the M4 is a poor desgien, would NEVER get pass FCC
i dont own one, but this topic just keeps coming up and up... the RFI generated by it
Mickz, finally someone with real tools can measure it...
just need a Power Supply engineer, to help design a better mousetrap
maybe this is in your skill set... its not in mine, while i can program microcontrollers
i dont have the full knowleged on complicated switching power supplies
It sounds like MickZ has the tools.
SpecAns, UHF, etc....
This is a design problem.
The solution is a new release. MTX-4b?
one thing, how it works, any wire connected to the power supply becomes the antenna for the noise!
so my guess is putting it in a metal grounded box, dont help
cause your gonna have wires coming out...
the should be a filter on each wire coming out, the reduce EMI/RFI
is there a published schmatic to these suckers?
Exactly. MickZ mentioned toroidal RF chokes on its inputs and output with slight change, so IMO that is on the right (or main) track.
But it will need more than a simple choke.....
And it should be internal - especially the real HF stuff that radiates before the external choke/filter.
Usually input & output LC filters suffice, but with some circuits, sectional isolation with their own filtering and shielding is used because the interference elsewhere is to complex to overcome (whether CPU or sense interference, or beat frequencies).
Certain internal PCB tracks may need surrounding by ground tracks, and wires replaced by coax etc.
EMI/EMC (RFI suppression) is a science on its own. Many circuit and PCB designers are unaware of it - until the problems show (in what is called their "prototype").
It usually requires a complete PCB redesign and layout change - even if extra components are not required!
Things can usually be placed in their own mini-Faraday shield and in/outs filtered. But that filtering can be very difficult and expensive... it depends a lot on frequency (rise-times) and magnitude. (EMP & TEMPEST apply - eg military shielding.)
right on, a high current PC switching supply needs to be designed by EXPERTS!!!!
and must pass FCC part 15 class B
there are some that do, they are about $200
and look like normal ATX supplies, but take 12v inputs
usually for industial uses....
and could be made better with a microcontroller, controlling when its on...
funny thing, HAM Radio people run into to this, and they been around alot longer
they might have some answers....
I was involved with telco inverters - up to 5kW outputs, the usual EMI/EMC stuff PLUS psophometric filtering (audio frequencies) - man, what a female canine that was!
And the came the SMPS versions....
But when you had good designers.... (IE, 20 years ago; 50kHz or 100kHz conversion - their "true-sine" UPS/inverter outputs had 0.2% distortion or less - I was going to use them for audio amplification).
And yes, the powerstream series seem impressive - especially for the price.
Maybe use them plus the building blocks I often refer to (normal relays with input-ORing or latching or low-volts cutout or timers or UPS-handshaking...) & thereby control your own destiny.
Just a quick update.
The problem with the M4 is the 12v booster/convertor circuit, the other rails are pretty much ok.
I modified the 12v output with some bypass caps and a few ferrite beads.
Got a lot of improvement, however a 200w ATX supply available from Jaycar, our local
electronic parts and kit supplier, is far superior in noise performance.
Iím going to trial this unit in the car is see how much it degrades GPS sensitivity.
This latest M4 PCB is nicely made and put together, itís such a shame it falls far short of its
claimed performance, and the standby timer programming is still buggy. This is so simple to
program, (basic micro 101 programming stuff) how they could still have troubles after all this
time is unbelievable.
To other comments:
Yes the case can/will make it worse as it close couples to the toriod booster inductors and
switching inductors/filters. These toriod inductors, carrying substantial currents, have a close
field patten and donít usually radiate far if kept away from anything that can and circuit layout is designed with this in mind.
As already suggested, any wire means the complete ATX lead set has to be filtered, however
in a car with no real ground reference it gets difficult to keep "all" of this really bad crap
suitability isolated Ė the car body is still an aerial of sorts one way or another.
I don't have your level of expertise, but I have yet to hear anything with a major set of expletives/adjectives for opus power-units so far, for those that have used them including myself. I still have an m4 in a drawer as it was nothing but a severe headache, hell, didn't even make it to the RF noise part. Opus, did the connections, set it, works fine to date 2 years later.
Impressive stuff MickZ. I'll have to stop knocking Aussies.... then again, you are here aren't you!
Besides, there is a big difference between selling forums with their blind (or trusting) sheep following, and the Jaycars, Oatleys & other local resources.
With the car's ground, it is a ground - the "lack of earth" is not an issue.
If you mean bad grounds and long ground paths, point taken.
And whilst the car body acts as antenna, that's only for external things - not anything internal to the car.