LED SMPS

[Crackle]

Has anyone had experience of using these types of power supplies for general electronic equipment.
Like amplifiers, car radios or transceivers.
https://www.ebay.co.uk/itm/254054057594?

I am thinking about noise on the output, noise of cooling fans, voltage stability under varying loads.
What the voltage adjust limits are, I have been told 11 to 14 volts.

thanks

Mike

[ppppenguin]

If I wanted a PSU like that for any purpose I wouldn't be buying from an unknown ebay supplier. There's a lot of fake CE marking and these are safety critical items. They may well work when you plug them in, but it's very hard to tell if they are really safe and do what they should under stress.

For nominal 12V operation you could do a lot worse than a car battery (or SLA battery) and charger. The battery ensures that the supply is smooth and will prevent any nasties from the charger giving you grief. This method was used with nicads in a few portable radios and TVs. See Graham Gosling's article about Pye TVs in the new BVWS Bulletin for example.

[BusyBee]

Not those. I tend to collect old computer power supply units (laptops and some external power units from miniature desktops). I have units, some of which are good for over 10A and of different voltage outputs dependant on original application. I tend to use additional smoothing etc in my circuits and have yet to come across any real problem except for switch on surge which will trigger the protection circuitry in the power unit if I am being too demanding.

Tracy

[ppppenguin]

Nice idea, Tracy. Old computer PSUs are junkbox items. I'll have a couple of the B&B stall this Sunday. A few small things to remember:

1: You will need to strap one of the pins to 0V (can't remember which one) to allow it to turn on when you turn on the mains
2: The 12V output current may be rather low for what Mike needs
3: You'll almost certainly need to draw some current (1 or 2 amps?) from the main 5V or 3.3V output to get a stable and sensible 12V output

Plenty of articles on using PC PSUs as bench PSUs. A very casual look gave me this: https://www.electronics-tutorials.ws/blo...upply.html

[Mark Hennessy]

Modern PC PSUs give most of their power on their 12V rails. It's been that way since the P4 came along in the early 2000s. The so-called ATX-12V standard. https://en.wikipedia.org/wiki/ATX#ATX12V_1.x

As mentioned, some minimum load is often required on the 3.3V and 5V rails - perhaps half an amp each as a starting point. The exact amount - if any - depends on the exact design. Suck it and see... Better power supplies have DC-DC converters to make 3.3 and 5V from the 12V rail, and that removes the cross-regulation problems of the traditional shared design (that usually relies on a mag-amp). That's increasingly the case today, as supplies need to be more efficient. If you have an "80-Gold" supply, it's likely using buck converters.

Some computers - like the HPs from Unreal Dave - are 12V only (ignoring standby rail - also 12V, unusually - and a perfunctory -12V at 150mA). The DC-DC converters are on the motherboard, meaning the power connectors as well as the data cables plug into the motherboard. An interesting trend - makes a lot of sense given the large number of Buck converters already on a PC motherboard.

The green wire (pin 16) is the one to short to ground on a standard ATX PSU.

Most have 2 12V outputs, but they are usually derived from the same place internally. A quick visual inspection is worthwhile. Combined, they could be good for around 20A or more.

Would much prefer a PC PSU to an eBay special of dubious merit. And easily replaced in the future - not tying you into a particular form factor, etc. Wouldn't want to use one as a bench PSU though - that's a worrying trend. Having had a PC catch fire - despite having very good quality parts inside - I can vouch for the amount of energy they chuck out on their 12V rails. On the bench, they're just not controllable enough.

Having just looked at the link in #4, I see that's just what they're doing. I haven't read it, but I really hope they recommend fuses at the very least.

[Mike Watterson]

PC ATX PSU The 5V is the main regulated O/P. Needs a 2A load!
Also the 12V is a bit low for ham or CB gear intended for running off a car battery. It's fine for stuff needing a REAL +12V. I've used a 12V car headlamp bulb as 5V load. Without sufficient load the +12V can be as low as +10.5V
The later PSUs with 3.3V O/Ps are not suitable.
Two best solutions are:
1) A proper PSU for 13.8V or 14V
Or
2) A smart charger and "Calcium tech" car/truck battery. The SLA gel batteries are not suitable for higher currents of transmitters, only receivers, unless you want a short life battery.

The 12V only type (used for Lunch box Form Factor PCs originally) are no use without modification. TX/RX gear intended for car/truck is OK on +12V only for RX but needs 13.7V to 14.2V for TX. Modern vehicles with engine running are closer to 14.2 than traditional 13.8 (older lead acid needing top up with distilled water).

[Mark Hennessy]

Only about 20 years out of date there, Mike

3.3V rails came in with the ATX standard in 1995 - all ATX PSUs have them. It was the previous AT standard that didn't have them.

As described, the ATX-12V standard came along with the P4 - all detailed in the link I gave. From that point on, the 12V rail became the main one.

Most have a pre-set trimmer internally, and it can be brought up a bit if needed. Resistor values can be altered if range is insufficient.

12V-only PC power supplies are where many think the industry might be heading - it certainly makes a lot of sense: 12 V only power concept

The 12 V Only System is not an industry standard yet but a proprietary solution, which is currently implemented by all Tier 1 Systembuilders like e.g. Dell, HP, Fujitsu!

Server PSUs are often 12V only. The last one I took apart gives 62.5A - it's an amazing power density (750 watts in an absolutely tiny package).

There's an argument to go to 24V or higher, of course. Perhaps that will happen one day. Laptops are almost there with 19V.

[Mike Watterson]

Yes, I'm wrong about the 3.3V on rechecking what I have. Lack of food?

However none of the ATX (3.3, 5, 12V) PSUs I have, have any adjustment on any rail . Nor has the 12V only DELL PSU, any laptop PSU or AIO PSU brick..

The point of the 12V only idea is to allow the Mobo to have the specific PSUs because the multi-rail PSU has to be overrated on some outputs to be general purpose to different models. Also a laptop style Mobo on an an All-in-one is basically a laptop without the Lithium battery charging circuits, so can use a lower voltage, such as 12V rather than 19V with little change to the circuitry. The true desktop PC is nearly dead as a product as sales are mostly server trays or laptops. The AIO format is a stupid purchase as it's often more expensive than a comparable laptop and basically is a laptop without the "free" UPS a laptop. The true desktop is more efficient with a PSU that has the rails required. Part of the issue is that the standard ATX doesn't have the 1.8V etc and there are onboard SMPSUs for RAM, GPU and CPUs. Certainly tray type "pizza" Servers, laptops and AIOs will be using single rail PSUs.

However, 14V is closer to ideal than 12V for CB/Ham/Marine RX/TX and many SMPSUs not designed for radio have too much RFI. I'd use adapted PC PSUs, (ATX, 12V, 19V etc) for all kinds of stuff but NOT radio gear designed for vehicles. Far better to have a designed for that use PSU or smart charger on a real vehicle battery.

[Mike Watterson]

There's an argument to go to 24V or higher, of course. Perhaps that will happen one day. Laptops are almost there with 19V.

Lots of stuff uses 48V.
Unless you are powering over a distance or more than one Mobo, I can't see any advantage to 24 over 12. If you are, then 48V is the better choice.

[Mark Hennessy]

Ask yourself why the P4 caused the move from 5V being the main rail to 12V. Hint: You're forgetting about current...

The same logic is what fuels the suggestion for even higher voltages. But there is a limit on how high to take it, largely based on MOS-FET ratings - which of course are getting better all the time. 48V is more feasible today than it ever has been, but Schottky diodes aren't so good at those sorts of voltages, leading to other, less efficient choices.

The drive towards 12V single-rail is primarily about energy efficiency.

A PC PSU making several rails from a single full or half bridge (or occasionally flyback) converter is always going to be an uncomfortable compromise that needs help from mag-amps or similar to achieve half-reasonable cross-regulation. Whereas a well-optimised converter that only has to make a single rail will be more efficient and inherently better regulated.

Adding a series of buck converters to step 12V down to 5V and 3.3V gives better efficiency and better performance than an "all in one". If you look at the "80-Bronze" PSUs out there, they tend to be the multi-rail compromises, whereas the "80-Gold" power supplies tend to use separate buck converters. These buck converters can run at much higher frequencies than the forward convert runs at, so component sizes can be minimised - and as they're working at very low voltages, very fast MOS-FETs with minimal switching losses can be used. Where "all-in-one" converters need Schottky diodes for rectification, buck converters can use synchronous rectification - at low voltages and high currents, the difference in efficiency is huge.

All the "12V Only" concept does is move these buck converters from the PSU to the motherboard. That can only be a good thing - the converters can be optimised for the actual load, which is more of a known quantity to the motherboard designer than the designer of a generic PC PSU, who can only guess at what provision each rail might need (and that's not at all easy today). There are already converters on the motherboard anyway - putting them all together makes several things easier, such as EMI. Since there's no longer a "northbridge" on motherboards, most board area on a motherboard is power supply anyway.

Basically, from every angle, single-rail wins.

Let's not confuse this with "AIO" form factors. The choice of form factor is actually irrelevant here - single-rail is more efficient than multi-rail, whether it's a laptop or a PC. There's no shortage of testing being done out there if you need convincing - just look at what these guys get up to: https://www.tomshardware.co.uk/how-we-te...33160.html