Hi X
You are right, it’s economical to use a Mx, it does an amazing job for filtering since it’s SIMULATING a huge capacitor bank.
How does the Mx behave when there is a huge sudden demand of power from the supply ? I know a REAL capacitor banks will do a great job.
Have you done some test and heard any differences ?
Thanks for sharing your experience
BR
Eric
Eric, take a look back at post #555. It'll jog your memory, that member xrk971 is using a pair of "Capacitance Multiplier" boards in the power supply path of his FirstWatt M2 amplifier. M2 operates 100% in Class-A which makes it an especially easy load.
Nelson Pass set the M2's output stage DC bias point at 1.3 amps (per channel), using his optocoupler feedback scheme from long ago. Patent granted June 1988 (!).
Nelson Pass set the M2's output stage DC bias point at 1.3 amps (per channel), using his optocoupler feedback scheme from long ago. Patent granted June 1988 (!).
Yes, I have compared it to the 8x 33mF CRCRC and for Class A, it is better. Class A has a steady draw of the same current no matter what. So for Class AB, you want the big cap bank to provide the reserves of current. But on M2, the Mark Johnson cap Mx sounds better to me than the 8 big 33mF caps. You get -50dB ripple filtering so it is very quiet. The stereo separation and imaging is also superior since I am using two cap Mx with one trafo. The cap Mx’s essentially isolate the rails from each other so there is no channel crosstalk.
It's pretty instructive to calculate the output impedance (vs frequency) of the two alternatives. ESR and ESL of the Big Old Bohunker Capacitor Bank prevent it from delivering all of the goodness that an ideal perfect "university chalkboard capacitor" would give you, in theory. While the "Capacitance Multiplier" benefits tremendously from high bias Class-A downstream circuits, and from modern power transistors with high fT. Crank the algebra, or run the sim. It'll be fun.
Hi prasi,
this morning i received an email from Sophia, with the answer all is ok with your files. The bug must be on their online viewer, and internally they check the designs with gerbv 2.6.2.
Usually the offered online viewer from JLCPCB should work, because when sending files for production, i don't like to start discussion with them.
i like the straight way
Create files- check them, send them , produce PCB and receive PCB.
BR
Günni
rectifier is anyway off-board. You can connect the V+ and V- from the rectifier to BR1+ and BR2- respectively. Center tap of the trafo to CT.
please refer stuffing guide.
regards
Prasi
I have been fighting to get my boards through the JLCPCB review process. The first 2 attempts were rejected, because of the drill file. I first uploaded the gerber from post #512. The second time I changed the Drill file to .DRD and it still said that it needed a drill file. I uploaded X's gerber, but it said it needed the board size, so I entered 80mmx90mm. It took it, but will need to wait to see if it gets approved.
I really want to try these boards in my next build, a M2 and Alpha 20.
I really want to try these boards in my next build, a M2 and Alpha 20.
Aha, #YouToo?
here are the gerbers that have been zipped in a single folder. they look perfect in gerberlogix viewer as did earlier gerbers. Only a slight change with the way outline layer is generated.
If JLCPCB can't handle these gerbers, I would suggest to use some other boardhouse like, pcbway, elecrow, etc.
Last edited:
Seeedstudio is another possibility. They don't have a problem with the Gerbers.
Thanks for the new zip-file, Prasi.