RAMPS 1.3 Powered from PC Power Supply

Hi all,

I have a spare modular 500W PSU I planned to use for powering my RAMPS 1.3 board; but I am confused as how to many yellow and black connections I need from the PSU to make the thing go:

I have just tried it with one pair of yellow and one black cables connected to each of the 2 power ports on this board (as I need a heated print-bed), but after roughly 30 seconds of connectivity a couple of the wires got REALLY hot so I turned it all off. (I noticed also that the PSU fan started to really speed up so obviously it was drawing quite a lot of current!)

How many should I use to stop this from happening again?

The PSU has the following connectors (with yellow and black wire counts listed):

1 x 20 pin ATX - for power on etc. (1 yellow)
1 x 4 pin (2 yellow, 2 black)
5 x 4 pin molex (1 yellow, 1 red, 2 black)

If my maths is correct this gives me a total of 8 +12v lines? I just read elsewhere in this forum that I should aim to split these equally between the 11A & 5A sockets on the board (saving one intact molex connector for a cooling fan?) - is this roughly correct?

Thanks,

Alex.

I have a 500 W ATX power supply: [julianh72.blogspot.com.au]

I just joined almost all of the yellows together, almost all all of the reds, almost all of the blacks, etc, and then run just one length of 20 A two-wire lead from the binding posts on my power supply to the terminals on the RAMPS. I just jumper the 5 A and 11 A terminals together with a short length of 20 A wire, so they are both fed from the same 20 A lead.

I kept one each of yellow, red, black, etc free, taped up and bundled up out of the back of the power supply box, in case I ever wanted to patch in any other circuitry, but so far I haven't accessed them, and I can't think of any reason that I will.

---

Follow my Mendel Prusa build here: [julianh72.blogspot.com]

It is also a good idea to load the 5V rail with about a 5R restistor if you are using a PC power supply they need a load on the 5V or 3.3V rails in order to get a stable 12V from the 12V rail the black and yellow cables are all joined up inside the P.S.

rcs Wrote:
-------------------------------------------------------
> It is also a good idea to load the 5V rail with
> about a 5R restistor if you are using a PC power
> supply they need a load on the 5V or 3.3V rails in
> order to get a stable 12V from the 12V rail the
> black and yellow cables are all joined up inside
> the P.S.
Yes, I should have mentioned that I have a permanent load on the 5V line - details are all in my blog entry, or on: [reprap.org]

---

Follow my Mendel Prusa build here: [julianh72.blogspot.com]

Would simply having a pc fan on the (+5V?) Red rail be sufficient to provide the load you mention?

No a pc fan will not draw enough current to load the 5V rail, a 5R resistor will draw about 1A and will get hot, I bolt mine to the side of the PC power supply to act as a heat sink.

I use a 10 ohm / 10 watt power-resistor on the 5 V circuit, tied to the metal chassis to act as a heat sink. Some people use something like a 10 W / 12 V light bulb, which can act as both a load on the 5 V rail and as a visual indicator that it is powered on. (I wasn't keen on the light-bulb idea, because if / when the bulb blows, your whole power supply will probably shut down under load, because of the loss of load on the 5 V line.)

---

Follow my Mendel Prusa build here: [julianh72.blogspot.com]

Perhaps you all can answer a question I have then - on your PSU there should be a sticker stating what Amperage each rail is capable of, and also a minimum value.

I compared 2 of mine and the newer one has a minimum value on the +5V rail - as you guys have described. On another much older (10 years+) PSU it doesn't have a minimum value - and it's this PSU I've decided to use; the modular one I mentioned above would have been too much work to wire up, although now that I think about it, I would have only needed to attach the cables I require to power the thing which would have resulted in a much neater package.. Damnit!

When I get home later I'll take a bunch of photos for comparison.

Quote

I wasn't keen on the light-bulb idea, because if / when the bulb blows, your whole power supply will probably shut down under load, because of the loss of load on the 5 V line

I think the life time of a bulb is proportional to the inverse 12th power of voltage. So if say it was a 1000 hour 12V bulb, then run at 5V it will last 4000 years!

---

[www.hydraraptor.blogspot.com]

And still blow when you need it most. Incandescents suck like that.


AlexFielder1980 Wrote:

> I compared 2 of mine and the newer one has a
> minimum value on the +5V rail - as you guys have
> described. On another much older (10 years+) PSU
> it doesn't have a minimum value - and it's this
> PSU I've decided to use; the modular one I
> mentioned above would have been too much work to
> wire up, although now that I think about it, I
> would have only needed to attach the cables I
> require to power the thing which would have
> resulted in a much neater package.. Damnit!

You should probably go with the recent modular one. Ten years ago, PCs drew most of their power from the 5V rail rather than the 12V which is mostly loaded by modern hardware, and the power supplies were designed accordingly.

nophead Wrote:
-------------------------------------------------------
> I wasn't keen on the light-bulb idea, because if /
> when the bulb blows, your whole power supply will
> probably shut down under load, because of the loss
> of load on the 5 V line
>
> I think the life time of a bulb is proportional to
> the inverse 12th power of voltage. So if say it
> was a 1000 hour 12V bulb, then run at 5V it will
> last 4000 years!


So why do all my "1,000 hour" incandescent globes and "6,000 hour" compact fluorescent globes keep blowing in a few months.

---

Follow my Mendel Prusa build here: [julianh72.blogspot.com]

1000 hours is only 41 days, so if it is on for a significant time each day that is only a few months. If your supply voltage was 6% high the life would halve.

A lot of the compact ones have crap electronics that blow before the bulb itself. Probably inadequate surge protection or capacitors with a lower lifetime rating than the bulb. Electrolytic capacitors are only rated for 2000 or 5000 hours. To make them last longer you have to de-rate the temperature and voltage specifications, a lot of people forget that.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> A lot of the compact ones have crap electronics
> that blow before the bulb itself. Probably
> inadequate surge protection or capacitors with a
> lower lifetime rating than the bulb. Electrolytic
> capacitors are only rated for 2000 or 5000 hours.
> To make them last longer you have to de-rate the
> temperature and voltage specifications, a lot of
> people forget that.


At the risk of going WAAAAY off-topic here (too late? ), how can you rate a globe for 6,000 hours if its integrated electronics won't last that long?!

Maybe someone should tell the manaufacturers and the retailers that the compact fluorescent globes which are labelled as being good for 6,000 hours - aren't!

---

Follow my Mendel Prusa build here: [julianh72.blogspot.com]

You can make capacitors last longer by running them at less than their temperature spec, or lower than their voltage rating. That is what is meant be de-rating. There are formulas to apply. So a good manufacturer can make them last the time they claim.

---

[www.hydraraptor.blogspot.com]