Just for my curiosity, I wanted to see if a Choke input supply could be designed.
I came up with a design. Not a good design, mind you.
This is out of PSUDII
28V 600VA Antec transformer, Diode bridge, Triad C-59u inductor and 360,000 uF caps.
The Good:
Rather than a 12A, 20% duty cycle current spikes, current is a sine wave from 100 mA to 4.75A.
Ripple is around 40mV and is a sine wave vs 200mV sawtooth
The regulation is marginally better at .1R Vs .115R
The Bad.
The inductor weighs 6.25 Lbs (need 2 for F5)
36 12,000 uF caps still has the Q of the power supply at about 1.5 (Multiply caps by 2 for =&-)
24 amp 1 sec turn on surge at the secondary. (Multiply by 2)
The inductors are $50 Us each.
The inductors are only rated for 12.5 amps AC+DC which works out to less than 8 Amps. After that, Ripple goes up to 220mV and has a saw shape. (assuming inductance goes to zero)
Any suggestions?
Comments?
From where I sit, A Zen V5 power supply looks like a much better option. I see why inductors are rare in power supplies.
Doug
I came up with a design. Not a good design, mind you.
This is out of PSUDII
28V 600VA Antec transformer, Diode bridge, Triad C-59u inductor and 360,000 uF caps.
The Good:
Rather than a 12A, 20% duty cycle current spikes, current is a sine wave from 100 mA to 4.75A.
Ripple is around 40mV and is a sine wave vs 200mV sawtooth
The regulation is marginally better at .1R Vs .115R
The Bad.
The inductor weighs 6.25 Lbs (need 2 for F5)
36 12,000 uF caps still has the Q of the power supply at about 1.5 (Multiply caps by 2 for =&-)
24 amp 1 sec turn on surge at the secondary. (Multiply by 2)
The inductors are $50 Us each.
The inductors are only rated for 12.5 amps AC+DC which works out to less than 8 Amps. After that, Ripple goes up to 220mV and has a saw shape. (assuming inductance goes to zero)
Any suggestions?
Comments?
From where I sit, A Zen V5 power supply looks like a much better option. I see why inductors are rare in power supplies.
Doug
Patrick,
I posted the details about the mini El-Pipo here:
http://www.diyaudio.com/forums/showthread.php?s=&postid=1643112#post1643112
Let me know if you have any questions or comments.
Cheers,
Owen
I posted the details about the mini El-Pipo here:
http://www.diyaudio.com/forums/showthread.php?s=&postid=1643112#post1643112
Let me know if you have any questions or comments.
Cheers,
Owen
DougL,
Just as a suggestion, I took all my measurements using a new generation Xantrex switch mode lab supply. It performs admirably better than the simple unregulated CRC filtered supply I use at home
I think if you were really looking for a proper solution, a well designed switch mode supply would be an excellent way to go. Since this is class A, you could get away with running it open loop, and you could probably land a finished efficiency in the mid 90's since your load is so constant.
It would be a good amount of work to design, but probably much more practical than 12 pounds of inductors and enough caps to start your car.
Cheers,
Owen
P.S. - Best of luck Vayu! It's a great amp to start with.
Just as a suggestion, I took all my measurements using a new generation Xantrex switch mode lab supply. It performs admirably better than the simple unregulated CRC filtered supply I use at home
I think if you were really looking for a proper solution, a well designed switch mode supply would be an excellent way to go. Since this is class A, you could get away with running it open loop, and you could probably land a finished efficiency in the mid 90's since your load is so constant.
It would be a good amount of work to design, but probably much more practical than 12 pounds of inductors and enough caps to start your car.
Cheers,
Owen
P.S. - Best of luck Vayu! It's a great amp to start with.
Those of you who have built the stock F5 would want to know that I have
adjusted the value of the 10K resistors in the current limiter circuit upward to
22K or so. The distortion measures slightly higher at high swing as a result,
but the harmonic content is lower at high swing and I think it sounds marginally
better.
adjusted the value of the 10K resistors in the current limiter circuit upward to
22K or so. The distortion measures slightly higher at high swing as a result,
but the harmonic content is lower at high swing and I think it sounds marginally
better.
LOL
I like that description.
I like the idea of not placing high current spikes on the power line.
While 12 Lb and $100 US is expensive, its probably worth it.
However, having a supply that rings because the required capacitors will not fit in a reasonable size box even if I chose to buy them is probably a red flag.
Doug
Nelson Pass said:Those of you who have built the stock F5 would want to know that I have
adjusted the value of the 10K resistors in the current limiter circuit upward to
22K or so.
What do I expect in terms of dc points when I 'll change this resistors ?
Will I need to set P1 and P2 again carefully from minimum value ?
Calvinhpk,
You answered your own question:
0.34C/W * 64W = 21.7C above ambient.
More than sufficient for one channel.
The F5 doesn't need anywhere near as much heatsinking as people seem imply. If you're ever in doubt, just mount one channel on the heatsink in question and try it out. If it wanders over 50-55C, then it's not big enough. I think you'd be surprised what you can get away with if you mount the heatsink properly.
If you're on the border, at say 55C, then here are a few tips which might help:
1. Mount heatsinks with fins running vertically
2. Leave a 1/2" or more between the heatsink and the floor to allow air flow
3. Don't obstruct the top of the heatsink.
4. Attach other large aluminum parts to it (eg 1/2" aluminum top/bottom)
Hopefully that helps.
Cheers,
Owen
You answered your own question:
0.34C/W * 64W = 21.7C above ambient.
More than sufficient for one channel.
The F5 doesn't need anywhere near as much heatsinking as people seem imply. If you're ever in doubt, just mount one channel on the heatsink in question and try it out. If it wanders over 50-55C, then it's not big enough. I think you'd be surprised what you can get away with if you mount the heatsink properly.
If you're on the border, at say 55C, then here are a few tips which might help:
1. Mount heatsinks with fins running vertically
2. Leave a 1/2" or more between the heatsink and the floor to allow air flow
3. Don't obstruct the top of the heatsink.
4. Attach other large aluminum parts to it (eg 1/2" aluminum top/bottom)
Hopefully that helps.
Cheers,
Owen
calvinhpk said:
Papa's article about F5 says:
"At 1.3 amps per channel, you will see idle heat dissipation of 62 watts. To keep the temperature rise of the heat sink to 20 deg C. above the ambient temperature, you will want a heat sink rated at about .6 deg C./watt for each transistor. An example of this would be a chunk of fi nned aluminum, with a series of 2¡± fi ns attached to an 8¡± by 6¡± base. You will need two per channel."
Therefore, we need a heatsink of about 50mm(2") fins on 150mm(6") x 400mm(16") base to secure the 20 deg C above the ambient temperature for one channel (two output MOSFETs). Consider yours comparing with this. . .
Cheers,
