Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › TPA6120a2, once again
New Posts  All Forums:Forum Nav:

TPA6120a2, once again

post #1 of 12
Thread Starter 

I'm trying to lay a small board for a tpa6120a2, just the amp board, no onboard PS, no pot. Here's a first attempt, at 50*50mm, dual sided. It's heavily based on jcx's suggestions in various threads.

 

Some comments:

 

- to ease layout and bypassing, the power supply is provided seperately for each channel.

- bypass caps are in 0805. I'm considering the use of 10uf/16V caps for all positions.

- the configuration has the tpa6120 in unity gain inside the feedback loop of a jfet opamp; it makes it much easier to interface with an external potentiometer if necessary.

- the bottom groundplane is just there to provide some extra cooling.

- instead of a 10r resistor, I'm planning to use an inductor from fastron at the output.

 

I'm still working out the various resistors values and cleaning up the board. Any input, about design or layout, would be greatly appreciated.

 

 

post #2 of 12
Done a few designs with this chip...

http://www.head-fi.org/t/555770/what-i-did-with-tpa6120

http://www.head-fi.org/t/586042/what-i-did-next-with-texas-headamp-chip-a-digital-control-desktop-portable

You can see quite a large hand wound inductor I used at the output, I also built it with a 1Meg ferrite, I had no problems with oscillation in either case (or without). The amps looked good tested with RMAA, but it's not possible to verify the performance absolutely without better testgear than I have at home.

I must have done a couple dozen different layouts for this chip, first with a modified feedback loop, then as I worked up better protection schemes and fixed the on-off clicks. If I were you I'd try to get some big (comparatively) alu or tant electrolytics close to the chip.

w
post #3 of 12
Thread Starter 
Quote:
Originally Posted by wakibaki View Post
The amps looked good tested with RMAA, but it's not possible to verify the performance absolutely without better testgear than I have at home.

I must have done a couple dozen different layouts for this chip, first with a modified feedback loop, then as I worked up better protection schemes and fixed the on-off clicks. If I were you I'd try to get some big (comparatively) alu or tant electrolytics close to the chip.

w

 

Yep, that's a common problem with good ICs and one I've run into too... my old version of a tpa6120a2 board was better than the ADC of my soundcard. :angry_face:

 

I'm going to see what I can do wrt bigger bypass caps.

post #4 of 12

For the thermal pad under the chip...

Are you going to be able to do this in an oven or are

you going to solder it from the bottom?


 

For Vias in EAGLE in general:

You can get them "tented" if you set up EAGLE correctly.

 

For anyone who doesn't know the term "tented", it means

it will be covered over with LPI and not left bare like a pad.


 

You can modify you board rules so that EAGLE does not generate

a mask below a certain pad size. It's in Design Rules > Mask > Limit.

I have mine set for 24 mils.

I make my vias with a drill size if 13 mils and a pad of 20 mils.

post #5 of 12
Thread Starter 

Hand soldered. I've no access to an oven. I've done a pair of time, it works if you fill the hole with a bit more solder than necessary and then heat from below (having soldered one side of the pins or the two bottom ones).

 

Thx for the tip on the tented vias btw.

 

 

edit to avoid multiplying the posts:

 

Four 220uF/16V pana FR can fit on the board. It required quite a bit of redesign, but it made the board cleaner overall.

 


Edited by 00940 - 5/6/14 at 10:01am
post #6 of 12
Cheaper than a soldering iron:-

http://www.ebay.co.uk/itm/Toaster-Mini-Oven-1000W-9-Lite-Oven-Broil-Bake-Toast-FREE-Baking-Tray-More-/271077604952?pt=UK_HGKitchen_SmallApp_RL&hash=item3f1d7bfe58

Obviously this eBay is not local to you, but you should be able to find one or arrange postage.

I make brass solder stencils by etching. I use toner transfer to mask the brass. The process is very forgiving. The solder paste can be quite ill-defined and still succeed because of the way that surface tension pulls solder, pad and pin together.

Parts have to be placed with tweezers. Some jewellers loupes in various powers may help.

My oven has top and bottom heating. I use them both at the same time.

I put the board in the oven, turn the power on full, and watch through the window. When I see the parts 'set up' and the solder change, I turn off the power and open the door. Gently, because it withdraws the grill tray, and inertia can cause the parts to slide. Actually there's a timer, but sometimes I have to interfere, because on an occasion a part will rotate, and if it's a chip with a lot of pins, it's better at this stage, to use the tip of a scalpel to move the chip, so I still use manual control of the heat.

Some afterwork is sometimes necessary. Monolithic components are sometimes skewed and occasionally solder bridges must be removed from ICs, this can often be done with a soldering iron alone. The hole under the TPA6120 needs to be filled or other through hole parts fitted.

w
post #7 of 12
Thread Starter 
Quote:
 Quote:Originally Posted by wakibaki View Post

Cheaper than a soldering iron:-

Indeed. I'll have to look into it.  And with ebay, availability usually isn't a big problem these days. I found a pair of tutorials on how to do stencils, doesn't look too difficult.  Thanks for the tip. :)

 

Recycling household items for electronics is nice. UV nails curers make great UV exposure tools for example.

post #8 of 12

I'm going to make my own re-flow oven, something like this: Toaster Oven Re-flow

post #9 of 12
Thread Starter 

Time to go to sleep... Here's the latest iteration of the pcb.

 

Some fresh questions:

 

- would it be worth the trouble to add some decoupling (small R or L followed by 100uF) to isolate the input opamps from the tpa6120 supplies ? I think I can make it fit.

- I'm usually not very picky about opamps. The good old opa134 is usually good enough. Are there more modern jfets input opamp I should consider instead ?

 

 

post #10 of 12

I'm kind of partial to LME49990 when paired with a buffer.

post #11 of 12
Quote:
Originally Posted by Avro_Arrow View Post

I'm going to make my own re-flow oven, something like this: Toaster Oven Re-flow

This is my second toaster oven, I bought the first for a place I worked at, everything was done on a shoestring.

While I think it's admirable in a way to try to mimic the profiling of a professional oven, I think it's a wasted effort for amateur use. I've had no problems I could attribute to a poor temperature profile. Industrial use is a bit different from amateur. Thousands of boards per hour may be being processed, and a run of defects can be very expensive. No-one wants the ultimate blame for a failure to fall on them, as there can be financial consequences and lawsuits. Precisely defined profiles are intended to reduce the likelihood of failure in mass manufacturing.

I just turn the oven on, wait until I can see that the whole board (sometimes multiple boards) have gone from grey solder paste to shiny, and turn off the heat. Job done.

w
post #12 of 12
Quote:
Originally Posted by wakibaki View Post

I just turn the oven on, wait until I can see that the whole board (sometimes multiple boards) have gone from grey solder paste to shiny, and turn off the heat. Job done.

w


I'll give it a try, I already have the oven I was going to use.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › TPA6120a2, once again