Powersupply:I'm using the LM317 as regulator for the TDA-tower. The voltage is set to 8.5V. The datasheet states that the LM317 will work without C10 at all, but a 25uF capacitor is recommended for better trensient response. All comparisons were made by using a switch to connect and disconnect the capacitors rapidly and easily to the circuit. By the way, these all were blindtests.
The Caps (repalcement for C10):
Results: Without any capacitor at all the balance between low, mid and high frequency is very convincing. When switching in a Black Gate or a Sanyo, sound is more dynamic and also less harsh, but only at mid and higher frequencies. Some of the bass attack gets lost. The balance turns a little to the bright side. The difference between BG and Sanyo is small. Maybe the BG is less bright. I was not sure what to prefer, with or without cap. Maybe bass performence is not weaker than without capacitor, but the change in relation between mid/top and bass lets the bass disappear a bit. The Panasonic FC 1000uF is defenetly an improvement. The sound is just as balanced as without cap, the before mentionened positive effects are recognizable over the whole frequezy spectrum. Bass is substantial, better than with BG or Sanyo and as textured as without Capacitor. The overall performance is very pleasent. Next I added (parallel to the Panasonic
1000uF) a Vishay 15nF capacitor. The highest frequencies are marvelous. No
harshness, no detail is smeared. Realising that adding a smaller capacitor
cleans and speeds up the sound and also supports a more colorfull performance (rigth now at high frequencies), I switched in a third one, a
Mundorf MCap-ZN 2.7uF. I was hopeing the midrange will follow treble. Well
that's just what happened. Every kind of music was played coherent, relaxed and
dynamic. The third capacitor is really needed for optimal performance. The last test was about finding out if a quite large Lelon 10'000uF elko could still improve something. It turned out to be worthwhile. Music is just more relaxed, more solid and sonore and there is no negative side effect. After these experiences I decided to use four capacitors in parallel in the supply for the chip tower. I tried to keep a constant factor between the capacitance values. My definitiv solution is as follows:
Parallel to the battery:
Results: The above listed capacitors I had around and they were place as close as possible to the DAC board. I did not experiment on exchanging one of the caps for another or adding one more. Switching them in had a positiv effect in terms of relaxation and queitness. There is no negativ effect when adding them. Maybe one could try the same solution as done above for C10 replacement or just use much more capacitance.
Coupling Caps:First some words to the test setup. The signal always had to pass through a volume attenuator (D.A.C.T., 10klog) and a coupling cap (1uF Mundorf Tinfoil) in the amplifer. The input resistance of the amp is rated at 47kOhm. (-3dB at 3Hz). The capacitor being tested was added in one chanel between the DAC output and the potentiometer and was compared to the direct signal path. A/B comparisons were made by mixing my favourite songs to mono and listening left and right chanel over the same speaker.
The Caps (replacement for C5/C6):
Results: The speaker I've tested the caps with, has a low frequency extension as follows: 70Hz: -3 dB, but slow roll of. In the beginning there where two interesting things to test.
With the Epcos 0.33uF the low end extension is clearly limited. There's nothing bad about the mid and high frequencies. They are precise and uncolored, but with the thinned bass response, the balance turns a bit to the bright side. A better overall job does the Mundorf 2.7uF. The treble resolution in not as good as with the small Epcos 0.33uF but it has the same balance as without coupling cap. The 4.7uF Mundorf MCap has the same tonal balance as the 2.7uF MCap-ZN. Maybe a 1uF capacitor would just be large enough in my system. Anyway, I most like the MCap-ZN among the first five caps listed above. None of them sounds harsh, but I realised, that as bigger the capacitance value gets (or as bigger the cap itself gets), the more precision you lose. This loss in precision can first be heared at higher frequencies, but also the bass isn't that textured anymore. That's why I can't say whether an MCap could perform just as well if it were of same value. When comparing the SCR to Mudorf (both 4.7uF), I prefered the Mundorf because of the more defined sound. The Epcos MKV sounds about the same as the SCR polypropylenes, no harshness but lower resolution than the MCap-ZN.
When compared to no capacitor at the output, the MCap- ZN is too slow and offers not enough bass authority. Best bass performance I obtained by using a Panasonic FC 1'000uF. Midrange could be more colorfull and high frequencies are more liquid with the Epcos 0.33uF or with the MCap- ZN. Paralleling a Vishay 15nF opens up the highfrequency spectrum and adding the MCap-ZN brings back the warmth in the midrange. Replacing the MCap-ZN with the 4.7uF Mundorf did not make a difference. This setup is hardly distinguishable from the direct DAC output. I decided to use the same solution as for the powersupply (except from the large BC elko).
|