DC-DC converter for a small boost

You could but then you'd have to rectify and that produces noise again. It can be muffled with caps across the diodes though. 2Vdiff/1A at 100Hz makes for rather chunky magnetics. What is your application?

I have done a lot of switchers and never had noise issues. This included ultrasound gear where you have to hear the grass grow. One trick to minimize magnetic coupling from the inductor into surrounding circuitry is to use a toroid.

Regards, Joerg

I'd look for a sine wave based circuit... using a low frequency square wave will stil lcreate harmonics...,,,,

If this is going on an aircraft, there are a LOT of RTCA TSO specs you better know about....

why do you need 13V, where is the 11V coming from?

Mark

Of course, you only really need an 11-volt-to-2 volt converter.

John

I've made switchers that powered superhet receivers, and second Joerg's comments -- well-made switchers can be quiet. Your problem is made easier by the VHF environment: switchers don't make much VHF.

One start might be to use a switcher with non-blazing switching speeds (slower rise and fall times). That'll reduce harmonics at their source, at the cost of a little loss in conversion efficiency. Linear Technology offers some devices, or you could roll your own.

Honestly I don't think you'll have much problem from there, but...

Basic Ideas o shielded magnetics are a must (toroids, or magnetically shielded inductors) o pay close attention to filtering and managing the current flows into such filters o minimize the loop area of heavy current paths o low

(ground and power planes make these easier to achieve)

o note that noise can be electrical or magnetic, and can be radiated, conducted, or induced in neighboring wiring.

Extraordinary Measures o extra filters on input and output o put the switching portions in a shielded box

Best luck, James Arthur

You can also add an RFI filter on both input and output, in addition to the bulk storage components. Think in terms of ferrite beads on leads and ceramic capacitors. And keep the flying node on the converter as physically small as possible to minimize its capacitance to the rest of the universe. Then enclose the whole thing in a metal box connected to the common node. Then those ceramic capacitors can be feed through types.

You may need small capacitors (0.1 uF or so) in parallel with the large ones to get low ESR at high frequency.

Another idea is a linear automotive type (12 VDC) audio amplifier with a sine wave oscillator at the input, and a transformer, rectifier, and filter at the output. You only need about 20 watts total, or you could just make a

Paul E. Schoen

Why not a 50-100KHz sinewave into a power opamp such as an L165 driving a small step down torriodial or potcore transformer to get an isolated 2v? Maybe not the best solution, but I just wonder if it would work.

Mike

Just using Large 1500uF caps will Not solve Noise Problems.

You Need Low ESR Type caps. Even a couple of 22uf caps in Parallel should reduce the noise very well if they are Low ESR.

The Whine is Probably just a poorly wound inductor..

Capacitors can make audible noise too.

You should be able to get the noise down to nearly nothing without having to go to extremes. I've read some of the other suggestions and basically they've already hit 3 main areas.

(1) Read up on doing the layout.

(2) Reduce the radiation from the inductor by having it be a torroid etc.

(3) Make sure every path in and out goes via an impedance and has an RF bypass.

I'm going to expand on #2 a bit.

Torroids are good from a radiation point of view but not perfect. They also can be mechanically harder to deal with than other types. If you don't use a torroid, you want an inductor where the gap is burried inside and not out at the surface.

Pot cores and EFD cores can have their gap at the center post. These are good. You can make them better by adding a copper foil over the outside.

Low cost inductors like Coilcraft's DS5022 can still be used if you provide some external shielding. This can often be done just by how you do the mechanical design. The leaked magnetic field comes straight out the top and bottom of these coils. You want a largish conductive object a short distance above and below them. This happens sort of naturally if you put your PCB into a box.

Hello Ken,

Pot cores in an aircraft can be iffy. I have seen too many shake apart, and that already happened after rough cargo flights of a few hours or long truck rides over washboard roads. Riveting might help but that ain't easy to do on ferrite.

I'd design it so you don't need an air gap. Make sure your DC balance cannot run away.

Regards, Joerg

I've used pot cores with the spring clip style mount in VERY shaky environments without problem.

Tested at length on the shake table at Kodak, Rochester. You know, the shake table where they warn you, "Don't stand on the shake table while it's running. The last guy who did lost his kidneys." ;-)

Very low frequency terms in there.

...Jim Thompson

That is no small feat. You need to obey lots of standards and regulations, plus very stringent EMI testing. If this thing inhibits a critical RF communication and something bad happens you'd be in hot water.

That's a pretty slow chip, around 50kHz IIRC. I have always designed my switchers above 100kHz and they were all quiet. Actually, one is running right here while I am listening to the frequency of our local airstrip. No noise whatsoever, even with no signals and the squelch fully open. As John and Paul wrote, stuff like a 1500uF cap does nothing to muffle anything above a few MHz. You need to spring for some good 0.1uF ceramics and a few 0.033uF ceramics. Nice common ground, plus ferrites on the lines in and out of the (hopefully) metal box. Good RF layout techniques are a must.

Regards, Joerg

Hello Jim,

Some of my stuff had to go on those impact testers. Slides a few feet up, then drops onto a concrete footing. They regularly replenished layers for the footing and one was said to be around 10ft into the ground by then.

It's the higher frequency rattle that kills parts. It's amazing what can be found at times. Ceramic caps, nuts, washers, pot core shards, tantalums sans legs. Africa was usually the worst destination in terms of shaking systems apart. Mostly washboard roads or what I believe they call sastrugi roads over there. Truckers drive as fast as necessary to reduce the shake in their spines but that leaves a high frequency rattle and sometimes it means they can't see a huge pothole in time. Then there are those flights on 50+ year old radial engine freighters where they don't use the synchronizer (usually because it quit working a few decades ago). Somewhere I have a photo where there are guys kneeling behind each engine during the start, holding up fire extinguishers with the pins removed. They wore eye protection, helmets and gloves as some length of flame out of the exhaust was considered "normal".

What I do in critical apps is use toroids, lay them sideways and tie them down with a good foam cushion around them.

Regards, Joerg

What always both fascinated and scared me was to view a CRT with a strobe... so you see the neck flexing :-(

...Jim Thompson

Hello Jim,

Scary. Surprisingly we never had a CRT fail, even though I bet many of those cargo flights were not pressurized. When selecting monitors we always made sure that there was not that huge conglomeration of parts on the socket PCB. Preferably there should not even be a socket PCB.

The worst we ever had was a huge and way oversized cast aluminum leg of an ultrasound machine that had been completely broken off. And this happened in London, UK. Most of the time I tried to convince my clients to go from assembled aluminum structures to a welded steel frame. It was cheaper, offered better EMI performance and was a lot stronger. Afterwards you could ram the systems through a 1/2" Hardibacker wall leaving them unscathed except for a few scratches.

Regards, Joerg

The problem there is you're inducing RF into circuits with -100dBm sensitivity. You need to think about shielding and distancing the power supply.

Yep, there are quite a few Pro video cameras (sony megabuck jobs)that you cannot hang a radio mic Rx close to the main body of the camera because the camera generates enough low level interefernce to substantially reduce the performance of very good radio mic kits

martin