Roender's FC-100 prototype and builder's thread

[Rudi_Ratlos]

Mihai,

I did some research on "parasitic and intrinsic non linear cross layers PCB capacitance" in the INTERNET and came across some interesting articles.

For example:

Parallel Plate Capacitor Capacitance Calculator

allows me to calculate the cross-layer capacitance that I introduced with the layout presented it post #38.

Provided that my GND-plane has a size of about 15 square centimeters, the capacitance that I have introduced by my layout is about 31.2 pf.

What is interesting as well is that the articles tell about effects introduced by any cross-layer capacitances.
It seems that you are misregarding your own recommendations, since your layout shows a cross-layer capacitance between
the output-track and the +/- backend power tracks with approximately the same value.

Allthough the value of about 30pF is very small, it means to have a capacitor in parallel to the speaker, "which may be risky".

METAL and I re-designed the tracks on the backend and will avoid any cross-layer capacitances in the new layout.

I have 2 questions to you:

1) The position of the GND-leg of the ZOBEL's cap in the middle of the frontend ("the small signal's) GND - plane:
Is it wise to put it there? Shouldn't I better define a third, isolated GND-PAD and run a wire from there to the STAR-Ground?

2) My second questions is related to the takeoff-point of the NFB.
Dr. Self recommends a point "that is in some distance from where the 2 half-waves come together".
In your RMI FC-100 thread somebody wrote that he has run a small shielded cable from the NFB takeoff to resistor R18.
Do you approve this idea?

@LUKE: your 12.000µF ELNA capacitor is very applicable! Do not forget that you have capacitors C7 and C8 in parallel to it,
summing up to a total capacitance of about 16.000µF per power rail. This should be enough!

I am looking to your answer, Mihai.

Best regards - Rudi_Ratlos

[roender]

Hi Rudi,

There are always some trade-offs in any good PCB design and you should know that there is no free meal :-)
If I had to choose between dirty ground plane and dirty power supply lines, which by the way are already affected by half waves output currents (when the amp is passing from A to B classe), I would always choose a clean reference ground. Hence I had preferred to have thick power lines capacitively coupled with the output plane.
Now, regarding your questions:
1. The zobel's network ground point is connected exactly where it should be, at the returning speaker's ground point, and, in fact, that point is the star ground.
2. The NFB take of point is designed in strictly accordance with Dr. Self recommendations. I am not convinced that a shielded cable will improve amplifier's performance or stability.
I wanna make myself clear again. I'm glad that you wanna redesign a tested PCB in order to accommodate various components but you have to test your RMI-FC100 amp based on newly designed PCB before offer it to others.

Best Regards,
Mihai

[Rudi_Ratlos]

My friend METAL and I did some work on revision 2 of the layout for Mihai's RMI FC-100 yesterday and - with some pride -
will show you the current stage:



http://www.abload.de/img/topviewrcys0.png

We tried to avoid any parasitic capacitance on the backend and did a tight but smooth layout of the frontend.

Moreover I inserted a decoupling capacitor at the input for those of you DIYers, who connect DIYed attenuators, source selectors, pre-AMPs, ...
to this AMP. The use of this capacitor can be deactivated - for the sake of sound - buy a simple jumper.

We will now check the schematics once more, compare it to the layout (this will be done automatically by EAGLE), do some additional cosmetics,
prepare the silkscreen, ...

The size of the PCB is: 173 x 73 mm.

Best regards - Rudi_Ratlos

[roender]

The speaker return point must be where label "Frontend Power GND" is placed, otherwise hum noise will occur

[Rudi_Ratlos]

I corrected this, Mihai. I sent you a PM.

Thanks and best regards - Rudi_Ratlos

[AndrewT]

Could you introduce a 1.5mm gap between the pair of fuses and the red trace?
This would allow the PSU PCB to be separated from the Amplifier PCB.

Does anyone see any technical problem with this request?

[Rudi_Ratlos]

I see, what you mean, Andrew:



http://www.abload.de/img/topviewlgbqv.png

If you do not like to use the PSU on the right side, you may cut the PCB at the indicated line.
I can ask and insert a "predetermined breaking point " there.

Question: Shall I put the rectifier-diodes on the right side of the PCB as well?

Best regards - Rudi_Ratlos

P.S.: Andrew: still 40 days to go.

[Sheldon]

I'd modify the traces around the mounting hole for the 2SC3423 transistors that are mounted under the board, so that the access hole can be sized generously. The 2SA1360 is good. Speaking from experience, you want plenty of room to insert the mounting screws and washers from the top of the board.

Sheldon

[pooge]

I would say "no" to putting rectifiers on board.

I would suggest adding a place to put in an optional rail-to-rail capacitor.
Bob Cordell has advocated using at least a 500uf cap for this purpose.
This would be especially useful if the main PS caps were taken off board.

A typical radial snap in cap of 560uf/100v rating would have a 10mm lead spacing, and be 22-25mm in diameter.

ECO-S2AP681BA Panasonic Electronic Components Aluminum Electrolytic Capacitors - Snap In

While there doesn't appear to be any room presently for a radial cap, even if shorter fuses were used, perhaps a plurality of smaller value axial caps could be mounted under the board and connected rail-to-rail around the area of the emitter resistors. This might not be practical, either, due to the diameter of the axial caps causing to great a spacing between the board and heat sink.

Whether or not this is provided, space could easily be provided for an axial cap along and to the left of the break line, especially if the break line was moved slightly to the right to accommodate the diameter of the axial cap. The star ground could then be shifted between the ends of the fuses. This would be really helpful to those who take the main caps off board, providing a shorter path for half-wave currents that would normally circulate back to the main caps off board.

Regarding the fuses, I would recommend providing options for different length fuses. Further, vias could be added for soldering in a fuse like this:

http://www.cdiweb.com/datasheets/cet/OC-2116.pdf

[Rudi_Ratlos]

@Sheldon: please have a look at this image:

RMI-FC100, a single stage audio power amplifier Post #2

This is, how it has to be done.

Best regards - Rudi_Ratlos