Compact Sized LME49810/11 +ThermalTrak Amp

[Adrculda]

Originally Posted by panson_hk

This is supposed to be the best option.

Perfect, as the 0.1uF is a Vishay MKT 1822 400V and the E-Cap is a Nichicon YXA 63V 100uF

[panson_hk]

Here is a plot for comparing THD of a 470 uF e-cap with and without a 0.1 film cap in parallel. The THD of the film alone is also shown. The curves associated with the 470 uF e-cap do represent the residual distortion of my AP. It meas that no additional distortion can be measured for the e-cap.

Well, do we really need to parallel a film cap with the e-cap in NFB path. Yes, we can since it does nothing wrong objectively. It may make subjective differences. I don't know for sure. However, it does make some people feeling better who believe e-cap must be bypass by a film cap for better high-freq characteristic. This is true in certain applications, but not universal. Simply copy and paste does not engineering.

[Adrculda]

i dont see plot....

[panson_hk]

sorry, here is the plot

[Adrculda]

Even though picture quality is poor, there is a slight difference.

the film doesnt seem to handle lower frequencies too well, yet the electrolitic does a good job at "dampening" that. These 2 caps seem to compliment each other, where one makes up for the shortcoming of the other one.

[panson_hk]

I probably have to make HD pictures.

The e-cap alone and in parallel with a film do not show any measurement difference. From the measured result, I can't see a reason to add a film cap.

[AndrewT]

to avoid using any e-cap in the NFB path, JLH used a pair of 4.7uF film caps in parallel and adjusted the NFB resistors up in value to bring back some bass response.
He did this even though the very high resistor values would increase amplifier noise slightly because he thought the sound quality improvement brought about by the no e-cap solution was a better compromise than reducing noise.

[klewis]

Originally Posted by panson_hk

I probably have to make HD pictures.

The e-cap alone and in parallel with a film do not show any measurement difference. From the measured result, I can't see a reason to add a film cap.

I just came across this article, I haven't looked for these caps, don't know if they are available. The article mentions "very good frequency response" that these caps are better for decoupling than standard alum electros.

http://archive.chipcenter.com/eexper...kruger007.html

[AndrewT]

Originally Posted by klewis

...... The article mentions "very good frequency response" that these caps are better for decoupling than standard alum electros.

but only for decoupling and particularly good for decoupling VHF. This is where they perform very well, in digital circuit with their very fast ON/OFF/ON/OFF going into the Gswitches/second.

[panson_hk]

I am working on a new layout for 3-lead power devices. KiCad is employed for the new board. Here is the schematic. Compared to the previous version, the following items are added.

• input DC blocking film cap
• 7.5 mm and 10 mm film cap footprints
• multiple footprints for NFB e-cap supporting ELAN SILMIC II
• 3.5 mm and 5 mm footprints for decoupling e-caps
• output coil and Zobel network
• separate rails for LME

Any comment is welcome.