Designing a "designer ghetto blaster" - using TK2050 maybe?

[Kremmen]

As there are bound to be audio enthusiasts/engineers around maybe someone would like to share info/opinions/insight on the following.

Recently i landed a deal to design a battery/power management system for a designer portable audio box. However, to make matters worse the audio guy walked from the project and now i am supposed to take over all of the electronics. I have designed and built a fair number of audio amps over the years and also i do power electronics as part of my work so my mindset is ok, let's do this.

There is nothing fancy per se; it is a stereo speaker amp box (1 woofer + 1 tweeter per channel) with mobile phone as the default signal source. Basic controls only (i.e. the 3 classical knobs for vol, bass, treble).  Target amp output is 2x100W into 4 ohm load.  Fair enough although there are extra challenges due to the constraints of the design. All circuit boards need to be custom designed due to the oddly shaped spaces available. Also the customer specifically excludes any ready made subassemblies or boards. It must all be designed and implemented "in house".

The customer, while not technically knowledgeable does have some ideas about the implementation. One suggestion they made was to investigate the TK2050 as the power amp. They have an idea to drive the woofer and tweeter using separate power amps, the tweeter amp could be based on TI's TPA3110. These are suggestions or ideas only, not hard requirements.

I don't have a strong opinion and both seem fine with me, only i cannot find TK2050 chips anywhere. I think Tripath was merged (into Cirrus?) years ago and i have no idea what chips/from where they use in those chinese "TK2050" amps from say Aliexpress and the like.
So that would be my question #1: Are TK2050 chips still available and from where? Neither Mouser or Digikey seem to know anything about them.

Question #2 could be about the separate amps for speaker elements. To me it makes sense in that you can do all the filtering/signal conditioning at the low level signal stage instead of the output power level. Any insight or points to note about this would be good info.

Question #3 then follows from the above: should TK2050 not be available then what would be a worthy replacement in the 2x100W power range? I have looked into TI's TPA3221 but have no practical experience with those. The design as far as circuit and layout are concerned seems straightforward and that i know i can manage. The data sheet will not of course indicate any quirks of pratical caveats and those would be very interesting, as well as the overall impression of sound quality. Although there the various design tradeoffs in implementations may make direct evaluation difficult.

Question #4: If not the above chips, then what? The design is wide open at this moment. The only thing existing is a rough 3D CAD model of the housing with speaker arrangement, and a blank sheet of paper for the electronics.

Question #5: any suggestions regarding volume/bass/treble control? It is of course doable the old fashioned way with analog filters and pots. However, i would't mind a handy chip to do it all in one or something like that. I will of course investigate but any ideas straight away are welcome.

The power management alone will be a bitch, i know that much already. Managing 10 or 20 18650 LiPo batteries in 1p or 2p, 10s without the whole thing catching fire one day, and simultaneously charging while running the amps will take some time to puzzle out. So i would not mind a direct, straightforward solution to the amp question...

[alexanderbrevig]

Some DACs has an EQ block that is programmable. Are the speakers determined? You should have a look at their frequency range and find an appropriate crossover.
Cirrus and TI has a selection of helpful ICs.
Sorry for not linking any particular part numbers...

[Kremmen]

Thanks for the response. Yes, the elements have been provisionally selected - Dayton Audio RS100-4 4 ohm and ND20FB-4 for the tweeter. At least i will have the speaker response curves available for amp tuning.

These:
http://www.daytonaudio.com/index.php/rs100-4-4-reference-full-range-driver-4-ohm.html
http://www.daytonaudio.com/index.php/nd20fb-4-rear-mount-3-4-neodymium-dome-tweeter.html

[alexanderbrevig]

Probably you know this, but none of those can handle 100W (what kind of battery are you using?). Also, that first one is a full range speaker which to me looks fine for a portable setup. I don't know your client nor their customer base, but if it were my product I'd rather try to please the bass hunters our there so it can satisfy their need to party at the beach That is to say, I'd rather pair the full range one with a single small sub element than trying to get the crossover right in the 3.5kHz area where we are sensitive.

Now for a more direct suggestion (which may or may not be totally off the mark for your specs):
I would have four of these guys https://www.maximintegrated.com/en/products/analog/audio/MAX98372.html/tb_tab0 at 12V for ca 13W into 4
Then you can use digital filters on the I2S stream for the crossover and maybe a few 'EQ presets'?

Is the main controller selected?

[Kremmen]

I am aware that the elements won't stand 100W. That is only a design target for the amp in case the elements are changed later in the design cycle. For now, they are fixed and it is not my decision nor do i intend to worry about it. The customer is aware of these things and they have their vision - flawed or not.
In the end of course i need to prevent the amp blowing the speaker cones away but there are ways to do that.

As of now, there is no "main controller" since the entire signal path with all controls is analog. This is the customer's expectation although i don't mind one way or the other. Any number of AVR or ARM MCUs would be fine with me. So with that in mind the MAX chip woud be a radical design paradigm shift. Your suggestion makes sense to me and i would seriously consider it, but i am not sure if it is sellable to the customer.

[alexanderbrevig]

Purely analog? So no iPhone then? I guess they could bring that silly 'aux' cable thing though.

One other thing to consider is the temperature and heat dissipation. I'm also still curious what kind of batteries this thing will have.

Maybe something as simple as the LM1875 would do it (if you convince them that 20W is sufficient)?

[Kremmen]

Well, it may not have dawned on them yet
As a matter of fact i better prod them a bit to do the homework. So far nothing more has been said than that the default source will be a mobile phone.  It may be that nobody has looked into what that implies in practice. It is a sure thing that iPhone owners are a target group so we may see a digital front end real soon now But we'll see.

I did mention the batteries in my first post. Circa 10 pcs 18650 in series, or worst case 2p 10s. Assuming the voltage is needed to power the 100W amps. And i also mentioned that it will be a bitch to design.
Somehow i don't see an LM1875 cutting it. My assumtion is they want those specific speakers and then i need at least 30+15W RMS power per channel plus headroom for momentary overloads.