I'm keeping the same trace thickness and layout as the original(somewhat). Since i asked that question i put every trace that i could on the other side which meant i had to use vias on the power traces but now i have a more or less uninterrupted ground plane.
If the vias have to carry any sizeable currents, I suggest using Z-wires on them. A Z-wire is a short piece of tinned single-strand wire inserted through the via and bent over on both sides of the board, onto the traces the via connects together. The wire is soldered to both traces. The wire ends up having a sort of Z-shape when used this way, hence the name.
Good advice. Make sure the via drill diameter is large enough that the via is not filled up during the finishing process and leave some blank areas in the solder mask, so you don't have to scrape it away afterwards.
Does anyone have gerbers for this power supply? I am looking for bipolar psu with LM317 on both sides.
I was browsing this thread and I cannot find anything updated.
Thanks
I was browsing this thread and I cannot find anything updated.
Thanks
I have skimmed through the many pages here and seen it stated several times that D-Noizator and its siblings are not LDO but I have by far not read all pages.
Is there reason to believe why is should not work for a pin-compatible LDO like the LD1084?
Is there reason to believe why is should not work for a pin-compatible LDO like the LD1084?
LDO's are a very different breed of cats regarding stability requirements, and there is no guarantee they are compatible with the X-noisers. At the very least, adaptations of the compensation would be required.
I think that here and there in the thread, other regulators, including LDO have been tested, in sim at least.
It would have to be tested, in sim first if a model is available, and then on a breadboard. It would be foolish to commit to a PCB without these precautions
I think that here and there in the thread, other regulators, including LDO have been tested, in sim at least.
It would have to be tested, in sim first if a model is available, and then on a breadboard. It would be foolish to commit to a PCB without these precautions
"At the very least, adaptations of the compensation would be required."
Pay attention to Elvee's statement.
The filter capacitor in an LDO is in the feedback loop, and its value and ESR affect stability.
Many of the Analog Devices (and Linear Tech) Low Dropout Regulators have test fixtures in LTSpice. You can test for stability by this method:
https://www.omicron-lab.com/fileadm...ote_Traditional_NoninvasiveStability_V3.0.pdf
Pay attention to Elvee's statement.
The filter capacitor in an LDO is in the feedback loop, and its value and ESR affect stability.
Many of the Analog Devices (and Linear Tech) Low Dropout Regulators have test fixtures in LTSpice. You can test for stability by this method:
https://www.omicron-lab.com/fileadm...ote_Traditional_NoninvasiveStability_V3.0.pdf
Here's a tutorial -- the "LDO Tunnel of Death" on page 19. https://www.tij.co.jp/lit/an/slva072/slva072.pdf another source of info at TI is application note SLVA026.
There was a National Semiconductor application note (before they were acquired by TI) which explained in more concise terms.
There was a National Semiconductor application note (before they were acquired by TI) which explained in more concise terms.
Short question - would this concept apply well on something like LM338 or something better??
(looking for 5A regulator with something like this implemented for my streamer/dac)
(looking for 5A regulator with something like this implemented for my streamer/dac)
Thank you...i don't have many equipment to test it but i am interested to make it and try it...
Any suggestions on values maybe.... for LM338
Any suggestions on values maybe.... for LM338
Hello,
Great question in this long topic. 🙂
Has anyone made one of the simple variants compared to the first scheme from the first post presented here?
Thank you.
Great question in this long topic. 🙂
Has anyone made one of the simple variants compared to the first scheme from the first post presented here?
Thank you.
I don't see exactly what you mean. All the notable, tested variants have been gathered in the index, at the end of the first post
Yes, it is the official, recommended version that has been built by the dozen, except for R4 which is normally 1K8, but 1K3 will also work with slightly reduced performance