I thought that I had made al the alterations to the power lines on my Sure Tripath TA2024 based amplifier modules. However over on DIYAudio.com Albin posted another suggested mod which I have to admit I was sceptical about. The modification affects the topology of the power lines on the Sure Tripath 2024.
Albins suggestion basically is to separate the VDD1 and VDD2 power lines going into the Tripath TA2024. The VDD1 and VDD2 power lines power the left and right channels of the amplifier. The theory behind this I suppose is that by separating the power to each channel further up the power line means that each channel has its own independent bank of capacitance; demands on one power line have less effect on the other. Albin achieved this by cutting a trace on the Sure Tripath TA2024 amplifier PCB and soldering on some wire to directly connect the left and right power lines at the DC input to the module.
I have to admit I was a little sceptical about this. The sure boards at first sight appear to have nice thick traces for the ground and power. Looking at the board; the power comes in through either the round DC plug, or through either of the +/- terminal blocks (top and bottom). The three power inputs are interconnected on the board. The power then makes its way through a Diode (D1). This diode is connected to the Left Channels bank of capacitors. It then is routed on the underside of the board. On the underside, it passes under the small bypass capacitors C9, C10, C18, C19 and to the right hand side of the board. Here it is routed to the topside where it feeds the Right Channels bank of capacitors.
Calling the banks of capacitors the "left channel and right channel capacitor banks" is not really fair. As the banks of
What Albin was suggesting was that you break the trace on the underside of the board, effectively you split the capacitor banks in half, and one half is now feeding VDD1 and the other VDD2. The power line is effectively being split at the input to the board. Albins suggestion also effectively bypasses the Diode labelled D1 on the board. As seen in my previous power line modifications this raises the voltage to the chip by 0.3V.
I had started to wonder if this raising of the voltage is what had improved Albin's perception of the sound. He had described his improvement as "as for the sound, to me it was like not working and working." I did know from my experience that bypassing the diode did make a difference to the sound.
Having a better look at the board, I got to thinking that Albin's modification could perhaps be improved on further. I decided that what I would do would be to separate the VDD1 and VDD2 power lines even further up the power line.
Sure have thoughtfully provided two sets of terminal blocks. This allows you to connect power at either side of the board. Stock they are linked with a trace on the underside of the board. I decided to split this link as well and have two cables (one per channel) bring power into the board.
You can see in the picture how the powerlines are ow to be configured, The banks of capacitors now each serve a channel of the amplifier. The powerlines are seperate right up the line to the PSU. Each channel of the amp now will work more indendantly than before and hopefully be cleaner for this. Theroretically two power supply units could be used, one per channel.
I started by splitting the power line on the underside of the board as per Albin's suggestion. I cut the trace with a knife twice. Then I removed the intermediate copper to remove a strip between the traces.
You can see on the photograph the correct place to make the cut. It's in the middle of the board, splitting the trace through the 8 vias that pass up to the bypass capacitors.
I soldered one side of my wire jumper to the positive post of the terminal block.
Next I scraped back some of the copper at the outside edge of the positive trace.
If you have to solder wire onto a plane of copper like this I find that it is very much easier if you melt solder to in the exposed copper plane before you attempt to solder the wire on. The solder flows more freely onto more solder than it does to the cold copper.
Once the copper plane had been "tinned" I then held the wire to the newly created solder pad and melted more solder to make a secure solder.
The next thing I did was to remove the barrel power connector from the board. This will be redundant by the time the modifications are complete so I thought I would get rid. This is optional really.
The next thing I did was to remove the diode bridge that I had soldered in earlier. As the power would be taking a more direct route to where it is needed, this now was also redundant. Again I could have left this in, but for tidiness sake I wanted rid. A loop would have been formed if I had left the bridge in place. This pad will also give me a 14v line should I need it later for extra circuitry.
As I am now using the existing terminals for my two separate voltage lines, I had to break the traces that join the power lines of the terminals. This I did on the underside of the board next to what will shortly be the right channels power in.
Lastly I soldered in a second cable to link the remaining positive pin of the other terminal block to the now completely separate power line. I used the same technique as before to solder the wire to the copper plane.
This board was modified to the same configuration as my other test board. It has an added 3000uf per power rail, and has Sonicap 4.7uf capacitors acting as the DC blocking Capacitors on the line input. So I have two amps which are identically configured apart from the power line modifications to be tested.
However I noticed straight away a difference between the two amplifier boards. The board with the original power line topology has a very smooth, but warm bass. Initial impressions of the warm bass are that the instruments have a richness, particularly large drums or double bass. It is I think quite pleasant to listen to.
Swapping over, the warmth of the bass was diminished in the amplifier with the modified power line topology. Initial impressions are that it sounds colder and cleaner. I also felt that the amplifier was more in control of the low tones than the other amplifier.
Listening to the amplifiers back to back with various pieces of music from the Berimbus Album by Renauld Garcia Fons, I found myself liking the amplifier with the modified power line topology more and more. Whilst the amplifier with the original topology had very appealing warmth about its sound, after listening to the new topology, this started to sound more and more as if it was adding a bloom to the sound that masks the detail in the mid range and bass regions.
I have a lot more listening to do to these amplifiers to make my mind up on this modification. Although it looks like it has made a good improvement to the sound, I am aware that I am comparing boards that have not fully burned in. On top of that the capacitors that I have newly fitted have a few hours play time difference between them. I will update this article once I have managed to put some burn in time onto both amplifier boards. Only then will I trust my own positive findings for the modification.
If you try this modification yourself I would appreciate your comments on whether you hear an appreciable difference in sound. I am hearing quite a difference, but I am sceptical that the power line modifications are to be credited for this. The difference just seems too unlikely!
Update
Today I spent a bit more time listening to the two amplifier boards. Previously this week I have been running the amplifiers into resistive loads with some music. There is perhaps another 12 hours been added to the play time.
Listening to them, I felt this time there was less difference between the two amplifiers in terms of tone, I think this comes down to the fact the Sonicap Gen I capacitor on the board with the modified power supply had zero listening hours on it when I tested them last week.
In the initial listening tests the newly modified board sounded colder than the unmodified board. This was less evident in the further tests.
I tried out several pieces of music, Renauld Garcia Fons: Entremundo, Mark Ronson: Version, Katie Melua: Call of the Search and Rem: Automatic for the People.
Although the tone may have altered since the first test on the modified amplifier, the extra detail is definately still there. It can be heard in several places in the frequency spectrum.
In the Bass regions, Drum kicks are more exact and the speaker sounds as if it is being damped more effectively by the amplifier. On the track Valerie, from the Version album, the bass line was more succinct, this gave the track more pace. On the Renault Garcia Fons album, the incredible skill Renauld's double bass playing was revealed in the improved lower end detail. On the track Aqâ Jân, Renauld plays at super high speed using what seems an unlikely number of bow strikes per second on the strings of the strings, each one of them was revealed to be near perfect. Amplifier just illustrated the supreme skill of the musician.
In the mid range, the ring from piano strings was inhanced on the modified amplifier, the origional power line topology added a slight haze to the notes which while not unpleasant, changes the impression of the piano.
The top end was unsuprisingly altered less so by doing the modification, however in busy tracks the impression was of more clarity here. I suspect this was more a byproduct of more clarity in the lower ranges.
On the negative side, whilst this added clarity enhances good production and excellent musicianship, it is less charitable to lesser recordings. In the Mark Ronson album, "Oh My God" sung by Llly Allen, the track plays well untill the point where Lilly sings. The vocal section sounds distinctly muffled and not particularly enjoyable to listen to.
Overall I stand by my origional impressions of this modification. I would consider it well worth doing, either in Albins simpler form or in the more involved manner I detail here. All of my Sure boards will now recieve this same treatment.
Listening to the two amplifiers I was surprised to hear a difference. I have to bear in mind though that the board with the original power line configuration has had a couple of hours playing time with the Sonicap input capacitors in place. It probably has an hour or two more play time on top of that. 26/09/08- Sure TA2024 Signal Line mod Pt 2
- Sure TA2024 Power Line Modifications
- Sure 2024, Signal line Mod Pt 1
- Sure TA2024 DC offset Mod
- Sure TA2024 Tripath, A first look
| Comments |
|
|
||||||||
|
||||||||
|
||||||||
) on the TA2024.
|
||||||||
|
||||||||
|
||||||||
|
||||||||
|
||||||||
