I would not pull the trigger at this moment.
Tobias stated at the beginning of the thread that he is aiming to achieve very low THD+N. Thus the solution which "just works" is not a sufficient condition to stop optimizing the design. It should work exceptionally good to be a satisfactory solution and this requires considering as many optimizations as feasible withing the design constraints. And if most of the indication circuitry can be left as is, the signal path and power supply lines should be optimized as much as possible.
Just my opinion... not necessarily right.
Oleg
Tobias stated at the beginning of the thread that he is aiming to achieve very low THD+N. Thus the solution which "just works" is not a sufficient condition to stop optimizing the design. It should work exceptionally good to be a satisfactory solution and this requires considering as many optimizations as feasible withing the design constraints. And if most of the indication circuitry can be left as is, the signal path and power supply lines should be optimized as much as possible.
Just my opinion... not necessarily right.
Oleg
I consider (positive) criticism as a base of inspiration and discussion. So it's not a question about who is more skilled but more about an exchange of experiences and ideas.
Anyway I will try to understand the change suggestions before implementing them, so I will face up with them anyway.
I just thought it trough again and didn't find the point. To find it, I will just explain, what I thought about this part of the layout and you can intermit where my thoughts are wrong or flawed:
At first, I want to consider both GNDs as seperated signals as they come from different areas of the base board so they could be on different levels through induced GND currents in the base board.
Each signal should only be connected to the appropriate GND (as far as the signal chain is concerned - not the LED control).
To keep it simple, just consider the two small GND planes as simple wires from pad 4 of the connectors to pad x-1 of the pot.
So the current loops are as follows:
For JP1 (left channel f.e.), the current comes from JP1 pad 1, through the DIP switch (switch 5-8), through the appropriate resistors, into pot pad 1-3, partly into pot pad 1-2 (signal 1 out) and pad 1-1 (signal 1 GND) and back to JP1 pad 3 (signal 1 out) and pad 4 (signal 1 GND).
For JP2 (right channel f.e.), the current comes from JP2 pad 1, through the DIP switch (switch 1-4), through the appropriate resistors, into pot pad 2-3, partly into pot pad 2-2 (signal 2 out) and pad 2-1 (signal 2 GND) and back to JP2 pad 3 (signal 2 out) and pad 4 (signal 2 GND).
So my thoughts are: It should work perfectly if the GND planes just were tracks connecting "signal 1/2 GND" from the pot to the connector. Shouldn't it work even better, if the tracks are whole planes, covering most part of the other part of the current path to protect from induction?
Maybe I could pull the GND plane to the left a bit more to the right in the top most area to cover the area under the 4 left most resistors, but then the current path still would go through the DIP switch and hover over a small area covered by the wrong GND plane (the middle part under the DIP switch). I can't avoid that completely.
You're right, turning the LEDs the other way round is dangerous. I will also try to optimise power rail routing, as I am completely reworking the right-most part of the board atm.
Yes, it's a tight layout now
But Eagle has very generous keep-out areas for small resistors and caps. It's getting a bit challenging if completely cutting the space down to the keep-out, but if obeying the keep-out I always got it soldered by now.But if it should fail, I also have the option for reflow soldering. Anyway I will have to test, which solder method is faster for this layout. If there are many identical parts I usually prefer reflow soldering, if not, hand soldering usually works well.
I will have to do it at some point for sure
I never thought the discussion about this project will get into these dimensions and I am very stunned about the participation in this project of everyone posting here.Atm I see a chance to get something done I would never have got done for myself. So I currently use the chance of all your participation to get the most out of the project.
At some point, it obligatory drives into hair-splitting and this is where I don't want to get into.
But I don't see this point far off anymore
But the power rails f.e. are still somewhat significant for the performance.
Very good hint! Just browsed through some parts of my BOM and they have very good pricing for their products.
There is an easy solution to the routing of the pregain section. Just swap the sides of the pot between the channels. This will allow you to have a dedicated GND plane for each channel without the need to compromise the layout. This way you could combine the leftmost part of S1 together with the rightmost pins of the pot and route the corresponding GND plane along the lower side of the pregain section. And similar for the rightmost part of the S1 and the leftmost pins of the pot. Just route the GND plane for it from the left around the upper rim of the board. I could have done it myself for illustration but am to lazy to draw the schematic
Also to gain more space you may consider using through hole version of the DIP switch S1. It has smaller footprint and will be stronger attached to the board mechanically.
I hope my explanation is not to cryptic.
Oleg
Also to gain more space you may consider using through hole version of the DIP switch S1. It has smaller footprint and will be stronger attached to the board mechanically.
I hope my explanation is not to cryptic.
Oleg
I just checked it twice. Should be alright, shouldn't it?
(Regarding mirroring: I always leave this over to Eagle, which should do it right.)
If you want to check, here's the datasheet:
http://www.bourns.com/data/global/pdfs/91_95.pdf
Hint: The position of the A.R. lug helps to identify the orientations of the drawings.
<edit>
I could have done it myself for illustration but am to lazy to draw the schematic
I sent you a PN with the project files. I would be glad if you could show your suggestion.
Last edited:
After an inspiring communication with OlegSh I can present rev.3 of the Interface PCB:
Some comments or advice?
Maybe just something I unintentionally worsened by implementing the latest changes?
<edit>
I just rotated C25 to be aligned with the others.
An externally hosted image should be here but it was not working when we last tested it.
Some comments or advice?
Maybe just something I unintentionally worsened by implementing the latest changes?
<edit>
I just rotated C25 to be aligned with the others.
Last edited:
- Project status update: heading for production release -
For me I'm contented with the current project state. Thus I want to give it into production as soon as possible.
Of course I'm still eager of your advice concerning flaws in SQ or general improvement suggestions.
So please warn me, if I have missed some relevant downsides or problems.
If there's nothing left, I will release the order for the PCBs at the end of this week (friday or saturday).
For the sake of clarity I will print both the schematics and layout of both boards at the current state again here:
At this point I meanwhile want to say a big thank you for all your great ideas, advice, know how and all your experience you shared with me!!
I will keep you informed about any progress of this project.
For me I'm contented with the current project state. Thus I want to give it into production as soon as possible.
Of course I'm still eager of your advice concerning flaws in SQ or general improvement suggestions.
So please warn me, if I have missed some relevant downsides or problems.
If there's nothing left, I will release the order for the PCBs at the end of this week (friday or saturday).
For the sake of clarity I will print both the schematics and layout of both boards at the current state again here:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
At this point I meanwhile want to say a big thank you for all your great ideas, advice, know how and all your experience you shared with me!!
I will keep you informed about any progress of this project.
Last edited:
Hello,
First of all I'm sorry for this long silence.
After I have started working for a new job at the beginning of 2016, I had to suspend the project due to a big lack of time. However, in christmas holidays 2017 I revived the project and finished it
Thanks again for all help! The sound is awesome and all the effort was totally worth it.
If you are interessted in some pictures, just PM me.
Best Regards,
Tobias
First of all I'm sorry for this long silence.
After I have started working for a new job at the beginning of 2016, I had to suspend the project due to a big lack of time. However, in christmas holidays 2017 I revived the project and finished it
Thanks again for all help! The sound is awesome and all the effort was totally worth it.
If you are interessted in some pictures, just PM me.
Best Regards,
Tobias
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.