SMPSU

Probably me.

The open frame ones are on ebay:

Theo

It probably was. ;-)

Yeah, I was looking at them earlier and on Amazon (for Prime delivery) but as it did manage to finish the print I have a bit more time to seek a replacement.

But in 12V form you mean?

And that's the thing. I've taken the original one out and it's top off and it looks like one of the caps has blown (top right of the group of

I think there was a fan in there (you can see the remains of the gloop on the lower side (on the pic) of the txfmr) but it failed and we took it out as the PSU ran cold (originally).

When I took it out earlier it was quite hot.

Where it is on the printer is open on all 4 sides and not in any way enclosed.

I wonder if a set of new caps and maybe add a second PSU to split the load might be a good repair / upgrade?

Cheers, T i m

p.s. From a quick Google it comes up as a 5V x 40A model but I'm guessing that's something they change in the factory (slightly different bits but the same mechanics / circuit)?

'Meanwell' PSU's seem to be pretty good quality and value. Lots on eBay.

That's an odd place to put the fan.

PA

If it's a 12V supply, you can use an ATX supply for that. The modern ones are double forward conversion, and the 12V section is separate from the 3.3V/5V section and the other sections don't need a load particularly (should have better cross-load characteristic than the old single-transformer ones).

ATX supplies can be very efficient, so less waste heat to remove.

If it's a 13.5V supply, then you're looking at some other form factor, like ham radio supplies (some of which are linear and lower RF noise, some of which are SMPS with a bit of hash on the wires). Now the efficiency can be poorer (like on the linear), and fan cooling is required for a good result. The SMPS ones will be more compact, because less heatsink is needed.

I think the very best ATX I've seen, was somewhere around

That's not an endorsement, but it shows what is possible. I'd still want some air movement through the poor thing.

There was a time, when some of these SMPS were only 60% efficient and the heat used to pour out of them, requiring a stout fan and associated noise. They've come a long way since then.

Paul

One +12v rail offering a lot of current as against two. That's how they started out until someone decided that so much demand on +12v meant that two +12v would be a better idea. Lower currents per rail. OK, but loading the two rails correctly didn't work out for some systems. So some have returned to a single, many ampere, rail.

I think that this really kicked off when all the PCIe connectors were loaded onto the same +12v rail of a two-rail PSU.

PA

Thanks Clive, they do seem to be a step above the basic offerings, or should do by the price of some of them. ;-)

The only one SP-200-12that looks like it's the same width as the existing is only 200W / 16A, and that's a bit less than the 20A of the original. However, the chances are the Meanwell is a real 16A versus the cheapo and given *that* lasted for a good 5+ years ... ?

We know the 3D printer heated bed draws ~10A (on 12V) but can also run on 24V and so would be nearer to 5A on 24V. If replacing the caps (ordered) doesn't sort the old PSU, I was thinking I could use a SP-200-12 for the 12V stuff and a 24V (10A indicated, assuming it's only 7) PSU for just the Bed Heater.

Originally is was being driven by a Mosfet on the main board but the FET got very hot so I used the FET to drive a 12V 40A automotive relay and that seems to have worked ok (bang-bang). That means I could easily set the heated bed up for 24V and use the same relay (still driven by 12V on the coil).

I tested all 4 caps and they all came out ok (on my little component tester) but I'm wondering if it might be different under real load / voltage (or why would one cap be bulged and the PSU not working properly)?

Cheers, T i m

Your tests were when it was cool? Closest to the toroid? Missing the fan?

PA

I do have one I've taken the main output wires from and just left the power on lines, with the idea of using it on the printer. And that would be a bonus as the setup can actively drive a PSU that powers the bed and extruder heaters and so shut them down in the event of a fault / overheat.

OK.

Noted.

The old linear supply was just so I could finish the print job for my mate Paul. Even though it did get fairly warm after the 90 min print, the duty cycle was probably 1:10 (after the bed had initially got to temperature).

Wow. Shame it's EOL. ;-(

Quite, and the printer isn't exactly quiet with 3 other cooling fans running. A 120mm thermally controlled fan would probably not be noticed. ;-)

Yup, many of my PC's are passively cooled, as has my WHS been for about 10 years now. ;-)

Do you know what they mean by 'single rail'?

I need to look back of the spec of this console PSU as I think that was pretty high power and could be switched remotely.

Cheers, T i m

Agreed.

Yeah, that ferrite was quite warm and one cap bulging but the caps tested out similarly, eg, VLoss ~1.8%, Cap ~1150 uF and ESR ~ .1?

Cheers, T i m

On ATX, they dabbled with a few ideas.

You could get some idea, from the length of some of the ATX PSUs. Two dimensions are standard, the third dimension (length) is variable.

A few supplies, really did have independent outputs. One PSU was really long (would have bumped into the DVD drive), and it had four transformers for 12V. There would be ATX12V1 and 12V2, and a couple rails for PCI Express connectors. The hardest part of using those, was finding the wiring diagram, so you could figure out how to load the thing (not overload any one transformer.

But the idea of doing that, quickly died, and they don't tend to do that any more.

Now, there's one huge transformer with 12V on it. What should happen, is sections of the +12V loom, are protected by current limiters. The main transformer might be 60A, but there will be three current limiters installed, each connected to the main output. The output wires then feed some part of the system (like, just the motherboard). The limiter might not be 20A, it might be a bit higher. It's there, to ensure a 20A rated wire, doesn't have 60+ amps flowing through it :-)

It's also possible they're selling some "let 'er fry" ones, where there is the 60A transformer, but only the 60A limit is checked, and maybe it can make some 12V wiring glow with the heat. Generally, the documentation is negligent, in that it does not give sufficient detail about the sections of the loom and what limits are in place. The marketing people want the message "it's got 60 amps, kids", to get through, and any other level of detail would be a distraction. Telling people there was a current limiter for sound reasons, would only elicit a customer response of "but I paid for 60 amps!" :-)

Since all the 12V current comes from the same transformer winding, joining some 1x4 Molex to some other pins, should not cause loop currents to flow. Loop currents would be a consideration on the supply with the four transformers.

Paul

Yup ...

Not when they were off the PCB and I was testing them. ;-)

They may have been. ;-)

But all the above are why I have already ordered some replacements and

I just was ready to see the one with the bulge show very different results (even cold) but it didn't?

Cheers, T i m

I'd check any capacitors in the unit, since one usually finds that depending on the design, these can be critical for its efficient operation. Brian

I'd check the rectifier bridge. One of the four diodes blown and it becomes a half-wave rectifier.

Interesting, thanks both.

I have 'converted' a few PSU's to 12V (mainly) PSUs over the years, the last for a 12V electrolysis tank (that worked very well).

If I already had a know suitable PC PSU that wasn't in use that offered at least 20A on the (single) 12V rail I might consider converting it for use with my 3D printer and having it free standing beside it (as it wouldn't fit underneath and I can't raise it easily as the filament spools sit on top and there is restricted height above it). I could mount it on the frame where it would be out of the way of the bed / extruder but it would end up requiring all the (heavy current) cables being longer (and potentially restricting access for maintenance of the printer etc).

So, a 'decent' replacement of the main 12V PSU to run the printer, controller and extruder and a second, probably 24V PSU for the bed heater. Higher voltage = lower current and so even easier to extend slightly (to the other side under the print bed).

Cheers, T i m

p.s. The whole thing is both quite heavy and has fairly fragile components on it (axis position sensors) meaning that you have to be fairly careful when moving it, especially re catching any of sticky-out-bits on anything or putting any stress / strain in the wrong places. That said, the whole design is very rigid (one of the key criteria when I was advising what printer we should build back then) and unlike many of the simpler 3D printer designs, requires no re-calibration after being moved (see above for cautions when moving it) or even if it's only sitting on two feet etc.

Understood and I (just) checked it (in circuit, diode test on my DMM) and it seems ok (10 ways). ;-(

I would have preferred to find something, rather than just 'hoping' it was a cap drying out etc?

That said. We have probably been working it fairly hard over a good few years, given it's 20A capacity may be fairly optimistic?

The heated bed draws ~10A, all be it with only a 1:5 duty cycle once up to temp (60degC), the extruder is either 20 or 40W (I can't remember now, assume 40) so another ~3.5A (worst case whilst at 100% when warming then maybe 50:50 once running (PWM)), 5 x stepper motors running at about 1A each plus whatever the Arduino Mega, the RAMPS interface board + display and 3 small fans may draw, lets guess another couple of amps:

So that's worst case 10+3.5+5+2= 20.5A (246W). ;-( [1]

ITRW, the extruder heats up pretty quickly, the bed takes a bit longer and only *after* that was done would the motors initialise ('holding' current etc), so I don't think we would ever see that 20.5A as such and certainly not for a prolonged period. And it has run for many years now and even after the PSU fan went noisy and failed, we checked the PSU temp and it was never much past warm?

That said, I think I'd prefer some margin and hence why a single 12V /

Cheers, T i m

There was one that had a budging top Brian that seems to test out the same as the others on my little automatic component tester but because of the bulge and the PSU not working properly, I've ordered Panasonic replacements and will test it again once done.

Luckily I also have a load tester that I use for testing batteries that I can put on it once the caps are replaced and test it at it's twenty amp rated load.

Cheers, T i m

I replaced the 4 output caps because of the blown one and sorta as predicted it doesn't seem to have fixed it.

I can go round all the passive components and check the smaller transistors but there are a couple of big devices underneath (TO220 but bigger style cases), one 3 pin (FET / Transistor) and a 5 legged one (SMPSU controller device)?

I can probably check for the 400V or so on the input caps (and any AC voltage present, suggesting those caps had gone dry?) but there is little point spending too much time (or money) on components, other than for the fun of possibly repairing it.

After changing the caps and powering it up with no load, I saw a 'weak' (fluctuating) 12V on the DMM but it wouldn't power up with a 5A load attached (it crowbarred).

Any thoughts?

Cheers, T i m

There looks to be a preset potentiometer, bottom right, in your photo. Might be worth noting its current position, rotating it from end to end of its range a few times, then restoring it to the original position. A spurt of switch cleaner before the above might help, too.

PA