12 replies to this topic

[georgepapa]

I've already mentioned in a previous post how horrible these amps can sound when pushed, not to mention the slightly grainy frying quality on strings and brass.
Since the whole thing cost me only 40 bucks I've decided to sacrifice it to some butchering type mods.
The output section on this amp is weird, consisting of a parallel LC combination which is basically a band stop filter and something like a zobel network across the terminals to address rising impedance.
I am considering two options.
First is bypassing the entire output filter stage and attempting to limit EMI by using simple clip on chokes on the speaker cables.
Second choice is to replace the existing filter with a more conventional B4 second order low pass filter with superior components and do away with the zobel network.
For the input section I simply intend to remove the high pass capacitors.
Naturally I don't expect these mods to improve distortion figure at high levels. I will at least expect a cleaner reproduction at normal levels.

[VanArn]

A Tripath amplifier is a class D amplifier and the output filter is essential in permitting only the de-coded audio signal to pass through and blocking the carrier(R.F) waveform from reaching the speakers.The speaker impedance is part of the filter equation and should be as required by the amplifier manufacturer's specification. Changing components as suggested appears to be a recipe for a poorer result.Perhaps a look at [url]http://www.diyaudio....050_2x100,might be worthwhile.

[georgepapa]

Just wondering if the zobel would be necessary with a simple lowpass rather than band stop filter. I understand an impedance correction network improves the function of the network by letting the filter see a flat impedance. What about he DC blocking filter at the input stage. Isn't this normally taken care of at the DAC output stage?

[georgepapa]

A Tripath amplifier is a class D amplifier and the output filter is essential in permitting only the de-coded audio signal to pass through and blocking the carrier(R.F) waveform from reaching the speakers.The speaker impedance is part of the filter equation and should be as required by the amplifier manufacturer's specification. Changing components as suggested appears to be a recipe for a poorer result.Perhaps a look at [url]http://www.diyaudio....050_2x100,might be worthwhile.

Thanks for the link. Some of the comments for the suggested mods were not positive. I think a simple quality of parts upgrade might make a subtle difference - as long as it doesn't cost me as much as buying a better amp.

[VanArn]

Class D amplifiers definitely require a zobel network above the audio band to remove all of the carrier signal before it can go to your speakers where it serves only to heat up the voice coils. The only filter variation that is an improvement,especially in preventing R.F. signals entering a conventional amplifier via the speaker cable is the Thiele modified type.If you are determined to make changes,it would pay to use a dummy load and have access to a c.r.o.(plus signal generator) to investigate what effects are getting through,otherwise ,without measurement you end up going nowhere in a hurry.Perhaps there are other issues causing the sound problems you have,starting with the recordings and most likely the actual speakers use.have you tried another amplifier in the system?

[georgepapa]

I have compared this amp to two different chip amps featuring 2020 and TAS5630 chips. I also have an old Hafler and an ME240 and a Yamaha Natural sound power amp. Both chips mentioned I believe produce higher fidelity results than the 2050, through the same speakers and with the same recordings.
Thanks

[georgepapa]

I have decided on a simple quality parts upgrade rather than messing with any of the circuitry. I will do this in three stages and my goal is to try and reduce this amps odd behaviour as the volume climbs.
I will firstly tackle upgrading the zobel resistors and caps. I have attached a photo of zobel network.
[ATTACH]42131[/ATTACH]
Look at R1 and C1 , the zobel network. What on earth are those miniscule components going across a full speaker load? In my experience with crossover filters a very decent high powered metal film resistor and an MCAP are mandatory in this kind of network.
If anybody can recommend high quality MKT caps for input and output filters would be appreciated.
Will report back on Zobel upgrade.

Edited by georgepapa, 11 March 2012 - 08:23 AM.

[georgepapa]

Those zobel components look like the tantalum variety. I have gleaned from other forums that tantalum capacitors yield very high distortion at high voltage levels and are not recommended for audio.

[gainphile]

... Watching this with interest

[Zaphod Beeblebrox]

Those zobel components look like the tantalum variety. I have gleaned from other forums that tantalum capacitors yield very high distortion at high voltage levels and are not recommended for audio.

They may LOOK like tantalums, but they are, mosty assuredly NOT tantalums. By all means replace them with different parts, but don't expect an appreciable difference. They will be some kind of film or monlythic cap.

You need to understand why the Tripath™ products were created, before you waste too much time on them. Tripath™ amps were created to:

* Deliver huge profits to the manufacturer.
* Make it possible for manufacturers to deliver reasonably priced amplification, in a compact package with decent MID-FI performance.

Tripath™ are fundamentally crippled, in comparison to decent quality audio amplifiers. It is not possible to acheive high end performance from them. In fact, I have always argued that better performance is possible from the decent quality chip amps from some of the linear amp manufacturers.

[andreasmaaan]

For what it's worth georgepappa, I have also compared the 2050 to a few other class d amps (sonic impact t-amp, bel canto evo 4) with various combinations of other gear and I have reached very similar conclusions to you. The 2050's sounded grainy and quite screechy, albeit very detailed. A long way behind the sonic impact and the bel canto IMHO.

[georgepapa]

I'm hoping that my mostly negative impressions of the amp are due poor implementation by Sure electronics who appear to have used the cheapest components possible.
I have read very positive reports of TK2050 based amps by Hifimediy and Virtue audio.
I have just ordered about 55$ worth of Vishay film caps and MO resistors from RS online feeling reasonably confident that a complete i/o parts swap to quality high powered components surely must make a discernible difference.

[georgepapa]

Okay! I have just completed the input / output stage cap and resistor upgrade. I shall hopefully post some photos this evening.The total cost of the parts upgrade was more than the cost of the amplifier and the work involved in removing the original components, painstakingly cleaning and opening out solder holes and then the excruciating job of squeezing in the much larger caps and resistors actually turned this into a pain in the arse of a job that I would not recommend you undertake. Having said that the amplifiers does indeed sound cleaner at lower volumes but the significant difference is at higher volumes where the amp not longer exhibits that shrieking unpleasant distorted sound. Rather than performing this upgrade I would recommend those seeking a TK2050 based amp to source out a brand which implements the design with higher quality parts. It's the more economical approach.
Thanks
George