Having complete failures with my homemade circuit boards.

[Refrigerator]

I recently dug out my box'o ICs and tried making something.
ICs i decided to use were:
1: TA2078P ( preset equalizer )
PDF :
http://www.szerviz-com.hu/php4/cat.php?action=download&ty_pe=TA2078P&group=T
2: D2025 ( power audio amps 2.3W/ch or 4.7W BTL @10% THD )
PDF can be downloaded here:
http://www.datasheet4u.com/datasheet/D/2/0/D2025_ShaoxingSilicoreTechnology.pdf.html

The material i used for the PCB was a copper clad board ( bakelite, not fr4 )
At the end of the day i had three boards and all of them failed, both of the D2025 amps gave out a loud oscillating noise and started heating alot, as the TA2078P gave out audio but it was very crappy and loud.
The method i use for making my PCBs is toner transfer method, although my toner transfer skills aren't the best i can still do pretty well ( not always ) and i always check my PCB with a multimeter to make sure there are no shorts.
Here's the photos. ( big apologies for all the flux and all the crap, i also reused some parts from the other D2025 board )

Please excuse the horrible soldering, i'm still using crappy solder and a plug-in soldering iron ( I've already ordered a new soldering station )

What do you think was bad in my circuit boards ? ( not including the horrible soldering of course  )
I want to make sure where's the problem before another one fails.
I'll post PCB layouts soon.
Thanks in advance.

[tautech]

I would increase polygon pour clearances to reduce any risk of solder bridges.

[senso]

It seems you have shorted pins to the ground(is it) plane, I would start by setting the expansion/isolation between the tracks and the plane to something like 20 or 30 mils, and make oval pads for the DIP's, so you have bigger pad in front and back of the pin and there is a smaller chance that you will short anything.

[Refrigerator]

I've tested the connections before powering them up and there didn't seem to be any shorts.
Here's the layouts.
D2025 BTL :

D2025 stereo :

TA2078P :

Btw, what do you guys think of my layouts, are they any good ? ( I'm a beginner at this kinda stuff )

[lapm]

I would pay attention to clearances. Seems you have some issues of them being quite narrow... Other then that i see nothing too bad...

[tautech]

Clearances, plane and between tracks need increasing. IC pads, change to oval. Increase annular ring size on pads where tracks pass through. And the big one for signals...remove all un-connected copper planes.
Study some PCB's to get some basic rules for your applications.

[Refrigerator]

Thanks for the help!
So it seems like the ICs were faulty ? Or could it be the caps/resistors ?

[IanB]

Thanks for the help!
So it seems like the ICs were faulty ? Or could it be the caps/resistors ?

Did you even read the comments above? 

[Refrigerator]

Thanks for the help!
So it seems like the ICs were faulty ? Or could it be the caps/resistors ?

Did you even read the comments above? 

Yes.

[Rudane]

Have you tested this design on a bread board first? Usually with a short, the failure is catastrophic. If it worked for a short while and then quit, check the polarity of your caps; one could be backwards. It's not likely the resistors unless you went over their wattage rating.

[Tac Eht Xilef]

What are you using to power them?

The TA2078 has ~5mA quiescent current, but in practice will require ~50mA with signal. The D2025 has a quiescent current of 50mA, but again will draw 10x that or more with signal.

The symptoms you describe - distortion in the 'equaliser', and audible oscillation/motorboating in the amp - are what you'd expect when running from a supply without sufficient current capacity.

BTW, your D2025 looks like a knockoff version of the TEA2025, which is known for stability problems / oscillation due to poor layout & suppy regulation. You might want to look up datasheets & app notes for that...

[Dave Turner]

You might want to post your schematics too.

[TheBorg]

I would pay attention to clearances. Seems you have some issues of them being quite narrow... Other then that i see nothing too bad...

For OP on this- Sparkfun has a fairly decent set of design rules that are really easy to implement for starting out into PCB making. Agree with lapm here, traces need to be wider and you need more clearance. Also make sure your part spacing is ok, some if it looks tight in EAGLE (but it looks fine on the board). Oh, and make sure your V+ and GND planes/connectors are fairly well isolated. I know from experience getting these too close is a pain in the butt, straight shorts across
That will definitely help in soldering at least, which would help a little there. For DIY boards they are pretty good! - I tried toner transfer and suck at it... iTead Studios is so much easier...

[Refrigerator]

I have totally forgot that i have a breadboard, i rarely ever do anything with it. Might be the time to do testing before actually making something. 

[Mechatrommer]

starting from a schematic picture might be a good idea... because some intelligent or unintelligent guys hate to solve problem from the bottom (in this case the manufactured pcb) where there are possibilities that the root cause might be from the top or the source (in this case the schematic or the design). ymmv.

[Refrigerator]

I don't draw schematics myself, i use the ones provided with the PDF, you can find all of them there.

[Tac Eht Xilef]

I don't draw schematics myself, i use the ones provided with the PDF, you can find all of them there.

No, you need to give us your schematics for the circuits as built.

There are several discrepancies between the datasheet schematics, your board patterns with overlays, and your photographs. For example on the overlay of your BTL board pattern/overlay you've marked a component of some type between pins 7 & 10.

That's not on any datasheet schematic - is it a mistake, something that you added, something that you left out ... what?

It's impossible for us to tell, because the only other clue you've given is photos of half-disassembled boards 

[Mechatrommer]

I don't draw schematics myself, i use the ones provided with the PDF, you can find all of them there.

No, you need to give us your schematics for the circuits as built.

i've checked he's good. 7-10 resistor is just extension which he didnt put. maybe the problem...
1) components values similar to DS?
2) components broken?
3) chip broken?
4) short or open tracks?

[Tac Eht Xilef]

i've checked he's good. 7-10 resistor is just extension which he didnt put.

Well, that's my point - it's probably OK, but (a) it looks like a resistor there without reason (which he hasn't given), and if that's the case then (b) his circuit isn't just "the ones provided with the PDF" like he said. On the other hand, maybe it was never installed - but since at least one of the boards has been half-gutted, we can't tell anything for sure...

maybe the problem...
1) components values similar to DS?
2) components broken?
3) chip broken?
4) short or open tracks?

Then there's also the small matter of being unaware of the necessary stability recommendations, due to the shitty datasheet for the knockoff D2025 (and from general knowledge of similar amp chips). For a start, the grounding of the single channel BTL board is poor - as far as possible it should be star-wired with hefty tracks back to a single ground point, and ideally pins 4,5,12,&13 would be connected by a single pour. Conversely, on the stereo board, the ground pour should be pulled back well away from both the input and output pins.

But I'm still betting he's trying to run it off a 9v battery or similar...

[Refrigerator]

Sorry for me being such a pain in the ass...
Well, since all the chinese pieces of crap are burned, i can continue on making other amps.
I still have some Toshiba and Philips amps, i also have a couple of russian TDA2030 clones.
Here's one i just routed out, it's a TDA7057AQ.
 ( don't expect to see anything amazing nor safe )

 ( PDF http://www.dectel.ru/components/ms/pdf/TDA7057AQ_4.pdf )


I did a small test and it seems like i can do traces 0.20 mm and up without much problems, 0.15 mm is where tracks get distorted and blurry.
In 7-10 ( in D2025 ) I had originally intended to put a 10k resistor, just for some testing.

[Tac Eht Xilef]

 

I wish you luck.