Welcome! Log In Create A New Profile

Advanced

Having trouble tracking down power/USB communication issue

Posted by remondo 
Having trouble tracking down power/USB communication issue
August 18, 2017 10:45AM
My Mendel90 (formerly a RepRapKitStore Prusa i2) has been working fine for many years up until a few days ago when it suddenly began powering down a few minutes into a print. The PSU would turn off and I'd have to wait a few minutes until it would turn back on again. I checked and tightened all terminals on the control board (Melzi) and had a good look over everything else but I can't spot a problem anywhere. I've tried different USB cables and different wall sockets around the house.

The original PSU is a cheap generic 500W PC power supply which I have altered as per the Mendel90 build instructions (several 12V wires soldered together, several ground wires soldered together) which has worked fine up until now.

Thinking this was a power supply issue I found a spare, much better quality Hiper 530W PSU and wired it up to the Melzi in much the same way, however the wires are much thicker so I've had to bundle several wires together and solder them to 1 single, thicker wire in a similar way to this so that it fits in the screw terminal on the Melzi board. Could this 'bodge' job be causing issues?

This newer PSU has now presented me with a new problem - as soon as I begin preheating the heated bed (PCB Heatbed Mk2), the Melzi disconnects and reconnects to Octoprint. The extruder heats up fine and all motors work. Now, the Mendel90 instructions explain that current can flow through the USB cable, disrupting communication if not enough ground wires are hooked up to the board but I have all 12 ground wires (save for 1 which I've reserved to power a Raspberry Pi via a molex to micro-USB connector I've fashioned) soldered together, albeit soldered to a thinner (but as thick as possible) wire to fit in the screw terminal.

The newer PSU has 2 12V rails. Is it OK to solder 12V/ground wires from the 2 different rails together? Could this be causing the disconnection problem?

Any ideas what could be causing this and/or how I can troubleshoot this would be very much appreciated. I'm out of ideas at this point.
Re: Having trouble tracking down power/USB communication issue
August 18, 2017 02:41PM
The first problem does sound like your old PSU was failing. The capacitors wear out and the ripple increases until one of the rails goes over voltage on the peaks and it shuts down.

The second problem does sound like a bad ground connection but it should be OK to have a short length of thinner wire. Or you could just use as many black wires as will fit if they are thicker. I.e. if you have roughly the same cross sectional area of copper it should be good enough. And a short section a bit thinner doesn't add much resistance because it is short.

All the ground wires should be connected together in the PSU regardless of it having a split 12V rail.

I think it should be OK to connected the two sets of +12V yellow leads together but I am not absolutely sure. I think they have one 12V supply inside and split it with two separate current sensors. If they actually have two completely separate 12V supplies then it would not be good connecting them together. If that is the case I would use one rail for the Melzi and connect the bed positive wire to the other instead of taking it from the Melzi bed connector.

So I am not sure what your issue is. I would doubly check the ground connection. With the USB reset jumper removed does the bed stay on when it disconnects?

It could be the PSU is not happy with the unbalanced load and is cutting out momentarily. The dummy load values were selected to work with the 500W PSUs we supplied in the kit.


[www.hydraraptor.blogspot.com]
Re: Having trouble tracking down power/USB communication issue
August 18, 2017 06:16PM
Thanks for the reply.

I keep the reset jumper removed (hanging on 1 pin) at all times except for when flashing. Is that not the correct way to set the jumper? I don't think the bed stays on as octoprint reports it at room temperature.

One thing to note is that a few seconds after I set the bed temperature, I see this in the Octoprint console:

Quote
Printer sent 'start' while already operational. External reset? Resetting line numbers to be on the safe side

Does this just mean the Melzi has been reset?

I've tried with and without a 10W resistor as a dummy load but have read in several places that the dummy load is not really necessary on newer PSUs. I've tried using just 2 wires, which is the max that will fit in the terminal but I experience the same problem. I even have a bigger terminal on the power input since I had one melt a while ago.

I might just try using wires from only 1 rail if all else fails but each rail supplies 18A max which I'm not sure is enough. I took the cover (carefully) off the PSU to fix a rattling fan a few days ago and saw that there are 2 clusters of ground wires and 2 clusters of 12V coming from the board. One set of the 12V wires has thin black stripes on the yellow coating, these wires were used for CPU 8pin and 4pin connectors.

Is there some way I can diagnose the issue using a multimeter? Seeing as it's only the heatbed that causes disconnects could it be a short somewhere?

Edited 1 time(s). Last edit at 08/18/2017 06:29PM by remondo.
Re: Having trouble tracking down power/USB communication issue
August 19, 2017 07:18AM
Yes it is best to keep the jumper off for normal use so that USB disconnects don't cause a reset.

If the jumper is off and the Melzi still resets then it must have lost the 12V momentarily. It could be a short on the bed connections but more likely the PSU doesn't like the unbalanced load and needs more dummy load on the 3.3V and 5V rails. You could look at the voltages on all three rails with a multimeter but if it shuts down very quickly you might not see what happens. The usual problem is the 5V or 3.3V rails go too high when the 12V rail is heavily loaded and the over voltage protection shuts the PSU down.

There is also a power good signal that you could monitor with an LED and a resistor. If you see it go out the PSU has detected a fault.

Yes 18A is enough, the machine should take no more than 16A with anything on.


[www.hydraraptor.blogspot.com]
Re: Having trouble tracking down power/USB communication issue
August 19, 2017 12:39PM
I seem to have solved the problem by taking both 12V and ground wires from only 1 cluster; specifically the striped 12V wires used for CPU power. I've just successfully heated the bed and started a print. I took extra care while soldering the wires together, tinning each end before I combined them to ensure each makes good contact.

Now to level the bed again after all this messing around!

Thanks for all your help as always.


Ok scrap that. The disconnects have suddenly started happening again. I managed to print for an hour+ then left the printer idle for a while after the print completed. I've just come back to it to start a new print and it disconnected as soon as I started the print. I really am stumped now.

Edited 1 time(s). Last edit at 08/19/2017 04:28PM by remondo.
Re: Having trouble tracking down power/USB communication issue
August 20, 2017 08:56AM
I've just tried printing from the SD card which seems to work. When I read the voltage on the heated bed terminals it reads 6.53V, is this correct? Shouldn't it be ~12V?

Edit: I'd been moving the jumpers around and just realised it was taking power from the USB cable to heat the bed! SD card printing does not work when powered from the PSU. I've ordered the aluminium clad resistors for the PSU; I will give that a go before I try anything else.

Edited 2 time(s). Last edit at 08/20/2017 09:02AM by remondo.
Re: Having trouble tracking down power/USB communication issue
August 20, 2017 08:06PM
Yes the bed voltage should be close to 12V.

I don't see how the USB cable can heat the bed other than providing the ground path if it is missing.

Are you sure the Melzi ground terminal is connected properly? It might be worth checking back of the PCB.


[www.hydraraptor.blogspot.com]
Re: Having trouble tracking down power/USB communication issue
August 21, 2017 06:32AM
I don't know how that is happening either. I've just had a look at the back of the melzi and it looks fine. I might solder on a new terminal just to be sure.

As I mentioned here, it's only when I heat the bed that it draws power through the USB cable, when JP15 is set to VREG. The extruder takes power from the PSU.

I've been powering the Melzi via the USB since I've had it without issues. It's only when I moved the jumper across that I noticed this issue but it does seem to be related to the disconnection problems I'm having.
Re: Having trouble tracking down power/USB communication issue
August 22, 2017 05:15AM
I always use VREG because it is more stable source of 5V. If you use 5V from the USB then it can reset if the PC drops the USB connection.

As DC42 said in the other thread the USB cable getting hot implies the ground current is flowing through it due to a bad ground connection to the PSU.

>it's only when I heat the bed that it draws power through the USB cable, when JP15 is set to VREG.

Perhaps when it is USB powered it always resets when the bed is on so you don't see the problem. When it has a stable 5V the Melzi doesn't reset, the bed stays on and the cable gets warm.


[www.hydraraptor.blogspot.com]
Re: Having trouble tracking down power/USB communication issue
August 22, 2017 07:41AM
Well I think I might have found the culprit. As I desoldered the power terminal to fit a new one the solder pad came with it, complete with the little copper ring inside the hole. I don't think it was making good contact because I was able to heat the bed by applying pressure to the screw terminal sometimes.

I don't suppose you have any Melzi boards for sale, or know of a reputable supplier to get one, do you?
Re: Having trouble tracking down power/USB communication issue
August 22, 2017 09:19AM
You can just scrape the green solder resist off the copper plane surrounding the hole and bridge from the pin to the copper with a short length of copper wire.

We don't have any stock but you might be able to buy one direct from our supplier here: htttp://www.werneth-electronics.com/home.php. We commissioned them and as far as I know they are the only ones with decent connectors and MOSFETs and thick enough tracks.


[www.hydraraptor.blogspot.com]
Re: Having trouble tracking down power/USB communication issue
August 22, 2017 09:33AM
Ok I will give that a try, thanks. Which side of the board should I scrape off or does it not matter?

Edit: Just realised that's a stupid question because I wouldn't be able to access the upper side with the terminal in the way.

Edited 1 time(s). Last edit at 08/22/2017 09:39AM by remondo.
Re: Having trouble tracking down power/USB communication issue
August 22, 2017 10:22AM
Sorry for the double post but I did as instructed (scratched the solder resist off) and everything is working as expected again. Fingers crossed it stays that way this time; I'll remain pessimistic for the time being!

It must have been that dodgy ground hole after all. Perhaps the old PSU is also fine but I'll stick with this one for now.

I get the full 12V now when JP15 is set to VREG and the USB cable is nice and cool!

Thanks a lot for your help!

Edited 1 time(s). Last edit at 08/22/2017 10:23AM by remondo.
Sorry, only registered users may post in this forum.

Click here to login