Hi all,
just tested the power supply of one of my (near) future Aleph-X´s. It seems to work as planned! Even the torobar chokes seem to work at a bit over 10A of current.
Below is a picture of the supply. It consists of: 600VA 2x18V transformer, 2 bridges made of discrete Schottkys, 30 x 3300uF/35V, 2x 2,2mH Torobars, 44 x 4700uF/25V Pan FC´s.
At 10.2A voltages are 22.2V for the first capacitor bank with 500mV ripple, 20.9V for the second bank with <10mV ripple.
The chokes get quite warm (13 watts of dissipation). The rectifiers get hot so I think I´ll have to do someting about the cooling although final bias will be around 8A (over 30% less dissipation!)
William
just tested the power supply of one of my (near) future Aleph-X´s. It seems to work as planned! Even the torobar chokes seem to work at a bit over 10A of current.
Below is a picture of the supply. It consists of: 600VA 2x18V transformer, 2 bridges made of discrete Schottkys, 30 x 3300uF/35V, 2x 2,2mH Torobars, 44 x 4700uF/25V Pan FC´s.
At 10.2A voltages are 22.2V for the first capacitor bank with 500mV ripple, 20.9V for the second bank with <10mV ripple.
The chokes get quite warm (13 watts of dissipation). The rectifiers get hot so I think I´ll have to do someting about the cooling although final bias will be around 8A (over 30% less dissipation!)
William
Attachments
HI William,
your PSU looks really impressive! Did you take this much caps to lower ESR and ESL?
I`m going to use chokes in my AX PSU too and therefore would be interested in knowing if the Torobar cores get into saturation which doesn`t seem to be the case with this low ripple @ high current but i would like to be shure. You could replace the Torobars with 0,12 Ohm(if the Torobar datasheet is right) Rs and measure the ripple again. If you don`t want to do this with the completed PSU maybe while building the second?
If the cores don`t saturate the dissipation should be higher than 13W
@ 10.2A (12,5W from real resistance + about 4W from inductive resistance - not shure about the complex part) but i assume you`ve calculated this value from 1.3V*10,2A?
illusionxx
your PSU looks really impressive! Did you take this much caps to lower ESR and ESL?
I`m going to use chokes in my AX PSU too and therefore would be interested in knowing if the Torobar cores get into saturation which doesn`t seem to be the case with this low ripple @ high current but i would like to be shure. You could replace the Torobars with 0,12 Ohm(if the Torobar datasheet is right) Rs and measure the ripple again. If you don`t want to do this with the completed PSU maybe while building the second?
If the cores don`t saturate the dissipation should be higher than 13W
@ 10.2A (12,5W from real resistance + about 4W from inductive resistance - not shure about the complex part) but i assume you`ve calculated this value from 1.3V*10,2A?
illusionxx
Hi xx,
no saturation! Look at the pic of my scope showing ripple before and after the choke. (one is 500 the other 50mV/div).
Yes I did calculate the dissipation from V x I. I even calculated the resistance of the torobar from this (giving 0,128 Ohms).
If you try Duncans Power Supply Designer you can change from choke to resistor. There you see that a resistor can never give this much ripple reduction.
William
no saturation! Look at the pic of my scope showing ripple before and after the choke. (one is 500 the other 50mV/div).
Yes I did calculate the dissipation from V x I. I even calculated the resistance of the torobar from this (giving 0,128 Ohms).
If you try Duncans Power Supply Designer you can change from choke to resistor. There you see that a resistor can never give this much ripple reduction.
William
Attachments
Yes the values in PSU-Designer are close to yours. That`s good news, i was already considering about winding my own chokes. Probably i`ll use the Torobars too.
Have you ever measured ESL and ESR of a single cap vs multiple parallel caps of the same design? Some people claim that the induktivity of the wire connecting the caps is too big to reduce total ESL. But most probably it´s only a matter of the number of devices in parallel?
illusionx
Have you ever measured ESL and ESR of a single cap vs multiple parallel caps of the same design? Some people claim that the induktivity of the wire connecting the caps is too big to reduce total ESL. But most probably it´s only a matter of the number of devices in parallel?
illusionx
Excellent work William. I have a question about the PSUD2 program. When you select a CRC layout, the resistance appears only on the positive rail. This is confusing to a neophyte like myself in that it implies that the resistors are only physically connected to the positive rail and not the negative. Is this so or does the negative rail get its own resistors before the second capacitance? If each rail requires its own resistance, why is only one resitance icon indicated in the layout?
wuffwaff
I think I understand now. After the bridge rectification, I will connect the +DC to the + pole of the first cap bank and the -DC to the - pole of the first cap bank. Then I will connect a parallel resistor bank to each remaining cap bank pole. The resistor bank output will then be connected to the second cap bank in the same manner as before (ie:+DC to + cap bank and -DC to - cap bank. Is this correct?
I think I understand now. After the bridge rectification, I will connect the +DC to the + pole of the first cap bank and the -DC to the - pole of the first cap bank. Then I will connect a parallel resistor bank to each remaining cap bank pole. The resistor bank output will then be connected to the second cap bank in the same manner as before (ie:+DC to + cap bank and -DC to - cap bank. Is this correct?
I just finished the second supply and took a picture from the side. Here you can see the transformer then the aluminium piece of extrusion where the Schottkys are mounted on and the veroboard with the softstart. One up is the first bank of caps then the torobars and (again one up) the Panasonics. Mass runs through a 6mm brass rod from the diodes directly to the main board. Plus and minus are also connected to such rods and are also directly connected to the (Hifizen) board.
The amp will consist of two easy seperable parts, the power supply seen here (without the print) with ac-connection and the front (not shown here).
The rest consisting of fets, main board, in- and output connectors, heatsinks and back.
The only electrical connections between the two are the three brass rods connected to the main board.
william
The amp will consist of two easy seperable parts, the power supply seen here (without the print) with ac-connection and the front (not shown here).
The rest consisting of fets, main board, in- and output connectors, heatsinks and back.
The only electrical connections between the two are the three brass rods connected to the main board.
william
Attachments
CLC or CRC, see what I mean?
wuffwaff
Don't apologize, it's most likely my failing. Okay, so if the opposing poles of the cap bank go to ground then both the positive and negative supply will require a separate resistance. Is this correct? I still have questions:
What if you want to add additional supplies, can you just add more wires to the secondaries and connect bridges to them?
illusionx
Does your post show full wave rectification and not a bridge? I thought that a bridge will produce both positive and negative output so can someone please explain the difference.
wuffwaff
Don't apologize, it's most likely my failing. Okay, so if the opposing poles of the cap bank go to ground then both the positive and negative supply will require a separate resistance. Is this correct? I still have questions:
What if you want to add additional supplies, can you just add more wires to the secondaries and connect bridges to them?
illusionx
Does your post show full wave rectification and not a bridge? I thought that a bridge will produce both positive and negative output so can someone please explain the difference.
You can use a single bridge for both rails but your transformer either has to have a single secondary with a center tap connected to ground or two identical secondaries connected in series with the series connection connected to ground. This provides a return for each rail.
yldouright,
I suggest the following web pages as good reading:
http://www.nzart.org.nz/nzart/exami...wer Supplies/STUDY NOTES - POWER SUPPLIES.htm
http://www.educatorscorner.com/media/Exp93c.pdf
http://technology.niagarac.on.ca/courses/etec1120/Files/Unit12.pdf
yldouright,
I suggest the following web pages as good reading:
http://www.nzart.org.nz/nzart/exami...wer Supplies/STUDY NOTES - POWER SUPPLIES.htm
http://www.educatorscorner.com/media/Exp93c.pdf
http://technology.niagarac.on.ca/courses/etec1120/Files/Unit12.pdf
kilowattski
Thanks for the links but they don't seem to have the topology I am looking for. None of the examples presented (yes I looked at all three) show a + and - rail result. To me, half the bridge needs to change direction and if that is the case, why do you need a center tap if you have a common ground point? The + and - rails are already in the wave so by pointing the diodes in the opposite direction, you should end up with a + and - stepdown and a common ground reference. At least that is what it seems like to me
Thanks for the links but they don't seem to have the topology I am looking for. None of the examples presented (yes I looked at all three) show a + and - rail result. To me, half the bridge needs to change direction and if that is the case, why do you need a center tap if you have a common ground point? The + and - rails are already in the wave so by pointing the diodes in the opposite direction, you should end up with a + and - stepdown and a common ground reference. At least that is what it seems like to me
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.