English
Login

Heatbed connection resistance analysis (MINI)

Relevant for

MINI
9 comments
Article is also available in following languages
English
Čeština
Polski
Deutsch
Français
Español
Italiano
日本語

We want to thank the community for bringing out potential issues and we address every safety-concern very seriously. We take full responsibility for our machines in case something happens, and we are insured for any such case. Every time we discover concerns from our users on the internet, we sum up our measurements and share them with you.

Summary

MINI’s bed (with 3.6A of current) is significantly smaller than the MK3 (with 7A of current), rendering the power flowing into the bed roughly at ½. The resistance of the full connection, including mounting bolts, are 7 m? for MINI and 4 m? for MK3S. It renders a power loss on the connection of 170 mW for MINI and 430 mW for MK3S (For comparison the resistor for the heatbed LED produces 250mW of heat). This shows that the heat generated at the connector is 2.5 times lower than on the MK3, providing even higher failure safety-margin than before. Terminals are secured with nyloc nuts, which do stay in place even in very rough conditions.

In conclusion, the only aspect we see as troubling is the appearance, which at first sight might look it’s missing the contact pads on the bottom side of the board, which are however not needed, as full connection must happen on the top side of the board. With our experience and safety track-record we stand by the MINI's design and it is safe.

Measurements

Average measured Power loss in the heatbed connection:

MINI is 170 mW, MK3S is 430 mW

Average measured resistance of one heatbed connection terminal:

MINI is 7 m?, MK3S is 4 m?

This fact is because power rises with a square of current and linearly with resistance. The normal running current of the MINI heatbed is 3.6 A, MK3S is 7 A, this makes a huge difference, For comparison, power loss on resistor for LED is 220 mW.

The following thermal-image shows a comparison of the LED resistor with the connection and its surroundings

Due to measured values, we can confirm that there is a slight increase in resistance but due to lower running current, we are certain that it's not an issue. We can confirm that even our oldest MINI printers from the early stages of development do not show any measurable connection resistance increase over time.

We used two different methods of measurements:

  • Standard 4 wire measurement of low resistance using Keysight measuring just the connection resistance

  • Specialized tester for pre-compliance testing of grid-connected devices REVEX plus measuring the whole path of one conductor from the bed connector to the copper pad (Which was showing larger values due to specification of the method and device limitations)

Measured values from method No 2:

 

 

Sample

SCREW

 

SCREW + WIRE

   
 

QR code

VCC / Ohm

GND / Ohm

Power / W

VCC / Ohm

GND / Ohm

Power / W

TOTAL / Ohm

Power / W

MINI

-137

0.004

0.0040

0.1037

0.0160

0.0160

0.4147

5.2400

67.9104

-160

0.008

0.0120

0.2592

0.0150

0.0190

0.4406

5.2100

67.5216

-153

0.004

0.0050

0.1166

0.0160

0.0130

0.3758

5.1600

66.8736

-226

0.012

0.0080

0.2592

0.0200

0.0160

0.4666

5.4100

70.1136

-328

0.005

0.0040

0.1166

0.0130

0.0120

0.3240

5.1600

66.8736

Average

0.007

0.0066

0.1711

0.0160

0.0152

0.4044

5.2360

67.8586

          

MK3S

-205

0.001

0.0010

0.0980

0.0080

0.0120

0.9800

2.7500

134.7500

-130

0.004

0.0040

0.3920

0.0100

0.0120

1.0780

2.7600

135.2400

-331

0.005

0.0080

0.6370

0.0160

0.0130

1.4210

2.7300

133.7700

-333

0.008

0.0050

0.6370

0.0130

0.0200

1.6170

2.7000

132.3000

-319

0.004

0.0040

0.3920

0.0120

0.0130

1.2250

2.6800

131.3200

Average

0.004

0.0044

0.4312

0.0118

0.0140

1.2642

2.7240

133.4760

 

7 comments

Log in  to post a comment
johnathan
Instead of taking time to publish this report, why not just move the connection point to the upper side instead? There is clearly room for the connectors there. 
krolarthur
I would still argue against this just for liabilty concerns. I do not believe this data would be enough to convince a court should a fire be caused due to a machine malfunction. Espically when the heating pad has specific areas to mate the connections.
David

Hi! That's why there is an emphasized note in the assembly guide about the importance of tightening the contacts properly :) If the contacts get hot, the printer will throw "preheat error bed" before starting a fire.

Joseph Tracy
Contacts were as tight as they could go and in the place that was shown to put them, but the printer never sent any Error code and began to melt the casing around the screw. You guys sent me a replacement part and I think I will try putting them directly on the plates this time. 
David K
Even so, I flipped the contacts from down to up, just for sure. It's possible and easy. Because the above mentioned didn't convince me, I'm an electrical engineer :-).That holding of ordinary probes by hand while measuring, as in the pictures, is funny. This way, transient resistances are not really measured if you are interested in the exact value.
Mr 3d Print Wizard
This is fine while the connection is tight but you've not taken the full community concern to heart. You're using a poor material as a conductor. The contact area is too small. The screw will loosen over time faster. You can still use the pads. There is no reason not to use the safest available connection method. Risking users safety isn't a good excuse for this connection method. That and this test doesn't prove anything. Only that the connection is fine when everything else is perfect.
enganear
Heat cycling will cause this connection to fail long before a properly designed joint where the terminal is firmly pressed against the pad. This study, while interesting, is very shortsighted. There is no good reason for the Mini bed termination to be designed this way. It does not serve Prusa well.