BT928 Headphone Amp Mod

Tharic-Nar

Senior Editor
Staff member
Moderator
I am not an audiophile, but I do enjoy listening to quality audio. For a while now I've been eyeing up an amp so that I'm not limited to my sound card for driving headphones. The majority of soundcards do not have amplifiers in them, so when plugging in high impedance headphones, the volume is often limited and can sound rather muddy - this is when an amp comes in handy.

There are a lot of cheap amps out there ($40-$50) and pretty much all of them just make things louder with little regard to clean signal paths, often resulting in noise from the power input. However, it's possible to improve these low end amps and turn them into something a little more appealing. So thus was born the perfect excuse to break out the soldering kit and experiment.

bt928_mod_desk.jpg


The amp I bought in was the B-Tech BT928 Stereo Headphone Amplifier. This particular model is quite cheap and well known for being modded. I won't be doing anything new, the following mod is quite straight forward and well documented as the 'PinkFloyd' mod. So let's have a look at what's involved.

bt928_mod_inside1.jpg


bt928_mod_inside3.jpg


The first thing to be done, and perhaps the most critical, is the removal of the tone control (50k potentiometer). By default, this works as means of suppressing high frequencies, but in reality all it does is destroy detail and makes everything sound washed out and muddy, even with it 'off' (max). Removing it and the two adjacent green capacitors effectively allows for the full signal to pass through unaltered (sort of).

bt928_mod_inside4.jpg


The next step is to change the op-amp. The NE5532N included is 'good enough', but it's always nice to have something with a little more character. Choosing an op-amp to replace it with is pretty much trial and error; to prevent future headaches, an 8pin DIP socket is mounted instead of resoldering each time. The op-amp I went with to begin with was a Texas Instruments (Burr-Brown) OPA2132. I could have gone with an OPA2134, an OPA627 is about three times the price, and an Analog Devices AD843 is about twice the price of the OPA2132. So as time goes on and I get a better pair of headphones, I may switch op-amp.

bt928_mod_inside2.jpg


With the op-amp replaced, it's time for the big one - switching all the capacitors out for something in the quality bracket (subjective of course). Supposedly for audio, the best caps are low impedance (ESR) with a high AC current ripple. I didn't know exactly what I was looking for at the time I got these caps in, so I will be switching these again later for some Panasonic FR and FC type electrolytic. For now, I am using HD series (high endurance).

The 16v 1000uF cap next to the power input was replaced with a 50v 2200uF. The rest remained as their respective 100uF, 47uF and 10uF - each replaced with a 50v higher grade cap. Why 50v? Because they were cheap at the time and as long as the cap is equal to or exceeds the input voltage, it doesn't matter (higher rated voltage also increases the lifespan of the cap too). Well, this is what I'm told :p. With all said and done, the board looked like this afterwards.

bt928_mod_inside5.jpg


So what did $15 of parts and some soldering do to the quality? Quite a lot actually. First of all, I'm only using a pair of Corsair HS1As at the moment, but even with these I can tell the difference. I will go through each mod and try and explain the differences in sound quality. This is subjective to an extent of course.

With the base unit, the tone control really does get in the way and is only good for killing high frequencies. General quality is rather low in general, coming from a direct input on the ASUS Xonar DX. There was a fair bit of noise on the line, so power filtering would be recommended. The unit comes with a 15v basic DC adapter. Getting a regulated supply would be beneficial. The unit can drive up to 200 Ohms impedance, so it's a lot more powerful than the soundcard.

Removing the tone control immediately released all those suppressed higher frequencies, but there was still noise on the line. Bass frequencies were not that punchy either, but things did improve.

Replacing the op-amp didn't actually have a huge impact. Noise was still on the line, but there was greater clarity in the mid-high range, but started to become borderline tinny.

Replacing the caps made a massive difference. Most of the noise had gone at this point. There is a little bit of click-based rumble, so further filtering is required either on the DC input or via better caps. The big change in audio quality though was a further increase in clarity, but also bass response became a lot more punchy and sharp. This is in fact much improved over what the Xonar DX could do by itself, despite the HS1A's being low impedance. Making EQ adjustments confirmed this and bass became much more alive without blowing out the mid-range. Bear in mind, these caps are not really that great for audio (technically), so when they get replaced with the FR and FC range, it should prove interesting what further improvements can be made.

Overall, the amp has become more powerful too, as I need to set the volume lower compared to stock. The last thing to bear in mind though is burn-in. I've only had the amp for about 10 hours and it'll probably take a week for the caps and op-amp to settle (quite common for audio equipment). So I'll see how things go in the future. Maybe the noise will clear up...

So what's next?

Later I will be fitting a switch so I can turn the amp on and off on the box, since it's always live when plugged in. I will also be switching out the volume control potentiometer too, since if I'm honest, is rather poor (imbalanced channels at different volumes, e.g. left channel can be louder than the right at on point, then even out as you go higher). The caps will be changed too at a later date (next month most likely).

So if you fancy a headphone amp and want to scratch that modding itch, this is definitely something worth trying - if you have the need and on a budget.
 

Tharic-Nar

Senior Editor
Staff member
Moderator
I quite literally hacked on a switch so I don't need to unplug the amp every time I want to turn it off. Had to drill a hole through the PCB, cut the tracer coming out of the DC input and bridge it with some leads from the switch. Yeah, I'm not elegant at all. :p

Pic of the Amp now with the switch on the front instead of tone controls.

bt928_mod_inside6.jpg
 

Rob Williams

Editor-in-Chief
Staff member
Moderator
This project is cool as hell; good work! Makes me envious that I've never tackled a project like this before, because it seems rather simple (though I suspect the amount of study isn't). Glad to know that aside from some minor issues, there has been a large improvement. Interested to see how you make the amp "finished", with no more hiss or other oddities.

Finished product looks good, too :D
 

Tharic-Nar

Senior Editor
Staff member
Moderator
It's nice to get the hands dirty every now and then and physically improve something. The main reason behind this mod was to just practice soldering. I have some understanding of what's going on, since the amp is very simple, but I really do get lost when it comes to balancing resistors and caps based on headphone impedance and the op-amp. My electrical knowledge is rather simple, but the best way for me to learn is to actually experiment.

The noise in the amp though really does seem to be based on the power supply. Moving the cable around the back changes the noise. I've resoldered nearly all the points to make sure it's not a loose connection (some of the joints were awful, gaps in the contacts and such). This helped a little, but not a huge amount. The grounding looks fine too.

I am wondering if there is an end to changes. But me thinks that I will hit diminishing returns soon. Changing the potentiometer (pot) and using a regulated power supply are likely to fix most issues. The further improved caps are not likely to improve audio much more, but worth a try. The only other issue is a 'pop' when turning the system on, and again, that's power related (read something about putting a rectifier in to help). Changing the op-amp isn't likely to yield much more either, again due to diminishing returns.

I'm sure some are thinking; why not just save the money and buy a better amp to begin with? Simple answer really; would I have learnt anything by doing so?
 

DarkStarr

Tech Monkey
No.... The real question is, could you build something of equal or better performance for the same price and put it in a cheap plastic case? Basically grab a breadboard and wire up something like this for less than the total cost.
 

Tharic-Nar

Senior Editor
Staff member
Moderator
Personally, I couldn't unless someone gave me specifics. At some point, sure, but I'd need to do a lot of reasearch first. Also, a plastic case would be bad, need grounding and lots of it (a lot of designs use the case as a ground-trunk). Don't get me wrong though, the temptation is there.
 

Rob Williams

Editor-in-Chief
Staff member
Moderator
Tharic-Nar said:
I'm sure some are thinking; why not just save the money and buy a better amp to begin with? Simple answer really; would I have learnt anything by doing so?

Nope, and that's the important thing.

This would make a good how to article. *hint, hint, hint*

That's an awful subtle hint! :D
 

Tharic-Nar

Senior Editor
Staff member
Moderator
When/if I get the other parts in like the power adapter, balanced potentiometer, and it all works with no noise, I'll write it up. Honestly though, there really isn't anything complicated apart from desoldering. Fixing the 'pop' when it's turned on will likely require an extension, but I'll see if I can figure it out (think someone's done it before).
 

Tharic-Nar

Senior Editor
Staff member
Moderator
Well, I've replaced nearly every single component on this board. The two 100k resistors were replaced with 10k to reduce the gain, so there's more travel on the volume control. It's not as loud, but there's finer control. By comparison, 1/4 volume was very loud with low impedance headphones. I am now running it at 1/2 volume.

The 1/4" jack has a further reduction in gain by replacing the adjacent resistors with 120ohm. This means I can drive low and high impedance headphones without blowing out the volme when switching. It is possible to reintroduce the 50K linear pot and swap the caps for resistors to turn it into a variable gain control - but I've already got a switch installed in its place.

All but the two pF caps were replaced with low ESR, high ripple caps from Panasonic, Rubycon and Nichicon. The 10uF upped to 47uF, 47uF upped to 100uF, 100uF upped to 1000uF, and the main power input cap upped from 1000uF to 3300uF.

All resistors were replaced with metal filaments, so are within 1% tolerance. This didn't make a huge difference to be honest, but they are really cheap, so I'm not too concerned.

Due to the extra travel on the volume, channel imbalance is less of an issue now. All that is left to fix is the damn noise. It's very faint... very very faint, but it's there. And after replacing everything bar the power supply, I'm fairly certain this is now the culprit. The problem is, I could just buy a new 15V supply and still have issues. So I'm looking into various rectifier designs and building my own supply. Where I'm going to put it after, I don't know, but one step at a time.

Final (near) board layout.

bt928_mod_inside7.jpg
 
Top