♫♪ My little cheap Circlophone© ♫♪

[danielwritesbac]

Originally Posted by Elvee

What do you mean exactly?

I do not know how much margin is required by the regulator board itself. A way to see the real question is: What is the minimum margin if not considering any other conditions? For example, what is the minimum margin for the regulator if fed by a source that would not have any significant rail sag?

P.S.
I can then add to the minimum, 1v estimated rectifier noise, and also add the pull down figure from my transformer's datasheet to total up a reasonable heavy load figure to use for the difference between input voltage versus zener?

[terranigma]

Ok, I think it is at finish for now. Amplifier itself with this layout hasn't noise pick up issue (at least with a clean source or nothing connected as source) despite its tremendous input gain.

Hope you enjoy the silence of Circlophone..
_____________________________________
This pcb layout has reported as faulty by its author. Please refer to post #780
♫♪ My little cheap Circlophone© ♫♪
Follow later posts for fixed and actual version.
_____________________________________

[heinz]

I like your layout. :-)

You could leave more space between the 2n3019 , to be able to mount a cooler. I my build the bd140 run quite hot too ...

[Elvee]

Originally Posted by danielwritesbac

I do not know how much margin is required by the regulator board itself. A way to see the real question is: What is the minimum margin if not considering any other conditions? For example, what is the minimum margin for the regulator if fed by a source that would not have any significant rail sag?

Under static conditions, the drop-out can be up to 3V, depending on the output current.

Paradoxically, with the ripple present, the dynamic drop-out will be somewhat lower, because of the peak rectification effect of D3 with C3.

[terranigma]

Originally Posted by heinz

I like your layout. :-)

You could leave more space between the 2n3019 , to be able to mount a cooler. I my build the bd140 run quite hot too ...

Thanks. You could see that there are lots of modifications.

Leaving more space between 3019's was in my mind too, but I considered its side effects. An alternative version that may have ~10mm gap is possible due to request.

BD140's are hot indeed but stays still in modest levels.

[odysseybmx414]

Terra, your PCB has the same amount of gain 10k/470.

[terranigma]

Originally Posted by odysseybmx414

Terra, your PCB has the same amount of gain 10k/470.

I meant open-loop gain.

Elvee's brief:

♫♪ My little cheap Circlophone© ♫♪

[danielwritesbac]

Originally Posted by terranigma

Thanks. You could see that there are lots of modifications.

Leaving more space between 3019's was in my mind too, but I considered its side effects. An alternative version that may have ~10mm gap is possible due to request.

BD140's are hot indeed but stays still in modest levels.

Suitable big inexpensive schottky diodes include SR504, SB540. The 50v and 60v editions have a less optimal voltage drop pattern, according to the datasheets. We're not utilizing the reverse voltage tolerance anyway and so a 20v model could do fine. For derating at this area, higher amperage tolerance is primary consideration. If you play as loud as I do, SB360 is a bad idea twice because it is only 3a with extraneous voltage drop, and so that sort of derating is not as helpful as an SR502, which is 5a with much less voltage drop. See the datasheets for forward voltage versus current charts.
For reference, consider that Polarized power caps on this same circuit have reverse tolerance of about 1v or less so it is fortunate that we're not consuming reverse voltage tolerances in this area.

At the area of 3019, there is heat pooling that could be resolved with a fan or with more favorable spacing or with adding drillings underneath the devices to allow airflow to remove the heat pooling. The resistors in that area also suffer heat pooling and should have a bit wider spacing.

At the area of BD140, there is heat pooling that could be resolved by adding 3 or 4 drillings underneath the heat spreader. However, we shouldn't use a heat spreader instead of a heatsink--A straight up wall gives minimalistic opportunity to conduct heat into air, since a heat spreader without a heatsink is a fast road to nowhere, and that plus no possible air movement, is why your drivers are hot.
Instead of so much heat, the BD140 heatsink should be a good size L-channel (or "T" channel) with holes in and under the L-channel to promote air movement, up, through, and around the heatsink. Then the drivers will stay cooler and decrease the risk of cooking the nearby passive parts.
Also, it would be great to have optional drillings so that the power caps could be located in such a way as to allow the BD140 heatsink to be wider than the board, and it would be nice to decrease capacitor heat exposure.

[danielwritesbac]

It seems that suitably low voltage drop SR, SB diodes (20v, 30v, 40v) are in short supply; however MBR735, MBR745 are widely available.

[terranigma]

Ok, How about this?

16mm diameter decoupling capacitors still possible if output transistors should solder from bottom of pcb which is optimum case for this layout. PCB dimension is 75x55 mm. If it cut out as 75x58-60, decoupling caps can be supported more efficiently. In this case ground plate can be widened with a permanent marker.