[germanium]

Quote:

Originally Posted by Bizarro /img/forum/go_quote.gif I did the measuring with my crappy multimeter and got 2-3mV offset for both terminals. The card was powered but not connected to amp. There was no signal transmitted. I connected the V to caps input and the COM to ground (screw that holds the card in place). Did I do it right?

Yes you did fine. That is typical D.C. offset. It is one of the lowest D.C. offsets you will find on any soundcard or componant output before the coupling caps.

 

[Bizarro]

My friend did the mod and replaced the caps with staples. The new staples sound great ;D
But seriously, the sound definitely improved. Bass is tighter and easier to follow. Sound is also more revealing. Altogether the music from line out sounds now livelier.

BTW, I'm using 49720HAs in IV, which I really like, and LT1057 in buffer (for now).

 

[germanium]

Quote:

Originally Posted by Bizarro /img/forum/go_quote.gif My friend did the mod and replaced the caps with staples. The new staples sound great ;D But seriously, the sound definitely improved. Bass is tighter and easier to follow. Sound is also more revealing. Altogether the music from line out sounds now livelier. BTW, I'm using 49720HAs in IV, which I really like, and LT1057 in buffer (for now).

Glad you liked the finished mod. try comparing the RCA outputed signal now through your amp to headphone & then try the headphone amp itself on the card & I think you will find that they sound much closer to each other in sound now especially if your external amp is D.C. coupled. Even if it is only direct coupled it should be closer in sound to each other.

They would likely sound even better with real copper or tinned copper wire though instead of staples. It would sound slightly fuller without loosing detail.

 

[ROBSCIX]

Do you prefer the bypass them all the time with such circuit or have you used higher quality caps?
I always figured if you had to leave them in place you could always replace them with higher qulaity units. I know in the big picture they are considered bad but for some they want them.

 

[germanium]

Quote:

Originally Posted by ROBSCIX /img/forum/go_quote.gif Do you prefer the bypass them all the time with such circuit or have you used higher quality caps? I always figured if you had to leave them in place you could always replace them with higher qulaity units. I know in the big picture they are considered bad but for some they want them.

If there is excessive D.C. offset I use mettalized film caps if possible as these typically do very little damage to the signal but these caps are quite large for thier given capacity & do not fit everywhere. They can also be more prone to pick up interferance that a very short wire used in the same place due to thier large surface. It is usually best to keep any leads as short as possible especially in a computer. This is impossible with metalized film caps but very easy with a short piece of wire if that is possible.

In the PM I sent you last time you will see that I have in fact used these caps where needed & appropriate. The SACD player I mentioned had too much D.C. offset to direct couple so I used some metalized film caps there.

 

[ROBSCIX]

Sure, I understand they can be needed for input and output sections depending on circuit parameters. Just wanted to gt your take on what you would use if you could only change the type and not bypass them totally.

 

[germanium]

Quote:

Originally Posted by ROBSCIX /img/forum/go_quote.gif Sure, I understand they can be needed for input and output sections depending on circuit parameters. Just wanted to gt your take on what you would use if you could only change the type and not bypass them totally.

Ok

 

[fotios]

Nice thread
The remarks of Germanium for metalized polypropylene (MKP) and polypropylene with tin foils (PP) capacitors are proper.
I have to add only one remark on this issue: although this type of caps is considered as the right stuff for using in hi-fi projects because presents almost zero dielectric absorption (especially the PP type) compared to other film - foil types and absence of inverse leakage current compared to electrolytic caps either polarized or non-polarized, unfortunately in the real world there is nothing without polarization. I have ascertained for this, because i use axial MKPs of the well known brand SCR for coupling various stages of my projects, between them. To my big surprise, i have discovered that a MKP capacitor can presents a significant DC offset in its output according to the direction of its leads! The one lead is connected to the internal foil, and the other to the external. I am not sure which lead is connected to the external foil, but i know that SCR offers the option of marking this lead by request. The better choice, as mentioned correctly by Germanium, it is the DC coupling which means the use of a single wire instead a luxury cap.
Another very important thing, for each one is interested for upgrades, it is the wires used for connections. It is better the use of single instead multi strand wires, as the last presents significant stray capacitance which is very important in stages working in the range of MHz like the clocks. In general, it is preferable the use of single strand silver wires (silver it is less resistive than gold, instead gold has better non corrosive features compared to silver) insulated with Teflon which presents the better dielectric absorption in the region of MHz or GHz.

Fotios

 

[germanium]

Direction of writing usually indicates direction of signal flow in metalized film caps. Outer foil connected to the lead at the tail end of the writing.

While in most coupling situation I found for myself little difference it is highly important to have the correct direction in the instance of negative feedback bleed caps as if the outer foil is not connected to ground the cap can pick up excessive noise & introduce it at very high gain depending on the gain of the amp through the negative feedback loop. At this point it may have some directional characteristics for blocking D.C. offset though technically it shouldn't. The D.C. offset may actually be due to the pressence of RF that has uneven positive & negative halves to the signal in this last situation.

 

[fotios]

Quote:

Originally Posted by germanium /img/forum/go_quote.gif Direction of writing usually indicates direction of signal flow in metalized film caps. Outer foil connected to the lead at the tail end of the writing.

I don't knew it. Thanks for informing. Are you sure that all companies use this rule of writing? To be sure i must contact with SCR to ask for this.

Fotios

 

[germanium]

Quote:

Originally Posted by fotios /img/forum/go_quote.gif I don't knew it. Thanks for informing. Are you sure that all companies use this rule of writing? To be sure i must contact with SCR to ask for this. Fotios

I would go ahead & ask but I'm pretty sure that is the standard procedure for labelling the cap & indicating the outer foil connection. Let me know what you find out either way by posting here so others can have a definative answer to this as well.

 

[leeperry]

Quote:

Originally Posted by fotios /img/forum/go_quote.gif Are you sure that all companies use this rule of writing? To be sure i must contact with SCR to ask for this.

it's true for the Mundorf white MKP, but their tech.support told me that these are not polarized and that it did not matter at all.

 

[fotios]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif it's true for the Mundorf white MKP, but their tech.support told me that these are not polarized and that it did not matter at all.

I had the same knowledge for MKP, until i used them in the coupling of the two main stages of my last power amplifier project.
The first stage, it is like a mini preamplifier inside the whole amplifier used as front end circuit. It serves in 3 things. The first is obvious, the conversion of a balanced signal to single ended. The second it is the level control of input signal so as the amplifier can be used directly with a source, say, a CD player. The third is more complex: Usually, power amplifiers have a Zin from 10KΩ up to 100KΩ, which is defined by a resistor connected across the input socket and the gnd. The higher the value of this resistor, the worse the frequency response of amplifier. From my research, to not infected the real possibility of amplifier in frequency response its input impedance must be 1,5KΩ as much. OTOH an amplifier with a such low Zin, it is a very difficult load to driven from preamps or other signal sources. The solution it is the use of a Z interface between the input and the main amplifier module, with high Zin and low Zout possibility. In my amplifier these interfaces are placed in the back plate. I have designed them with ultra low noise and big GBP discrete devices working with +/-26Vdc supply, to not reduced the overall slew rate of amplifier. The input of these modules is DC coupled. I use as well trimers to eliminate their DC offset in output.
Parenthesis: Because their flexibility, i have in my mind to design in the near future a true symmetrical power amplifier from input to output. True symmetrical designs is the new "toy" in Hi-End implementations.
As i told, i have in practice zero offset in the output of input modules. In the main power modules, i use as well a trimer to eliminate the DC offset in the main outputs. Although the input and the main modules can be DC coupled, for preventing a malfunction of the input module which can present a DC in its output i placed these big MKP caps (4,7μF each) between them. The one leg of this cap is connected at the wiper of the 1KΩ level pot, and the other leg at the input of the power module. Although the input module is trimmed for 0V offset in its output, thus in the MKP cap is applied 0V, at the output leg of cap (either is connected or not at the input of main module) is presented a significant DC offset (from memory almost 0,5V). Of course, with the trimer of the main module i can re-adjust the speaker output at 0V again. That is the case.
Because i was suspicious that MKP axial caps of so big value have a significant surface exposed in any interference, i tried and a lower value radial MKP (2,2μF) of WIMA but the result was the same. Again DC offset. IMHO, axial or radial MKP are suitable only for AC coupled inputs or for decoupling supply rails or as speed-up (bootstrap) devices across bias circuits composed from resistors and for forming input filters in conjunction with resistors. And of course, for passive crossover inside speakers. From memory, i have never seen such caps for coupling different stages inside Hi-End devices. Instead, i have seen big value NP electrolytic caps used for this purpose.

Fotios

 

[fotios]

There is a funny side on this issue with these big luxury axial MKP caps. In the first unit (which is presented in the pictures on my web page) that i made, there are these big SCR caps. Inside my amplifier exists a smart analog computer which offer every kind of protection against malfunctions. Thus, if the input module presents a DC component in its output which in sequence can be applied in the input of power module, the protection is activated immediately by disconnecting the load from the output, and by driving the DC in the gnd. This causes the fuses to be blown preventing any other part from destruction. Consequently, i can omit these big coupling MKP caps. I did this in the second unit, but the customer asked me where are those beautiful Hi-End class MKP caps? To not worry him, i removed from the pcb the short links and resoldered the caps in their place

 

[johnzz4]

Germanium, I'm guessing that the DC offset of the ST should be very close to the STX you tested. Safe assumption?

If so, I will be bypassing caps since I can't go back to the line outs with the lesser soundstage and detail. Right now, I'm favoring the headphones by sending the headphone out to my headphone amp, and the line out is feeding my monitors... I would like my monitors to have the same quality signal that I have going to the headphones.

Let me know your best guess on the DC offset of the ST in comparison to the STX and I'll get soldering.