Slow heating up on MK3 dual power heat bed

Hello

I purchased a MK3 dual power aluminium heat bed from [www.cubic-print.com] .
It ships without NTC so I took a "spare" thermistor of my other 3d printer (QU-BD TwoUp) which is a default EPOCS glass bead. I attached it in the middle hole of the bed and secured it with Kapton tape.
I use a Ramps 1.4 board with Repetier firmware 0.9x.
I did solder 2 wires to the "-" pad and connected both wires with the pin D8-negative and the other wire from "+" to the positive port on D8 aswell. (I am writing this just to make sure all my doings are correct/replicable).
My PSU is an old PC PSU with 520 Watts.
The (yelllow) 12 V rail can provide 22A. This is at least what the decal on it's backside says.
I connected both 12V rails (+GND of course, just to make sure I did not foget anything) to one of each input pins on the Ramps board.

When heating up the bed to 100°C for my very fist print in ABS it takes about 5-10 minutes to reach 70-75°C and from there on it only increases in 0,1°C steps.
There must be something wrong but I don't know what. This is my first print with a heated bed on my brand new Wallace.
PLA without heat bed printed just perfectly fine.

---

Blog: Capotexl
Filament Factory - How to build your own cheap filament extruder
Frankenstein Laser Engraver
Make the world a better place

I glued a cork board under the MK3 heatbed and it went up to 90°C after about 30 minutes.

I am currently testing another power supply. This one has 150W and 15A on the 12v rail. Measuring says it has 12.34V on the ramps input clamps and 12.01 on the heatbed clamps.

Edited 1 time(s). Last edit at 09/14/2014 01:50PM by ianmcmill.

With the new power supply it went up to 90°C after 30 minutes again. So with both power supplies I got the same result. Any ideas ?

What size wires go to the bed? 16Ga wires are not enough.
I used two 16ga and while it worked, it wasn't great. I switched to 14 and while it heated fast, it eventually killed the PSU because it drew more than the PSU could handle.

Which is the next issue, 20 amps may not be enough, I killed a few 20amp psu's this way.

4 motors = Nearly 4 amps
Hot end = More than 3 amps
Heated bed = 10-11 amps, some as high as 12 depending on resistance.
Fan = 1/3rd amp ea.
You still have to account for anything else needing power and the fact that China uses their own math for rating power supplies, which is usually about 20 (sometimes 50%) lower than rated, and that you shouldn't run a power supply at peak all the time.

Granted, not everything runs full power all the time, so 20 will get you going, it's not enough.

1) Put cork or (corrogated cardboard for testing) under the bed for insulation.

2) Put a plastic-wrapped magazine on the glass for insulation as it warms up.

3) Buy an SSR relay and 24V power supply.

My MK2 heated bed had the same warmup times on 12V, as does my current 300mm round Onyx heated bed.....

Heat beds are always problematic.

Thoughts from me:
1. What voltage do you get on the following places with the heatbed off (referenced to Gnd on the PSU input to the RAMPS 11A connector):
a. 11A PSU input to RAMPS?
b. D8 + pin (nearest the 11A connector)?
2. What voltages do you get on the following places with the heatbed on (referenced to Gnd on the PSU input to the RAMPS 11A connector):
a. 11A PSU input to RAMPS?
b. D8 + pin (nearest the 11A connector)?
c. D8 - pin (the other D8 pin)?
3. What voltages do you get across the connector of the heatbed with the heatbed on, at the heatbed end?

From comparison of 1 and 2, you can see if your PSU is drooping under load (eg: putting out less than 12V).

From the values given by 2 and 3, you should be able to map the rough voltage drops in the circuit, which may help identifying further issues if the PSU is not drooping under load.

Note: There are always going to be voltage drops in PCB tracks, cabling, components, etc. It's just a matter of mitigating issues to see if you can improve things.

PS: A standard RAMPS setup will run happily on 13.8V (such as a CB/HAM radio PSU), and the extra 1.8 volts can sometimes make all the difference. The "loss" in your wiring and components (the fuse and the mosfet) will be offset by the extra voltage. I would avoid going over 14V without precautionary measures being taken, though a standard Arduino Mega should be OK up to at least 14.5V, maybe even 15V. As always, YMMV.

Edited 1 time(s). Last edit at 09/14/2014 09:47PM by Cefiar.

Quote
sheepdog43
What size wires go to the bed? 16Ga wires are not enough. I used two 16ga and while it worked, it wasn't great. I switched to 14 and while it heated fast,

It changed because you wanted it to change. Switching from 16 to 14ga should have very minimal effect at 12v/10amps.

Quote

it eventually killed the PSU because it drew more than the PSU could handle.

Then you had a very cheap power supply. A power supply is only going to produce the current that it can produce. It's not like the heated bed can "suck" more power out of the power supply than it can provide. If the bed can consume more than the power supply can produce, then the power supply's overload protection circuitry should kick in.

Quote

Which is the next issue, 20 amps may not be enough, I killed a few 20amp psu's this way.

If you killed a 20 amp power supply drawing < 20 amps, then it wasn't a 20 amp power supply.

Quote

4 motors = Nearly 4 amps
Hot end = More than 3 amps
Heated bed = 10-11 amps, some as high as 12 depending on resistance.
Fan = 1/3rd amp ea.
You still have to account for anything else needing power and the fact that China uses their own math for rating power supplies, which is usually about 20 (sometimes 50%) lower than rated, and that you shouldn't run a power supply at peak all the time.

Granted, not everything runs full power all the time, so 20 will get you going, it's not enough.

It's amazing how 7.3+ amps is able to continuously flow through a RAMPS 5 amp PTC without blowing it. Either the PTC is way out of spec on all the boards, or you're not drawing anything over 5 amps combined. And your fans draw 1/3 amp, EACH?!?!? What are you using, a furnace blower to cool your hot end and/or prints?

Thanks guys for all the good information ! Really appreciate your knowledge sharing !

Quote
Cefiar
1. What voltage do you get on the following places with the heatbed off (referenced to Gnd on the PSU input to the RAMPS 11A connector):
a. 11A PSU input to RAMPS?
b. D8 + pin (nearest the 11A connector)?

1 a. 12.37v
1 b. 12.37v

Quote
Cefiar
2. What voltages do you get on the following places with the heatbed on (referenced to Gnd on the PSU input to the RAMPS 11A connector):
a. 11A PSU input to RAMPS?
b. D8 + pin (nearest the 11A connector)?
c. D8 - pin (the other D8 pin)?

2 a. 12.35v
2 b. 12.25v
2 c. 0.18v

Quote
Cefiar
3. What voltages do you get across the connector of the heatbed with the heatbed on, at the heatbed end?

3. 11.88v

Quote
Cefiar
From comparison of 1 and 2, you can see if your PSU is drooping under load (eg: putting out less than 12V).

From the values given by 2 and 3, you should be able to map the rough voltage drops in the circuit, which may help identifying further issues if the PSU is not drooping under load.

BTW I am using the ALU MK3 without glass. Cork is already attached.
I am using ordinary ceiling lamp wires (german ones with 3 braided wire). They are 0.75mm² which is like something inbetween AWG 20-21. Maybe they are too thin ? I've just read that the description says "Not for connecting cooking and heating devices" .

I changed the wires from 21 AWG (0,75mm²) to 16AWG (1.767mm²).

The heatbed has a resistance of 1.9 ohm.

On the pads directly on heatbed I measure 11.99 V.
The temperature increases in 0.5°C steps. After 10 minutes I get 80°C. From there on the temp only increases in 0.1°C steps.

Quote
ianmcmill
I am using ordinary ceiling lamp wires (german ones with 3 braided wire). They are 0.75mm² which is like something inbetween AWG 20-21. Maybe they are too thin ?

.75mm is inbetween 18 and 19 AWG and would be right at the very edge which what you should use. They likely are getting warm if you're running them non stop.

What is the resistance of your bed? The specs claim 1.4-1.6ohms configured for 12v, which is on the higher side for heated beds. That translate into lower power requirements and less heat given off. That, combined with the Al core means the less heat it does produce gets spread around a lot more, as well as radiated more. The metal core is nice as it helps keep even heat and can temper any minor momentary fluctuations in heat...however it takes and more to heat up and to maintain that heat.

...

Edited 1 time(s). Last edit at 09/15/2014 01:57PM by ianmcmill.

Quote
cdru

Quote
ianmcmill
I am using ordinary ceiling lamp wires (german ones with 3 braided wire). They are 0.75mm² which is like something inbetween AWG 20-21. Maybe they are too thin ?

.75mm is inbetween 18 and 19 AWG and would be right at the very edge which what you should use. They likely are getting warm if you're running them non stop.

What is the resistance of your bed? The specs claim 1.4-1.6ohms configured for 12v, which is on the higher side for heated beds. That translate into lower power requirements and less heat given off. That, combined with the Al core means the less heat it does produce gets spread around a lot more, as well as radiated more. The metal core is nice as it helps keep even heat and can temper any minor momentary fluctuations in heat...however it takes and more to heat up and to maintain that heat.

Oh yeah I mistook diameter for area.
Okay so after 25 minutes it reached 100°C. The heatbed has 1.9 ohm resistance. 0.3 more than they say in the datasheet.

Edited 1 time(s). Last edit at 09/15/2014 02:17PM by ianmcmill.

The issue with thin wire tends to be that as it passes current, it heats up and as it heats up, the resistance goes up slowly, not by much, but after a while it becomes a limiting effect.

FWIW: I would personally dump the PTC fuses and use real fuses, which might gain you a small amount. If you want to get even more out of the board, I'd lay a bit of solder or even a piece of wire on top of any existing tracks on the RAMPS PCB related to D8. This will reduce the volt drop between 2a and 2b.

You also definitely want to heatsink and to actively cool the mosfet. Longer operation with high current could lead to the mosfet heating up a bit over time. Lots of on/off switching will also lead to it heating. The hotter the mosfet is, the higher it's internal resistance will be, leading to higher voltage drops across it. This will reduce the volt drop at 2c.

Notes:
- You probably want to repeat these readings as the temp rise levels out a bit, as that seems to be when the system is in stress.
- Are you running BANG-BANG or PID in firmware. For a heated bed, you really don't need PID control for the bed, and it usually causes issues. You might find that the firmware defaults to PID and this is why the temp increase slows down so much.
- PTC fuses don't have well defined trip points, and they take AGES to trip (20secs is fairly fast for a PTC). The closer they are to their trip current or max current (these are different values), the higher the resistance across them, leading to a larger voltage drop. IMO they're ill-suited as protection devices and in fact cause a LOT more problems than they solve. With a high current PSU and a high current short, it's quite possible to make a PTC fuse catch fire, as it gets so hot internally that they explode before the fuse itself trips.

Edited 1 time(s). Last edit at 09/15/2014 07:02PM by Cefiar.

Quote
ianmcmill
The heatbed has 1.9 ohm resistance. 0.3 more than they say in the datasheet.

Get a better heater. Compared to spec, you're heated bed is producing up to 25% less heat than it's suppose to, and nearly 50% the heat other heated beds (1-1.2 ohm) produce. The Alumnium core acts as a giant heat sink. This works for you when you want even heat for your entire bed...but bad when you're already struggling to get it to temperature let alone maintain it.

I ordered the heatbed here [www.cubic-print.com]
It's B-stock with just one little scratch on the alu surface. The tracers are fine.

The seller does not give specs but links the reprap wiki site of the MK3 Alu which says

Quote

Resistance between 1.4 and 1.6 ohm for the 12V
Resistance between 5.0 and 5.4 ohm for the 24V

I switched back to PLA. Heating to 60°C takes just 4 minutes.

I've found energy saving self-adhesing transparency. It is made from PE (http://www.hornbach.de/shop/Fensterfolie-Isolierfolie-fix-o-moll-transparent-1-7-x-1-5-m/8247094/artikel.html).
Could this help ? Or is it possible that it insulates too much so the PLA/ABS won't stick to its surface ?

Lots of people on the Ormerod forum complained about long bed heating times to ABS temperatures. 120W isn't quite enough power. The solution most of us have adopted is to use a 12V PSU with a voltage adjustment allowing it to be turned up to 13 or 14V. These PSUs cost very little on eBay, less than GBP20 here in UK for 25A one. I printed a cover to go over the mains terminals, see [www.thingiverse.com].

Another thing that helps is to put an insulating mat on top of the bed while it is heating up.

---

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].

Also, which MOSFET is being used to drive the heaters on yout PCB? Current wisdom is that IRLB8743PBF or for extreme cases IRLB3036PBF be used, and the schematic I looked at on RepRap spec'd out STP55NF with a higher RDSon, plus if a non-Logic Level part was used by accident, that'll make it even worse. The IRLB8743 works nicely with about 3mOhm vs 15mOhm resistance when on and with high current stuff (IRLB3036 costs moire bit is rated for a full 60V & RDSon of about 2mOhm).

Simpe quesdtion, Do you have insulation on the bottomside of the heatbed? Some people make the mistake of ignoring that and loose a lot of head down.

I do agree, most of your problem is probably PSU & wiring.

Just trying to think outside of the box on other things that might have been overlooked.

The three MOSFETS on my Ramps 1.4 are P55NF06L ; 7SABF ; V6 ; PHL 243. None of the ones you named there. Might this be the bottleneck ?
I insulated the alu plate with a 3mm cork sheet.
I just received my Teensylu v0.8. Gonna give this a shot as soon as I get another hotend.

Edited 1 time(s). Last edit at 09/21/2014 06:33AM by ianmcmill.

Quote
ianmcmill
The three MOSFETS on my Ramps 1.4 are P55NF06L ; 7SABF ; V6 ; PHL 243. None of the ones you named there. Might this be the bottleneck ?
I insulated the alu plate with a 3mm cork sheet.
I just received my Teensylu v0.8. Gonna give this a shot as soon as I get another hotend.

P55NF06L is how STP55N06L is marked which is the 0.016 - 0.020 Ohm part spec'd out for use and the L at the end means you have the proper Logic Level version of the part, so you have the part that is spec'd for RAMPS 1.4

ref: http://www.st.com/web/en/resource/technical/document/datasheet/CD00002690.pdf

Compared to:
http://www.irf.com/product-info/datasheets/data/irlb8743pbf.pdf
http://www.irf.com/product-info/datasheets/data/irlb3036pbf.pdf

Let me repeat myself from the first post ... the power supply and possibly wiring is most likely the factors in your problem then the MOSFETs. The MOSFETs cat improve how it works while heating up less. You could easily be loosing 0.3V or more with the current MOSFET's (depending on how much current the power supplies can actually supply). The printers I've helped build we've always used IRLB8743 combined with 30A power supplies and heavy gauge wiring and insulation below the heatbeds to maximize heating and minimize heat times. The order to check things is Power Supply, Wiring & Connections, then MOSFET's. When your on a limitted budget sometimes the power supply moves to the end unfortunately.

Full disclosure: I do sell the MOSFET's on eBay mostly to US & Canadian buyers, but I do honestly think the spec's for the RAMPS 1.4 is substandard for hotbed needs. Other boards like the Sanguish have changed to the IRLB8743 which is how I got interested in those MOSFET's to begin with. They both work better at switching high currents and heat up less (when MOSFET's get hot the let less power through)

IRLB8743PBF 30V 150A on eBay
IRLB3036PBF 60V 195A on eBay

Edited 1 time(s). Last edit at 09/21/2014 07:23AM by Learner.

Okay okay. Still I went from 1st PSU to 2nd wirings and now ended up with 3rd the MOSFETS.
Switching the computer PSU (22A on 12v line) to the PSU I used with my Filament extruder (16A on 12v line) did increase the overall temperature (despite it has 6A less on the 12v line) I went from 80°C to 100°C combined with using new wirings from an old flat iron (which should really do the job!). All of the above replacements helped but most likely the MOSFETs (IMHO) are the bottleneck (if I got you right on this... )

Edited 1 time(s). Last edit at 09/21/2014 07:27AM by ianmcmill.

Others should comment a well because I admit I'm biased. The fact that changing from a 22A to a 16A improved your temps does say the first PSU was having problems despite it's ratings.

You really should have a good quality PSU with > 20A on the 12V rail or a dedicated non-ATX PSU for the heatbed for good ABS temps. The MOSFET change can help, but I think you're still not getting enough power from the PSU. If you're hesitant on getting a better PSU, then trying the MOSFET change and making sure all the wires are overly large and solidly connected can help.

When the heatbed is heating, how hot does your MOSFET get? Remember the hotter a MOSFET is, the higher the resistance (it can easily double or triple). Thats part of the reason the MOSFET change can help, it starts out with much lower resistance, heats slower, still has less resistance even after it warms, and never gets anywhere near as hot.

You have already said that you are getting 11.99V across the heatbed terminals and that you are using a 12V power supply. This means there is nothing wrong with your mosfets or cabling. The problem is that at 12V, your bed heater is not powerful enough. You can solve this in 3 ways:

1. Improve thermal insulation.

2. Increase voltage.

3. Replace the heater by a more powerful one.

See my previous reply for some suggestions regarding the first two. In your position, I would replace the PSU by one that you can turn up to 14V. Heater power increases with the square of the voltage.

I am assuming that firmware is not causing the heater power to be reduced to soon. You can verify that by commanding a very high bed temperature and seeing whether the time to get to the temperature
yiuvactually want improves significantly.

---

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].

Can someone link me to the 14V power supply on ebay (I can't seem to find one that doesn't cost a few hundred bucks)? I'm having the same problem with my Prusa i3. I attached an aluminum plate to the bed and now I can't get it over 80c really.

Edited 1 time(s). Last edit at 07/17/2015 01:55AM by therippa.

Quote
therippa
Can someone link me to the 14V power supply on ebay (I can't seem to find one that doesn't cost a few hundred bucks)? I'm having the same problem with my Prusa i3. I attached an aluminum plate to the bed and now I can't get it over 80c really.

The standard 12V LED power supplies on eBay have a voltage adjustment pot, which typically lets you go up to 14V or a little more. I use the 12V 300W one in this range [www.ebay.co.uk]. But here are a couple of warnings:

1. Make sure that the power output is high enough. Increasing the PSU voltage will increase the heated bed and hot end current, but reduce the rated output current of the PSU. For example, a 12V PSU rated at 20A (i.e. 240W) is only good for 17.1A when you turn it up to 14V, because it is the output power that matters.

2. An Arduino/RAMPS combo is OK at 14V input if you don't connect anything else to the 5V rail; but if you have an LCD backlight powered from the 5V rail then do not exceed 12V input to the Arduino, or you will overheat the 5V regulator.

---

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].

I'm not sure if everyone has gotten to the bottom of this issue yet, but I thought I should weigh in a bit.

I have purchased my the MK3 Aluminum heated bed from Robotdigg, and I've been noticing that a lot of them are not made according to the specifications listed. It says that the resistance between the 12V traces should be about 1.5A and for the 24V traces 3.0A, however, I get double that. This means that you are essentially cutting down your power output by exactly half, so the 96W heater is actually a 48W heater. This is most likely the reason why you are experiencing such slow heat up times.

If you have the option of using 24V (either switching your board to a 24V compatible board) or using an external 24V power supply with a solid state relay, you could simply just wire the 24V to the 12V connections on the heated bed and it should heat up much faster.

You might want to check up with your provider to see what the specs are. Robotdigg just changed their latest version of the MK3 Aluminum bed so the new ones might actually be the correct resistance.

I don't know if I can add anything to the thread, but I currently had a very similar discussion at the G+ group.

For the general problem of taking to long to heat, it seams, lies with the dealers. Almost no one lists or even knows, when asked, how many watts the plates have or the top resistance.
What happens is that most of us ends up with 90w plate that works at 40w (as someone else said previously) when what we really wanted was 150w plates.
I have a very decent power supply now and still it is impossible to reach 110º for properly printing ABS. I'm trying to buy a better heat plate, but I have no idea how to choose anymore

What made A LOT of difference for me was to change the heated bed from bang bang to max pid on Marlin. After running auto calibration, I can now reach 95º while pre heating and stable 90º while printing.
Insulation helps heating up, but makes no difference to the top temperature while printing.

What bother me as hell and no one seams to have a good explanation, is why temperature drops to 75º every time the printing fan turns on.

Quote
aFrazatto
What bother me as hell and no one seams to have a good explanation, is why temperature drops to 75º every time the printing fan turns on.

Simple: the print cooling fan cools the bed as well as the print.

There have been many reports of PCB bed heaters providing insufficient heating power. One way of increasing the heating power is to turn up the voltage. This is easy to do of you are using a LED-type PSU with a voltage adjustment pot. However, I do not suggest you try this if you are using RAMPS electronics, because it can't take the extra voltage - in particular, if you are powering a graphical LCD display from it, then you are probably already overheating the voltage regulator chip on the Arduino, and increasing the supply voltage will make things worse.

---

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].

Quote
dc42
Simple: the print cooling fan cools the bed as well as the print.

Sorry but this makes no sense....
If it was so, the temperature wouldn't always be the same and it would make little difference on tall objects.

The temperature ALWAYS drops to 75º :3

Quote
aFrazatto

Quote
dc42
Simple: the print cooling fan cools the bed as well as the print.

Sorry but this makes no sense....
If it was so, the temperature wouldn't always be the same and it would make little difference on tall objects.

The temperature ALWAYS drops to 75º :3

Does the indicated temperature drop immediately the fan turns on, rather than over a period of a minute or more? If so, then that suggests a noise, grounding, or firmware issue.

---

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].

Quote
dc42
Does the indicated temperature drop immediately the fan turns on, rather than over a period of a minute or more? If so, then that suggests a noise, grounding, or firmware issue.

It's not immediate, since the heat must dissipate, but it is quite fast.
I found another person with the exact same problem, if it was firmware, it would be more common.
I believe it's something with the generic/cheap chinese RAMPS I bought.