SMPS, issues: (9-21-2016, rev.4)
WIKI on SMPS
You are here for 1 reasons, $250 linear supply costs more than $50 SMPS.
(ham oriented) short as can be on such a complex issue of EMI.(RFI){a near infinite topic if ever there was, besides, RF propagation)
Switch mode Power supplies, (not linear)
The topic is complex and doing the tests correctly 10 times more so.
The SMPS has many advantages and only one important down side:
· Cheaper, due to the lack of 35 lb copper and BIG IRON transformer.
· Weighs up to 20 times less, due to the above fact.
· 20 times less expensive to ship, from China.
· Vastly more efficient, 2 to 3 times. 95% versus 30% of linear.
· It is way more complex, but the parts are cheap and light.
· It is vastly smaller, up to 10 less physical size.,
· It can have noise on its output, but quality designed filters solve that problem. Or a noisy fan.
· It’s GREEN, and ENERGY STAR like. it uses way less energy. And less CO2 to mother earth!. Best is to focus on FCC LINE filters. (IMO)
I’d guess, the transformer is 50x less cost too. (price them out and see)
Mouser lists them from $65 to $140 each, just a raw transformer.(linear)
All Engineering is based on compromise.
· AC Line conducted noise outward, in to your homes, AC power feeds.
· DC line conducted outward (to a radio).
· Radiated (mostly electromagnetic from the supply transformer out the sides of the case. (mine radiates 60khz noise about 1 foot out)
· Acoustic noise, be that a fan or cheap loose SMPS internal parts.
The SMPS supply, is not some new kid on the block, we had them in 1975. and were full of noise then. Even APPLE had them in 70s.
The FCC rules are clear, if it passes Class B reg.’s. (truly does)
Then it will be low noise.
On the output port DC, and noise out on the AC line. That is because it uses a LINE filter on both ends, or it will never pass the FCC test. (ever)
All SMPS use a chopper, it’s just a large huge square wave power driver that
Allows the maker to use cheap tiny (like 20 pounds less IRON).
Transformers are way more efficient when you avoid 60hz frequencies, even 400 Hz , as used on military Aircraft are vastly smaller transformers.
Wiki on Square wave noise.
No lie keeping those nasty square wave harmonics inside the box, is not easy!
FILTERS ARE DISTRIBUTED !
This means the new radio, has a proper filter and so does your Class B SMPS power supply. That work as TEAM, cascaded. In the total, the noise can be decreased to over -100DB “down”.
The proof of a good filter is on the radio end is easy see graphic #1 , my ICOM.
Graphic #1 this is the Ferrite core I use on all power supply DC cables.
For $1 Insurance. So do many makers of your gear use one.
Back in the dark ages, my SB104 radio is not FCC rated for noise and the radio has its own, power supply, neither of which are with line filters.
The cure is add a line filters.
I Added brute force filters (3 caps in parallel) and the Ferrite.
Or add a filter just like the ICOM above has. They sell them at mouser.com
I mention mouser, as they have a very fast shipping system that I love.
Cures:
Brute force filters, or use real FCC line filters. (or both)
Ferrite filter, for $1
Facts on noise:
Examples of designs.
See the DC noise before the low pass filter. (in the TI.COM SMPS designers page)
Source noise is about 90 DBuV/m
They (Ti.com) teaches you Texas Instruments how to get noise from the above gross noise levels to the below FCC spec, for CLASS B.
TI.COM states: This is a fast glimpse at Class “B”:
The tests vary by noise class, and circuit impedance and is complex.
Class B FCC: rules, TI.COM
“this case, in order to pass CISPR 22 Class B requirement,
the |Att| dB = 45 dBuV/m.” )
My conversion to volts that is 178 micro volts , I presume 600 ohm line load.
End (ti.com, methods)
This next schematic is the old (year2000?) SS30, see the L4/5/7 and cap filters?
0
The next scope photo is the noise on my SMPS “Mean-well”, 30amp supply 14vdc.
Comments : measuring noise is not easy.
Best is measure it as it enters your RIG, your radio, on the 14vdc
Connector as it enters the radio, or using coax test lines .
Or better on the output side if the radios class B filter. After all they work as team, so let them, and meaure team performance.
Make sure the test leads are well shielded as noise is everywhere, and can skew the test results, the above is at rear of my ICOM. DC power jack.
Its about 10mv.
It’s a cheap china supply from mouser, 99% sold are from china, few to none made in USA. (not now.)
The above, is the older SS30M, and the ferrite donuts are clear.
I don’t like meters in my power supplies, (more to fail or may add inductance to the power feed, Newer SS30, (2016 made) has cleaner outputs.
It’s hard to measure. It is so low.
The case is steel, and no case side radiation can be measured beyond 6” inches.
So, do not set any SMPS on top or bottom of your radio, ever.
Let me say this , almost all China supplies, (DC 30amps,example)
Not one maker spec.’s out the noise, nor its signatures in the RF spectrum.
Not one shows a chart like as seen on the last pages here.
Except the BEL supply maker.
Most supply makers just cop out and say complies to FCC part B.
This is really bad for them and you.
If they (makers) were smarter they would spec, it fully, and get more customers.
And you can read it and be sure, it’s not making noise on your favorite band.
Imagine that, facts and a nice chart showing spectral content.
End.
------REFr. Sources------
REfr, reading.
Try AN118(some of the more difficult supplies, to make clean)
To see how common mode filter works see this page. (they are attempting FCC class B. certs)
This page, shows putting real CM filter on the output and proof.
The next photo shows both the USA FCC and EU limits on and MIL.
1mhz in dead center here. And 50 in red, is 50dBuv.
The USA rules start at 9000cps (Hz)
50 dBuV is equal to -57dBm (below 1mW, 50 ohms) that is 300uV (50ohms)
So they are saying class “B” is below 1mv noise, using CISPR 22 rules.
A nice conversion table to Dbm is here.
The non CISPR 22 conducted limit is. 250 uV (see radiated too).
The rules are best summarized here.
I do see a 30amp supply at digikey.com for “power-one” co. 400watts
The power supply is CISPR 22 compliant and 120mV max ripple.
(professional datasheet, unlike many China versions of same.)
Keep in mind this is noise on both ends, DC end and AC line end!
Conclusions:
I’m happy to use both types of power supply , linear or SMPS.
But I use my scope to be sure it’s really clean, and the scopes FFT feature.
Radiated noise, keep the SMPS at least 1foot from gear. 3foot is better.
Keep the feed line DC less than or equal to 3 feet. ,use the $1 donut.
Many hams can’t be bothered with these racket makers, and who can blame them?
A linear with good 60hz filters and best regulated, is noise free. Nearly so.
Manyy shacks have one huge 14v power supply running the whole ham rack system.
This is a wise idea. Get a 100 amp 14vdc reg linear and forget all this smps worries .
There is great post on one great forum, post 10 seen here, on the topic of inverters. The 3 types, and why some work and others don’t.
The words you get what you pay for, are so true here.
No lie a sine wave drive is best , bet that 50/60 or 400hz or more.
For sure.
Read this link post 10.
Under revision until end of Sept.2016