Dual LM4780 BPA-200 ExpressPCB.com Design. Opinions?

[los_ren]

hi

for a class im building a bridged/parallel LM4780, using the schematic from a national semiconductor application note for the bpa-200.

http://www.national.com/appinfo/audi...ation_Note.pdf



for the PCB im using the expresspcb.com miniboard service, three 2.5"x3.8" double sides PCBs for ~$60 (shipped). because of the size of the board, the design will be split into two identical bridged pairs, individually buffered, to be paralleled at the speakers. im sorry the PCB isnt labled, but referencing the schematic should make figuring it out easy as the design is extremely simple.



for part values im following the spec, tho im using metallized polyester instead of national recommended metallized polypropylene coupling caps. elliot sound guy says it doesnt really matter at audio freq, im thinking a 5% cap is a 5% cap, and the alternative part is huge and way more $$$. i trust the elliot sound guy. also, instead of LF411ACN opamps for the buffer ive chosen the LF411CN chips, and i am going to use %5 5w 0.1ohm wirewounds for the output caps instead of the %1, again due to $$$ or lack of availability. all 1/4w resistors will be %1 metal film.

power for the LM4780 will be from a 500va +/-25VAC toroid through a pair of braingt's rectifier boards for about 37VDC output. though slightly underated for the maximum power output by nationals standards, i dont plan on pushing the amp near its full capacity. decoupling at the power amp will likely be done with panasonic NHG or M series 50V 1000uF(x2), 470uF, and 0.33uF caps for each the V+ and V- sides. space for a 10000uF TS-UP series cap instead of the two 1000uF M/NHG series caps has been provided. power for the buffer opamp will likely come from two 9VAC power tansformers, arranged center tapped through a pair of bridge rectifier packages, decoupled with 50V 470uF caps at the rectifier, and 50V 0.33uF and 25V 47uF caps on the board. the buffer rectifier will likely be my first experiment in p2p construction =D

in this pcb img, NC pin holes have been left out for better power routing, however since making decision more has been learned and the NC pin holes will be put back in.

just wanted other opinions, maybe youll spot something ive done critically wrong. i am getting graded on it, heh, so its not all fun and games if i blow it up (at least not after the initial novelty of it...). most of the inspiration for the design came from the amp i built on braingt's LM3875 boards this summer; reason for the obvious similarities. if this amp sounds 1/10 as good as that one ill be happy (this ones probabaly gonna do bass amp duty, unbelievably high clarity would just be a bonus, and maybe not so great cuz i mostly suck at playing).

thanks for your time.
rmr

[azamkita]

Hi Los Ren

What a beautiful thread.

Have you finish it ?. What is your DC out put reading ?.

I think of the same idea, but the input opamp will be the OPA627 bias in class A & using the regulated supplies as well.

Now, I am collecting parts.

Very hard to find as per NS spec 0.01% tol . I may go for Dale 1%.

Good Luck and best regards

[los_ren]

i woke up yesterday and juggled the parts list a bit, have ordered 1% 5w wirewounds and 2200uF panasonic FC caps, using .1uF tantalums now also. just kinda made it neater and labled it.

[los_ren]

i'd have to check again to make sure, but +/-37v might be way high for the LM3886/LM4880 chips running into a 4ohm load (unless all three boards recieved are built and paralleled intead of two, but the parts were orded and i have no intention of paying shipping again...). i probabaly wont check tho, its not an issue, theyre going into an 8ohm load, same as the LM3875 units i built running the same load. i dont see how these would perform worse.

[azamkita]

Hi,

I prefer non inductive 5w resistor

37 volt WOW !!!

My plan is 24 trans, so that after rectification I am getting about 33 volt DC. I can use it without regulator.

If I want to regulate, I can use LM338 and bring it down 4 more volt.

My trans is smaller 275 VA for each mono block. Well, just 5.5 Amp per rail as the max of LM338.

most regards

[digi01]

los_ren,wonderful layout.I am looking forward to your result.

good Luck and best regards

digi

[los_ren]

yes, +/-37v. 500va toroid with parallel 25v secondaries, 25v(1.414)= 35.35v rectified voltage. think the ac wall voltage is a bit high tho, always tests bit over 37v.

[azamkita]

Hi Los Ren

How about putting t-network ?. Will it give better sound quality ?.

regards

[los_ren]

okies, this is the finished design:



i made the star GND prettier. i named it Squid GND... the only GND traces not wired directly into it are the GNDed + input of the inverting section and the nearest GND pin, which are connected to each other then to Squid GND. and the PSU GNDs, which will all collectively have their own wire to the chassis star GND (is this bad? this doesnt seem bad to me tell me if its bad).

i managed to minimize large power trace crossings (power/output) with low power signal type traces (input/GNDs) to one, where the + non inverting input crosses over the +35v trace underneath the LM4780 pins. i think ive at least done as well as the national reference board for the LM4780 in that respect.

the majority of the parts have arrived, ive adjusted the physical layout and checked everything for fit. all fits except for 2200uF cap holes, need about 1mm of adjustment.

any assistance for obvious mistakes would be appreciated... i caught myself wiring a trace meant for a mute pin to a -V pin this morning =[

since funds for the projects are low for a couple weeks, i have time to make adjustments and any input is welcome.

thanks,
rmr

[cdoggy81]

Looks good, let us know how it all turns out!