Powering heated bed kills PSU [Gen7-1.2]

I am trying to add a heated bed to my printer. Testing on the second +12V rail showed that it worked so I hooked it up to the Gen7 board. Then things went down hill a bit...

When I enable the bed itself, it powers on, pulls 8 to 10 amps and heats up quickly. But if I enable the stepper motors, the PSU craps out on me and the printer sits there, probably without its required 12 volts to run everything. I looked at the specs and I guess that the single +12V rail of the PSU with a maximum output of 15 amps is hitting its maximum.

Now I know that the 1.3 version of the Gen7 electronics have a secondary power connector to supply power to the heaters, probably for this very reason.

Looking at the schematics, am I correct in concluding that I could take +12V from the second rail and connect it to the GND / mosfet input to prevent overloading the first rail? I mean, the ground within a PSU is connected between all rails so this sort of creates the same situation as the electronics in version 1.3, right?

I double checked the schematics and concluded that it would be safe to use the +12V from one rail with the ground from another.

This sort of works now.

Sort of.

The problem is that when the heated bed is turned on, the whole system restarts. So the PSU turns off and the Arduino sends its 'start ok' message again.

Now I know I am still missing the 1N4004 diode to dissipate the back-EMF when the heater shuts down. But this should not prevent me from turning it on in the first place, right?

On the odd test that it actually turns on properly, as soon as it reaches its target temperature and the PWM kicks in (at least I think its PWM) the PSU shut down then and the system resets.

I tried switching the +5V supply from the PSU to the FTDI board, in case the +5V crashes when the heater turns on but that does not seem to help.

Any suggestions?

What kind of load do you have on 5V? (Should be an amp or two)

You mean from the PSU? None I'd say: the jumpers allow the Arduino to be powered from the +5V from the USB port and the +5V is not used anywhere else: the rest of the logic is using +12V afaik.

I tried with a Schottky diode over the second +12V rail which made no difference.

I also tried the more powerful PSU from my normal computer, which did not work. The latter has 28 ampere +12V rails, stabilized and all that (800 watts total): if that does not work, its not a lack of power thats making the Arduino reboot (I think).

Try adding a load to the 5V rail. A 50W 12V accent lighting bulb will give you a couple amps and stabilize the power supply quite a bit. That's why Gen3 has a power resistor built in. ATX supplies don't behave well if the 5V line is unloaded.

I use a 20W 5Ohm sand resistor to load my 5V on my bench supply (well I did till age killed it) The psu not the resistor. yeah if your using a psu power supply you need to load the 5V rail or your screwed I'll be using a 12V server supply that doesn't need a 3.3V load (and it dun have 5 volts the regulators are on the board of the thing this came from )
this think can deliver a stable 70 amps on 110 105A on 220 but regular ATX psu's need a 5V load or the 12V will be all over the place.
terramir

Cheers for the suggestion: I'll try that solution next.

In the past few days I've been trying things like:

• adding more resistance (works on power-on, but the resistors are 5W so they are smoking within a minute, not sure if it would help while PWM kicks in)
• powering the bed using the +5V, this will pull 3.5 amps resulting in a 17 Watt heater. The bed tops out at 50 degrees, making it usable with PLA
• using a 2N3055 power-transistor with heatsink to limit the current rush using a RC circuit at switching times. This might work if the bed has more resistance: it prevents the PSU from shutting down but due to the internal resistance of the transistor, it looks like its burning more power than the bed. After 10 minutes, the bed reached 35 degrees and the heat sink of the resistor was reaching 120+ degrees...

I now have two more options on my to-test list:

• Use a dummy-load to put on the +5V rail
• Buy a +12 volts switching power supply from ebay, the bigger ones cost 40 dollars and are rated for +12V with 30 amps. Perhaps this is a more simple solution as I would probably be able to power both heaters and all motors using a single cable. The downside is that if it fails I'm stuck with a 12 volts DC power supply I won't use...

Quote

I now have two more options on my to-test list:
[*] Use a dummy-load to put on the +5V rail

You must use a dummy load on the 5V rail if you are using a PC PSU. It isn't optional. You can't draw a big load from the 12V rail without a sgnificant load on the 5V rail.

At best case the 12V rail will drop to about 10.5V, at worst case the 5V rail rises above 5V and the PSU shuts down because of the over-voltage protection.

Edited 2 time(s). Last edit at 01/24/2012 10:45AM by nophead.

---

[www.hydraraptor.blogspot.com]

I've just tried with a number of power resistors to see if can get it to work.

I tried with anything from 12.5W (2.5 amps) up to 62.5W (12.5 amps). The only thing I get is a second heater next to my printer and as soon as it starts toggling the power, the both PSUs stop again.

I can only conclude that an ATX power supply is unsuitable for use with a ~100W heated bed which will be PWM-ed to the proper temperature...

Edit: just ordered a +12V 30A switching power supply from China - I'll update if that solved my problems (and I'm assuming it will)...

Edited 2 time(s). Last edit at 01/25/2012 04:35PM by Cyberwizzard.

ehhh your power supply may not be able to handle the sudden jump in current flow some atx psu's if made cheaply can't hack that kind of power draw especially the inrush which could be about 2x the regular current. terying the 800 W was a good step but you may need to try that with a dummy load as well. this is why I'm thinking the bed design is flawed, it should constantly be powered on 5V let's say 10-20w and when the thermostat says to do it there should be a 12V 80w (or more current to the 12V traces of the bed. I am working on designing one right now. But I need some links to the mk1 and mk2 designs in order to do the designs.
Will look into it further but for you a quick fix might be slapping the 800W psu in and put that 12W resistor on the 5V line.
it's wasteful, but other than that I just figure something is wrong with your electonics, because others have the heated bed working (did you substitute any parts?)
terramir

Update: I finally received my new 350 Watt 12 volt power supply. I tried removing the ATX power supply and replace it by the new one with a molex connector screwed in.

Now I run into a problem: nothing moves. I do not have a +5V power supply so I used the jumper to get +5V from the serial interface. At this point, one LED lights up and I can talk to the Arduino.

However, issuing commands does not result in any movement.

I looked at the schematics and it looks like the +5V can be pulled from the ATX connector, molex connector or serial interface. Am I wrong here? Does the molex connector require a +5V?

I think you need to connect 12V to both molex connectors as the 12V feed for the heaters is separated from the motors.

---

[www.hydraraptor.blogspot.com]

A feature added to version 1.3 of the electronics while my board is version 1.2

I only have one molex connector and it is powered. Well, the +12V and the ground, the +5V line is empty.

Edit: looking at the schematics it seems like the power supply to the Arduino, I2C connector and SPI connector are on an internal +5V volt rail, which can be powered from the serial port, ATX cable or the molex connector. The +5V provided to the stepper drivers and opto-switches comes from the molex connector.

So while I currently power the processing part of the PCB, the business end is dead without a +5V voltage source on the molex connector.

Tomorrow I'll get a LM7805 to generate the +5V needed to run the rest of the board...

Edit 2: Using a second jumper I could connect the internal +5V (from the serial board) to the molex rail, which would effectively power up the Polulu's and opto's using the USB +5V. I somehow doubt if thats a smart solution...

Edited 2 time(s). Last edit at 02/24/2012 05:17PM by Cyberwizzard.

Update: I soldered a LM7805 to the +12V power supply to power the rest of the printer. I even added a capacitor to stabilize the +5V line.

Once again, reaching the target temperature results in PWM-ing and crashing the Arduino...

So.... can I use PID with PWM for the print head and bang-bang on the bed? If so, how?

You might want to have a look into this thread, page 3: [forums.reprap.org]

I think I've found a solution which will be documented before too long.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY