So how much current the chips need? I never read specifically the requirements......I heard T2024 power up using a small size battery....would it sings with around 800mA ?

With a small pea sized transducer, it might just 'squeek'..

I tried with 12V 1.2A/20Hr. Play for about 6hr none stop. I didn't try until it is totally flat because i think that is bad for the amp.
If get bigger one like car 12V battery, I think you can last for days but the charging and gas release from it make you think twice if you want to put it indoor.

This post has been edited by kww: Jun 19 2010, 01:48 PM

No.

I think you get what I meant wrong.

There is a 5V regulator onboard. But I think Its rather weak cos it has to supply quite a few circuits like the digital and analog processing units.

What I did recently was to cut leg 30 off and use 7805 to supply 5V to pin 2 and 8. From the scope there were no more ripples to those pins.

Soundwise, I don't wanna say much cos I was hoping someone else could do a similar mod to verify my results.

Sakaic, Can you take a picture? Tq

You did not explicitly expressed that you were replacing the built-in 5V regulator with an external one!

Surely with over 88% efficiency, Tripath has optimised the design of the circuitary/chip.. I have not come across anyone who uses an external 5VDC supply, so you're the first!
IF it does improve the performance of the chip and the SQ, well good for you!

Waiting for your final verdict/review on such a modification..

The efficiency of 88% comes from the class D design. Mind you, it is supposed to be theoretically 100% efficient.

The Linear regulator is built in. It supplies power to the digital circuitry that is the integrators and the other gain and filter circuitry as well as the analog input section.

In the datasheet its the op-amp on the input side and that big black box they label processing and modulation.

Will upload photos a bit later but basically its just to cut off pin 30 from the pcb and solder the output of the 7805 to that hole to replace the internal 5V regulator.

Preliminary tests show oscillation which is a sign of under current on pin 30. Those I got from the scope with 1khz output and 8 ohm dummy load. Only thought to try it out after some guys in diy audio made a custom board which not only supplies 5V from an external regulator but also has separate regulators for each of the power supplies.

This post has been edited by sakaic: Jun 21 2010, 03:24 PM

As promised, here are some photos of what Ive done so far.

The addition is half of what I want to do to the amp.





the first half is to make sure that power supplies are as clean and steady as possible.

The second is to return the amp to datasheet spec.

I am curious as to how it sounds as it was intended. After all all the THD and performance figures that we are given come from that config.

So far I have changed the diodes at the bottom and reverted the 0.47u output cap to 0.22u as per the datasheet.

Then next is to change the inductors and the 0.1u caps back to original spec.

Also, building a linear power supply at 14.1V using a KA78T12 with 3 diodes at pin 2.

Sakaic, Mind to share what is you impression before and after each mod?

Well.....there are some that I will share cos of the expected results and some that I wanted to ask ppl to verify as a double blind.

The ones I wanna share.

- changing the inductors to the low noise ferrites and 15uH has opened up the sound quite a bit. The sound staging is now better. Echos I guess

-changing the .47uF caps definitely made a big change. On my Usher s-520, the details were much better. Hearing new things. On the downside, there is so much of everything else that the sound might be a tad bright on some speakers like AE Neo 1 and also makes the bass less 'there'

The 7805 mod I am hoping for someone to confirm the results. But I have heard much better drive. It doesn't saturate so easily any more. It still will (low power) on something fast like Viva La Vida but on something vocally strong like jazz etc. it sounds better.

p.s. me currently listening to paradiso girls to destress helps in evaluating the amp. Damn burnt fingers..........

This post has been edited by sakaic: Jun 23 2010, 10:25 PM

wau nice ferrites, those are expensive right?

I think so......got them "illegally" from a friend who took some from the engineering bench

He told me they use that on the tester as these are the type that have low noise emission. And by tester I mean IC tester

The ultra fast diodes really makes different in this little amp ???

Though it only works on digital devices....????

The thing is that it is a digital amp.

From the scope at idle with no inputs already switching at 875kHz to get the 6V each output terminal.

the overshoot is about 10% I haven't scoped it yet but I hope it helps.

I think the bad thing ppl always use the term class-T is that they forget it is still a class D.
Just a very advanced class-D.

As promised here's more or less the final mods I wanna do to the amp minus the linear power supply.




All the output caps and inductors are now datasheet spec for 8ohm speakers.

It sounds VERY different now. very dynamic and detailed. Musical but maybe a tad too bright for some peoples taste.

Has that tube like sound.

This post has been edited by sakaic: Jun 24 2010, 08:02 PM

The term Class T is by Tripath themselves.
Class D is by everyone else.

sakaic, thanks for sharing, looks cool.
A few questions:
-That is a huge power buffer cap on the right side. 0.5F?
-your 7805 doesn't need heat sink?

I found vocal on 2020 to be thin compare to 2024. I change the input cap and it sound a bit warmer but still lacking.

kww,
The strength of 2024 has always been in vocals. I do agree with you on that.
The 2020 will win in terms of lower frequency performance (and more dynamic) as this is where the 2024 is lacking.

I love vocal, can't help to always refer to 2024 for any amp I listen to.

kww, the cap is a 10mF. The 7805 is a bit warm but ok.

The biggest change in performance came from changing the 0.47u out to 0.22u and the inductors.

Makes a lot of difference. Now I can safely say that my 2020 is almost on par vocally with my 2024.

The thing is that this Helder 2020 board is poorly designed and made. The routing for the outputs are a bit weird as far as the placement of components and this causes small differences in the output offset voltage as well as very minor differences in the output (from a scope). To understand what I mean go through the datasheet regarding component placement and pcb design.

BTW I think I found a potential replacement for the tripath series from Sanyo. Not yet released but in terms of architecture very similar.

This post has been edited by sakaic: Jun 25 2010, 12:18 AM

i believe the input cap and the output inductor make the most difference.
changing those other caps do improve but not as much, unless u mean changing to other values such as ur case

but really nice mod there, i tot this board kinda like other design....