PSU Cutting out...

Hello All

I don't seem to be having much luck here...
I got to the "Orthogonal Axis Compensation Test" stage where it did a few layers and then the PSU cut out and only the 1 LED was lit on the ATX power pcb.
I had to unplug the mains and wait for a few seconds before plugging it all back in where it all started to work again until I told the printer to home where it then cut out again.

So it sounds like I may have an iffy PSU or the printer is taking too much current from the PSU.
I have also noticed that when the heaters are on the little fan slows down a little, I don't know if anyone else gets this as well.

So the main issues that I have so far is I have a short on AD7 to Ground on the Duet board and now the PSU, what else can go wrong

PS the 10mm bearing change didn't help much, the whole hot end still sags and needs to be tilted by another 0.8mm

Paul

---

RS Ormerod No 436

Bad luck Paul, I had the same problem, see this post http://forums.reprap.org/read.php?340,287773

I solved mine with second hand PSU from a scrapped PC. I don't think the Alpine PSUs are the best quality...

Tim

---

RS Ormerod #472

Quote
TMD_RS436
So it sounds like I may have an iffy PSU or the printer is taking too much current from the PSU.
I have also noticed that when the heaters are on the little fan slows down a little, I don't know if anyone else gets this as well.

Sounds to me that you need to ask RRP for a replacement PSU. But the fan slowdown when the bed heater comes on is normal, the 12V supply on cheap ATX PSUs isn't well regulated (nor does it needs to be in this application).

Quote
TMD_RS436
PS the 10mm bearing change didn't help much, the whole hot end still sags and needs to be tilted by another 0.8mm

Yes, my unit really needs an 11mm bearing too, both with the x-carriage part I received with the kit and with my own replacement printed part. In both cases, to solve it I had to elongate the slot and the nut trap as well as fit a 10mm bearing. I think RRP needs to adjust the design of that part.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Not sure about the PSU, it could be a fault or a short. But you could try increasing the load on the PSU 5V and 3.3V supplies as the built in resistors are just below the minimum load spec and just could be the issue as the PSU could turn off it these are sensed as being too low a load current.

for the hot end sag, make sure the slot where the bearing screws into the nut is clear and you can slide the nut fully in for adjustment.
Also check the bolts that hold the hot assembly to the sliding part are in a good position as this is a socket head bolt in a countersunk recess in the printed part. Washers may help. Have a good look round the assembly to see what is parallel with the arm to try and be sure of the specific area causing the problem. Post a photo if you would like us to examine.

---

Ormerod #007 (shaken but not stirred!)

Hello Tim

Quote
tim_h
Bad luck Paul, I had the same problem, see this post http://forums.reprap.org/read.php?340,287773

I solved mine with second hand PSU from a scrapped PC. I don't think the Alpine PSUs are the best quality...

Tim

Well the PSU supplied to me was made by Ace and says the +12V line can handle up to 24A Max.

I haven't got any spare PSU laying about.

Paul

---

RS Ormerod No 436

Hello dc

Quote
dc42

Quote
TMD_RS436
So it sounds like I may have an iffy PSU or the printer is taking too much current from the PSU.
I have also noticed that when the heaters are on the little fan slows down a little, I don't know if anyone else gets this as well.

Sounds to me that you need to ask RRP for a replacement PSU. But the fan slowdown when the bed heater comes on is normal, the 12V supply on cheap ATX PSUs isn't well regulated (nor does it needs to be in this application).

Yeah I thought it would be something like that, the fan doesn't slow down that much, but enough to hear the difference.
Like I said to Tim, its only rated at a max of 24A on the +12V line, not knowing the max current the printer takes with everything going so I cannot say if 24A is enough.

Quote
dc42

Quote
TMD_RS436
PS the 10mm bearing change didn't help much, the whole hot end still sags and needs to be tilted by another 0.8mm

Yes, my unit really needs an 11mm bearing too, both with the x-carriage part I received with the kit and with my own replacement printed part. In both cases, to solve it I had to elongate the slot and the nut trap as well as fit a 10mm bearing. I think RRP needs to adjust the design of that part.

Well I don't think you can get a 11mm with an inner size of 3mm, I checked, so the first thing I was going to do was print a new version with the hole moved closer in, or maybe an adjuster etc, but I didn't even make it passed the 4th layer of the Axis compensation print.
One thing that I was tempted on trying was using my dremel on a blank 0.8 mm PCB and cut them into strips and glue them to the back giving it the extra 0.8 mm it needs.

Paul

---

RS Ormerod No 436

Quote
TMD_RS436
Well the PSU supplied to me was made by Ace and says the +12V line can handle up to 24A Max.

Interesting, the Alpine one says 32A max. But 24A should be more than enough,

I'm wondering whether the sudden reduction in load when the bed heater turns off causes the output of these PSUs to momentarily rise to a point at which an overvoltage circuit cuts in and shuts the supply down. If that's the case, then adding a few thousand uF on the 12v line at the ATX board or the Duet input might help.

Edited 1 time(s). Last edit at 01/07/2014 04:11PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Hello Treth

Quote
Treth
Not sure about the PSU, it could be a fault or a short. But you could try increasing the load on the PSU 5V and 3.3V supplies as the built in resistors are just below the minimum load spec and just could be the issue as the PSU could turn off it these are sensed as being too low a load current.

How much load are we talking here on the +5 V line ?

Quote
Treth
for the hot end sag, make sure the slot where the bearing screws into the nut is clear and you can slide the nut fully in for adjustment.

Do you know I had forgotten about that, I didn't even notice that I could slid it, even when I changed that bearing to a 10mm one
Just removed the bearing and cleaned up the hole so it now has less of a sag, I may need to make that hole a tad bit longer towards the nozzle and then recheck.

Quote
Treth
Also check the bolts that hold the hot assembly to the sliding part are in a good position as this is a socket head bolt in a countersunk recess in the printed part. Washers may help. Have a good look round the assembly to see what is parallel with the arm to try and be sure of the specific area causing the problem. Post a photo if you would like us to examine.

Will take a look at it tomorrow and might take a few pictures also.

Paul

---

RS Ormerod No 436

Hello dc

Quote
dc42

Quote
TMD_RS436
Well the PSU supplied to me was made by Ace and says the +12V line can handle up to 24A Max.

Interesting, the Alpine one says 32A max. But 24A should be more than enough,

Well I didn't know the power requirements of printer so I wouldn't know.

Quote
dc42
I'm wondering whether the sudden reduction in load when the bed heater turns off causes the output of these PSUs to momentarily rise to a point at which an overvoltage circuit cuts in and shuts the supply down. If that's the case, then adding a few thousand uF on the 12v line at the ATX board or the Duet input might help.

Yeah, but that wouldn't answer the issue where it cut out when I reconnected it after 5 seconds after last cut out and I only told it to home itself.

As for adding components to the duet, I didn't really want to do that just yet, due to I may have to return it for a replacement due to that short of AD7 to ground, but I am still waiting to hear from Ian.
But the ATX Power board I could add some on, I have a bag full of them somewhere.

Paul

---

RS Ormerod No 436

Quote
TMD_RS436
Well I didn't know the power requirements of printer so I wouldn't know.

These are my attempts at worst-case figures:

Bed heater: 10A
Extruder heater: 4A
4 x stepper motors x 2 windings @ 800mA: 6.4A peak (average should be much lower)
Fan 0.2A
5V and 3.3V supplies: 0.2A

Total less than 21A peak, unless I have missed something important.

Quote
TMD_RS436

Quote
dc42
I'm wondering whether the sudden reduction in load when the bed heater turns off causes the output of these PSUs to momentarily rise to a point at which an overvoltage circuit cuts in and shuts the supply down. If that's the case, then adding a few thousand uF on the 12v line at the ATX board or the Duet input might help.

Yeah, but that wouldn't answer the issue where it cut out when I reconnected it after 5 seconds after last cut out and I only told it to home itself.

Good point, with the heaters off, the current will be very much lower.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Its been mentioned before on other threads but a hacksaw blade (break a bit off the end to fit) and doubleside taped to the x-axis works a treat, still using 9mm bearing.

---

Regards

Chris H

I've had a few reports of PSUs cutting out, though they reset after a couple of minutes. At the moment, we're a bit mystified what might be causing this, as the PSUs (despite other reported problems) should cope with the current draw. We could just send out replacement PSUs, but it would be better to find the cause. I have spoken to the boss, and wonder if one of the reasons this might be happening is because of the higher rated resistors in the PSU power PCB, as we talked about in this thread: [forums.reprap.org]

Could this be the cause?

Ian
RepRapPro tech support

Ian

It is possible that not achieving the specified minimum load could cause shutdown, I had thought that it was more to help with 12 v regulation.

RRP supply two brands of ATX psu , it seems the one shutting down is an ACE in this case , are all such of this brand? If so they could be replaced with Alpine which would solve the problem.

Rory

It seems this also happens with Alpine units, see other thread.

It is worth checking the JP9 link on the ATX board is making good contact.

Edited 2 time(s). Last edit at 01/08/2014 02:45AM by Rory166.

The minimum load spec should certainly be checked on two counts.

1. If the minimum current is not enough they will reset and turn off, this used to be on the 5V rail, but now appears to be a 3V3 rail requirement.

2. The 5V rail voltage determined the voltage on all the other rails, not sure how this works with the 3.3V rail and 5V rail combination. The output voltage would ring on the supply would ring if it operated below this minimum load and transfer through the transformer to all rails. I believe they introduced the cutting out of the supply if the minimum load is not present as this ringing generates high voltage stress internally in the regulator and could destroy the power supply (and load).

In my earlier post I suggested increasing this current and perhaps the easiest place in on one of the free power leads. The connector could be cut off and replaced by an electricians choc block connector allowing the additional load to be placed by any user and there are no mods near the Ormerod to worry about.

---

Ormerod #007 (shaken but not stirred!)

Well I thought I found a good PSU laying around in the back of the cupboard was even a 700W PSU with 2 separate 12V rails each supporting up to 25A so I plugged it into the mains only (not in the printer) and the power trips, so sadly its a dud PSU which was probably why it was at the back of the cupboard so no joy there.

On a side note the sagging issue seems to be better now that I filed out the bearing slot a bit more and moved the bearing as far as needed to get it level.
Not calibrate it all up again.

Paul

---

RS Ormerod No 436

Hello All

I just though I would try and print something that wouldn't take over an hour to do which was the snowman and funny enough the PSU didn't cut out this time.

I did have an issue the first time I tried to print the snowman where the filament wouldn't stick to the bed so I stopped it and reset Z and the second time it worked.



It looks ok to me, but I have nothing to compare it with.

Do we really need to print out the Axis Compensation parts?
The reason that I am asking is it tells me that it may take around just under 3 hours to complete according to the following:
Estimated duration (pessimistic): 356 layers, 2:43:33
That's on average 45 seconds per layer.
And just under 3 hours I don't want for it to get halfway and the PSU to cut out.

I do understand that printing most still will take ages to complete, but at this stage with the PSU I don't really want to try large prints until the PSU issue is resolved.
Hopefully changing the load resistors to a slightly lower values may help, just need to find some time to work out the values.

Paul

---

RS Ormerod No 436

Paul, well done. As a first print it looks absolutely fine. My first attempts were a mess of stringy bits and 2 layer thick prints where the z axis was too low the duet board kept stalling!

I don't think the axis compensation parts are not essential, they are only really important when you come to do big prints with dimensions that need to be accurate.

---

RS Ormerod #472

Power supplies can be quite tricky sometimes. That goes for many other devices as well. I bought myself a 40A plasma cutter yesterday. Fairly cheap as well. The spec clearly says that you can cut at full power for 5 minutes and then you must let it rest and cool down for 5 minutes. The same for my heavy duty 90W Scandia gun-type soldering iron. I can use it for 90 seconds at full power and then it must cool down. My small HP printer spec-ed that it could print 8 pages a minute but also stated that more 1500 pages per month equals abuse. It all boils down to duty cycle. You have access to full power until a certain point and then it needs to cool down.

Now PC PSU designers, especially for desktop use, know that you will never actually use your PSU at 100% all the time. Just as the car manufacturers knows that you won't go 160 miles per hour all the time. The PSU is designed to handle peak power when all the harddrives, MOBO, CPU, RAM etc... uses max power at the same time. But somewhere demand drops. We shipped 2x servers to the UK end of last year. These machines are designed to run 100% CPU power and drive 4x HD in RAID10 configuration. For this we used dual PSUs at around $600 each. They are designed for full load 24/7 BUT still require an air conditioned server room.

Then, with these cheaper PSUs, you sit with huge tollerances on where it trips. And when it trips the first time it probably trips easier the 2nd time. In my case I live in a warm country. When I started having my failures the ambient temperatures were between 25C to 28C. So in theory the bed would heat easier but then the PSU was also under more strain. Thus my ABS attempt failed as I can only get it near +-90C before the PSU caves in. So some of the clients might have received pretty stable PSUs while others ended up on the other side. Mine is somewhere in the middle.

The PSU is also the only thing RRP does not have control over. There is control over all the mechanical parts and the controller but RRP cannot guarantee that a PSU, in the box, would work everywhere. Mine dips and resets the controller with just the bed and extruder heater on. Not even moving a motor. So how does the PSU check over-current? The cheap way is to check it over a resistor. Assume the resistor is 1ohm and the current limit is 1A then you would be looking at a trip voltage of 1V. Now as the PSU heats up the resistor also heats up and it increases in resistance. Lets say it is now 1.1ohm then the trip current will be around 0.9A.

I look at these PSU specs very much in the same light as these cheap music systems which claim 1000W PMPO. When needed it can deliver 1000W into the speakers, but not RMS or continuous. Thus my feeling is that this particular PSU cannot maintain 20A even if it states that it can do 30A and even higher peaks. I am fed-up with mine so I am switching over to the lab PSU (which is @#$%^& huge compared to the PC PSU and can also only supply 30A).