Finally i integrated power with pre.
I had funny story with the hum in the speakers. After integration the hum was noticeable hard. I read many threads about star ground and related, but nothing really works for me. I was really upset.... After some time 'eureka'! I forgot I had undervoltage from the low voltage PSU (if you remember from previous posts I had 11V after 12V regulator). So I downgrade a LCD backlight even more (to 3.8V) to do not overload transformer so much. This gives me 15V before 12 V regulator. The hum is gone !
I am really enjoy, the sound is fantastic! I feel is even better with 2x15000 per rail PSU, but due to the size I couldn't fit this in my case. I am using 2x10000 per rail. I need to fit everything in the case, so more mechanical works need to be done yet....
I had funny story with the hum in the speakers. After integration the hum was noticeable hard. I read many threads about star ground and related, but nothing really works for me. I was really upset.... After some time 'eureka'! I forgot I had undervoltage from the low voltage PSU (if you remember from previous posts I had 11V after 12V regulator). So I downgrade a LCD backlight even more (to 3.8V) to do not overload transformer so much. This gives me 15V before 12 V regulator. The hum is gone !
I am really enjoy, the sound is fantastic! I feel is even better with 2x15000 per rail PSU, but due to the size I couldn't fit this in my case. I am using 2x10000 per rail. I need to fit everything in the case, so more mechanical works need to be done yet....
Great!
Yeah, regulators have "dropout voltage" specs. The input voltage, i.e. the BOTTOM/DOWNWARD PEAKS of the input's ripple voltage (just before the regulator), MUST STAY HIGHER than the regulator's output voltage, by AT LEAST as much as the dropout voltage spec, always. OTHERWISE, the output volage gets really ugly, probably with large ripple, because the regulator stops regulating.
Also, the Dropout Voltage changes, depending on how much current your load draws. So you usually need to look at the plot of that, in the datasheet, and design for the worst case. You should also design for the worst-case (lowest) AC Mains voltage, and also take into account the transformer regulation (i.e. transformer's loaded vs unloaded output voltage change).
About the caps' size and fit: Possibly, you could use multiple smaller caps, in parallel, instead of one large cap, which might enable them to fit better. Paralleled capacitances result in a capacitance that is their sum. In that case, you might also be free to move some of them closer to the point of load, if that was a possibility in your case, and if it might help them fit.
Yeah, regulators have "dropout voltage" specs. The input voltage, i.e. the BOTTOM/DOWNWARD PEAKS of the input's ripple voltage (just before the regulator), MUST STAY HIGHER than the regulator's output voltage, by AT LEAST as much as the dropout voltage spec, always. OTHERWISE, the output volage gets really ugly, probably with large ripple, because the regulator stops regulating.
Also, the Dropout Voltage changes, depending on how much current your load draws. So you usually need to look at the plot of that, in the datasheet, and design for the worst case. You should also design for the worst-case (lowest) AC Mains voltage, and also take into account the transformer regulation (i.e. transformer's loaded vs unloaded output voltage change).
About the caps' size and fit: Possibly, you could use multiple smaller caps, in parallel, instead of one large cap, which might enable them to fit better. Paralleled capacitances result in a capacitance that is their sum. In that case, you might also be free to move some of them closer to the point of load, if that was a possibility in your case, and if it might help them fit.
Last edited:
Still it was a good experience how the regulators make a great work for us. And due to my 'strange setup' (where I can smoothly modify the current from transformer and impact a dropout voltage in loaded condition) I would never feel that 'on my skin'. This was extremely educational.
The smaller caps... I never though about this. I was 'charmed' by lots of chip-amps designs with the big one two per rail. I spend some time to search correct PCBs for special 35x40mm or 30x40mm sizes. Most common high quality caps are 30x50mm or 25x50mm (the height was a problem). At least on my local market.
Lets do some math:
1 x 30x40mm 10000uF cap -> takes 30mm x 30mm space on PCB -> 900mm^2
2 x 18x40mm 4700uF cap -> takes 18mm x 36mm space on PCB -> 648mm^2
I am not sure what about capacitor ripple current, but from the space perspective this looking good.
The smaller caps... I never though about this. I was 'charmed' by lots of chip-amps designs with the big one two per rail. I spend some time to search correct PCBs for special 35x40mm or 30x40mm sizes. Most common high quality caps are 30x50mm or 25x50mm (the height was a problem). At least on my local market.
Lets do some math:
1 x 30x40mm 10000uF cap -> takes 30mm x 30mm space on PCB -> 900mm^2
2 x 18x40mm 4700uF cap -> takes 18mm x 36mm space on PCB -> 648mm^2
I am not sure what about capacitor ripple current, but from the space perspective this looking good.
Usually (always?) the ripple current rating requirements will go in the right direction, too, since it gets divided among the multiple parallel caps. I don't see a downside. The more smaller caps used, the lower the total ESR and ESL.
I guess that eventually physical space becomes a problem again so there is probably some optimum size and number for each application.
I guess that eventually physical space becomes a problem again so there is probably some optimum size and number for each application.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.