 + and the 2 - ) is more challenge to fight currents via offset than simple non-inverting parallel?

also, could someone throw up the offset adjustment technique used on the inverting side? (use of trimpot)

[Gcollier]

Originally posted by traw
i'd thought about sending off a layout of the an1192 implementation of bpa200 to get fabbed... was going to be reasonable cost wise. still may do...

attaching photo of eagle layout i slapped together, cannot remember where i left it so may have some errors in it. let me know if should attach eagle file as well.

on another note... i know thomas is versed on the bpa200 impl woes (das module?). app notes in an1192 mention if 0.1% match on resistors, could leave servos off. really possible?

would like explanation as to why 2 parallel halfs ( the 2 + and the 2 - ) is more challenge to fight currents via offset than simple non-inverting parallel?

also, could someone throw up the offset adjustment technique used on the inverting side? (use of trimpot)

Thats a pretty nice board, but I am not sure I like the routing of the V+. The trace might be a bit thin...and kind of long. Unless you use some HEAVY copper on the board that is. It might be better to route the power individually and just jumper from chip to chip to tie the supplies together. That way you could also have room for decoupling caps at each chip. If you are interested in 0.1% resistors you could try PRP (Precision Resistive Products). Sorry that I am not much help on your other questions...but that would require some thought on my part.

[Scribble]

Thomas,

I think this may be the project you refer too (sounds similar to me)

http://shine7.com/audio/bpa300.htm

[Scribble]

traw,

I like your board, it looks nice, simple and clean. I wouldnt know where to start scrutinising the layout thought

Thomas if you ever do find the time to undertake such a project let me know if i could help in anyway possible as im sure there would be many people interested in a "universal" board as such.

[traw]

i do see i forgot to run the input to the middle buffering opamp... but left room to run a blue trace to it. i'd likely switch the TO11 footprint to circle holes. that one is the library component for 3886 from package off cadsoft site and i had single boards fabbed with them and drill holes were too small (doubt it was fabricators fault or at least need to componensate if it is).

the ballast resistors a little sloppy and had intended to mount them underside. also need room for standoff holes.

but, in my novice, maybe ignorance, could be a disaster grounding wise.

i actually have all the parts from the nat. semi parts list on hand as i was going to give it a whirl someday. i just need to convert this to gerber files send off to ourpcb.... would run about 100 bucks to get 10 made, but could get 100 (50 pair) at 300... haven't tried them yet though.

i'd also been tinkering with idea... on das module, used a special format version to lay chip flat. but, i've also happen to have a das audio module from proamp that is bpa300 basically and saw clever technique. lay chip flat, bend chips up and from underside. the front row becomes back and back the front... gives extra room to lay thicker power traces. then can also cheat (risk?) and run through the NC pins where needed... attaching rough start of that... was trying to make single side-ish such that could prototype/etch at home. but also was going to try 0.1% match resistors, bump up the ballast size to 0.3 or so... and use bipolar electro caps for the 4.7uf throughout for size initially (blackgate or muse - cheap for the size needed) the red trace, grounding, as going to be drawn with pen (have a load of double side board to use up) --- this one very rough in progress, and was intended for servo-less IF EVEN POSSIBLE