Since everything important in my radio system operates from 12V,
except for my laptop PCs,
I like to have one heavy duty, well regulated, with serious over
voltage protection and effective
RFI suppression. I was recently given two "new old stock" 120V/240V:
12V @10A average
15A peak linear power supplies. Since my primary power supply for the
last 30+ years has
been a surplus 8A Lambda linear power supply that had undergone a
serious of upgrades
over three decades. It is stating to show it's age and the ripple on
the DC has doubled in
the last year or so. It is still below 20mV, but the caps are nearing
"end of life" and given
the proprietary parts Lambda uses, it would be very difficult to
replace the electrolytics
without serious modifications, up to and including migrating to a new
chassis or case.

I post this as an email to some friends.
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- --
A friend sent me this and I thought you guys might enjoy the
Milatary's take on power supply
RFI/EMI. [linkhttp://www.tscmplus.com/MIL-HDKBK-241b.pdf]

Given my limited usage of my right hand, it is like I now have two
left hand, I have been limited
to fairly simple tasks. In an effort to understand how much RF noise
comes from a 'normal' liner
power supply. I choose a new, old stock, AC-ONE 12V 8A unit. This is a
typical industrial/commercial
unit. Center tapped output transformer with 2 normal, 10 year old, Si
20A 100PIV diodes, and a uA723
with it's own center tapped winding for the higher voltage required.
This section uses 1N4004.
There is no RF bypassing anywhere in the unit.

Using a 100uH home made RFI sniffer probe built into a Sharpie Pen
case and a battery operated
Radio Shack DX398 LW/MW/SW receiver , and loading the power supply
with a fan cooled 2 Ohm
load, I found the main diodes to be nasty RFI sources. A 100 ohm in
series with a 0.01 snubber
across each heavy diode reduced the noise to levels below my detection
threshold.

The low current 1N4004 diodes produced a small amount of RFI, and a
snubber of 100 ohms and
0.01uF across both removed that noise source.

The uA723 had a very slight broadband RF 'white noise' that was
removed by the addition of 0.01
bypass caps on the Vcc, reference, voltage set, and Vout pins. The
2N3055 driver and outputs
produced no detectable noise.

I next connected the AC-ONE to my home made LISN and found that some
noise was still present
on the AC mains from the unit under test. I added a 0.1 and 0.01
across all the secondaries and a
0.01 across the primary and this reduced the noise to right at my
detection threshold. By adding a
single state Corcom RFI filter, there wasn't room for my favorite
Curits 2 stage filter, all the noise
was gone.

This unit is built in a nice heavy duty aluminum case and I replaced
all the fastening hardware with
stainless steel an used star washers under both the bolt heads and
nuts to insure a good conductive
bond.

I added a Reverse and Over Voltage Crowbar to the output and added a
100uH 20A RFI choke in
series with the output and used 0.1 and 0.001 to bypass both the input
and output of the choke.

I also added a Gel Cell 'port' and relay to allow for instant power
fail transfer for emergencies.

Even though it wasn't needed, I added a 10uF tantalum, with a 0.1,
0.01 and 0.001 across the
filter caps. I also added a 10uF tantalum across the Vcc and reference
out of the uA741.

This unit had a DC brushless fan that was thermostatically controlled.
When it kicked in the noise
was unbelievable! In spite of all my tricks I could not quieten the
fan. I used common mode cokes,
added DM chokes, tried bypass caps and still the noise was bad. I
finally substituted a 3' AC
Panmotor fan with a separate heatsink mounted thermostat for control.

I don't know if this 'hall effect', solid state commutation fan is
typical or an isolated noise source.
I checked the few that I have and most have significant noise, but can
be quietened by a CM
choke and capacitors. When I gain some more dexterity I plan on
disassembling the fan and
try bypassing the switching transistors, or adding a LC circuit to
slow down the turn on/turn off
and see if that will tame it.

In addition to the data I gathered from this project, I have a nice
power supply that will power
all of my 12V equipment! My 35 year old Lambda is still chugging along
but it is nice to have
something more recent and uses standard parts. Lambda uses very odd
parts that don't cross
to anything common.

Terry
-- -- -- -- -- -- -- -- -- -- -- -- --
Since ua723 2N3055 power supplies are still very common, companies
like Astron still market them
and they are a staple of the DIYS ham, I thought others might benefit
from my experiences.

Terry

On Dec 3, 11:04 pm, Telamon
wrote:
In article
,



wrote:
Since everything important in my radio system operates from 12V,
except for my laptop PCs, I like to have one heavy duty, well
regulated, with serious over voltage protection and effective RFI
suppression. I was recently given two "new old stock" 120V/240V: 12V
@10A average 15A peak linear power supplies. Since my primary power
supply for the last 30+ years has been a surplus 8A Lambda linear
power supply that had undergone a serious of upgrades over three
decades. It is stating to show it's age and the ripple on the DC has
doubled in the last year or so. It is still below 20mV, but the caps
are nearing "end of life" and given the proprietary parts Lambda
uses, it would be very difficult to replace the electrolytics without
serious modifications, up to and including migrating to a new chassis
or case.

I post this as an email to some friends. -- -- -- -- -- -- -- -- --
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- A friend sent me this
and I thought you guys might enjoy the Milatary's take on power
supply RFI/EMI. [linkhttp://www.tscmplus.com/MIL-HDKBK-241b.pdf]

Given my limited usage of my right hand, it is like I now have two
left hand, I have been limited to fairly simple tasks. In an effort
to understand how much RF noise comes from a 'normal' liner power
supply. I choose a new, old stock, AC-ONE 12V 8A unit. This is a
typical industrial/commercial unit. Center tapped output transformer
with 2 normal, 10 year old, Si 20A 100PIV diodes, and a uA723 with
it's own center tapped winding for the higher voltage required. This
section uses 1N4004. There is no RF bypassing anywhere in the unit.

Using a 100uH home made RFI sniffer probe built into a Sharpie Pen
case and a battery operated Radio Shack DX398 LW/MW/SW receiver , and
loading the power supply with a fan cooled 2 Ohm load, I found the
main diodes to be nasty RFI sources. A 100 ohm in series with a 0.01
snubber across each heavy diode reduced the noise to levels below my
detection threshold.

The low current 1N4004 diodes produced a small amount of RFI, and a
snubber of 100 ohms and 0.01uF across both removed that noise source.

The uA723 had a very slight broadband RF 'white noise' that was
removed by the addition of 0.01 bypass caps on the Vcc, reference,
voltage set, and Vout pins. The 2N3055 driver and outputs produced no
detectable noise.

I next connected the AC-ONE to my home made LISN and found that some
noise was still present on the AC mains from the unit under test. I
added a 0.1 and 0.01 across all the secondaries and a 0.01 across
the primary and this reduced the noise to right at my detection
threshold. By adding a single state Corcom RFI filter, there wasn't
room for my favorite Curits 2 stage filter, all the noise was gone.

This unit is built in a nice heavy duty aluminum case and I replaced
all the fastening hardware with stainless steel an used star washers
under both the bolt heads and nuts to insure a good conductive bond.

I added a Reverse and Over Voltage Crowbar to the output and added a
100uH 20A RFI choke in series with the output and used 0.1 and 0.001
to bypass both the input and output of the choke.

I also added a Gel Cell 'port' and relay to allow for instant power
fail transfer for emergencies.

Even though it wasn't needed, I added a 10uF tantalum, with a 0.1,
0.01 and 0.001 across the filter caps. I also added a 10uF tantalum
across the Vcc and reference out of the uA741.

This unit had a DC brushless fan that was thermostatically
controlled. When it kicked in the noise was unbelievable! In spite of
all my tricks I could not quieten the fan. I used common mode cokes,
added DM chokes, tried bypass caps and still the noise was bad. I
finally substituted a 3' AC Panmotor fan with a separate heatsink
mounted thermostat for control.

I don't know if this 'hall effect', solid state commutation fan is
typical or an isolated noise source. I checked the few that I have
and most have significant noise, but can be quietened by a CM choke
and capacitors. When I gain some more dexterity I plan on
disassembling the fan and try bypassing the switching transistors,
or adding a LC circuit to slow down the turn on/turn off and see if
that will tame it.

Snip

Keep in mind that the efficiency is due to the switching speed. If you
slow it down the transistors spend more time between being in the
saturated on or off state and spending more time in the partially "on"
state will get hotter. This is a general rule for any switching circuit.
And yes if should help reduce the switching noise.

--
Telamon
Ventura, California

This is a linear regulator so there wasn't any switching to speak of.
Most of the noise was from the "switching" action of the Si diodes
and the, I am guessing here, broad band noise from the internal
Zener diode.,

I try to avoid SM power supplies at all cost. I am more then willing
to trade a few BTU of waste heat for less RF noise.

Terry