Heating the Hotend and the Bed
-
I have an Anet A8( Prusa clone), with a ramps board, and my large CoreXY driven by Duet Wifi.
Both powered by 12 volts.
The Anet is powered by a 360w power supply.
The CoreXy is powered by a 450w power supply.
Both with Mk2 hotbeds and E3D hotends.
Both have SSR's on the bed and hotend.I have recently become suspect of the temperature readings and it's accuracy for the hotend on my CoreXY, though the bed readings are accurate.
That discussion is occurring here:
https://www.duet3d.com/forum/thread.php?id=4512
if you care to know.
In short, I don't have anyway to read the actual temp of my hotend on my CoreXY, but I suspect that the reading from the web interface is different than the actual temp because I find myself having to set the hotend temps much higher than the filament specs would call for.
Example, I have to heat PETG up to 255-260c to eliminate adhesion issues. Anything cooler and my prints look bad.My Anet A8 has always heated up pretty quick. I can go from room temp to 60c on my bed in about 2 minutes.
I have gotten it up to around 100c very easily before but I rarely go over 60c.
It has 220 x 220mm bed, no insulation except for a buildtak sheet on top, and then a piece of glass on top of that.My CoreXY on the other hand, struggles to reach 50c on the bed. It is 300 x 300m, with cork insulation underneath, buildtak sheet on top and a piece of glass.
It usually takes around 15 minutes to get to 50c and is my default temperature for most things.
I have gotten it as high as 60c but that is about 25-30 minutes and barely gets there.
If I need anything hotter it just wont go past 60c.
I have become very talented with hairspray on a glass bed, as I don't generally have lifting issues. I only print PLA and PETG.
The hotend builds quickly but not as quickly as the Anet.I always attributed the hotbed heating so much slower because of it's size. But it got me to wondering if 80mm more is really that much bigger considering how effortlessly the Anet A8 reachs setpoint, and that I should at least be able to achieve better than 60c, even if it were still slower.
I have performed a PID tune on both the hotend and the bed but there is no change in performance of either.
Right now I am fine with how the bed is performing, even if it is slow and won't build past 60c.
But I was wondering if what I am experiencing is normal, and if not, if it is related to the temp problems I am having with my hotend?Thanks in advance.
-
Maybe its a 24v bed @ 12v supply?
-
1. Is it a PCB bed heater? It's very common for them to be under-powered.
2. Measure the voltage appearing on the bed heater when it is running at full power. If it is significantly less than 12V, then you have excessive voltage drop in your SSR or cables. I hope you chose a good low voltage drop SSR, not a SSR-40DD.
3. You can turn up the power supply voltage to the CoreXY to get more heating power. Unless it is an ATX power supply, there is likely to be a potentiometer at one end of the terminal strip.
4. Why are you using SSRs on the hot ends? They are never needed.
5. If you suspect the hot end temperature readings are wrong, check that you have the correct thermistor parameters in the M305 command.
HTH David
-
David,
I am embarrassed to admit I told you I had SSR's when I have mosfets. I always thought that an SSR and a Mosfet were the same until now. Thank you Google.
And to add insult to injury, I bought them for the Duet but I never installed them.
I just recently rewired my Prusa clone, which does have both Mosfets installed and when I was typing this post out I had it pictured in my mind that they were installed on my Duet. Imagine my surprise when I took the door off the case for my Duet and it was all alone in there. If you're shaking your head right now in disbelief, it is well deserved. I am not usually this inept.When I commissioned my first El-cheapo 3D Printer the integrated mosfet on the control board fried on the very first print.
I had to wait a month for a replacement to arrive from China, and during that time I read everything I could about the problem and the consensus among the folks who had the same problem swore by having them on the hotbed and hotend. In my defense, one mental giant adding to the conversation posted a picture of a Mosfet that he installed on his, but he called it an SSR. Lesson learned, don't believe everything you read on the internet. I suddenly feel the urge to give up 3D printing and partake in something a little less intellectually demanding, like basket weaving. Again, forgive me.Nevertheless, I swiftly installed them along with the replacement control board onto my Prusa clone.
The investment I made in the Duet board, seemed to justify doing the same.
Little good they are doing in the box of wiring stuff in my cabinet.So a revision: No protection of any kind installed on the heaters of my Duet Wifi.
I am indeed running an ATX.
My heat bed is a PCB.Shall I continue with measuring the voltage for the bed? Or is it still necessary now, knowing nothing is between the Duet board and the bed?
Shall I hurriedly install the Mosfet (aka SSR) to my duet?For the record, here in the deep south of these United States, we call all carbonated drinks, 'Coke'.
A perfectly acceptable conversation in my area would go like this…Scott- Billy Bob, you want a Coke?
Billy Bob- Sure!
Scott- What kind?
Billy Bob- Diet Pepsi.
Scott- Coming right up!I could talk about ya'll too.
You guys spell the word "color" wrong, and you don't pronounce 'Aluminum" correctly either.
(You notice my attempts at being witty and humorous to distract you to the fact that I am an idiot? I hope it worked)Thanks for the help. Truly.
-
No need to install the MOSFETs. It's still worth measuring the voltage on the bed terminals - take great care not to short the multimeter probes together while doing this. Then compare that voltage with the voltage on the PSU output terminals.
If there is a difference of more than about 0.2V between the two voltages, consider using thicker wires, and check that the terminal block screws on the Duet are tight. Otherwise, turn up the voltage on your 12V PSU, then retune the heaters.
-
I am indeed running an ATX.
I know you can adjust psu voltage, but from my research you can't adjust atx voltage without modifying it.
Isn't this correct?Maybe the atx is the problem for ALL my heating problems and I should just replace it with a psu.
Or maybe a bigger atx? I thought 450w of which 360w is dedicated to the rail running to the board would be plenty. -
You may find that your ATX PSU is providing somewhat less than 12V on its 12V output. The less expensive ATX PSUs use a single feedback loop that averages the error on the 12V and 5V outputs. So if you have a low load on the 5V output and a much bigger load on the 12V output, the 5V rail ends up being high and the 12V rail ends up being low. The better branded PSUs don't have this problem.
-
Thank you, David.
Is there an ATX that you'd recommend that would power everything adequately?
I thought I had bought a decent one, but apparently not. -
What brand of PSU did you buy?
I only have experience of two ATX PSUs powering a 3D printer. The Alpine PSU (550W AFAIR) one that came with my Ormerod kit had the problem I referred to above. I then purchased a Corsair CX430M, and that one was much better - the 12V rail remained close to 12V under varying loads and the bed heating time was less. But after that I bought a cheap 300W fanless Chinese LED/CCTV PSU and enclosed the business end of it like this https://www.thingiverse.com/thing:441081. It's been powering that printer for about 3 years.
-
APEVIA Venus 450W from Amazon.
It had good reviews and I never thought the power drop would be that much despite the higher wattage.The only reason I wanted an ATX was for the remote power on option, the ability to navigate DWC without powering up the whole printer, and the 5v rail to power my Pi Cam, all in one box and having one power cord for everything.
I am of the mind now to just get good PSU and turn it on the old fashioned way, while separately powering the RPI.
-
I suggest you measure the 12V output of the PSU, with the bed heater both off and on, to see how good that PSU is.
-
ATX PSUs aren't just 12V. A 450W PSU will be the 3.3V, 5V, and 12V rails added together. On the Apevia Web page the current on the 12V rail is 32A, i.e. 384W. Therefore the Anet and CoreXY should be similar on the paper. If the ATX supplies several 12V rails, the current on one f the rails might be even lower.
Have you tried to perform a PID tuning (M303) wit the P parameter? The P parameter is the PWM fraction to use.
I have recently migrated my PSU to 24V with a 50W heater and I noticed that with standard PID calibration the M303 without the optional P parameter produced a temperature overshoot. The temperature it raised very quickly to the desired temperature, but the risk of fire in the case of an issue is high. I returned the temperature adding the P parameter with a final value of 0.3 that does not produce the temperature overshoot warning but on the downside it takes longer to reach the printing temperature. It is not a big problem for me, because the print has to wait for the bed anyways.As I understand from the Wiki, the default value for the P parameter is 0.5. Try re-tuning the PID with a specific P parameter on higher values and check the results.
-
This is the specification for you PSU on the Apevia Web page: http://www.apevia.com/ProductsInfo.asp?KEY=Venus%20Power%20450W%20(ATX-VS450W)
-
As I understand from the Wiki, the default value for the P parameter is 0.5. Try re-tuning the PID with a specific P parameter on higher values and check the results.
That's only true for old versions of firmware. On recent versions (1.18 and later AFAIR) the tuning algorithm is different and P1.0 is the default.
If you are using a 12V heater cartridge with 24V power, I strongly advise you to buy a 24V cartridge and use that instead.
-
Thanks for the clarification. I just re-read the Wiki and for V1.18 it mentions the default value of P=1.0 (my bad)
My cartridge is from Aliexpres and specifies 24V/50W. However I have measured a resistance of 8.1 ohms, which would mean 20W if powered with 12V and 71W if powered with 24V. That got me thinking of the quality…...
