Best thermistor location in heated bed (thermal imaging)

[Hansen]

Having issues with my Rev5 heated bed not being able to reach more than 80-ish degrees, I decided to try something fun. I have access to a thermal imager, and I used it to determine 1) the real temperature of the bed and 2) see if the heat is evenly distributed over the entire bed face. I heated the bed to 60 degrees and took this picture:

I used the predrilled hole in the center for my thermistor, but it is seen that the center is actually "cold" (~55 deg.) relative to 4-5 inches away from the center (max. ~61 deg.). I guess three conclusions can be drawn from this:

1. The temperature across the entire face of the heated bed varies 5-6 degrees with the center being "cold".
2. The thermistor measurement is "off" by about 5 degrees (the printer thinks the bed is 60 degrees, but it is actually 55 at the location where the thermistor is located).
3. The best (average) temperature reading is achieved by placing the thermistor in the "green zone" in the picture, which is about 2 inches from the center (I believe as already suggested somewhere in this forum - You would have to drill a new hole though avoiding the tracks in the PCB).

I am somehow lucky, because the error introduced by "2." makes up for the error in not placing the thermistor at the correct (average temperature) location

Have a nice day
Hansen

[guanu]

yup, you must have got one from the first round before we found out the same thing and created a second different hole for the thermistor....

see http://forum.seemecnc.com/download/file ... &mode=view for the new thermistor location... and rev 6 coming here soon for testing also.

Guanu

[Hansen]

It is always good when two parties come to the same conclusion independently

I do still have the issue with not being able to reach bed temperatures above some 80 degrees, so I think I'll have to contact SeeMeCNC though

[RegB]

yup, you must have got one from the first round before we found out the same thing and created a second different hole for the thermistor....

see http://forum.seemecnc.com/download/file ... &mode=view for the new thermistor location... and rev 6 coming here soon for testing also.

Guanu

Ahh yes,
I just got my kit, peeked through the plastic to see if I have a rev 5 and saw that I have a rev 6 - - Wadda SURPRISE ?
I hadn't expected yet another rev level.
Is there any change to the assembly & set up ?

[guanu]

nope, same solder pads as rev 5... just different heating element design... I brought a rev 6 here at the shop up to 120 in under a half hour and it held it for a few hours no problem

Guanu

edit: I take that back, the only real difference between rev 5 and rev 6 for soldering is that you dont need the + wire for the heated bed LED, its tied into the power of the bed, and since all + are shared on the rambo, its not needed anymore, so doesnt matter if its hooked up or not, but the - of the hot end led is still required...

the thermistor can go in the center of rev 6, does not need the extra drilled hole like rev 5

[geneb]

GAAAAAAAAAAH!

*headdesk*

g.

[guanu]

theres no problem gene, the solder pads look identical to rev 5, so theres no need for anything different, if people put on the + for hot end led, it will make no difference, so theres no changes to how it can be soldered and hooked up.. all is good...

[geneb]

Hrm. Ok.

g.

[RegB]

I opened the onyx package, not that I'm there yet, just curious.
There is only ONE thermistor hole, so that is simple enough.
Presumably the cold center spot is fixed with the new heating wire track, so the thermistor can go back in there ?

I was about to peel the backing paper off the snowflake when I noticed that one of the "ears" has a different profile to the other five.
If it had been done with power tools it would be a router with maybe a 1/4" bit - result is a double cutaway in guitarspeak.

If this is to clear something, e.g. wiring, I think we should know which mounting hole to align that particular "ear" to.
So, maybe a manual tweak for a different reason than the rev 6 ?

BTW, I received a 550Watt power supply. Is this because the new Onyx draws more current ?

[Hansen]

BTW, I received a 550Watt power supply. Is this because the new Onyx draws more current ?

Can you, by any chance, measure the resistance of the heating element in the Onyx? It can be difficult, because the resistance is probably lower than the internal offset of your measurement device, but you could:

1. Measure the offset by placing both probes onto one of the solder pads.
2. Measure the resistance by moving one of the probes to the other solder pad.
3. Subtract the offset from the resistance measured.

I am one of the people that are having difficulties reaching bed temperatures above some 80 degrees, so I am very interested in the result, because it could mean that your Onyx does indeed draw more current than my Onyx (rev. 5 - approx. 0.8 ohms).

[RegB]

I will try, but my equipment is limited to hobby level tools these days.

I think the real answer can only come from SeeMeCNC.
If they tell us directly WHY rev 6 and WHY 550 Watt supply - these may be only loosely coupled (-:

I was speculating based on 1) guano's statement about a rev 6 board coming up to 120 in 1/2 hour
2) my 550 Watt supply, I had expected the 450 Watt version.

[RegB]

About 0.5 to about 0.6 with the standard disclaimer about Harbor Freight precision/quality tools - - we used to say Heath Kit, or Radio Shack (-:

So, if I can remember how to do this;
0.8 Ohms draws 12 Volts/0.8 Ohms = 15 Amps for a 180 Watt draw.
0.6 Ohms draws 12 Volts/0.6 Ohms = 20 Amps for a 240 Watt draw.
Ignoring internal resistance of the supply, which is OK if it is reasonably well regulated.

Guess, guess, guess... That might well justify the next size power supply and would allow significantly faster heating.
Cold spots ? I don't know, that probably depends more on how evenly the heating wire track is spaced across the area - spaced with respect to heat loss.
Edges and center are likely to dump more heat per sq cm., so need closer spacing.

[Eaglezsoar]

BTW, I received a 550Watt power supply. Is this because the new Onyx draws more current ?

Can you, by any chance, measure the resistance of the heating element in the Onyx? It can be difficult, because the resistance is probably lower than the internal offset of your measurement device, but you could:

1. Measure the offset by placing both probes onto one of the solder pads.
2. Measure the resistance by moving one of the probes to the other solder pad.
3. Subtract the offset from the resistance measured.

I am one of the people that are having difficulties reaching bed temperatures above some 80 degrees, so I am very interested in the result, because it could mean that your Onyx does indeed draw more current than my Onyx (rev. 5 - approx. 0.8 ohms).

I believe the switch to the 550 Watt power supply is because the manufacturer of the power supply quit making the 450 Watt unit, forcing Seemecnc to use a 550 Watt.

[Hansen]

About 0.5 to about 0.6 with the standard disclaimer about Harbor Freight precision/quality tools - - we used to say Heath Kit, or Radio Shack (-:

So, if I can remember how to do this;
0.8 Ohms draws 12 Volts/0.8 Ohms = 15 Amps for a 180 Watt draw.
0.6 Ohms draws 12 Volts/0.6 Ohms = 20 Amps for a 240 Watt draw.
Ignoring internal resistance of the supply, which is OK if it is reasonably well regulated.

Guess, guess, guess... That might well justify the next size power supply and would allow significantly faster heating.
Cold spots ? I don't know, that probably depends more on how evenly the heating wire track is spaced across the area - spaced with respect to heat loss.
Edges and center are likely to dump more heat per sq cm., so need closer spacing.

You are absolutely right on the power calculations. However, the 12V line drops about 1 volt, when it is loaded, because this kind of power supply expects the 5V line to be loaded, which it isn't by default. I've loaded the 5V line with some fancy LED lights, so the power supply is somewhat more stable. This is besides the point... it seems your rev. 6 Onyx has a lower internal resistance than my rev. 5, so it will draw more current and get hotter. Of cause there will be some tolerances on both your and my resistance measurement, but it seems like the higher resistance is indeed the problem with my Onyx.

Regarding the 550 W PSU: Mine is the "old" 450 W, and it can provide 28A from the 12V line. I believe the current consumption from a 0.6 ohm Onyx will be too close to the PSU limit (the hot end requires about 3.3 A, and then we have the steppers in addition to that). So, I could be wrong, but I think SeeMeCNC made an active choice, when they choose a bigger PSU with the new Onyx.

It will be interesting to know how they will solve my and others problems with the Onyx, if a new PSU is required in addition to the new Onyx. I wrote them a couple of days ago, but I have yet to hear from them

[RegB]

Tangent alert;
I suspect that I will add a light ring at some point...
IF that is made of LEDs and IF those draw from the +5 Volt line then that "might" help to stabilize the +12 Volt supply.
All Iffy/maybe (-:

Rolling changes, ongoing improvements, retroactive change fixes for bugs and original shortcomings, etc.
Interesting questions indeed.

[RegB]

It turned out that the "odd ear" in the snowflake is to clear the two LEDs, so not new to V6, but as far as I know undocumented for V5.
Also, the diagram Fig 4-37 shows Tee nuts at 1, 3, 5, 7, 9 and 11 O'Clock.
They changed that a week ago to 2, 4, 6, 8, 10 and 12 O'Clock, they are now at and half way between the towers.

Bahh, the trouble with that is that the 6 O'Clock one fouls the front support.
I have left it out for the time being, but I may grind an edge off it and put it back.

Danggit this mechanical assembly stuff is SLOW )-;
When do I get to something "electronic" ?

[Hansen]

Tangent alert;
I suspect that I will add a light ring at some point...
IF that is made of LEDs and IF those draw from the +5 Volt line then that "might" help to stabilize the +12 Volt supply.
All Iffy/maybe (-:

Rolling changes, ongoing improvements, retroactive change fixes for bugs and original shortcomings, etc.
Interesting questions indeed.

I found that I needed to draw about 1-1.2 amps from the +5 supply, for the 12V supply to drop only 0.3-0.4 volts instead of ~1.0 volts. So to eliminate the voltage drop entirely, you would have to load the 5V with waaay to much power for a light ring I only put a few watts of LED lights on my printer, so the 12 V is not completely stable, but it is better than nothing I guess...