Damir, thanks for the excellent work and report.
Whenever OIC is used it is important to see the loop gain around the OPS as well as the outside loop.
That would be with the Tian probe where the value "10" for R41 sits on the feedback loop.
Could you show this or post the ASC so I can look?
Best wishes
David
Last edited:
This is a starting point of a board for this project. I am seeking "positive" input on it, concentrating at this juncture on the basics. As mentioned in the other thread, a star ground and symmetry are two of the major criteria of layout correctness. So please indicate pros and cons of this layout and or suggest alternative layouts.
I am aware that this layout does NOT match the current schematic. The current focus is on layout topology.
Thanks to Dadod for all of his brilliant work and many hours of effort to get this project to this point.
I hope the board is worthy of the CFA design it will carry. I will strive toward that goal.
Please comment on putting a Cap Mul on board or not, I have 1000uf 100V caps as an alternative.
As indicated on the board, it is only for development and testing. Thanks
I am aware that this layout does NOT match the current schematic. The current focus is on layout topology.
Thanks to Dadod for all of his brilliant work and many hours of effort to get this project to this point.
I hope the board is worthy of the CFA design it will carry. I will strive toward that goal.
Please comment on putting a Cap Mul on board or not, I have 1000uf 100V caps as an alternative.
As indicated on the board, it is only for development and testing. Thanks
Attachments
This is intended to be a one sided board and the "second" side will be done with wire links?
Best wishes
David
In a word "yes".
BUT, if it is agreed upon then it can be double sided plated thru. With a ground plane etc.
It is up to Dadod and the input such as yours. I am just trying to help.
With the comp caps of such low values, a ground shield/plane might have significant affect on operation (via stray shunt c from trace to grnd plane).
I would be OK with RC network with 1000mfd/100V. Just to keep it simple.
Thx all for showing the CMA can be no more difficult and maybe less difficult than a VFA.
-RNMarsh
I would be OK with RC network with 1000mfd/100V. Just to keep it simple.
Thx all for showing the CMA can be no more difficult and maybe less difficult than a VFA.
-RNMarsh
Attachments
It's your board and you've made the effort.
I think this decision should be made reasonably early.
If it is double sided then there is more possibility to minimize current loops and their associated problems.
Some of the possibilities can probably be mimicked on a one sided board with careful placement of wire links and/or bus bars.
I have little experience with multi layer boards and planes so can't comment informedly about that aspect.
Best wishes
David
You are very correct.
That is why I am putting it up now, so final solution had with out tearing up the whole board. I am concentrating on the OPS rails and output trace and their relationship to the Ground. I am sure the eagle eyed members here will point out any and all short comings I may make in the layout quickly and a live test can be made to savor the results of the endeavor.
I agree that a double sided board is likely to be the better of the choice, I will turn it on and post a layout.
That is why I am putting it up now, so final solution had with out tearing up the whole board. I am concentrating on the OPS rails and output trace and their relationship to the Ground. I am sure the eagle eyed members here will point out any and all short comings I may make in the layout quickly and a live test can be made to savor the results of the endeavor.
I agree that a double sided board is likely to be the better of the choice, I will turn it on and post a layout.
Yes, that's exactly what I meant, thanks
Hm. 10 MHz.
As you practically certainly noticed, I have discussed this with Ric Lee and wonder what the practical limit is.
I know you have built some amps with TPC/TMC/OIC and that would provide a check with reality.
Have you checked this with any of your completed TPC/TMC/OIC amps?
Best wishes
David
I have a preference for a wider layout with the OPS all on one wide side.
It fits my heat-sink layout better and I have worked out how to minimize the current loops of all the heavy current circuitry and the rest falls naturally into place.
It's probably do-able with your layout too but I've never really worked it out.
Do you have a definite preference for deep but skinny layout?
Best wishes
David
You are too quick for me. Had it posted before I finished my reply.
You had some practice with PCB layout?
What do you use?
Last edited:
Double side is a way to go with no ground plane.
Be careful where to connect the feed back point, I think that the suggested layout is doing that wrongly, look the picture and what D. Self says.
I think also that is better to have output transistors protrude out of the PCB (not to be kept below the board) for easier mounting.
Predriver transistor shoud not be on the same heat sink but on own small heat sinks.
BR Damir
Attachments
This is from my VFA TMC TT amp singing each day with no stability problem. It is not 10 MHz but 4 MHz, I think with CFA you can go higher.
BR Damir
Attachments
I actually mean driver transistors, as here no predrivers exist, Q5, Q7, Q9 and Q10 should be mounted on own small heath sinks not on the main heath sink with the output transistors.
Q6 and Q8 are correctly with own heath sink.
It is needed to change Q17 and Q19 to higher voltage transistors(2sa970, 2sc2240 or similar) as BC550/560 will not withstand 63 V.
Last edited:
Thank you once more.
When the simulated PM is healthy then 4 MHz I can definitely believe.
Since the CFA speed improvements are in the front end I don't expect they will raise the limits for the OPS very much.
But MOSFETs should help. Fun if someone builds this and we find out
Best wishes
David
Thank you once more.
When the simulated PM is healthy then 4 MHz I can definitely believe.
Since the CFA speed improvements are in the front end I don't expect they will raise the limits for the OPS very much.
But MOSFETs should help. Fun if someone builds this and we find out
Best wishes
David
Yes, you are right, it is not CFA but OPS with lateral MOSFETs. I did not try vertical MOSFETS yet.
BR Damir
FYI -- double-sided can be useful when you need to control trace impedances and the like. For a DIY project, single-sided with the traces on TOP side has advantages in doing mods and trouble shooting. You can cut trace, add parts and do upgrades and trouble-shoot all from the top side without tearing the pcb/mosfets out each time you want/need to do any of these things and need to get at the underside to do it.
Works for me.
Thx-RNMarsh
Works for me.
Thx-RNMarsh
