Power Supply failure  

Re: Power Supply failure

One note on the meanwell 350 I got. It works great (just have it wired up directly to the harness).

However, it has a fan. It's a heat triggered fan, and about once every 30 seconds or so it fires up for 1-2 seconds. It's not that loud, but the contrast is annoying... I was hoping that since it was overpowered by 1.5x it would run cool enough to not trigger the fan.

I'm printing a case for it anyway (just because) while I wait for my replacement supply. I want to check a few things:

Then I'll probably switch back to the replacement stock supply and keep the 350 watt one for projects.

Be aware that the supply is not necessarily overpowered for your application. You are only using 2 of 3 channels and depending how they deigned it, you could have *less* amps available per channel. (14/3 < 10/2). You would have to open it up and see what they did. I also wouldn't assume it runs cooler.

I don't believe the power output terminals are designed as channels in respect that they restrict current draw from each terminal screw other than a max load of each terminal screw. If you plan on taking full advantage of having all the supply rated 14A (350W) power available then consider having all the three positive terminals tied to each other. This is because each terminal screw can only support supplying a limited amount of current before it may overheat (6 amps or so is my guess). You wouldn't want to pull the full 14 amps thru only one terminal screw on each the positive and the ground (-) buss. Not saying that the MK3 will pull 14 amps. So you could create a 12awg to 16awg wire jumper to connect up the third buss terminal to one of the other same two terminals on the positive buss and the same goes for the ground (-) buss. This way the current will be better distributed between each of the positive and ground DC connections. The PS circuit board most likely has large copper buss traces to equally distribute power to the terminal block. In some cases there are metal jumper plates supplied that can be screwed down under two adjacent terminal block screws essentially doing the same thing.

Re: Power Supply failure

One note on the meanwell 350 I got. It works great (just have it wired up directly to the harness).

However, it has a fan. It's a heat triggered fan, and about once every 30 seconds or so it fires up for 1-2 seconds. It's not that loud, but the contrast is annoying... I was hoping that since it was overpowered by 1.5x it would run cool enough to not trigger the fan.

I'm printing a case for it anyway (just because) while I wait for my replacement supply. I want to check a few things:

Then I'll probably switch back to the replacement stock supply and keep the 350 watt one for projects.

Be aware that the supply is not necessarily overpowered for your application. You are only using 2 of 3 channels and depending how they deigned it, you could have *less* amps available per channel. (14/3 < 10/2). You would have to open it up and see what they did. I also wouldn't assume it runs cooler.

I don't believe the power output terminals are designed as channels in respect that they restrict current draw from each terminal screw other than a max load of each terminal screw. If you plan on taking full advantage of having all the supply rated 14A (350W) power available then consider having all the three positive terminals tied to each other. This is because each terminal screw can only support supplying a limited amount of current before it may overheat (6 amps or so is my guess). You wouldn't want to pull the full 14 amps thru only one terminal screw on each the positive and the ground (-) buss. Not saying that the MK3 will pull 14 amps. So you could create a 12awg to 16awg wire jumper to connect up the third buss terminal to one of the other same two terminals on the positive buss and the same goes for the ground (-) buss. This way the current will be better distributed between each of the positive and ground DC connections. The PS circuit board most likely has large copper buss traces to equally distribute power to the terminal block. In some cases there are metal jumper plates supplied that can be screwed down under two adjacent terminal block screws essentially doing the same thing.

I should have been more clear. I was just cautioning against making assumptions. I gave a made-up example. You are likely correct. I don’t think my fail was due to pulling too many amps, just a bad component. My replacement is supposed to arrive today. Prusa wants the fail back so they can do post-mortem.

Re: Power Supply failure

My status:

I've got (most of) a holder for the MeanWell 350. Base and base-plate nicely done, test fit in PLA. I will switch to PETG for production but PLA prints faster for prototyping. I had to shift it higher several times to make room for the power-panic board in the base.

The connection to the Y-frame is still a sore-point. The lower PSU hole is between the two holes in the y-frame, and the upper one is significantly above the top hole. I'm not convinced I'm going to get a solid connection with my 'shim' idea unless I drill it out of a 1/4" piece of steel (one option). Another quick option is to drill/tap the side of the PSU case where the y-frame holes are, but that feels like an awful hack. I'll have to disassemble the PSU to make sure that can be safely done (no components near those spots), and that the PSU case is sturdy enough (I don't know if the existing holes are reinforced in some way).

I am also going to open the MeanWell PSU up to see if the fan is 12v or 24v. If it's 12v, I can swap it with a 80mm noctua I have. If it's 24v, no such luck.

Yes, I realize opening the MeanWell will void it's warranty. Yes, I'll be careful.

My new Prusa PSU is on it's way. That is the fallback if I'm not satisfied with my MeanWell 350 attach.

Experiments:
- see if the clicking is the same on the new Prusa PSU
- compare to clicking (if any) on the Meanwell
- measure voltage droop on both supplies at full bed/nozzle heat

Re: Power Supply failure

I'd like to use a second PSU just for the heatbed to eliminate clicking until the new PWM scheme is released. I can just hook up 24v to the middle two power terminals on the EINSY right? The ones directly below the 15a fuse?

Re: Power Supply failure

That may work, but I'd be nervous having two different supplies with the same board. You'd potentially have a ground loop (variation in ground level higher than desired). Probably not, but something to be aware of.

What second supply are you using? Why not get a big enough 'non-clicking' supply and feed the whole board off of it?

Re: Power Supply failure

Thanks--I'll be sure to bond their grounds if I do use two PSUs.

What second supply are you using? Why not get a big enough 'non-clicking' supply and feed the whole board off of it?

Good question. My understanding was that the stock PSU had extra big caps in it for power panic functionality (which I want to retain). So if I used the other PSU for the whole machine I fear I lose that feature. Might be worth testing that assumption though.

Re: Power Supply failure

I don't believe they used a special power supply, although I suppose its possible. I suspect they used a standard, reliable 24v supply, but verified that it had the ability to maintain voltage to keep the Einsy/steppers (only) running for a few seconds (enough to lift the hot end a bit).

I'm using the 350W MeanWell for the whole system right now waiting for my replacement supply. I have it wired to mains at this point.

At some point this weekend, I may hook the power panic up to it and see how it works.

Re: Power Supply failure

Paul, here’s what I found on closer look. My fuse was blown too.

8FA6DCBB-69AF-4142-80EF-75CF8AB394BC.jpeg

same resistor blown here. for anyone looking for it, top right (if facing the plug side of the psu) big capacitor, below is a heatsink, bottom left of the heatsink, green gold silver black brown resistor. might not be the only blown part but it is blown.

for anyone wondering, prusa ships ups 2nd day air for replacement parts, only delay is clearing customs on their end which might add a day or so. However, ups over there appears to work weekends since a part sent on friday got to me on a monday

printing time on mine is probably somewhere over 200 hours since 12/18 might be as high as 300-400 i had been printing abs (255/100) for nearly 24 hours straight when it died (did not hear anything as i had on NC headphones, fuse was blown, replaced it, big loud pop and completely dead, replaced power supply, working great.

Re: Power Supply failure

They are pretty quick for me as well. I talked to them on Thursday, shipped Monday, arrives today. I'd be more upset if I hadn't picked up a spare supply on Amazon (ordered Wednesday, arrived last Friday).

Re: Power Supply failure

For me, the failed part appeared to be a small black 4-pin DIP part spanning the two boards, labelled U3. Melted down pretty thoroughly!

Re: Power Supply failure

Replacement Prusa PSU installed. It clicks, just like the old one. Darn it.

The Meanwell 350 worked great, and it makes a nice backup PSU, but it has two issues that mean I won't stick with it:

So, MeanWell stays as a backup.

I'll be looking for a alternative supply that doesn't click, though. If I can find one with the correct holes and no fan, I'll grab one.

Re: Power Supply failure

The Meanwell 240W and the Delta 240W 24v supply without the fan I posted earlier might fit better on the frame. Delta has been around for years and always seemed like a good supply. Maybe they also make a 300+ watt version.

Delta 240 watts without a fan: https://www.digikey.com/product-detail/en/delta-electronics/PMF-24V240WCGB/1145-1101-ND/4833510

Delta 320 watts with a fan: http://www.deltapsu.com/products/download/Datasheet/PMF-24V320WCGB

Thanks for the feedback! Enjoy.

Re: Power Supply failure

Replacement Prusa PSU installed. It clicks, just like the old one. Darn it.

I spoke too soon. Replacement supply isn't reliable. I made it through self-test and the beginning of a PETG print (heat up, first 4 layers or so) and it reset.

Debugging, it appears that when the PSU is warm the voltage sometimes droops to 21V during self-test nozzle heat. Very excessive for a 20-30W draw on a 240W supply (my MeanWell never drops >0.15V, much less 3V). During bed heating (self-test or preheat), it will usually immediately drop to <5v, the Einsy reboots, and it pops back up to 24V again (since there is no load). [Sometimes it will make it partway through the bed warming before it reboots]

Working with Prusa. Backup supply reinstalled. It's been rock solid for a week so at least I'll be able to get back to printing. It seems fine, passed self-test and voltage never waivered. Reproduced all the tests that failed with the flakey supply with mine, no issues, so I have reason to believe it is the PSU, not something else wrong with the printer.

I need to think about a quick-connect power-supply system. The more I mess in the Einsy box connecting/disconnecting power, the more likely I'm going to bump or bend something.

Re: Power Supply failure

Can you share the case you made for the meanwell?

Re: Power Supply failure

I'll find a way to post it. It's in SCAD, and it's a work in progress. It fits around the MeanWell 350, has a tab for the y-rail, and room for the switch/plug/power-panic-board, but I don't have a way to secure it to the vertical frame yet. Right now, I push the PSU in there and leave it sitting on the desk next to the printer.

The SCAD should be useful for anybody else trying to do their own design though. I'll see if I can figure out how to github it.

Since it looks like I'll be another week with the MeanWell, I'm likely going to try a different approach: put in a sturdier printed brace between the PSU screws in the vertical frame and the y-rail. I'll then attach the PSU to that rail.

The current file is here. First time posting publicly on github, so let me know if you can get it.

https://github.com/paulmeyer/3dprint_public.git

Re: Power Supply failure

FYI, if you plan to mount your PSU elsewhere, here is a reinforcement brace to use in its place.

https://www.thingiverse.com/thing:2772947

Re: Power Supply failure

I built my own brace and attached the PSU to that. This is a mockup out of low-infill fast-PLA so I can check the fit. I need to extend the tabs for the rail a bit and shorten the back where it hits the back plate. I'm also going to thin it down 7mm, I overdid it the first time.

It's got openings to hit five of the 4mm bolt holes on the PSU, and it braces between the frame and rail.

The only downside is that the position of the terminals on the PSU, and the fact that the PSU is shifted left about 2cm and up a few mm means that the original wire harness doesn't fit. I had to grab some 14 gauge wire and make my own.

However, at least I'll be safely printing again until my second replacement supply comes. I confirmed my second fail on Friday after Prusa support had gone home, apparently, so no contact yet.

Re: Power Supply failure

I would be curious to know if most of the power supply failures have been on 110v or 230v ?
Maybe there is something in that, don’t know.

I’ve not yet received my printer (week 44) so kind of hoping that Prusa has a bead on the root cause of the failures,
and ships me a good one.

Jason.

Re: Power Supply failure

I’ve not yet received my printer so kind of hoping that Prusa has a bead on the root cause of the failures,
and ships me a good one.

Jason.

You and me both this PSU issue is my main reason to worry right now, seems like the most detrimental to operation of the printer bug ive seen so far, hard to print anything without a PSU.

Re: Power Supply failure

I’ve not yet received my printer so kind of hoping that Prusa has a bead on the root cause of the failures,
and ships me a good one.

Jason.

You and me both this PSU issue is my main reason to worry right now, seems like the most detrimental to operation of the printer bug ive seen so far, hard to print anything without a PSU.

Wouldn't worry too much, how many blown PSU vs Units shipped? Unfortunately it's the nature of the beast, I've had a PSU's blow on me on another printer. These Power Supplies were never designed for 3D Printers. If anything the Prusa PSU, has a an advantage as it appears to be a customized part for their load requirements vs a generic LED PSU which some aren't always equipped for the load. Not to mention (unlike most Chinese printers), there is a warranty on the part.