Allen,
I see that Terry has already answered one of your questions. Thank you Terry!
The layout you site is very close to the final one. I made a subtle change to it when Hugh found an error. However the final layout is 99% identical to the one that is posted.
The output devices are close together. However I've used similar spacing on previous projects without problems. Please note tho that it is important to use a robust heatsink as the heat will be concentrated in a small area and must be wicked away.
I see that Terry has already answered one of your questions. Thank you Terry!
The layout you site is very close to the final one. I made a subtle change to it when Hugh found an error. However the final layout is 99% identical to the one that is posted.
The output devices are close together. However I've used similar spacing on previous projects without problems. Please note tho that it is important to use a robust heatsink as the heat will be concentrated in a small area and must be wicked away.
pheonix358 said:
Therefore to replace any components(upgrade or testing) you need to remove the pcb from the heatsink?
The decoupling caps(DEC) on the output trannies are earthed via the 470uF (on the -55v) and the 1000uF +55v ground point?
allan
Yes Carl saw that, thanks
Sorry also
How about a reg'd front end possibility?
It does make a difference on the standard one.
The resistor across the (LTP adjust) pot?
It is necessary, I found out the hard way.
Whoops just saw it there.........470ohm..
Allan,
If I properly understand your question (I can be thick headed!) all bypass caps are between rails and ground, and are in thick traces.
The current layout does not lend itself to splitting out the front end for regulation. Studying the layout, if we lose the 'StaCon' connectors that are redundent to the Phoenix connection blocks I could add in StaCon connectors for voltage regulation. Builders not wanting to use voltage regulation would have to solder in a jumper. Should I make this change?
Who wants to contribute a proven voltage regulation circuit? We could do a group buy on 'general purpose' power supply PCBs concurrent with the ULD amplifier. The power supply PCBs would be useable on other projects as well.
If I properly understand your question (I can be thick headed!) all bypass caps are between rails and ground, and are in thick traces.
The current layout does not lend itself to splitting out the front end for regulation. Studying the layout, if we lose the 'StaCon' connectors that are redundent to the Phoenix connection blocks I could add in StaCon connectors for voltage regulation. Builders not wanting to use voltage regulation would have to solder in a jumper. Should I make this change?
Who wants to contribute a proven voltage regulation circuit? We could do a group buy on 'general purpose' power supply PCBs concurrent with the ULD amplifier. The power supply PCBs would be useable on other projects as well.
Carl
The 470uF and 10ohm were there to form an RC circuit to help PSRR on the - rail,
On the standard pcb SC put the - rail running under the output inductor.
You have it running next to the inductor.
The output +/- rails use the 1000uF, when driven hard these caps will have heavy current pulses on them, which we don't want interfering with the low current front end.
There are also 2 earth connections for the +/- rails
allan
ps
need to go to work
The 470uF and 10ohm were there to form an RC circuit to help PSRR on the - rail,
On the standard pcb SC put the - rail running under the output inductor.
You have it running next to the inductor.
The output +/- rails use the 1000uF, when driven hard these caps will have heavy current pulses on them, which we don't want interfering with the low current front end.
There are also 2 earth connections for the +/- rails
allan
ps
need to go to work
Allan,
Attached is the layout adjusted to allow for front end voltage regulation. It is more obvious on this version of the layout as to how the negative rail connects to the 10R resistor and 470uF capacitor. They do form a RC network. When looking for what you described I did find an error in the layout! The decoupling cap associated with the 470uF cap was not connected to ground. It was floating. I have fixed that. Please notice that this version of the PCB can be 'front end regulated'.
As to your comment about grounds ...
Yes the ground from the POS rail is bonded to the NEG rail. Are you suggesting ground loop problems may result? I invite others to make comments.
Attached is the layout adjusted to allow for front end voltage regulation. It is more obvious on this version of the layout as to how the negative rail connects to the 10R resistor and 470uF capacitor. They do form a RC network. When looking for what you described I did find an error in the layout! The decoupling cap associated with the 470uF cap was not connected to ground. It was floating. I have fixed that. Please notice that this version of the PCB can be 'front end regulated'.
As to your comment about grounds ...
Yes the ground from the POS rail is bonded to the NEG rail. Are you suggesting ground loop problems may result? I invite others to make comments.
Attachments
I just found two mistakes in the silkscreen of the above PCB layout. The label for the StaCon that is for negative regulation is marked 'M VREG'. It should be marked 'N VREG'. Also the plus sign for the 1000uF 63V cap is on the wrong side of the cap.
I also just noticed that in the schematic posted back on post #167 (current schematic minus the vreg changes) both the 1000uF and 470uF caps for the negative rail are upside down as well. I seem to have 'capacitor issues' of late.
I also just noticed that in the schematic posted back on post #167 (current schematic minus the vreg changes) both the 1000uF and 470uF caps for the negative rail are upside down as well. I seem to have 'capacitor issues' of late.
Carl_Huff said:
Carl, that's cool. The only other thing would bee to try and reduce the 3 and a bit " dimension down to as low as you can realisticaly go. This is the dimmention that matters for your heatsink size.
Some may want to mount it in a 2U rack knowing that they will never need full power but want headroom. Many heatsinks in Aus have 75mm as one of the measurements so going under that would be good. You would need to save 1/4 inch and I'm not sure you can get there.
Terry
Terry,
The current PCB layout is 3.25 inches (82.55mm) tall. Add to that another .75 inches (19mm) for the output devices. This layout when mounted vertically needs a heatsink that is at least 4 inches (101.6mm) tall. A 5 inch (127mm) or greater heatsink would be a better idea. 3U is 5.23 inches (133mm). There is no good way to shrink it.
If you want to go with a shorter chassis I suggest that you mount the PCB horizontally to a right angle flanged heatsink. Conrad sells them.
The current PCB layout is 3.25 inches (82.55mm) tall. Add to that another .75 inches (19mm) for the output devices. This layout when mounted vertically needs a heatsink that is at least 4 inches (101.6mm) tall. A 5 inch (127mm) or greater heatsink would be a better idea. 3U is 5.23 inches (133mm). There is no good way to shrink it.
If you want to go with a shorter chassis I suggest that you mount the PCB horizontally to a right angle flanged heatsink. Conrad sells them.
Some of the best power supplies are shunt.
What about using a TL431? These are stellar devices, cheap, and very quiet. It only needs to shunt about 15mA, not too much, though you might have to stack it to handle the higher voltage as they will only tolerate 37V.
The application notes have good start circuits.....
Hugh
What about using a TL431? These are stellar devices, cheap, and very quiet. It only needs to shunt about 15mA, not too much, though you might have to stack it to handle the higher voltage as they will only tolerate 37V.
The application notes have good start circuits.....
Hugh
i used the old SC100W reg'd power for test.
LM317/337 based, set to +/-55v
Glen Kleinschidt has a reg'd power supply.
There are various others around here too.
The SC layout used top and bottom tracks to counsel out each other's noise/pulses.
These's ideas goes back to there 15Watt class A layout.
allan
ps date format.
LM317/337 based, set to +/-55v
Glen Kleinschidt has a reg'd power supply.
There are various others around here too.
The SC layout used top and bottom tracks to counsel out each other's noise/pulses.
These's ideas goes back to there 15Watt class A layout.
allan
ps date format.
AKSA said:
awpagan said:i used the old SC100W reg'd power for test.
LM317/337 based, set to +/-55v
Glen Kleinschidt has a reg'd power supply.
There are various others around here too.
The SC layout used top and bottom tracks to counsel out each other's noise/pulses.
These's ideas goes back to there 15Watt class A layout.
allan
ps date format.
Couldn't modify post so......
There are a variety of different reg'd psu around.
It's really up to personal preferences of which you use.
TAD,
Those .22R output resistors are assuming vertical mount cement Xicon resistors from Mouser. Nothing else will fit.
Does anyone have comment on the power and ground traces? They are above and below one another on both layers bonded by vias. At this point the 'front end' traces are separated from the output traces. I could probably add a few vias to the ground trace that runs along the bottom of the the PCB between power connectors. Thoughts anyone?
Now as to the date format ...
06.01.09 as in June (6th month), first day, year 2009. You Aussies have got to learn to read American!
Those .22R output resistors are assuming vertical mount cement Xicon resistors from Mouser. Nothing else will fit.
Does anyone have comment on the power and ground traces? They are above and below one another on both layers bonded by vias. At this point the 'front end' traces are separated from the output traces. I could probably add a few vias to the ground trace that runs along the bottom of the the PCB between power connectors. Thoughts anyone?
Now as to the date format ...
06.01.09 as in June (6th month), first day, year 2009. You Aussies have got to learn to read American!
Carl,
The gap between the Q9 heatsink and the output coil can we assume this is pretty close to scale?
The preliminary prices you quoted earlier in this thread are for the most part excellent. Are you satisfied enough for someone to make some protos and check this out. I for one would like to build a pair. I have a beautiful set of heatsinks all ready for this size project. The Regulator boards Bob Ellis was selling would work here nicely. I have several of them.
Any room on the pcb for a board mounted fuse holder? That is always a nice location for this type of protection.
Tad
The gap between the Q9 heatsink and the output coil can we assume this is pretty close to scale?
The preliminary prices you quoted earlier in this thread are for the most part excellent. Are you satisfied enough for someone to make some protos and check this out. I for one would like to build a pair. I have a beautiful set of heatsinks all ready for this size project. The Regulator boards Bob Ellis was selling would work here nicely. I have several of them.
Any room on the pcb for a board mounted fuse holder? That is always a nice location for this type of protection.
Tad
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.