LINEAR PSUs

[sinus trouble]

Greetings necks!
The linear PSU is not everyones cup of tea! They are bulky, produce lots of heat and weigh an absolute ton!
I fully understand the 'switch mode' PSU concept and even repaired some in the past but never used one in any of my projects! Not saying there is anything wrong with them? Merely just a reliability issue to me?
Anyways I have been experimenting recently with the regulated type rather than just the basic
unregulated type!
I managed to get a handful of 2SK2690 power FETs and used a simple LM317 to drive the gates!
The results weren't too bad? I fear they will need forced cooling? But under load they seemed pretty stable!
Cant post any pics as yet??

[Analyser]

Greetings Sinus!

The reason you can't post photos is because you use too many question marks? There is secret code on the forum preventing people with question and exclamation mark syndrome which prevents photos from appearing?!

To answer your question though, linear power supplies win for me and this is because they are more reliable and less noisy than SMPS.
Sure linear are heavy and less efficient but all I seem to be spending most of my time on these days is fixing bloody SMPS or trying to get rid of the noise. And it's pissing me off.
To get a quality SMPS you need to spend much more than the equivalent linear one. Still need to watch for the noise though.

2 rants over. Thank you.

[sinus trouble]

Greetings Sinus!

The reason you can't post photos is because you use too many question marks? There is secret code on the forum preventing people with question and exclamation mark syndrome which prevents photos from appearing?!

HeHe!!!! I was in denial of my condition, but are now seeking rehab! Have you ever thought of a career in counselling Mr Analyser?

Obviously to test this kind of regulator, I thought a purely resistive load would be too easy
I decided to use a DC motor which should give me plenty of noise to deal with and filter accordingly!

[Analyser]

Always happy to provide counselling Sinus.

Using a resistive load is fine, makes it actually easier to see what's coming out of the power supply rather than getting distracted by the noise coming out of the motor. Use a scope to look at the noise and try to keep the ground wire as short as possible. Some scope probes come with a small springy thing which most people throw away because they don't know what it's for. It's to replace the ground lead and reduce the chance of the scope picking up ambient noise in the room.

I bet it will be fine though, unless your FETs are oscillating or something.

[Albert H]

I have always preferred linear supplies. It's easy to get nice toroidal transformers made to your own specification with a bridge rectifier and smoothing cap for about the same price as a reasonable quality SMPSU. Also, I've been known to change the windings on a transformer, or add an extra winding - if you need various voltages in your rig, it's an easy way to do it.

The linear supply is bigger and heavier than the equivalent SMPSU, but it will be infinitely cleaner! I usually don't bother with regulation for PA stages - I get the right secondary on the transformer! For my MRF151 and BLF177 boards, I use a 36V secondary on the transformer - rectified and smoothed, that gives me very close to 50V off load, and about 48.5V at 140 Watts RF out. The current drawn is about 5A, and I usually work on the 10000µ / Amp "rule-of-thumb", so you'll see either a single 47000µ (expensive) capacitor or (more usually) three 22000µ 80V caps smoothing the PA supply. I usually use either a separate transformer or an extra winding on the main one for the 15V supply for the driver. This is separately rectified and smoothed, and uses a 7815 regulator to stabilise it. I've also been known to add an extra secondary for a fan supply!

Some of the bigger SMPSUs on Ebay look tempting, but watch out - some of these put all sorts of high frequency crap on to their output, and this will be added to your output signal!

[sinus trouble]

Great advice, Thankyou both!
Its defo a bit of trial and error at this stage and creating a power oscillator would be an epic fail!
Also the idea of adding a low power auxillary winding to a toroid is interesting too? providing it doesnt overload the primary?
I do see alot of so called 'intelligent battery chargers' on a daily basis which mostly use switch mode technology! To be fair most are housed in plastic housings to comply with double insulation so shielding is minimal! they cause havoc with my workshop radio!

[thewisepranker]

Just because it's in a plastic box doesn't mean it's not shielded.



However, just because the inside is coated with something conductive, or the case is made entirely of something conductive, does not mean that the screening is effective. Lots of these commonly available polymer enclosures with internal conductive coatings offer very little screening effectiveness over those that are uncoated. The reason for this is in the fine detail - no provision for conductive gaskets, the bending moments between clamping and/or mounting screws, lack of conductive conversion coatings, moulding flash, inclusions... The traditional "London" style rig (the folded box) is about as ineffective as they come - but it's not really relevant, as you're interested in generating high field strengths anyway.

Besides, interference at around 100 MHz will be coming from the input and/or output leads, not from directly inside the case. It may have been generated inside the case, but it won't directly radiate from there very well. Stick a ferrite or two on the leads and see if that improves it.

If something of yours is interfering with your radio, take apart and find out whether it's screened, and whether the input and output leads are filtered (at all). You might be glad you did, as you might find some horrors in there and may never want to turn it on again for fear of burning your house down.

[sinus trouble]

Nice 1 Mr Pranker!
I guess we learn something new everyday? I have never seen them enclosures before?
For a primary frequency that id imagine would not exceed 10khz max? The assumption would be that the VHF spectrum is virtually untouched?
Reputable companies like hitachi, bosch and panasonic must have to stick to CE regulations? But still cause noise if close enough to a reciever?
I think they do have strict rules and comply with them? Seems unavoidable with the high sensitivity of recievers and EMI floating around!

[thewisepranker]

The switching frequency is usually over 20 kHz so that people can't hear it, with noticeable power in the harmonics required to generate nice sharp corners up to about 1 MHz.

The CE thing is pointless - it pretty much means China Export. It's not hard to machine the CE mark into your mould tool and sell it on eBay - I don't see anyone checking. Perhaps they check £5000 laptops and TVs, but not £30 PSUs.

The likes of those manufacturers do have to comply with standard EMC limits, yes, but you can't stop emissions, only attenuate them. Therefore, if you get close enough with something sensitive enough, with a low enough noise floor, you'll be able to measure the emissions.

Any SMPSU you've bought off of eBay for not much money is not going to have been EMC tested. If it's noisy, you can keep the leads short and filter it to the point where the filtering components cost more than the PSU itself. They're not all incredibly noisy, though - there's plenty of RF equipment out there using switch-mode power supplies, operating at much higher frequencies and using more complex modulation methods than simple FM.

If you have a noise problem with a switch-mode PSU, try filtering it properly. You'll eventually find one that is clean enough - then you just keep buying that one.

[radionortheast]

remember afew christmases ago whole wifi in house was been knocked out by one of these smps whatever they call them, thought of getting a new one from ebay but you don’t know what your getting, so i used one of mine, theres no easy way to get rid of the interference, can’t beat a good old meaty transformer afews diodes and capacitor works like a charm, even just a diode and cap is good enough for powering many a thing..