0
\$\begingroup\$
  • I'm trying to build a controllable switch for a heavy load (3D printer), controlled by a microcomputer.

  • I don't want to use relay because of slowness, sound and durability. So I've built a switch based on MOSFET IRLB3034 datasheet.

  • And here is my circuit diagram: circuit diagram

The main issue is, the transistor getting more than 150 °C after 20 sec of working at full power.

  1. Vgs = 4.5v or even 11v (that makes no difference) of constant DC
  2. Vds = 1v and rising
  3. Current is about 15A (max of power supply is 20A)

What could be my possible mistake?

\$\endgroup\$
  • 1
    \$\begingroup\$ Have you checked the actual VGS at the pins of the mosfet? \$\endgroup\$ – JG97 Jul 24 '19 at 8:52
  • \$\begingroup\$ Please specify what is your Control signal. Is it constant DC or is it PWM? Also, as @JG97 asked, what Vgs are you seeing? Your Vds of >1v at 5A is 15w, so definitely will get hot. We have to understand why Vds is >1v \$\endgroup\$ – Kripacharya Jul 24 '19 at 9:06
  • \$\begingroup\$ @Kripacharya Measured right now and Vgs=11v exactly and Vds was rising from 0.5v to 1v quickly and more. \$\endgroup\$ – Tolyas Jul 24 '19 at 9:30
  • \$\begingroup\$ Do you have a spare mosfet? Maybe another model? Sounds like yours is broken or bad \$\endgroup\$ – JG97 Jul 24 '19 at 9:33
  • 1
    \$\begingroup\$ See Note 5 on Sheet 2: Pulse width ≤ 400µs; duty cycle ≤ 2%. infineon.com/dgdl/… Seems to apply to all cases where Vgs = 4.5V. Pinouts are show in little tiny print on page 8:1 is Gate, 2 is Drain, 3 is Source. \$\endgroup\$ – CrossRoads Jul 24 '19 at 12:12
0
\$\begingroup\$

I believe the consistent view is that your MOSFET is suspect. If you were doing pwm, then maybe the large C on the input ((10.3nF) was causing problems. But because you are testing at DC this excessive heat should not take place.

Try sourcing the same/ similar MOSFET from a more reliable source.

\$\endgroup\$
  • \$\begingroup\$ Ok, I replaced transistor with the same model but from absolutely-sure-trusted store. And now it's works fine, Vds is 0.4v when Vgs = 4.5v and temperature never risen above 100 °C. So it's acceptable. Now I planning to replace R1 to less resistant to improve it a little. Thanks to everyone for help ^^ \$\endgroup\$ – Tolyas Jul 28 '19 at 15:32
1
\$\begingroup\$

Assuming you've actually measured Vgs at the MOSFET pins (it gets only about 90% of the drive voltage, and the drive voltage is heavily loaded by the LED), then one would tend to conclude that the MOSFET is not actually a genuine part of the type indicated. The certainty would increase to near 100% if the MOSFET was sourced on a platform that hosts bad actors.

The dissipation should be no more than about 0.7W which will get hot without a heatsink, but not excessively so (150°C is excessive). If your traces or wires going to the MOSFET are thin you may also be getting heating from those sources. Tja is 62°C/W so the rise should be the order of 40-45°C. Rds(on) increases with increasing junction temperature, maybe 50%. But not 125°C rise worth.

\$\endgroup\$
  • 1
    \$\begingroup\$ I think the pinout is wrong, pin1 might be gate. Double check the pinout and double check the transistor orientation even if the pinout is correct. \$\endgroup\$ – Justme Jul 24 '19 at 9:01
  • \$\begingroup\$ Yes, actually it's Vds = 11.1v, when supplying 12v. But according to datasheet it should be enough open even on 3v... And I measured temperature on MOSFET package and after 20 sec there was a little smoke. \$\endgroup\$ – Tolyas Jul 24 '19 at 9:02
  • \$\begingroup\$ 3V is pretty iffy, but definitely 4.5V. Do check the pinout as Justme suggests. Normal numbering is 1-2-3-tab = Gate-Drain-Source-Drain (not 100% clear on the datasheet). Though gate to load would not cause it to get hot unless the MOSFET was damaged internally. \$\endgroup\$ – Spehro Pefhany Jul 24 '19 at 9:07
  • \$\begingroup\$ @SpehroPefhany I check the pinout and it's working. If there was a mistake it wouldn't work. But it heating itself like a heater. \$\endgroup\$ – Tolyas Jul 24 '19 at 9:20
  • \$\begingroup\$ Was the MOSFET ordered from an authorized distributor or is it an eBay, Ali or other such find? \$\endgroup\$ – Spehro Pefhany Jul 24 '19 at 9:27
1
\$\begingroup\$

the transistor getting more than 150 °C after 20 sec of working at full power

If I understood correctly, this is being used to control a 3D printer heated bed.

Its not clear from your post what "full power" means, but if it is really on a 3D printer application, I suspect its not on all the time but rather in PWM due to PID control of the bed temperature.

Considering that the gate capacitance of that FET and that the gate resistor is relatively high (1k Ohm), it could be that a lot of the heat you are generating is from turning it on and off (High RdsOn periods).

If that is the case, you can try to lower the gate resistor a bit and/or find a part with lower gate capacitance.

\$\endgroup\$
  • \$\begingroup\$ Yes, I'm issue that in 3D printer, but it controls whole printer as a power button. And "full power" means printer heating bed and hitend simultaneously. \$\endgroup\$ – Tolyas Jul 24 '19 at 12:02
1
\$\begingroup\$

As @Kripacharya said the figure 8 in datasheet shows that I'm operating beyond transistor's capabilities. So that might be the main reason (as far I can't test with suitable MOSFET), so I need to find another one or another way to reach my goal. fig. 8

P.S. Could anyone suggest transistor searching tool for similar cases?

\$\endgroup\$
  • \$\begingroup\$ The drain to source voltage in the figure is not the voltage when turned off, but when current is flowing. The of state would be at 12V but 0 A. In that case you are even with 1v VDS and 15A inside the soa \$\endgroup\$ – JG97 Jul 24 '19 at 17:24
  • \$\begingroup\$ No, the SOA does not come into play here. Big misunderstanding of MOSFET specs. That would only be the case if you deliberately drove the MOSFET into linear region on a DC basis, and even if you did the current would be half and voltage half worst case, since it's a resistive load, so SOA is fine (power dissipation is not, which is what you are seeing). The MOSFET is most likely a counterfeit and does not meet the data sheet worst-case (let alone "typical") specifications. \$\endgroup\$ – Spehro Pefhany Jul 24 '19 at 20:21
0
\$\begingroup\$

The mosfet you picked should be able to handle 15A with no heatsink without any issues.

The datasheet says that at Vgs=10V the Rdson should be 1.4mR. P = I2*R = 15*15*0.0014 = 0.315W. At 20A that would be 0.56W

The datasheet also says that Junction-To-Ambient thermal resistance is 62degC/W. So at 560mW that gives 35degC temperature rise over ambient.

Did you measure the Vds over the MOSFET? At 15A and 1.4mR it should be only 21mV. If you see anything higher, the part is bogus.

btw in your schematics you are running the load current through a jumper (J5). A standard jumper will have a resistance of 20-50mR. So at 15A that will dissipate at least 4.5W - that should be enough to melt the jumper.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.