TA2024 amp improvments.

[onis_uk]

HI all, I've just purchased two of these cheap and crusty amps from ebay. I've always loved the sound of T class amps so thought I'd upgrade my current 1 amp setup to a bi amp config.

I've read a lot on upgrading these little amps on forums such as DIYAudio etc... but many folk over there seem to be pseudo-electronics experts! Offering the classic range of conflicting "information" So, after many years of watching Dave on YouTube I figured this is the best place to come for electronics advice!

The standard practices seem to be, to swap out the input caps for larger 2.2uf ones, the word "audiophiler" on the OEM caps scares me a little!... then just swapping out other caps on the board for bigger/better alternatives.
https://orronoco.blogspot.co.uk/2015/03/tripath-ta2024-components-upgrade-less.html

On the output stage there are 2 inductors and a filter cap, they don't seem to be swapped much but I think they are there to filter out PWM from the digital amp chip?

So basically I'd like to know what are the function of this board that can be improved on?

Are the input caps there to filter out DC going to the amp.. correct? if so what values do you want to select and why?! I've seen people say that best improvements in sound are when this cap is just removed and bypassed altogether...safe??

What can be done to improve the output stage filtering?

Also I see folk use Elna 4700uf caps across the 12+ power to help filter the ripple from the SMPS and help improve bass response etc

[Zero999]

Swapping perfectly good capacitors sounds like BS to me. There is some truth that capacitors can cause distortion in audio circuits but only if the wrong type of capacitor is used poorly. The article talks about changing AC coupling capacitors which is nonsense. The AC coupling capacitors should have a fairly constant voltage across them and no audio frequency signal present. When you think about it logically it makes sense. We want the AC coupling capacitor to have a low impedance to audio frequencies, so all of the AC voltage is dropped across the amplifier's high input impedance.

Is the TA2024 still made? I thought Tripath went out of business years ago?

[onis_uk]

yes tripath isn't about anymore, but I think the land of knockoffs has still been making the chips. These amps sound amazing with some fairly high sensitivity speakers.

So you don't think there are any advantages to changing out components on the board? I would assume form the kind of price point that they would be quite inferior to what can be purchased?

How about focusing more on the 12+ rail and providing the amp a cleaner 12v?

I've seen some guy from Japan (via google translate) making a small 5v+ supply to replace the 5v output from the chip (see pin 1)

http://www.datasheetcatalog.com/datasheets_pdf/T/A/2/0/TA2024.shtml

[DimitriP]

but many folk over there seem to be pseudo-electronics experts! Offering the classic range of conflicting "information"

Information is rarely conflicting, opinions are.
Just like "here", there are some great people at diyaudio and changing forums  doesn't lessen the "opinion noise", it just changes it.

[onis_uk]

but many folk over there seem to be pseudo-electronics experts! Offering the classic range of conflicting "information"

Information is rarely conflicting, opinions are.
Just like "here", there are some great people at diyaudio and changing forums  doesn't lessen the "opinion noise", it just changes it.

yes but I feel more likely to contact someone that actually knows what they are talking about, on diyaudio they just suggest things without much proof/reason on why we should be doing such mods.

[Zero999]

How about focusing more on the 12+ rail and providing the amp a cleaner 12v?

I doubt a clean supply is that important. According to the datasheet, the minimum supply rejection is 60dB, with 80dB being typical.

I've seen some guy from Japan (via google translate) making a small 5v+ supply to replace the 5v output from the chip (see pin 1)

And I've seen people put magic stones on top of their speakers because they believe it will enhance the sound quality.

[onis_uk]

Lol you just reminded me of this.........



So, in all seriousness you don't feel there is anything worth doing that could improve the output from this amp?

I take it connecting the inputs together is fine on this amp? I have two of them to run a bi-amp set up, one for each speaker, then 1 output for each of the two drivers.

[Hideki]

So, in all seriousness you don't feel there is anything worth doing that could improve the output from this amp?

In all seriousness, there is nothing you can do to improve it until you figure out for sure what its shortcomings and problems are. Until you know (believing is not enough), any change is only going to be audiophoolery.

[onis_uk]

oh, that's disapointing, I was hoping for a more in-depth discussion on how the amplified audo signal is recovered from the high frequency switching signal...

I can see from the board it passes through some caps and inductors just wondering how the output stage works?

Im guessing to test amplifiers like this i would need a DOSC and an expensive signal gen?

[Hideki]

oh, that's disapointing, I was hoping for a more in-depth discussion on how the amplified audo signal is recovered from the high frequency switching signal...

Well, this is not a dedicated "audio hacker" forum, so I'm not sure why you expected that. You would have to find someone familiar with that exact chip. I assume you have downloaded and studied the datasheet, which (very briefly) explains the function of the different components.  I assume the circuit board you linked to is more or less exactly the same as the application/test citcuit in the datasheet?

I can see from the board it passes through some caps and inductors just wondering how the output stage works?

It's really just a lowpass filter with series inductors, then caps to ground.

Then there's a zobel network which terminates the output filter at high frequencies. The combination of RZ and CZ minimizes peaking of the output filter under both no load conditions or with real world loads, including loudspeakers which usually exhibit a rising impedance with frequency.

Then there's a Differential Output Capacitor. Differential noise decoupling for reduction of conducted emissions. Must be located near chassis exit point for maximum effectiveness. All this information comes straight from the datasheet.

Im guessing to test amplifiers like this i would need a DOSC and an expensive signal gen?

You will likely need more test equipment than you already have. But I fail to see the point in trying to improve this kind of circuit. Take the input capacitors for example. Why are they there? Well, if you study the datasheet, it specifies that the inputs are at a 2.4V offset (that's the voltage smoothed by the BIASCAP at pin 16). So... unless you want to send this offset to the output of whatever is connected to the inputs, it would be normal to have a capacitor there, and it would help avoid loud clicks if you plug something while the amp is powered up. 20k won't let a lot of current flow and the output might already have output capacitors, so it's unlikely that you really damage anything by leaving them out. 2.2u is straight from the datasheet. With 20k resistors you get a cutoff at about 3.6Hz.

Do electrolytic caps generate some distortion at low frequencies when the voltage across them changes? Yes, they are ever so slightly nonlinear if the cap value is low enough. Increase it and the problem mostly goes away.

<rant>
BUT!!! Does doing this actually IMPROVE anything? Can you hear it? Are you absolutely sure that you don't PREFER the sound of slight low frequency distortion? You have to be very careful about such effects. Most tweakers fail this test miserably and spend their life tweaking various components until it sounds better. Sometimes that is a valid "measurement" and sometimes others will agree that it actually sounds better... But human hearing is not a reliable piece of test equipment, so it's very easy believe instead of know, and fall into the audiophool crowd. It's also much much easier to see what tweaks other people have done, where they report great improvements, and just copy that, without much understanding and with absolutely no objective measurements.
</rant>

[Audioguru]

The cheap amplifier has two huge red capacitors rated at 400V! They were probably ripped from a vacuum tube amplifier 55 years ago. Film capacitors today are small and rated at 50V or 63V.

[Tac Eht Xilef]

The cheap amplifier has two huge red capacitors rated at 400V! They were probably ripped from a vacuum tube amplifier 55 years ago. Film capacitors today are small and rated at 50V or 63V.

Nah they're just bog-standard 400V (allegedly) axial leaded MKP film capacitors from China. Their purpose in the circuit is to be Big and Red, not small and yellow (or worse, white) like appropriately-rated caps from other manufacturers.

The usual suspects ascribe almost mystical qualities to them, except in reverse - where other caps are 'like lifting a veil from the sound', Audiophiler caps are described as "almost like if a blanket was covering my speakers" (an actual quote). Some audiophiles suggest they might be OK for audio coupling, 'but I'd never use them in a power supply' .

[innkeeper]

There's your problem, its got no tubes!

on a more serious note... what is wrong with it that your trying to improve....
if you want to improve on an amp, characterize it, find its flaws and work on improving those.
if you cant measure the difference you darn sure cant hear it, except in your imagination.

[onis_uk]

oh, that's disapointing, I was hoping for a more in-depth discussion on how the amplified audo signal is recovered from the high frequency switching signal...

Well, this is not a dedicated "audio hacker" forum, so I'm not sure why you expected that. You would have to find someone familiar with that exact chip. I assume you have downloaded and studied the datasheet, which (very briefly) explains the function of the different components.  I assume the circuit board you linked to is more or less exactly the same as the application/test citcuit in the datasheet?

This is exactly the kind of moronic crap you get on forums, you come along with your jumped-up attitude..... this as an electronics forum so im asking electronics based questions.

Thanks for nothing.

[onis_uk]

The cheap amplifier has two huge red capacitors rated at 400V! They were probably ripped from a vacuum tube amplifier 55 years ago. Film capacitors today are small and rated at 50V or 63V.

Yes exactly, they are just there to dc couple to ground? so the AC going through them will only be 6vpp from any source equipment?!

[gildasd]

The cheap amplifier has two huge red capacitors rated at 400V! They were probably ripped from a vacuum tube amplifier 55 years ago. Film capacitors today are small and rated at 50V or 63V.

Yes exactly, they are just there to dc couple to ground? so the AC going through them will only be 6vpp from any source equipment?!

On a totally different project (amplifying a signal maybe x100 fainter than your input), it took me maybe 100 hours to filter the amplification to ground on an oscilloscope to the level of "noise free-ism" I needed.
I highly doubt that one BIG cap does a better job than three or four properly rated (but less sexy) caps that have been tested in circuit.
Secondly, using the same oscilloscope and a clean input signal (a sine is great), I chased all the noise sources I could find (notably between power and ground on ICs) methodically.
With this knowledge, I made a custom board on Kicad, got it made in China, populated it then retested/corrected/bodged to an other magnitude of extra cleanness...

If you want to make a cleaner amp (that might be technically superior - but worse sounding to then human ear...) there is no free lunch, you are going to have to work on those solder burns and sell that magic cable for a 100Mhz oscilloscope.

[onis_uk]

Yes this is the main problem, with audio equipment, as noted before the human ear isn't flawless.... a much "cleaner" signal may sound horrid.

I had a DAC once... didnt like it, too harsh and "bright" sounding with a very subdued bass.

On a different note, would it be possible to join both the speaker outputs to one speaker? 8ohm load?

[gildasd]

Yes this is the main problem, with audio equipment, as noted before the human ear isn't flawless.... a much "cleaner" signal may sound horrid.

I had a DAC once... didnt like it, too harsh and "bright" sounding with a very subdued bass.

On a different note, would it be possible to join both the speaker outputs to one speaker? 8ohm load?

Do a test with a dummy, if your 8 ohms resistor (correct voltage class of course) catches fire, then no...
Once that is done, with an "el cheapo" scope, see what difference the in signal has with the out.
A basic 10£ signal gen would help a lot.