S
spamhog
- Jan 1, 1970
- 0
I need 6 amps @ 12Vdc, quitet and stiff, 24/7.
I have two nice new-old-stock ATX PSUs rated 8A on the 12Vdc rail, and
found they can actually be trimmed to produce 12+ Vdc instead of the
customary 11.5 or so.
I want touse one as DC source, impose RF silence and ruthlessly
repress the AC hum.
Having noticed that both ends of a SMPS contain RFI filters, and that
the AC-DC rectifier-filter has a role in hum filtering, I am
contemplating the following surgery, aimed at PASSING BOTH AC AND DC
THROUGH MORE THAN TWICE THE USUAL AMOUNT OF PASSIVE AC AND RF
FILTERING.
1) Grab these two PSUs of identical rating. Bolt together the tops of
their cases (i.e. the parts that the boards are bolted on)
2) Get the board of the 1st one off the case, split it along the AC-DC
galvanic divide.
3) Get the AC side, remove components to leave just AC input filter,
NTC inrush limiter, rectifier, caps & ancilaries (bleeder, reverse-
plarized diodes, etc etc.)
4) Reinstall this bit in its box.
5) Route the +-300Vdc wires via an added RF toroid choke in the first
case, through a copper pipe to the other case, via another choke in
the other case, to the AC (!!!!) input of the other PSU (obviously
severed form the AC socket).
There the DC will be RF-filtered again, throttled by the second NTC,
and presented to the second set of caps with the correct polarity by
the second rectifier bridge.
6) Arrange the other board to be completely insulated from its case.
7) Route the required DC wires (in my case, 0V/Ground, +12V, +5V)
backward to the 1st case via the same toroid choke - copper pipe -
toroid choke. All windings in parallel.
7) Pipe to be insulated from second case, not carrying any AC or DC
current, and grounded to a Single Grouding Point inside the 1st case,
together with everything else, incl. the board in the other case
thhough the bolts.
8) SGP will be a thick copper plate bolted inside the 1st case, where
the DC-DC converter of the 1st board once was, All grounds will
converge here.
This should help tame hums and avoid any magnetic saturation in the
chokes from high DC currents, as all cores will see zero net charge
flow.
8) Reuse components from the 1st, broken up board, or add big
electrolytics, to make a further high-DC-current capable RFI fiiter.
9) DC will also go to whatever required ballasts load(s) housed in the
1st case, which will retain its cooler to remove their heat.
10) DC output (both wires, in my case just +12V and 0V) through
another toroid choke.
10) Yet another choke for the AC ground.
Choking the grounds should help keep any RF or noise from entering or
exiting the case.
DOUBTS I ALREADY HAVE
- Core Saturation in the AC filter fed with DC
At least the first RF choke on the AC RF filter of the "preserved"
board apears to have windings that are out of phase - typical of
chokes for transferse mode. This should imply that it won't go crazy
by feeding it DC instead of AC. Is this sound reasoning? The second
choke is potted, but I assume it's also transverse mode, so it should
be OK. If not, I'll live with it.
- Power Factor Correction
PSUs are rumored to have some form of PFC, some claim to have an
active one. This perversion of feeding DC to an AC circuit might or
might not subvert it. I assume nothing will go kaboom... but I
haven't done the smoke test yet. Should I worry about this, or focus
my worries on this horrible financial crisis instead?
DOUBTS I DON'T HAVE JUST YET
By definition, I don't have any doubts I don't have.
Please suggest what other doubts I should have.
COMMENTS? ANY REASON NOT TO DO THE ABOVE?
TIA
Filippo
I have two nice new-old-stock ATX PSUs rated 8A on the 12Vdc rail, and
found they can actually be trimmed to produce 12+ Vdc instead of the
customary 11.5 or so.
I want touse one as DC source, impose RF silence and ruthlessly
repress the AC hum.
Having noticed that both ends of a SMPS contain RFI filters, and that
the AC-DC rectifier-filter has a role in hum filtering, I am
contemplating the following surgery, aimed at PASSING BOTH AC AND DC
THROUGH MORE THAN TWICE THE USUAL AMOUNT OF PASSIVE AC AND RF
FILTERING.
1) Grab these two PSUs of identical rating. Bolt together the tops of
their cases (i.e. the parts that the boards are bolted on)
2) Get the board of the 1st one off the case, split it along the AC-DC
galvanic divide.
3) Get the AC side, remove components to leave just AC input filter,
NTC inrush limiter, rectifier, caps & ancilaries (bleeder, reverse-
plarized diodes, etc etc.)
4) Reinstall this bit in its box.
5) Route the +-300Vdc wires via an added RF toroid choke in the first
case, through a copper pipe to the other case, via another choke in
the other case, to the AC (!!!!) input of the other PSU (obviously
severed form the AC socket).
There the DC will be RF-filtered again, throttled by the second NTC,
and presented to the second set of caps with the correct polarity by
the second rectifier bridge.
6) Arrange the other board to be completely insulated from its case.
7) Route the required DC wires (in my case, 0V/Ground, +12V, +5V)
backward to the 1st case via the same toroid choke - copper pipe -
toroid choke. All windings in parallel.
7) Pipe to be insulated from second case, not carrying any AC or DC
current, and grounded to a Single Grouding Point inside the 1st case,
together with everything else, incl. the board in the other case
thhough the bolts.
8) SGP will be a thick copper plate bolted inside the 1st case, where
the DC-DC converter of the 1st board once was, All grounds will
converge here.
This should help tame hums and avoid any magnetic saturation in the
chokes from high DC currents, as all cores will see zero net charge
flow.
8) Reuse components from the 1st, broken up board, or add big
electrolytics, to make a further high-DC-current capable RFI fiiter.
9) DC will also go to whatever required ballasts load(s) housed in the
1st case, which will retain its cooler to remove their heat.
10) DC output (both wires, in my case just +12V and 0V) through
another toroid choke.
10) Yet another choke for the AC ground.
Choking the grounds should help keep any RF or noise from entering or
exiting the case.
DOUBTS I ALREADY HAVE
- Core Saturation in the AC filter fed with DC
At least the first RF choke on the AC RF filter of the "preserved"
board apears to have windings that are out of phase - typical of
chokes for transferse mode. This should imply that it won't go crazy
by feeding it DC instead of AC. Is this sound reasoning? The second
choke is potted, but I assume it's also transverse mode, so it should
be OK. If not, I'll live with it.
- Power Factor Correction
PSUs are rumored to have some form of PFC, some claim to have an
active one. This perversion of feeding DC to an AC circuit might or
might not subvert it. I assume nothing will go kaboom... but I
haven't done the smoke test yet. Should I worry about this, or focus
my worries on this horrible financial crisis instead?
DOUBTS I DON'T HAVE JUST YET
By definition, I don't have any doubts I don't have.
Please suggest what other doubts I should have.
COMMENTS? ANY REASON NOT TO DO THE ABOVE?
TIA
Filippo
