Bed temperature

[Stephen Toms]

Hi Everyone,

Sorry I couldn't come along yesterday, got called into work, hope it was a good one.

I'm getting towards the end of my 3D printer build now and have come up with a little issue and I wondered if anyone had any advice?

I am struggling to get my heated bed up to temperature, and what temperature it does manage happens very slowly. My set up is as follows ... RAMPS 1.4, MK3 Alu Hotbed connected to 12v, latest Marlin setup for PID bed temp, ATX supply capable of 25A on the 12v line and running via either pronterface or repetier. I have measured the resistance of the bed and found it to be about 1.8 ohms. When I measure the current running through it its registering about 3.3 amps which is obviously about half what I would expect, so I am assuming some form of PWM with a 50% duty cycle is at play. Through the serial port I am getting data like this ...

< 17:38:17: ok T:23.7 /0.0 B:76.9 /100.0 T0:23.7 /0.0 @:0 B@:127 >

Is the figure at the end (B@:127) a PWM value? I am targeting a temp of 100 degrees and as you can see its topping out at about 77. The PID autotune set at 60 degrees went fine and these were the figures I used in Marlin. The MAX_BED_POWER is currently set at 255 so not really sure why (if PWM is at work) the host isn't send a longer duty cycle to the bed.

Any thoughts appreciated, and any stupid mistakes pointed out are greatly received ;)

Steve

[Andy Kirby]

Steve

Nice work on the printer, is that your own frame design?? Just seen it
on Witter.

Re Heated build plate.

Something quick to try that will make a difference. Insulate the
underside and if space, the edges of the plate as well so that you are
not loosing as much heat.

You can use corrugated cardboard at a pinch.

I use ceramic fibre from a kiln. I have some spare if you want some. It
does dust off a little so when you have made it so size it could do with
the edges taping (Kapton) or spraying with a heat compatible binder of
some sort. Might be worth trying brushing on a coat of thinned PVA just
to bind the surface. it melts at 160 which is a temperature you are
unlikely to want to reach with a heated build platform.

Aluminium is a little too effective as a conductor sometimes.

On duty cycle there is no better way than taking a look with a scope. If
at full power you are only getting as you think 50% then take a look at
PWM max. in the config. Although I did think these came defaulted to 255.

Re rated wattage.

Lets see 1.8 Ohms at 12 Volts gives ummmmmmm

6.66666667 Amps

or

80 Watts

http://www.rapidtables.com/calc/electric/ohms-law-calculator.htm

Too lazy to take my socks off.

So yes 3.3 Amps is 50% or thereabouts. It does look like your PWM Duty
Cycle. Umm dumb question time you did check the PWM value for your bed
and not your extruder heater ????? Sorry gotta ask.

If you hadn’t measured the resistance (did you do it with the leads that
connect back to the ramps in circuit) I would be advising checking out
the connections and or wires. They can add a bit more on particularly if
they are making poor contact or too thin.

So you are running at 40 Watts, a lot will get radiated and convected to
the air with an aluminium plate. Get the remainder of your drive wattage
and it should easily make 110 Degrees C. With the underside insulation
fitted.

> Is the figure at the end (B@:127) a PWM value? I am targeting a temp of
> 100 degrees and as you can see its topping out at about 77.

Dunno but I would guess same as you here. 127 being 50% or so. Chris
Palmer would have the answer to this one or maybe the forum for the
Marlin developers is worth a go....

Wonder what your Proportional term is set at ??


All the same good work.

Cheers

Kirbs

> --
> You received this message because you are subscribed to the Google
> Groups "Sheffield Hardware Hackers" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to sheffield-hardware-...@googlegroups.com
> <mailto:sheffield-hardware-...@googlegroups.com>.
> To post to this group, send email to
> sheffield-har...@googlegroups.com
> <mailto:sheffield-har...@googlegroups.com>.
> Visit this group at
> http://groups.google.com/group/sheffield-hardware-hackers.
> For more options, visit https://groups.google.com/d/optout.

--
Andy Kirby
http://www.kirbyand.co.uk/
http://www.kirbyand.co.uk/serendipity/

[Stephen Toms]

Morning Andy,

Sadly the frame isn't my own design - that was the original plan (or at least to cut it myself) - but impatience got the better of me and I brought one. Most of the rest of the hardware and components (as well as a lot of helpful advice!) came from Darren at WeDo3D.

I picked up some cork from Darren yesterday, as I had read that insulating the bed would help, so I am ok for the ceramic fibre but thank you for the offer.

After a bit more probing around I was suspicious that it was a dodgy PSU so sourced another yesterday and bingo! Bed heats up much faster now (probably assisted by the insulation as well no doubt) and can get to 100 degrees.

I think my assumption of PWM values was a bit of confirmation bias on my part, saw a low amperage reading, saw a number that MIGHT have been a pam value that look about tight so put 2 and 2 together and came up with 45. Never mind, at least its sorted now.

Got my first print out of it last night, although it did take a few attempts to get it to stick to the bed. I have made up an ABS slurry, but think I'll experiment a bit with a few more layers of that, glue stick, hair spray, etc. I've heard a PVA wash works quite well so might hunt around for some mixing proportions and "borrow" some PVA from my little one ;)

Cheers,

Steve

> an email to sheffield-hardware-hackers+unsub...@googlegroups.com
> <mailto:sheffield-hardware-hackers+unsubscribe@googlegroups.com>.

> To post to this group, send email to
> sheffield-har...@googlegroups.com

> <mailto:sheffield-hardware-hac...@googlegroups.com>.

[nop head]

I think the B:@127 is a display issue. I think PWM is actually 7 bits although it is set in the config in 8 bit format. So for 100% you set 255 in configuration.h but it displays 127.

1.8R is actually a bit high for a 12V bed. The Prusa Mk2 design is 1.2R for 10A.

Adding an Amp meter in series is likely to add significantly to the resistance and give a low reading. A DC current clamp is more accurate.

The resistance of copper increases with temperature so at 100C the current falls to about 70% of the starting value.

To unsubscribe from this group and stop receiving emails from it, send an email to sheffield-hardware-...@googlegroups.com.
To post to this group, send email to sheffield-har...@googlegroups.com.

[Andy Kirby]

With your PSU have you put loading resistors on the 3v3 and 5v lines.

PSU's have a lower loading limit below which they shut down or semi
shutdown. They restart quite quickly to see if the issue has been fixed.
The net result is poor regulation (The smoothing caps hold some of the
voltage up) over large ripple and a bunch of noise. With the knock on
effects that this produces.

Bearing in mind that the Arduino/Ramps combo regulates 12V to 5v there
is enough regulated headroom (7V) that the micro will not drop out or
see what is happening and carry on regardless.

Given that the heaters are PWM therefore switched there are points in
the cycle at which there is next to no load followed by full load.

This can cause problems with SMPSU's that are more sensitive. It does
vary from PSU to PSU even under the same make/model.

Putting chunky power resistors on and loading up the 3v3 and 5v rails
stops this occurring as there is a constant load on the rails it is
sensing. It is wise to heatsink the load resistors as they can get hot
if their heat is not dissipated.

The 12V is very rarely used for sensing in the more modern PSU's (ie
ones with a 3v3 rail) It used to be on the old PSU's though.

Sorry if you already have this one covered.

Cheers

Kirbs

On 16/09/14 10:41, nop head wrote:
> I think the B:@127 is a display issue. I think PWM is actually 7 bits
> although it is set in the config in 8 bit format. So for 100% you set
> 255 in configuration.h but it displays 127.
>
> 1.8R is actually a bit high for a 12V bed. The Prusa Mk2 design is 1.2R
> for 10A.
>
> Adding an Amp meter in series is likely to add significantly to the
> resistance and give a low reading. A DC current clamp is more accurate.
>
> The resistance of copper increases with temperature so at 100C the
> current falls to about 70% of the starting value.
>
> On 16 September 2014 08:43, Stephen Toms <stevet...@googlemail.com

> http://www.rapidtables.com/__calc/electric/ohms-law-__calculator.htm

> sheffield-hardware-h...@googlegroups.com
> >
> <mailto:sheffield-hardware-__hackers+unsubscribe@__googlegroups.com>.

>
> > To post to this group, send email to

> > sheffield-har...@__googlegroups.com
> > <mailto:sheffield-hard...@googlegroups.com>.
> > Visit this group at
> > http://groups.google.com/__group/sheffield-hardware-__hackers
> <http://groups.google.com/group/sheffield-hardware-hackers>.
> > For more options, visit https://groups.google.com/d/__optout
> <https://groups.google.com/d/optout>.

>
> --
> Andy Kirby
> http://www.kirbyand.co.uk/

> http://www.kirbyand.co.uk/__serendipity/

> <http://www.kirbyand.co.uk/serendipity/>
>
> --
> You received this message because you are subscribed to the Google
> Groups "Sheffield Hardware Hackers" group.
> To unsubscribe from this group and stop receiving emails from it,
> send an email to
> sheffield-hardware-...@googlegroups.com

> <mailto:sheffield-hardware-...@googlegroups.com>.

> To post to this group, send email to
> sheffield-har...@googlegroups.com

> <mailto:sheffield-har...@googlegroups.com>.

>
> Visit this group at
> http://groups.google.com/group/sheffield-hardware-hackers.
> For more options, visit https://groups.google.com/d/optout.
>
>
> --

> You received this message because you are subscribed to the Google
> Groups "Sheffield Hardware Hackers" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to sheffield-hardware-...@googlegroups.com
> <mailto:sheffield-hardware-...@googlegroups.com>.

> To post to this group, send email to
> sheffield-har...@googlegroups.com

> <mailto:sheffield-har...@googlegroups.com>.

[Stephen Toms]

Hi Nop and Andy,

Thanks for the replys and info.

Nop, thanks for the hint that 1.8 ohms was a little high. As I'd never tried to test something so low value before I hadn't considered any loss in my probes and when I shorted then I got a reading of 0.5 so I assume the actual reading of the bed is closer to 1.3 so much closer to spec. The PWM value being displayed as a 7-bit value seems to make sense as well - interesting.

Andy, I did have power resistors providing a dummy load (10 ohms and 4.7 as I recall) on the 5v and 3v3 lines and both were also tied into their various sense lines, but thanks for asking as that really could have been a "wood for the trees" moment. I think the main issue with the supply was that it was a bargain i couldn't resist. Won it in an auction on eBay for £7 as a factory recon unit from somewhere I've have another ATX from in the past. Superficially (both inside and out) it looks great but obviously there is something that didn't get re-conditioned as well as it should have! A project for another day I think ;)

The printer is up and running now, and I am chuffed with the quality of the stuff that it is already producing - always room for improvements though!

Regards,

Steve