5

I just recently upgraded my Printrbot Simple Metal with a heated bed (and longer x-axis). I looked up some tutorials, and all of them placed the thermistor as in the picture on the left below, so I did too. However, the design of the heat plate seems to strongly suggest thermistor placement as in the picture on the right, inside the small hole near the center.

enter image description here enter image description here

I can certainly see the upside of that. The reported temperature may overshoot the average bed temperature (the reasoning used in the tutorials I read), but most of my prints are built in the center of the bed, and the center placement would surely make the temperature control system more responsive.

What are the pros and cons of these placements? And what would be the proper technique for putting the sensor in the center? Should I still use kapton tape? Do I need to make sure the thermistor makes physical contact with the aluminum of the print bed?

mhelvens
  • 941
  • 2
  • 12
  • 21

2 Answers2

2

Depending on thermistor type you can get values from around 1ohm/C to 0.01ohm/C so placing it on particulat place on heatbed (HB) could be very important. If objects are placed in the center then it's quite obwious that thermistor should be placed also in the middle.

Another issue in terms of thermistors are the wires. Because thermistors can be so sensitive then wires should have as less influence as possible on its temperature capacity. That's why wires and thermistor legs are so thin. Generally thiner than regular element's legs.

Answering your question - I would say it's better to put thermistor in the center hole, to use thermal grease in the hole so thermistor would have contact with HB. And personally I would use special thin wires to connect thermistor to the board. Kapton tape is then usefull to position thermistor head in the hole and in the thermal grease. Have a look on the picture.

enter image description here

I have such configuration. (In fact I got additional insulation as mentioned in this thread).

I would not recommend to use circuit board tracks as a thermistor connectors at all, but it's my opinion of course.

[edit]

Please be careful - thermal grease can be electrical conductor or insulator. You should choose insulator because short circuit thermistor legs will cause measuring impossible.

[edit2]

As you know HB is not even necessary when you print PLA so in this case it can be even better to not have heating at all than to overheat the bed. Because if you put your thermistor near to the edge of HB then (as said by Leo Ervin) the difference can be at level of 20C so if you heat around 70C and you get 90C your model can become soft on the bed surface and it will loose dimensions or you'll get elephant foot.

As stated in the comments - it's not really big deal whether you measure the temperature in the center of somewhere around. The issue is what level of perfection is ok for you.

Community
  • 1
darth pixel
  • 3,438
  • 1
  • 11
  • 20
  • Thanks for the advice. But why would you not use the circuit board tracks? They are obviously there for that very purpose. – mhelvens Jun 07 '16 at 16:42
  • 1
    I disagree with this answer. It's nowhere near this critical, and thermistors are by far not this sensitive. Using the bed tracks is perfectly fine. The purpose of giving a thermistor very thin legs is so that it has less heat capacity (and thus responds to temperature changes more quickly) but given the bed as a whole already has an extremely large heat capacity, anything the thermistor's leads add is completely irrelevant. – Tom van der Zanden Jun 07 '16 at 19:12
  • @TomvanderZanden I'm a bit confused about what you say. You do confirm that legs are thin to reduce heat capacity, which implies it's is relevant parameter but then you say it's fine to stick thermistor to heating element and you say it's not relevant. It's contradiction, isn't it? ;) – darth pixel Jun 07 '16 at 19:26
  • It's not a contradiction. The legs are thin so the thermistor can respond quickly when the temperature is changing quickly. As the temperature is not changing quickly in this application (the heat capacity of the bed is too large for the temperature to change quickly), whether the legs are thin or not becomes irrelevant. – Tom van der Zanden Jun 07 '16 at 19:40
  • But hot legs and hot wire can be significant component of a measurement. Thermistor wil not measure where it's supposed to. And again you've said it has thin legs (and wires) to respond quickly but you suggest to connect it to thick and capacious tracks (which are hot in fact). IMO this will have influence on the process (of measuring). – darth pixel Jun 07 '16 at 19:56
  • 1
    The temperature of those tracks is exactly what you're trying to measure. Well, strictly speaking you want the temperature of the bed surface but the two are coupled closely enough that it doesn't matter. Many people just place the thermistor on the underside of the heated PCB (with the PCB separating the thermistor from the build surface) and seem to have no trouble with that. You don't need measurements that are accurate to 0.1 of a degree. – Tom van der Zanden Jun 07 '16 at 20:04
  • I would use a heatsink plaster instead of thermal grease. Plaster solidifies and sticks to what it is applied to, grease doesn't. – Leo Ervin Jun 07 '16 at 20:06
  • The tracks or wire only matter because (a) they contribute a little resistance to the circuit, and (b) that resistance changes with temperature. But that resistance is in the ballpark of an ohm, and copper changes resistance by about 0.39% (that is, 0.0039) per degree C. The thermistor is about 100,000 ohms. Thermistor resistance change varies a lot, but is typically 100 to 1000 ohms per degree C. A change of 0.39% of 1 ohm in a circuit changing by hundreds of ohms per degree C simply doesn't matter – TextGeek Jun 08 '16 at 15:16
2

It doesn't make a great deal of difference where you place the thermistor; do whatever is most convenient for you. The temperature of the heated bed is not very critical, and some inaccuracy in the measurement is perfectly fine. The point is to keep the print from cooling down too fast, and whether the bed is at 95C or 93C makes little difference.

The pros and cons of thermistor placement are exactly as you describe: in either case the center of the bed will be hotter than the edges. Placing the thermistor in the center the temperature of the center will equal the setpoint, while the edges are cooler than the setpoint. Placing the thermistor at the edge the edge will be at the setpoint while the center will be hotter than the setpoint. Placing the sensor at the center you would need to use a slightly lower setpoint to get the same result compared to having the sensor at the edge.

It is possible the holes in the PCB are meant for if the PCB is used with a different printer. In any case, soldering the thermistor to the pads and placing it in the hole is acceptable, kapton tape wouldn't be required then (but can't hurt either). It would be best if the thermistor made good contact with the heated bed surface, using (non-conductive!) thermal paste would be convenient for this (but not strictly required).

Tom van der Zanden
  • 14,588
  • 2
  • 33
  • 60
  • Yep, the difference of temperature from the edges to the center is around 20C, however in the center area it is pretty much the same. Here are some thermal images to prove this: http://www.sd3d.com/not-all-heated-beds-are-created-equal/ So unless you decide to put the thermistor very far from the center, it doesn't matter. – Leo Ervin Jun 07 '16 at 20:15
  • How fast do temperature changes in the center propagate to the edges? My main argument for placing the thermistor near the center would be that its reading will more quickly reflect changes from the heat plate, making the whole control system more responsive, resulting in the bed keeping its temperature more consistent. – mhelvens Jun 07 '16 at 20:37
  • @mhelvens That's where your PID tuning comes in. If you place the thermistor near the edges, the system will have a greater tendency to oscillate (by the time the edges reach the setpoint the center will have overshot by a great amount) but that's exactly what PID tuning is for: the tuning accounts for the fact there is some delay. – Tom van der Zanden Jun 08 '16 at 06:13