Texas Instruments TAS5261

[BWRX]

Now that I see the boards and the teeny weeny tiny components in real I'm getting second thoughts.

Don't get discouraged. I'm still in the process of writing up a couple documents that should help with assembly. One document is about the tools/materials you'll need to build the boards and the other is a surface mount soldering tutorial. I've got most of the pictures taken, it just needs some explanation to go along with the pictures.

 

[theAnonymous1]

GeWa, I know the task might look daunting at first glance, but it's really not that bad.

The very first PCB I ever put together was mostly surface mount components. At first I didn't think there was a chance in h*** I would be able to do it. I was wrong, and after a very short while I had perfected my technique and got the PCB completed without any mistakes.

The best piece of advice I can give you is BE CONFIDENT. If your not confident in what your doing and you are constantly worrying about making a mistake, you probably will.

Good luck!

 

[BWRX]

The best piece of advice I can give you is BE CONFIDENT. If your not confident in what your doing and you are constantly worrying about making a mistake, you probably will.

That applies to more than soldering!

Now to work on fixing them to a board and wiring them up... correctly

 

[fokker]

If you can come up with a picture or schematic, etc. for this please share it with us. (I know a guy who designed some very interesting Class-D amps that have no feed back loop at all = and it sound pretty good and runs cool ... I'll try to get his input.)

my original concept was to use a pwm generator (audio signal applied to the FB pin, and PWM comes out of the OUT pin) to test / drive a mosfet driver / mosfet ouput stage. so the set-up would be openloop.

the pwm generator is a typical dip8 device (ucc2803), with compensation on the error amplifier to widen the range of pwm duty cycle. I selected the ucc2803 primarily for its speed and for its ease of use.

one draw back with such an approach is that most (all?) duo output pwm controllers (3523 for example) are designed for full bridge rather than SE applications so their duty cycle is limited to 50% and that does not work for audio - you can force such pwm controllers into 100% duty output but then you lose one output.

In my case, the pwm generator was designed to figure out why my mosfet drivers fail to work but it in theory can be used just as how the tpa2000 is used in this project as pwm generator. Except that in my case both input and output are SE and yours are fully differential which works out well with your driver / output chip.

Another random thought: we should be able to make this into a UcD type amp: the tas5261 (?) is a combination of mosfet driver and output devices. so if we use a comparator chip (ths4120 for example) that has differential output, we should be able to get this to work. In this case, the output of the comparator is ref'd to the ground and the input of the tas5261 is ref'd to the ground so no need for level shifting.

 

[pinkmouse]

Fokker, very interesting thoughts, but I'm not sure I entirely grasp your UcD lookalike idea. Could you post a block diagram? All this class D stuff is very counterintuitive to me.

Well I've psyched myself up for some soldering tomorrow, but my fine bit is pretty gnarly so I'll have to pop out and get a new one first thing in the morning, but I should have some pics for you guys later.

 

[BWRX]

one draw back with such an approach is that most (all?) duo output pwm controllers (3523 for example) are designed for full bridge rather than SE applications so their duty cycle is limited to 50%

That is also true of the TPA2000 series chips. It actually sounds quite good for being marketed as a lowly filterless amplifier for portable devices.

In my case, the pwm generator was designed to figure out why my mosfet drivers fail to work but it in theory can be used just as how the tpa2000 is used in this project as pwm generator. Except that in my case both input and output are SE and yours are fully differential which works out well with your driver / output chip.

Your modulator could definitely be used in place of the TPA2000. The TAS5261 is marketed as a bridged output stage but it is basically two completely separate driving stages and output bridges. You could theoretically use the two output bridges separately. This thought had crossed my mind before I designed the board as I wouldn't mind turning it into a moderately powered stereo amplifier. That's the main reason I designed the two output stages with separate supply connections and gate drive supply circuits (for better stereo separation should this board ever be used as a stereo amplifier).

Another random thought: we should be able to make this into a UcD type amp: the tas5261 is a combination of mosfet driver and output devices. so if we use a comparator chip (ths4120 for example) that has differential output, we should be able to get this to work. In this case, the output of the comparator is ref'd to the ground and the input of the tas5261 is ref'd to the ground so no need for level shifting.

A UcD implementation requires feedback taken after the output filter. Assuming a 50% duty cycle at idle, the output would be floating at half the supply voltage. You would need a large electrolytic capacitor to AC couple the driver if you used a single supply rail. I wouldn't want to do that and it's just a clumsy implementation anyway. If you used a split supply you could make the output float at ground potential and the input circuitry would now be referenced to the negative rail. If you can get an input stage to work off of the split supply rails, level shift control signals with respect to the negative rail, and scale them down to 3.3V, then you could make it work.

 

[BWRX]

Now to work on fixing them to a board and wiring them up... correctly

So far so good. Now I just need to find a way to hold the heatsinks on the chips' heat slugs so I don't go welding metal...

To those who will be building these - please take the time to properly mount the power supply components and the boards as I have sort of done Also be sure to properly secure the heatsinks.

My boards are mounted to a block of MDF and the power supply components are mounted to another piece of MDF. My heatsinks are "secured" by rubber bands at the moment, but please be sure to attach them with screws and ensure that they won't short anything out. There are lots of things that can and will go wrong if you aren't careful - especially when dealing with AC mains voltages and high current DC power sources like the ones meant to be used with these boards.

 

[pinkmouse]

I've spotted the deliberate mistake! You have four standoffs, but only six screws for them on the BOM and in the kit.

Do I win $5?

 

[pinkmouse]

Oh, and I hope those rubber bands are UL certified.

 

[BWRX]

I've spotted the deliberate mistake! You have four standoffs, but only six screws for them on the BOM and in the kit.

Do I win $5?

Four of the screws are of course for mounting the standoffs to the PCB. The other two are for connecting a heatsink. I left it up to you guys to decide what type of screws to use to fasten down the standoffs to whatever case, block of wood, etc. you wanted. They're standard 4-40 threads so hardware should be cheap and easy to find.

The rubber bands I was going to use were not UL certified so I just decided to cover the exposed areas of the heatsink with paper and tape and let them sit on top of the chips. This approach was only used for a "can't wait to hear it" listening session but I was afraid any slight movement would cause or Luckily I didn't do anything stupid and the amps are still fully functional.

 

[pinkmouse]

They're standard 4-40 threads so hardware should be cheap and easy to find.

Good job I've got a box of strange American screws pulled out of old kit!

 

[BWRX]

Attached is a zipped file that contains the following 4 files:

"bot part location 1.gif" - actual bottom side silk screen, refer to this drawing for proper orientation of electrolytic capacitors
"bot part location 2.gif" - bottom side part designators moved for clarity
"top part location 1.gif" - actual top side silk screen; refer to this drawing for proper orientation of diodes
"top part location 2.gif" - top side part designators moved for clarity

The items in the parts kits are labeled with numbers. Those item numbers are used to identify the part designators using the BOM. For example, the strip of components with item number 1 are parts C1, C2, C3, C4. Look on the board (or on the part location drawings) to see where these parts need to be populated. It's that easy. The surface mount capacitors and the plastic box film caps are non polar so don't worry about their orientation. The only parts you need to pay careful attention to are the diodes (items 13, 14), the electrolytic capacitors (items 2, 6, 8, 11), and the LM317s (item 30). The diodes have a stripe on their package that must be oriented the same way as shown in the part location drawings. The electrolytic capacitors are polarized and must be installed correctly or they could explode when you try to apply power to the board. Note that the negative lead of electrolytic caps is marked by a stripe on the capacitor's insulation. The positive lead of the electrolytic caps is marked by a "+" sign on the PCB and in the part location drawings. The LM317s (three terminal TO220 regulators) must be installed with their tabs facing the edge of the PCB. If you put them in backwards they will be damaged if you try to apply power to the board. ALSO, do not install the through hole parts on the wrong side of the board. Parts get installed on the side of the board that has their silk screen outline and/or part designator. I've made this mistake before and no one is safe from moments of sheer stupidity

 

[BWRX]

Attached is the Full Assembly BOM in pdf format. I tried zipping it with the other files but it was too big for me to attach, so here it is again.

 

[FastEddy]

BWRX: pdf files are the best for us ... The Apple world can't deal with zip files easily = a three step routine that is fraught with errors, loss of data & other grief. (FYI: Zips really are not much use anymore = email spam filters often reject 'em, virus checkers can deal with 'em but there is a performance penalty.)

You could simply post the gif files (or links to such) on the WiKi page ... (If you need some space on a web domain, I have several that I'm not using. ... hot about http://24bithifi.com ?)

 

[pinkmouse]

I'll pdf 'em up.

 

[BWRX]

BWRX: pdf files are the best for us ... The Apple world can't deal with zip files easily

What do you mean? You can create zip files directly in OS X. Right click on a file or group of files and go down to 'Create archive of "files"'. In fact, that's what I used to do it. All you have to do to unzip the file is double click on it and the files are extracted into a folder in the same directory. OS X has no problem handling these types of files. gif is a better means of compression for images like the part location drawings that only have solid colors. pdfs are better for text based documents and mixed image/text documents.

 

[pinkmouse]

Brian, if you want to export those layouts as EPS files, (via CAM Processor), and mail them to me, and indeed anything else you want in the manual, and I'll get everything packaged up nicely for you. The ones you posted are a little too blocky to pdf well.

 

[BWRX]

You could simply post the gif files (or links to such) on the WiKi page ... (If you need some space on a web domain, I have several that I'm not using. ... hot about http://24bithifi.com ?)

I didn't know you could upload files to the wiki pages. I'll have to look into that. Thanks for the offer to use your webspace but it would be nice to keep all the files with the thread if possible.

Brian, if you want to export those layouts as EPS files, (via CAM Processor), and mail them to me, and indeed anything else you want in the manual, and I'll get everything packaged up nicely for you. The ones you posted are a little too blocky to pdf well.

Let me finish up my assembly notes then I'll send you the files (all in pdf format) to package up and post here.

 

[FastEddy]

" ... What do you mean? You can create zip files directly in OS X. ..."

" ... I didn't know you could upload files to the wiki pages. I'll have to look into that. Thanks for the offer, but ..."

Too late!! = This is almost as easy as dealing with the zip files: http://3dotaudio.com/diytas5261.html



(I could have activated 24bithifi.com domain = but if I wait until next year, I save about US$80 on domain fees.)

 

[BWRX]

Well there you go! Thanks Ed.

Just so you guys know, I've had my pair playing some nice music since late last night but I just got some proper heat sinks installed about 30 minutes ago. It's nice not having to worry about knocking the heatsink off and killing the amp.

And you'll be happy to know they sound just fine with my 8 ohm speakers. All frequencies are there and there's plenty of power on tap that I can't use with these 90dB @ 1w/1m monitors. As you'd expect with a dual mono configuration the imaging is very good. I wish I had some harder to drive speakers to test these with... the worst I have here is some larger Klipsch speakers that are still only 8 ohms but do like to be driven with more power.

There is a bit of a high pitched whine though, and that's due to the inductors of the two amps being so close together (it's the biggest scrap piece of MDF I could find!) and the switching frequency of the two amps being unsynchronized. If I turn one amp off the whine goes away.