Clearing out wall warts

[rjwaldren]

I have most of my net work hardware installed in a walkin closet off of my kids playroom. I've been cleaning up the MDF/Media cabinet and getting everything cut over to N and gigabit. I also just added an Atom D510 based NAS to the room in addition to the Laser MFC thats always been there.

My problem is wallwarts. There are about 7 of them, all low output with Voltages ranging from 6 to 12V. The peak combined output of all of them is only ~45W. I have a couple spare 72W IBM laptop adapters that output 16V@4A so it should be plenty of juice to run a couple of 12V, 9V, 7.5V, 6V regulated outputs. The idea is to use one AC adapter to feed a small box with with multiple regulated outputs within the cabinet. Something similar to the attachment (The laptop adapter would eliminate the A/C and rectifier side of the circuit).

Does anybody have any better ideas? I've been looking around and can't find a COTS solution - Just adjustable single output products with interchangeable tips.

[woodturner]

Does anybody have any better ideas?

Why not just buy a power supply with multiple outputs and wire up the connectors? You will get a much more reliable solution at a lower cost.

Power One is one manufacturer of these products.
http://www.power-one.com/chassis-mou...switchers.html

[LCHIEN]

your solution will be much cheaper than a commerecial solution from Power One.
But your circuit design is horrible. Using 78XX series linear regulators will be terribly inefficient. Depending upon the current loads 72 Watts may not be enough to power 45 watts of loads. your power effiiciency will run as low as 36%. Do you have any idea of how to compute the power dissipation of the 78xx devices when fed with 16V? You wil also have some likely heatsinking problems,. Do you know how to properly select, size and apply heatsinks for the calculated power that the devices will be dissipating? OTOH, the current draw may be much less than the max current listed oin the wall wart - that may reduce the problem but you won't know unless you measure. Don't to forgert to measure both the operating current and the startup initialization current ususally the latter is 50% higher than the first.

Also placement of the capacitors is critical. Now a schematic is not necessarily indicative of the circuit layout but if you build it approximating your layout then you will have oscillation trouble. 78xx regulators work well when properly laid out but layout is important.
You might wish to change from linear regulators to the newer buck boost regulators which will solve much of the power problems i identified above.

[jackellis]

I'm not an electronics guy but I know a little about the cost of electricity. If, as Loring suggests, an energy inefficient setup requires 3 watts out of the wall to provide 1 watt to your gadgets, you're wasting about 90 watts per hour, or around 700 kilowatthours per year. At PG&E's tiered rates, that wasted electricity is going to cost you at least $100 and perhaps as much as $200 per year. Power supplies are never 100% efficient, but the extra cost of a more efficient power supply might easily be recovered in lower electricity bills.

It does add up. Turning off a server made from PC components saves us about 100 kWh per month, worth about $240 per year.

[rjwaldren]

This is a new idea and I haven't had a chance to take any live measurements. It's not likely worth pursuing unless I want to go with a pure solar solution. The initial goal is to clear out the mess.

Loring, you answer was actually part of the question I forgot to ask... The laptop adapter is a switched model so it's efficiency should be considerably better than a straight transformer brick. Fortunately more and more AC/DC bricks are moving this way. The 45W is the total output capacity of the power bricks currently in place - not the actual load that they support. I knew the added circuit, regardless of design, would create additional loses, but didn't realize it would be as significant as you predict. The sample circuit I posted was from a guy doing something similar for guitar pedals, with a much lighter duty cycle than my always on application.

I was looking to use scrounge parts rather than having to spend any money on it, but with what I have around I'm sure it would have been excessively heatsinked. In light of the advice, I think I'd better off just hiding the bricks as the cost of doing it correctly would likely never be recovered in the energy savings.

Oddly I'm a electronics tech by training, but haven't worked in the field for 15+ years, so I've forgotten more than I realize. Most electronics these days is swaptronics that requires no application of real knowledge... Don't use it and you'll lose it

[rjwaldren]

Jack,

That's another concern, though not the primary concern in this situation as the components in question do need to be always on. It seems that the current setup is the best I can do without spending more than it's worth. I'll take measurements later, but the cost of a high quality PS will likely never pay for itself. The only real solution is using truly low powered devices or completely removing power. Power efficiency vs performance has been a major deciding factor in recent purchases, but I've found Power efficiency vs actual Cost usually just means paying for the hype. I'm sure you've seen over your way, PG&E using rate hikes as a motivator to get us to buy more efficient product's, too bad the products rarely deliver. Over the last 3 years, I've replaced my A/C, refrigerator, Dishwasher and cut my entire house over to CFL's, yet my bill keeps going higher. But when I look at my usage on the nice website they have, my usage is the same or lower. I also find it interesting that the "baseline tier" that they use for billing is about 60% of the "typical household usage" they use to rationalize rate hikes.

Sorry Jack, the subject quickly sends me into semi-rant mode.

The new NAS is about as efficient as can be, less than 30W at full load. It's using the newest Intel Atom and chipset with all acpi options optimized. It has a better power footprint that any dedicated NAS solutions I found, plus has the added benefit of being a full and quite capable PC. Where I could really see a benefit is in peripheral standby, I've been looking for a good current or usb controlled power strip to take peripherals completely off while the main component (PC, TV, etc) is in standby. Most of what I've found is either unreliable or the cost isn't recouped in the savings. And manual solutions require action on the part of the user which I've been unsuccessful with which has proven even less reliable. I was recently reading and article about "green hard drives" recently where they compared different green hard drive to the same makers non-green version. Over the expected life of the drives you still paid more for the green label. So I don't really buy into a lot of the marketing hype.

I do keep the main desktop in sleep so I can get to it and the NAS is always on for right now. I use port knocking to remotely open router ports and my router is set up to watch for attempts to connect to sleeping computers and wake them as needed. I haven't added the NAS to that pool yet, but I hope to get the NAS working that way and use it as a trigger to completely remove power to the laser printer when sleeping. The NAS is acting as a single storage point for my wife at home and work as well as secured a photo share, etc for family across the country.

[jackellis]

Sorry Jack, the subject quickly sends me into semi-rant mode.

Just so you know, I don't work for PG&E. If you want to know why their rates are messed up, send me a PM.

[rjwaldren]

I didn't think you did... It's just an area that gets me worked up. PM sent...

I did a little checking and though I can't do any load testing right now, no Kill-a-watt and exposed A/C and a 2 year old are a bad combination. Based on the data plates on the individual transformer the ratings put them in the 45% eff range at full rated load. Of course, the actual load conditions could vary dramatically from the those used to rate the brick and are likely better than the full load numbers. I was impressed that the new gig switch included a switched brick and rates in the 79% range, hopefully the new router will have one two. I thought it was funny that it has a Efficiency Level V label. I could find a Efficiency Level definition that spec'd any thing beyond IV, other than saying "not yet defined". The old linksys is the worst in both power draw and inefficiency at the power source, but to be fair I'm running it way out of spec. You could cook eggs with the current router and modem bricks.

[woodturner]

your solution will be much cheaper than a commerecial solution from Power One.

It might be, but it also might not.

Realistically, on the retail market, he's going to pay $40 to $50 for a decent switcher that will provide the voltages he needs.

He will have probably $20 in parts for his design, plus a couple of hours of time to build it. Then he will have the energy cost of the wasted energy. Over time, he will pay more for this solution.

In addition, he will not have the circuit protection of a commercial power supply - no real short circuit protection, no real overvoltage protection. It's a little like comparing a bicycle to a car. Both will get you there, but the car is likely to get you there more safely.

[LCHIEN]

It might be, but it also might not.

Realistically, on the retail market, he's going to pay $40 to $50 for a decent switcher that will provide the voltages he needs.

He will have probably $20 in parts for his design, plus a couple of hours of time to build it. Then he will have the energy cost of the wasted energy. Over time, he will pay more for this solution.

In addition, he will not have the circuit protection of a commercial power supply - no real short circuit protection, no real overvoltage protection. It's a little like comparing a bicycle to a car. Both will get you there, but the car is likely to get you there more safely.

on those points i agree with you. His home-grown, under-engineered supply may well cost him a lot more in the long run. I was just thinking initial cost when i made that statement. And BTW, for a switcher providing the odd assortment of voltages (12, 9, 7.5 and 6V) is going to have to be a modular unit semi custom, quad output voltages 40-50 watt; I don't think he can touch a Power One quad unit (according to my Digikey catalog) for under $150.

[rjwaldren]

Loring - stopping bagging on the design. I was just a starting point. I'm kidding, your advice was what I was hoping for and invaluable as always.

Honestly I was surprised not to find a commercial solution designed to replace multiple warts at once. Just a bunch of single outputs with interchangeable connector, or A/C spider cables.

I did look at the PowerOne selections:
For the quad outputs, none provide all of the required outputs or the right combo of variable outputs to support my requirements. They were all variable on 1 or 2 outputs and fixed on the remaining 2-3. So additional lossy conversions would have to be added. And your right, the 55W started at about $150 and the 80W jumped to $200. That holds for similar units across manufacturers. I wish I held on to some of these that I scraped at my last job.

So the original direction has been aborted... Whatever solution I chose would not be as simple or cheap as I hoped and would only save the difference between the new and old config. That would never be significant enough for the trouble. I have looked for switched replacements for the transformer based warts and may change those out someday. But that will be more for the size and heat issues, so the can be hidden better rather than the power savings. It's funny that changing a single 60W bulb with a 13W CFL is more significant than a small rack of network/media devices.

The new focus is to use the NAS to control complete power removal from the Laser Printer. Per the manufacturer it pulls 18W at sleep. I know, flip the switch - Not wife approved and she does use it a lot. Anyway the new router should be here today and will have it's warranty voided immediately after I verify functionality. Then I can move my WOL scripts over, I've been waiting on this to finish configuring the NAS.

[jziegler]

Let me chime in a little bit here. Loring is right that a switching supply is the way to go, but unless you can find complete switching modules to suit your needs, switchers do not make a good hobbyest project, IMHO. As an EE, I deal with switchers on most of the boards that I deal with, and even with good PCB layout people working on them, sometimes a board needs to go through a revision just to clean up the noise on them. Most of the parts are also small surface mount packages which aren't easy to solder without lots of practice.

Modules would be a solution, but probably won't be a whole lot better than what you have. You'd still have several modules to hook up to the single AC/DC module.

Actually designing a multiple output power supply for stuff like this would not be difficult for a EE, I wonder how much of a market it could have.

Jim

[rjwaldren]

I agree, switching is not an economical homebrew project. I don't know what kind of market it would have but my searching shows there is some interest in eliminators. All the available solutions are simple move the warts further apart. Most of the DIY stuff I found was related to audio or guitar pedals.

I did find that Jameco has a wide selection of "slim" switching-regulated warts designed to fit side-by-side in power strips for reasonable prices. there's no reason I can't just get a overrated one at each voltage and split it to multiple devices. Still not free but it would clean up the spider web and provide cleaner power at the output to boot.

[HarmsWay]

I use these at work to replace wall warts in delivered instruments:

http://www.bb-elec.com/product_famil...=Sub&Trail=695

http://www.meanwell.com/search/mdr-10/default.htm

Probably too expensive for home use, but convenient for us and much cheaper than in-house designed power supplies. Sometimes we run multiple devices off a single unit but generally one for each to allow remote power-on switching. The DIN rail mounting makes a group of them very compact and secure.

Bob