The DIY 2.1 amplifier. Should I build it? Would you buy it?

[neildavis]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

The design is ready to go. I am headed to Taiwan (again) and will solicit some more feedback on the best layout and fabrication houses there and we will get going shortly.

Regarding Rev 2: Any chance a future iteration could provide this circuitry in place of the analog processing: http://www.dafx14.fau.de/papers/dafx...budget_ope.pdf ?
It's basically the Analog Devices reference circuit for the ADAU1701, with through-hole parts (except for the ADAU1701 itself).

I think it would be possible to have an "open-source" initiative to write code for the Arduino interface that would allow real-time interaction with the DSP. And it would be possible to have a collaborative effort to develop a nice GUI to compete with the miniDSP stuff. Just a thought at this point, but since you are finalizing the board layouts, I thought this "next step" might be worth considering.

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Neil - Greetings from 37,000 feet.

That looks totally awesome and it exactly what I was trying to accomplish in the first place. However, costs and GUI concerns were just too much for me to tackle for revision 1.

If we get to a rev 2, I'd love to go in that direction. Or, if you plan to do something with that chip, let me know and I'll help out as I can.

Also - if you have any feedback on the current layout/schematics, I'd love to hear it. Things are really cramped and the fixed layout positions (dictated by the case holes), have limited the "best" layouts....

[neildavis]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Or, if you plan to do something with that chip, let me know and I'll help out as I can.

I've got software that will control the ADAU1701 in real-time, but it's still pretty rough. The software is targeted to active loudspeaker builders, as it lets you read in the measurement files (FRD) and tweak the filters to create up to 6-way systems. The code assumes that the ADAU1701 has already been configured as a general-purpose DSP using the Analog Devices SigmaStudio tool, and it processes the SigmaStudio output files to determine where the volume registers, filter coefficients, muxes, etc. have been assigned. The software is similar in scope to the miniDSP plug-ins, but it also allows redefining the DSP architecture using SigmaStudio. This software still has some bugs and quirks and needs lots of polish, but it works!

And like miniDSP, there is a micro between the software and the ADAU1701 chip that receives the commands from the software and converts to the low-level I2C reads and writes at the ADAU1701. The micro also programs the ADAU1701 on power-up, and does local volume control and input switching. I use a $3 Freescale micro and assembly code to do this, but it could be done with an Arduino board such as the SparkFun RedBoard, which sells for $20. The Redboard uses the FTDI USB interface, which my software supports. The Arduino code would need to be developed, but I could fairly easily translate the code in my micro (with some time, of course).

So it wouldn't be that hard to make this work--all of the risk areas have already been addressed, and what's left is a lot of refinement and some "standard" hardware. With your motherboard for cheap amps, this could be a very interesting "product".

Also - if you have any feedback on the current layout/schematics, I'd love to hear it. Things are really cramped and the fixed layout positions (dictated by the case holes), have limited the "best" layouts....

It's hard to comment on the layout without having the amp boards to play with. The schematics look OK.

[devinkato]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

So excited for this project - If I understand correctly, in its current iteration, the 3 outputs are derived from a single stereo input.

Is it possible for the 3 outputs to be derived from 3 inputs for those of us with external crossovers? Sub + L + R. i.e. Sound cards w/ sub outputs + miniDSP.

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

So excited for this project - If I understand correctly, in its current iteration, the 3 outputs are derived from a single stereo input.

Is it possible for the 3 outputs to be derived from 3 inputs for those of us with external crossovers? Sub + L + R. i.e. Sound cards w/ sub outputs + miniDSP.

Your current understanding is correct - a mono signal is derived from a 2-channel input.

Your wish to use a subwoofer output from a dedicated sound card is possible, but it would require some modification to the chassis and the board. You'd need to provide another RCA input (dedicated for the subwoofer in) on the chassis. On the board, you'd have to cut the "LEFT" and "RIGHT" traces that lead into the subwoofer volume pot (R12). Once the traces are cut, you'll need to solder the mono signal's RCA jack to the volume pot (where LEFT and RIGHT were cut).

Keep in mind, however, that the subwoofer output on a computer's sound card generally only receives signal from the LFE channel of a multi-channel stream. Therefore, most music will not take advantage of this output unless the sound card is specifically set up to allow for this.

[devinkato]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Keep in mind, however, that the subwoofer output on a computer's sound card generally only receives signal from the LFE channel of a multi-channel stream. Therefore, most music will not take advantage of this output unless the sound card is specifically set up to allow for this.

A lot of the newer sound cards actually will let you define a crossover point for the sub - it's rudimentary, but not simply LFE. For my particular application, I want to utilize the amp with a MiniDSP for ultimate flexibility. I imagine I could also bypass the subwoofer pot completely and rely on the DSP for volume control.

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

A lot of the newer sound cards actually will let you define a crossover point for the sub - it's rudimentary, but not simply LFE. For my particular application, I want to utilize the amp with a MiniDSP for ultimate flexibility. I imagine I could also bypass the subwoofer pot completely and rely on the DSP for volume control.

If that's the case - things get a lot easier. You would not install the subwoofer pot (which provides you a lot more space to install an RCA input for the mono signal), and change a few values/etc, but it's doable.

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Just as a quick update - these have been sent out for manufacturing. I hope to get them back in about 2 weeks.

[Silver1omo]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Awesome!

[neildavis]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Just as a quick update - these have been sent out for manufacturing. I hope to get them back in about 2 weeks.

I've been doing some "planning ahead" in my software for an ADAU1701-based version. Maybe it will happen, maybe it won't, but at least it won't require major code re-writes if it does.

But it got me thinking about that common clock problem that was discussed earlier in this thread. Do you or anyone else know whether these amps really need to use a common clock when multiple devices share the same supply? I know that NXP and ST recommend using the same clock for multiple amps to avoid audio beat frequencies, and Tripath recommends setting up the clocks frequency for multiple channels to be greater than 40KHz difference. TI provides a sync capability for the TPA3116, but they don't provide guidance for when it is needed. I don't know whether there is a power level for which the different clocks starts to be a concern, or whether it can be addressed with better power supply filtering. Using the amps for different frequency bands as you do in this 2.1 amp might eliminate any potential "beating". I've just read what it says in the datasheets, but I don't know how much you can actually get away with in real-world designs. Also, I'm not clear on whether the "beating" is a power supply issue or an artifact of the residual switching frequency in the outputs.

So I was wondering whether you have been able to experiment with the stock amps or whether you went ahead and tried to synchronize the clocks.

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Neil - one of my good friends is the applications manager for TI's amplifier group. I asked him about the "beat tone" issue, and he said it so rarely shows up, they actually pulled the SYNC pin out of one of their products. However, a large customer complained without reason, and they kept it in on most all products.

Basically, if the board has any reasonable layout and bypass caps on the supply, the beat tones won't be a real issue. These boards are actually quite well laid out, so I'm not expecting any issues.

Regardless, that wasn't something I could plan on with the base board I designed. It would have to be blue-wired in any case, and it would be different for the two boards I designed for. We'll see what really happens when everything shows up and I get debugging.

[neildavis]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Basically, if the board has any reasonable layout and bypass caps on the supply, the beat tones won't be a real issue. These boards are actually quite well laid out, so I'm not expecting any issues.

Thanks for the response. I've never gotten any real concrete information about this issue--I've just seen the warnings in the spec sheet, but I've never seen anything else written about it describing real-world experiences. So I've wondered what it takes to make it show up and whether there are other ways to prevent it without using a common clock (such as inductors to isolate the power supplies, extra filtering, etc.). The TPA3116 amps use closed loop outputs, resulting in a relatively high PSRR. So the TDA7492 amp would probably be the one to watch for that problem.

Just a "warning": I'll probably try to twist your arm a bit to go ahead with an ADAU1701-based version of this design :o . Charlie Laub had posted about that "open source" miniDSP effort from Germany, but I'd really like to see someone on this board come up with a better version of that design. The ADAU1701 hardware is fairly simple--there isn't anything really difficult about it, other than the "normal" complexity of making a board suitable for production. What's a lot harder is the software, but I've got all of the pieces for that working in various forms. I can put together a simple GUI that I would make available for free.

I'd also like to branch out and use the TI versions of the ADAU1701, but I need to spend some more time with their catalog to pick a good part to use. I wanted to use the TLV320AIC3256 and even bought one of the evaluation boards. I started to add it to a prototype board I was making , and then realized that it uses a QFN package with .4mm lead spacing. I've been able to solder .5mm chips, but .4mm was new territory for me that I didn't want to explore.

[Pete Basel]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Nice work.

I was recently looking at Linkwitz' Pluto and LXmini thinking that the cost of the
amps for an active 2-way really add up fast. This, even without your board could
provide 4 ch of amplification. It would be nice if Linkwitz' active crossover board
would fit in place of yours - I'll have to check the dimensions.

I was also thinking that it would be nice to have a 2nd order high pass in the
subwoofer path to provide a peaked response for 6th order alignments. I think
that it might work hacked into the first sub path OP amp section.

[Pete Basel]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

20V 6.75A 135W Lenovo power adapter, not so cheap though and I don't trust the cheap clones:
http://www.ebay.com/itm/Original-OEM...item462db240fa

Have you considered using 2 one for each amp board?

Did you have a particular one in mind?

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

20V 6.75A 135W Lenovo power adapter, not so cheap though and I don't trust the cheap clones:
http://www.ebay.com/itm/Original-OEM...item462db240fa

Have you considered using 2 one for each amp board?

Did you have a particular one in mind?

Pete - given that the board was designed to fit into the SURE case specifically, I was forced to work with only one power connector. I've tried my best to make the design capable of handling as much current as possible, but the DC jack and the front panel switch are two limitation points.

Now, actually seeing 6-7 amps of real, sustained current seems very unlikely. The amps will likely start to go into thermal shutdown before they get to that point (and your ears will be bleeding).

The DC jack was designed for the CUI PJ-082BH, which is a 2.5mm x 5.5mm standard jack capable of handling 10A.
http://www.digikey.com/product-detai...HCT-ND/3102547

The switch was much tougher to source. It was designed for a funky vertical toggle type, which very few suppliers carry. I ordered this 6A version off eBay (and I hope they'll hold up):
http://www.ebay.com/itm/390432567752...%3AMEBIDX%3AIT

For the power supply, I believe I mentioned it earlier in the thread, but I recommend a 150W Asus adapter, like this one:
http://www.ebay.com/itm/Genuine-Orig...item1c2053421f

That lenovo one you linked is nice, but I don't know where to source the mating connector from, and I'm not sure the board would be designed for that mating connector.