sureelectronics Tripath boards?

[erikb1971]

so that would be these:


and while I am at it... where to find the low pass filter situated here on the old board:



Tnx

Erik

[v-bro]

It isn't to hard to improve the input section on this amp with better input capacitors, but as for the output section, which is clearly the weakest link in the chain there isn't much possible to improve it. Problem is the high switching frequency. Here the board, which is THE main component of the amplifier is impossible to improve without a re-design.

Though, without being all negative you may find some improvement from using better input capacitors. You do seem to have spotted them there indeed...

To elaborate on the above, as soon as one understands there are parasitic components of L, R and C in any board design and even in components these are not an issue as long as they do not present too much negative influence on the main functioning of the intended circuit. Here the output filter capacitors chosen are also ceramic. These are nice and small and do function, but are not even half as good as a polyester capacitor.

The problem arises when we would stick a bunch of polyester caps on the surface of the board and extend the leads, as I've seen many mods done. The extended wire length presents a series-connected parasitic inductor (L) and resistor (R) to the capacitor (C). The inductor will then start to radiate AND pickup magnetic energy, which is abundantly present in such an output filter. Usually the board designer has taken care of that in his design by a proper lay-out where the influences are for instance shielded by ground planes, or if that presents unwanted capacitive coupling he can place the component in a way that the energy is not influencing (see bifilar winding techniques for instance, or the placement of loudspeaker filter inductors)

THE problem here is the high frequency. As you may have noticed the higher frequencies we want to filter the lower the values we need. That can be an advantage, but with parasitic components it is a disadvantage. In these switching amps we are speaking of fairly high switching frequencies of up to 2,5Mhz in some models.

Reading the Tripath application notes there are some tips and tricks mentioned regarding optimal board lay-outs and these are not even always working out optimally. Bare in mind that the lay-outs ideally separate the effects of the output section as best possible from the input section, in order to prevent high frequency feedback and the resulting distortion and noise.

Now to improve the performance of such an amplifier topology takes a whole different way of reasoning than good oldfashioned amplifiers, though 99% of modifications the online crowd are on about directly project the knowledge of analogue audio circuitry on these switching amplifiers. Using air core inductors is one of them, and the use of polypropene capacitors in the in and output section. The fact that they are superior in speaker filters doesn't automatically mean they are the better choice in the signal path of these amps.

I have worked a great deal with TA2021B amplifiers using ceramic capacitors all over and I must say they sound pretty good due to the fantastic Tripath chip, but after experiencing a design which accomodates film capacitors in the output filter and as input capacitors it is clear which one is the winner.

[v-bro]

P.S. I don't want to try to discourage modding this little amplifier, it's a great learning experience. Just don't expect nirvana in the end. All I want to give is some guidelines in the considerations of choosing parts and design lay-out of such an amplifier. And I want to warn those who plan on sticking on expensive "boutique" parts it is a waste of money and effort and just hope to prevent disappointments.

[Malefoda]

Well said. What and why, I like that VBro!
In the other end of time line in audi history, some vaccum recievers had film/paper caps sheilded to prevent noise catching, beside the fact some polyester caps such as tiny MMK from Evox Rifa sound good, do a small MKP, such as a Wima MKP10 1.5µF in its lower voltage rating, can be sheilded with copper foil, linked to a ground plan point of a pcb, then avoid antenna effect and benefit the good polypropylene land?

[QT0815]

Hi,
have just seen that there was a response to my post. I have modified it some days later, which could be seen here : SureTA-2024 Board integrated in Lepai LP-2020 - Quad Channel amp

The sound of this amp is really good, can't say if it helped to change these caps.
My problems are, that there are a lot of distorsions in my radio. I can't hear radio without noise anymore, i think it is acting in as a jammer somehow. I think the fast switching is the cause. Had it with the lepai too, but i can't disable it anymore hehe..

[v-bro]

Originally Posted by QT0815

Hi,
have just seen that there was a response to my post. I have modified it some days later, which could be seen here : SureTA-2024 Board integrated in Lepai LP-2020 - Quad Channel amp

The sound of this amp is really good, can't say if it helped to change these caps.
My problems are, that there are a lot of distorsions in my radio. I can't hear radio without noise anymore, i think it is acting in as a jammer somehow. I think the fast switching is the cause. Had it with the lepai too, but i can't disable it anymore hehe..

It's called EMI, and it happens with poorly designed switch mode equipment...

Problems can vary from noises in the audio to no phone reception, distortion in video equipment etc. etc. It can even screw your car if it is imposed on the wiring of your car mate... (messing with the ECU)

One reason for all that happening is the Lepai uses a single layer board and virtually no output filter. For the Sure just about the same reasons, especially the inductors in the lowpass output filter are crap de la crap....

[QT0815]

Yeah, i know its the EMI, but i thought it won't be much worse than with the Lepai. Maybe the cables to the speakers are acting like an antenna now. But they are sending the EMI to my radio. I am also using an noise suppressor on the input of the Sureboard. I heard the whine of the generator in my loadspeakers, this whine was dependent on the rpm of my engine - really annoying...

[v-bro]

Ah the always present and irritating whine...

It is nothing more than electromagnetically induced noise. These moving fields trapped in signal path involved circuitry that act as an inductor will super impose it on the audio signal.

You can place a ground loop isolator in the signal path, but it may also make matters worse as that's an electromagnetic operating device just the same.

The art of good circuit design is to prevent these matters from happening in an effective way at the source, without disturbing the audio signal. In the power supply... If you take any well designed high powered audio amplifier for car audio you'll notice the operating voltage of the amp is considerably higher than 12V, it's been boosted to that higher voltage by a DC-DC converter and the nicer designs have an isolated converter. Read: Input and output ground DO NOT CONNECT!

The usual whining noise problem is due to ground loops, which your modded amp is a grand example of. Namely you have split audio in to two circuits each operating on the same power supply and thus share their grounds. If you look upon this circuit closer you'll notice there is a massive loop there. To emphasise the fact that a ground loop is just what it really is, a loop of metal, an inductor 'winding' which can pick up moving fields and current will flow in the loop in the same pattern as the field and this will 'induce' the noise...

The whole point of grounding and shielding is that it reduces signal conductors inductance to virtually zero to make them unsensitive to such noises. Its not even necessary to completely surround the conductor with ground connected metal, but just keep it close by, hence the fact that twisted pair wiring works just as well. On a circuit board this is what's often done on good designs. At the same time many shielding ground planes/wiring can pick up noise just the same and when this shield is also used as the reference of the signal conductor (unbalanced wiring) we are in trouble. Ideally you would disconnect one end of the ground connected shield and break the loop, but you can't because there will be no reference any more.

It doesn't matter where you break the loop as far as eliminating the noise is concerned, it's just a matter of choosing the most smart and feasible place in your application.

The simplest and cheapest for most common men is to take the 'ground loop isolator' that will sell for a buck or two at your car audio grocers. The little metal drum usually fitted with male/femal RCA connector wiring attached to it actually contains two audio transformers, a wonderful way of providing isolation and breaking the loop. Though for a buck or two we can't really expect any form of quality. I have measured many audio transformers with my audio analiser and also with square waves at the scope and even though there are transformers that hardly deteriorate the signal quality these cheap ones do and present terrible ringing, lack of bandwidth and add an abundance of distortion. Nice quality transformers, often found in pro audio equipment of good quality cost at least 20 bucks a piece and for stereo asymmetric signal we need two....

As said the best way to provide isolation with NO influence on the signal path is the use of an isolated power supply. Another advantage is there is the source (MP3 player connected to lighter socket charger) will also not be able to present a ground loop problem. This is the path to bliss guys, really bliss and more endless bliss.....

[prezden]

Well I hope someone is still subscribing to this thread. I have the newest version of this amp and am wondering whether it is still recommended to remove C3, C24, R3 and R16. The configuration around C3 and C24 is quite different with a resistor on either side of the caps. I this something Sure has done to correct the problem with these caps.

[QT0815]

You might read my post SureTA-2024 Board integrated in Lepai LP-2020 - Quad Channel amp , there you could see how i have modified my board. Should be the same revision...