**BIG** 24V supply

[Curious.George]

I'm looking for ideas for (surplus) sources of really hefty (regulated) 24V supplies. Currently, I've been using modified 12V supplies from servers (disk spindle power) to get ~24V @70A.

I'd like to get closer to 100A (realizing I'll have to power it from a 220V branch circuit if I want to preserve the nominal safety margin)

[goontron]

Re: **BIG** 24V supply

https://www.mouser.com/ProductDetail...GEhaOXBzc74%3d

[Curious.George]

Re: **BIG** 24V supply

https://www.mouser.com/ProductDetail...GEhaOXBzc74%3d

Sorry, I wasn't clear. I know I can BUY damn near anything that I want/need.

What I'm looking for is ideas as to what I might be able to repurpose to meet my needs -- in much the same way that I've been, thus far, rescuing 12V supplies from servers.

But, that voltage/power level seems to sit outside the normal bounds of most other bits of kit. I'd thought of an electric welder but that wouldn't be regulated...

[Sparkey55]

Re: **BIG** 24V supply

Think about using two microwave oven transformers. Very large wattage capacities. Wire the 120 VAC primaries in series and phased correctly for a 240 VAC input. Cut out the HV windings and wind new secondaries for 12+ VAC then connect the secondary windings to give a 24+ VAC center tapped output. These transformer are most efficent at near the rated wattage. The rectifiers rated for 50 to 100 VAC at 200 AMPS or better. Use two 75K to 100K uf 25VDC or better filter caps. The regulator circuit is up to you to figure out.

[RJARRRPCGP]

Re: **BIG** 24V supply

I'm looking for ideas for (surplus) sources of really hefty (regulated) 24V supplies. Currently, I've been using modified 12V supplies from servers (disk spindle power) to get ~24V @70A.

I'd like to get closer to 100A (realizing I'll have to power it from a 220V branch circuit if I want to preserve the nominal safety margin)

"24V Magic" LOL.

[Curious.George]

Re: **BIG** 24V supply

The regulator circuit is up to you to figure out.

The regulation is entirely the problem -- otherwise it's just a big transformer and rectifier!

But, at ~2400W, even a 90% efficient power supply (which would be an effort to design well) needs to rid itself of ~240 watts (as heat).

[Stick 4 60W light bulbs in a box and see how quick it heats up!!]

[Sparkey55]

Re: **BIG** 24V supply

The regulation is entirely the problem -- otherwise it's just a big transformer and rectifier!

But, at ~2400W, even a 90% efficient power supply (which would be an effort to design well) needs to rid itself of ~240 watts (as heat).

[Stick 4 60W light bulbs in a box and see how quick it heats up!!]

Well, at an 2.4KW output power that you said that you are looking for that is NOT going to fit even in floor standing server cases with redundant power supplies. A brute force supply like I suggested or even a SMPS will most likely need a seperate housing to control the cooling be it air or liquid circulation.
Why do you need so much power for?

[Per Hansson]

Re: **BIG** 24V supply

It can be done using two 1200w server power supplies:

http://studio939.blogspot.se/2016/01...ystem-for.html

[brethin]

Re: **BIG** 24V supply

Maybe take a look at large UPS units.

[Curious.George]

Re: **BIG** 24V supply

It can be done using two 1200w server power supplies:

http://studio939.blogspot.se/2016/01...ystem-for.html

Hmmm... I'd have to run 220V to each supply in order to get 1200W @12V from each (2400W @24V from a series-wired pair). I've been doing the same thing but with 70A supplies, also from HP servers (which seems to be what the above power supply is rated at for 120V mains).



(I should check the nameplate rating on my units to see how they would fare at 220V line)

I have been unable to find anything larger -- even the power supplies in my Blade Server (dual 2200W?) don't cut it (cuz they primarily provide power to logic as the server only has ~30 2 inch SCA drives). The power supplies in my disk shelfs are also inadequate (~400W even though 15 3" spindles!)

The idea of soldering binding posts to the card-edge fingers (in the URL you provided) is interesting; on my first version, I opted to solder Anderson connectors on short pigtails for quick disconnect as well as their high current handling capability.

In the newer version, I have two sets of dual-redundant supplies wired in series (hopefully some current sharing as well as "hot spare" capability). The redundant supplies slide into a "power supply card cage" (also pilfered from the HP servers) so I can remove/replace one easily. The cages, in turn, slide into card-edge connectors that I've hard-wired to provide the desired output voltage (and control).

But, the real advantage to doing so is that it allowed me to skip all the steps of having to disassemble the power supplies (they are amazingly dense!) to isolate the board from the metal standoffs that tie it to "chassis". I.e., the two redundant supplies in "cage #1" share a chassis ground (from the cage itself and the skins of their individual power supplies). Similarly, the two redundant supplies in cage #2 share a chassis ground (of THEIR cage) -- but, this cage is electrically isolated from the other cage (so one "chassis" can float at +12V while the other is at "0V")

It looks like the DPS-1200FB is a pull from DL580's. I'll have folks keep an eye out for them on my behalf (if I'm not around when a server comes in, it gets stripped (for recycling) before I ever get a chance to see it (and rescue it's components).

Thanks!

(Meanwhile, time to check MY nameplates to see if I can just change the mains voltage to get what I need...)

[Curious.George]

Re: **BIG** 24V supply

Maybe take a look at large UPS units.

I don't understand; how would that help me? Modest UPS's (24V batteries) only provide a relative low charge current (at ~24V). Bigger UPS's (48V batteries) are the wrong voltage.

The online possible win would be a large on-line/double conversion UPS -- and, I'd have to hope the DC supply was nominally 24V (and not something higher/more efficient)

[Curious.George]

Re: **BIG** 24V supply

Well, at an 2.4KW output power that you said that you are looking for that is NOT going to fit even in floor standing server cases with redundant power supplies. A brute force supply like I suggested or even a SMPS will most likely need a seperate housing to control the cooling be it air or liquid circulation.

No. My Blade server has dual redundant 2200W supplies and it's just 8U. Modern power supplies can switch at such high frequencies that they can be much smaller (magnetics) and, thus, waste less heat. But, designing them is an art form in itself! (not my area of expertise).

Why do you need so much power for?

A prototype for a piece of industrial equipment that uses lots of fractional horsepower (primarily 1/3 and 1/2) motors for its actuators. I could put each on a separate power supply (adequate to handle the few hundred watts for THAT particular motor). But, that means an (oversized) power supply for each motor -- even when a motor is largely quiescent.

By putting them on a single LARGE (shared) power supply, I can track total instantaneous load on the supply (in software) to know when I can add/shed loads and live with a lot smaller power plant.

Appearances are important. Put a dozen power supplies in a box and folks start to wonder what the hell they are getting into! ("Surely we'll never be able to manufacture or support something like this, economically!")

It also makes it easier to handle "errors" -- if THE power supply shuts down, you know it immediately. OTOH, if motor #4's power supply shuts down, you wouldn't necessarily see it as soon (you'd have to explicitly monitor the output of EVERY individual supply instead of just the one)

[goontron]

Re: **BIG** 24V supply

Speaking as someone who has done proto: If this is for anything that will have a decent production run, it may be more cost effective to implement OTS than to spend the time to salvage...

[Curious.George]

Re: **BIG** 24V supply

Speaking as someone who has done proto: If this is for anything that will have a decent production run, it may be more cost effective to implement OTS than to spend the time to salvage...

I won't be doing any of the production design. My role is typically to build "proof of concept" prototypes (to prove an idea is workable) and "patent proofs" (to document patent claims).

For PoC, you typically are trying to do as little as possible to convince folks that an idea could be fleshed out and actually work -- no "big potential problems" that are likely to come up (you address those, specifically, in the prototype). As no one is sure it's gonna fly, money is usually tight and no one will ever see the prototype after the dog-and-pony for management.

For patent proofs, a similar set of issues apply -- but for different reasons. It's "the first" device to make the claims laid out in the patent application so it isn't expected to be "polished".

In each case, design engineering and/or manufacturing engineering move the "product" forward into production.

I've built prototypes (cases) out of wood, in the past. Other prototypes were pieces of perf-board hand wired and mounted ON a wooden plank. You expect the "audience" to be able to imagine a (high priced) Industrial Engineer's rendering of the product into a "work of art"

using salvage power supplies gets me to check off the "powersupply" box much quicker than if I have to dig through catalogues, check lead times, get a purchase order cut and approved, wait for shipping, etc.

[Curious.George]

Re: **BIG** 24V supply

(Meanwhile, time to check MY nameplates to see if I can just change the mains voltage to get what I need...)

I can apparently get 82A if I move to 220V mains. Not quite what I was hoping for, but an improvement, nonetheless.

I will start looking more carefully at the power supply "pulls" and see if I can find some that will buy me more margin at 220 (instead of using a 220V circuit as a pair of 110's as is more customary, here)