So the reason I stared this post was to figgure out if I could use the caps I have on hand to build a PS for my F5. I used a gigaworks PS and had problems so now I am back to designing my own point to point PS for my f5. These are the caps I have.
2 X Nippon Chemicon 82da 22000uf 63VDC, Max MAX ESR @ 120hz = 14.3mOHM, Max Ripple current @ 120hz = 9.98A
2 x BHC ALS10A 1037CC 40/085/56, 68000uf 40VDC. ESR and Ripple current unknown due to being obsolete. Closes specs I could find were for ALS30 which had an ESR @ 100hz of 4mOHM, impedance of 4mOHM and a ripple current @ 100hz of 46.5A
From what I see I think I am ok with both of these caps as far as ripple current. Am I missing something? Which cap should be closer to the rectifiers and which should be closest to the amp board? I am guessing that the BHC should be closest to the amp board due to it's lower ESR (if the ALS30 ~ ALS10A on this spec).
Considering bypass caps, would it be advantageous for me to put bypass caps right on my F5 board, rather then attaching them to my PS caps? I think I am going to get the amp working perfectly then consider this because it is so easy to do after the fact.
2 X Nippon Chemicon 82da 22000uf 63VDC, Max MAX ESR @ 120hz = 14.3mOHM, Max Ripple current @ 120hz = 9.98A
2 x BHC ALS10A 1037CC 40/085/56, 68000uf 40VDC. ESR and Ripple current unknown due to being obsolete. Closes specs I could find were for ALS30 which had an ESR @ 100hz of 4mOHM, impedance of 4mOHM and a ripple current @ 100hz of 46.5A
From what I see I think I am ok with both of these caps as far as ripple current. Am I missing something? Which cap should be closer to the rectifiers and which should be closest to the amp board? I am guessing that the BHC should be closest to the amp board due to it's lower ESR (if the ALS30 ~ ALS10A on this spec).
Considering bypass caps, would it be advantageous for me to put bypass caps right on my F5 board, rather then attaching them to my PS caps? I think I am going to get the amp working perfectly then consider this because it is so easy to do after the fact.
You don't need this massive amount of capacitance for a 25W /channel class A amp. Use the 2 x 22K uF, put parallel bypass caps right after the rectifier stage, at the output wire connector, at the amp board wire connector input, and as close to the BJT and mosfet transistors as possible... yes, use several bypass capacitors, it's not a matter of having one as a token gesture, it's a matter of bypassing at every active component and noise entry point. That should be more of an improvement than an additional 68K uF in bulk PSU capacitance for a 25W/channel amp.
Thanks for all the help guys. I just re-read your post Gootee and it is actually starting to sink in. I am not ready for analog, digital design just yet but I now am starting to get it a bit.
! I ended up using those caps because that was what I had on hand.
I originally used the 68000uf as the first cap. I recently tried measuring some things on the amp and changed it to the 22000uf as the first cap after the rectifier. I am not ready to play with bypass caps just yet but I appreciate all the help so far. I have one last question that I am unsure of. Does it matter if I put the bleeder resistors accross the second bank of caps or the first? I would just put it on the first but it is much easier to put it on the second. Any reason to change this?
! I ended up using those caps because that was what I had on hand.
I originally used the 68000uf as the first cap. I recently tried measuring some things on the amp and changed it to the 22000uf as the first cap after the rectifier. I am not ready to play with bypass caps just yet but I appreciate all the help so far. I have one last question that I am unsure of. Does it matter if I put the bleeder resistors accross the second bank of caps or the first? I would just put it on the first but it is much easier to put it on the second. Any reason to change this?
I'm not sure if the bleeder's location matters. So in that case I would put it after the second set of caps.
But you really should see if you can fit at least modest-sized decoupling caps very close to EACH active output device's power and ground pins, and some small bypass caps right at the pins. The minimum size I would try for decoupling caps would be 220uF but 1000uF or more would probably be better. But that's just a guess.
To see how to actually figure out what values the decoupling caps need to be, go to the link below!
http://www.diyaudio.com/forums/power-supplies/208579-30vdc-10a-psu.html#post2942537
Also follow the link from there to the posts listed.
The main thing, for enabling the most-accurate transient response that your circuit is capable of, is the PSU impedance AS SEEN BY THE PINS, i.e. right at the point of load. Any caps more than a few centimeters away can not help enough with the worst-case fast-transient current demands, without possibly causing (too) large disturbances to the power rail voltage, and even then the current might not get there in time. This is due to the inductance of the conductors (traces or wires) between the caps and the actual point of load, where the current is demanded.
So anyway, it turns out that using several smaller decoupling caps in parallel instead of one big one will lower their apparent total inductance, and thus the PSU impedance seen by the device, which is a good thing. And it could also make it easier to get them all closer to the pins, compared to using one larger one.
For high-frequency stability, you probably DO also NEED some small bypass caps, right across the power and ground, within a millimeter or two of each point of load. If you don't know which ones to use, it's still important to use SOMETHING, there, anything from 0.01 uF ceramic to 1 uF film, or whatever. Physically-smaller caps will usually be better, for high frequencies, becuase they usually have lower inductance (possibly only because the overall conductor path length is shorter simply because they are smaller).
P.S. If anyone wants a really-good but practical way to come up with OPTIMAL snubber component values, check out post 292 at this link:
http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-30.html#post2828689
Cheers,
Tom
But you really should see if you can fit at least modest-sized decoupling caps very close to EACH active output device's power and ground pins, and some small bypass caps right at the pins. The minimum size I would try for decoupling caps would be 220uF but 1000uF or more would probably be better. But that's just a guess.
To see how to actually figure out what values the decoupling caps need to be, go to the link below!
http://www.diyaudio.com/forums/power-supplies/208579-30vdc-10a-psu.html#post2942537
Also follow the link from there to the posts listed.
The main thing, for enabling the most-accurate transient response that your circuit is capable of, is the PSU impedance AS SEEN BY THE PINS, i.e. right at the point of load. Any caps more than a few centimeters away can not help enough with the worst-case fast-transient current demands, without possibly causing (too) large disturbances to the power rail voltage, and even then the current might not get there in time. This is due to the inductance of the conductors (traces or wires) between the caps and the actual point of load, where the current is demanded.
So anyway, it turns out that using several smaller decoupling caps in parallel instead of one big one will lower their apparent total inductance, and thus the PSU impedance seen by the device, which is a good thing. And it could also make it easier to get them all closer to the pins, compared to using one larger one.
For high-frequency stability, you probably DO also NEED some small bypass caps, right across the power and ground, within a millimeter or two of each point of load. If you don't know which ones to use, it's still important to use SOMETHING, there, anything from 0.01 uF ceramic to 1 uF film, or whatever. Physically-smaller caps will usually be better, for high frequencies, becuase they usually have lower inductance (possibly only because the overall conductor path length is shorter simply because they are smaller).
P.S. If anyone wants a really-good but practical way to come up with OPTIMAL snubber component values, check out post 292 at this link:
http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-30.html#post2828689
Cheers,
Tom
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.