Rod Elliott P03 variant with TTA1943/TTC5200

[muresan_c_dan]

This is my first attempt at designing boards, but I would like to give it a shot.
It looks easy on the schematic, but I´m not so sure about the board design, and I would like hear your opinions.

Thank you!

Please disregard the PCB design, because it is full of errors.

[madtecchy]

Good luck with the P3.. The layout is very very important and not that easy . ROD did an awsome job . do you think you can do better????

[muresan_c_dan]

Originally Posted by madtecchy

Good luck with the P3.. The layout is very very important and not that easy . ROD did an awsome job . do you think you can do better????

Not better in the audio quality, but at least the same.
Obviously not better in board size, this one is huge, but it'll do.
And buying the boards is not out of the question, it's just that they are no longer made (replaced by P3A), and if they were, the Toshiba package is different. (I think)

[SpeakerScott]

Without studying it for a long time:

1. Power traces are not wide enough.
2. Star ground routing combines power and signal ground on the board.
3. That speaker output line winds for quite some distance and under sensitive stuff.
4. Not enough supply capacitance on the board.

Study Doug Self's section on grounding, or Bob Cordell. Ether book is good.

[madtecchy]

Any idea as to why Rod released the P3A ..as an update to the p3 maybe????

[muresan_c_dan]

Originally Posted by SpeakerScott

Without studying it for a long time:

1. Power traces are not wide enough.
2. Star ground routing combines power and signal ground on the board.
3. That speaker output line winds for quite some distance and under sensitive stuff.
4. Not enough supply capacitance on the board.

Study Doug Self's section on grounding, or Bob Cordell. Ether book is good.

Thank you Scott,

1. 5mm traces would suffice?
2. Have to make separate input ground connection.
3. Very hard to trace between r14-r15-r16. I would like to have that Spk+ closer to the middle of the board.
I will try to route differently.
4. Supply caps are off the board, 60,000uF per side, single 500VA torroid, 30V secondary.

[jaycee]

Before you lay out the board, a few comments on the schematic:

* Use a current source for the LTP instead of R3
* Don't use the diodes for biasing
* Don't use the trimpot in the VBE multiplier in this way - configure the trimpot as a variable resistor rather than connecting the wiper to the base of the transistor
* Use better driver transistors than the BD139/140 - MJE243/253, MJE15032/3 or 2SA1837/2SC4793 will be far superior

* Put some decoupling capacitors on the board. At least 100uF + 100nF on the supply rails. Without these, the amp will be unstable.

Also with just 35V supplies, you only need one pair of output transistors. Although later you mention using a 2x30V transformer which would give you 42V supplies...

Now for the board:

The output trace is far too long. Keep this short as possible. Put the SPK+ terminal in the middle of the board ideally.

Use at least 2.54mm traces for power and output.

The low value resistors will need to be rated at least 3W - make sure the part size you have chosen on the board has the right dimensions, they look too small to me.

Don't bother with a SPK- on the amplifier PCB. Return the speaker ground straight to the star ground at your supply capacitors.

60,000uF really is overkill. Use them if you've got them, but 2x6,800uF per supply rail will be more than enough.

[muresan_c_dan]

Originally Posted by jaycee

Before you lay out the board, a few comments on the schematic:

* Use a current source for the LTP instead of R3
* Don't use the diodes for biasing
* Don't use the trimpot in the VBE multiplier in this way - configure the trimpot as a variable resistor rather than connecting the wiper to the base of the transistor
* Use better driver transistors than the BD139/140 - MJE243/253, MJE15032/3 or 2SA1837/2SC4793 will be far superior

* Put some decoupling capacitors on the board. At least 100uF + 100nF on the supply rails. Without these, the amp will be unstable.

Also with just 35V supplies, you only need one pair of output transistors. Although later you mention using a 2x30V transformer which would give you 42V supplies...

Now for the board:

The output trace is far too long. Keep this short as possible. Put the SPK+ terminal in the middle of the board ideally.

Use at least 2.54mm traces for power and output.

The low value resistors will need to be rated at least 3W - make sure the part size you have chosen on the board has the right dimensions, they look too small to me.

Don't bother with a SPK- on the amplifier PCB. Return the speaker ground straight to the star ground at your supply capacitors.

60,000uF really is overkill. Use them if you've got them, but 2x6,800uF per supply rail will be more than enough.

I'm a total electronic noob, so I don't quite understand everything.
The first 3 points you state, are described by Rod as easy and ok. The trimpot/BC550 is a bias servo... Maybe you could elaborate?

The transistors you mention do sound good though...

Decoupling caps on the board, well I think I understand this one, and I will include them.
The supply caps are at hand, so are some of the resistors, including the 0,5W rated 0,1 and 0,5 Ohm. Only resistor rated higher is the 560, at 5W.

Also, I am about to make the traces thicker, ground and power 3,81 min, rest 2,54.

Thank you for the help!