IRLZ44N MOSFET big voltage drop with DC-brushed motors, LiPo 3.7 V

Other Hardware Robotics
1

My project is a mini-drone controlled with Arduino nano over PWM-> mosfet config. I know my Schottky diode is an overkill. Other thing i think is that too much energy is wasted on MOSFET.

EDIT Two types of measurements are posted below. The first are following this scheme and are most likely wrong (comments).
Second measurements have Amp meter on the drain, directly between VCC and electric motor.

image
image720×576 61.4 KB

Datasheets and links
IRLZ44N
SB840
(sorry, new user == two links max, motor and lipo links in comment

Measurements during 100% duty cycle:
Headings define type of component(s) wired between battery and D of MOSFET.
|| means parallel

10k resistor

  • V_LiPo = 3.71 V
  • V_R = 3.71 V
  • I = 0.310 mA

10k resistor || SB840 (SB840 = Schottky)

  • V_LiPo = 3.7 V
  • V_R_SB = 3.69 V
  • I = 0,367 mA

ELM720

  • V_LiPo = 4.1 V
  • V_ELM = 2.3 V
  • I = 2.68 A

ELM720 || SB840 (ELM = electric DC brushed motor, 720 are dimensions)

  • V_LiPo = 4.2 V
  • V_elm_sb = 2.2 V
  • I = 2.5 A

ELM720 || SB840

  • V_LiPo = 3.62 V
  • V_elm_sb = 1.8 V
  • I = 2.2 A

ELM8250 || SB840

  • V_LiPo = 3.91 V
  • V_elm_sb = 2.45 V
  • I = 2.08 A

ELM8250 directly on battery (no mosfet and diode)

  • V_LiPo = 4.17 V
  • V_elm = 3,16
  • I = 2,8 A

ELM720 directly on battery (no mosfet and diode)

  • V_LiPo = 4.15
  • V_elm = 2.99 V
  • I = 3.3 A

EDIT New measurements following recommended instructions (measuring current on drain, Amp meter between VCC and ELM). I can't post new scheme, new user
ELM 720 || SB840

  • V_LiPo = 4.11 V
  • V_elm_sb = 2.11 V
  • I = 2.5 A

ELM 720 directly on battery (no mosfet and diode)

  • V_lipo = 4.15
  • V_elm = 2.99
  • I = 3.3 A

ELM 8520 || SB840

  • V_lipo = 4.1
  • V_elm_sb = 2.45 V
  • I = 2.3 A

ELM 8520 direclty on battery (no mosfet and diode)

  • V_lipo = 4.07
  • V_elm = 3.14 V
  • I = 2.67 A

EDIT 3rd measurements. Turning on the mosfet directly from VCC (lipo battery).

ELM8520 || SB840

  • V_lipo = 4.04 V
  • V_elm_sb = 2.5 V
  • I = 2.13 A

ELM8520 (no schottky)

  • V_lipo = 4.05
  • V_elm = 2.54 V
  • I = 2 A

edit 4th measurements. Turning on the mosfet directly from voltage regulator set to 7 V. It can output 2 A max

ELM8520 || SB840

  • V_lipo = 4.03
  • V_elm_sb = 2.42 V
  • I = 2.29 A

What i think is the problem is that R_DS(on) of the IRLZ44N MOSFET is too big for chosen motors. Or is this normal and i will have a hard time finding a better MOSFET?

I looked up for the lowest R_DS(on) and V_GS(th) around 2 V MOSFET but then R_DS(on) is measured with some obscure currents like "R_DS(on) = 1.6 mOhm @ V_GS = 4.6V && I_D = 172A.

I need recommendations on how to make my circuit more efficient. How can i choose a better MOSFET and schottky?
How can i calculate or at least predict the R_DS(on) when only 3 A would be "flowing"?

Thank you very much in advance. TTŠ

2

Missing links
Motors
LiPo

3

The R of your Ameter is higher than that of the MOSFET. Put that meter in between the battery and the motor. Then you can exclude that voltage drop in your further measurements.

1 Like
5

Thank you DrDiettrich, i will edit the post with new measurements.
EDIT - i added new measurements

6

At 3A of current, the Vds of an IRLZ44 is less than 0.2V which is roughly 0.07ohms Rds. I don't think the FET is the problem. If you read the datasheet carefully, you will see that the switching behavior charts were created with a gate resistor of 4.6ohms. At a 5V drive, that gate current is considerably more than the 40mA you have coming from the Arduino.

If you have a scope, put it on the Drain terminal and I'm sure you'll see that it's a slowly changing slope instead of a quick on/off edge.

tl;dr: you need a stronger gate drive to get good switching behavior.

1 Like
7

Now i finally understand how to interpret this graph.
If i understand this correctly, there are 0.2 V wasted on MOSFET when V_GS = 2.5 V @ I_D = 3A (red dot).

image
image1227×1064 66 KB

And for example, if i had V_GS = 4 V (purple dot) there should be practically no energy wasted on MOSFET @ I_D = 3A?

EDIT: So probably MOSFET drivers would solve my problem?

8

Why bother with Vgs less than 5V, the operating voltage of your Nano?

9

Try this: Connect the source directly to ground. Measure the voltage on the gate, it needs to be 4V or greater. Put the amp meter in the drain lead. Reason your amp meter has an internal resistance that would raise the source voltage lowering the Gate Source voltage to a non acceptable and make the circuit prone to oscollation.

1 Like
10

It was just an example to check if i underatand this graph correctly.
i have voltage regulator set to 7 V, powering nano, so nano output is exactly 5 V and around 40mA.

11

I added 3rd measurements, where i turn the mosfet on directly from lipo battery VCC.
They dont seem that different from arduino driven ones

12

Check out 2nd, 3rd and 4th measurements please. All are made with amp meter on the lead on drain

13

Where???
Please do not go back and edit old posts with new data.
Your post #1 is a now a mass of confusing information.

You have got the gnd of the Lipo connected to the gnd of the Nano?

NANOMOTOR
NANOMOTOR1197×845 98.4 KB

Can you please post an image(s) of you project, so we can see your component layout?

IN A NEW POST.

What are .. ELM720, and ELM8250?

Thanks.. Tom.. :smiley: :+1: :coffee: :australia:

14

The issue is that when you are applying PWM, the gate is being switched on and off and you need enough current to do that quickly. This problem will not occur at DC!

15

Maybe this Video is helpful:

1 Like
16

IMG_20230122_191908
IMG_20230122_1919081920×864 151 KB

This is how things look on protoboard (don't worry, sheets are non sintetic and i moved things here because of the light source)

My scheme looks exactly like the one you've made. Btw. can you please tell me which program you used to draw that? Thank you :slight_smile:

My Arduino ground and LiPo ground are connected with blue wire on the edge.
Green wire goes from pin 9 to 220 Ohm and 10k Ohm resistors. The 220 goes to gate, 10k goes to gnd. Positive pole of huge schottky diode is connected to red wire, negative to the black wire and drain of mosfet. Red and black from schottky go to electric motor.

ELM = dc electric motor
ELM8520 = 8.5x20 mm electric motor dimensions
ELM720 = 7x20 mm electric motor dimensions
SB840 = schottky diode SB840
|| = in parallel

Here are measurements made based on the scheme you've drawn (amp meter on the drain lead):

ELM 720 || SB840

  • V_LiPo = 4.11 V
  • V_elm_sb = 2.11 V
  • I = 2.5 A

ELM 720 directly on battery, through protoboard (without mosfet and diode)

  • V_lipo = 4.15
  • V_elm = 2.99
  • I = 3.3 A

ELM 8520 || SB840

  • V_lipo = 4.1
  • V_elm_sb = 2.45 V
  • I = 2.3 A

ELM 8520 direclty on battery, through protoboard (without mosfet and diode)

  • V_lipo = 4.07
  • V_elm = 3.14 V
  • I = 2.67 A

Measurements turning mosfet on directly from VCC lipo battery:

ELM8520 || SB840

  • V_lipo = 4.04 V
  • V_elm_sb = 2.5 V
  • I = 2.13 A

ELM8520 (no schottky)

  • V_lipo = 4.05
  • V_elm = 2.54 V
  • I = 2 A

Measurements turning on the mosfet directly from voltage regulator set to 7 V. (Regulator can output 2 A max):

ELM8520 || SB840

  • V_lipo = 4.03 V
  • V_elm_sb = 2.42 V
  • I = 2.29 A

And sorry if editing original post made everything unreadable and confusing. Im learning.

17

Another pic of regulator

IMG_20230122_191840
IMG_20230122_1918401920×864 69.1 KB

18

How battery comes in

IMG_20230122_194650
IMG_20230122_1946501800×4000 498 KB

19

I have huge findings to share with everybody.

ELM8520 directly on battery connected through protoboardb:

  • V_lipo = 4.06 V
  • V_elm = 2.65 V
  • I = 3.94 A

ELM8520 directly on battery bypassing protoboard:

  • V_lipo = 4.05 V
  • V_elm = 3.55 V
  • I = 5 A

Protoboard is screwing with me so hard. I need to solder things together and then see whats actually going on. I'll keep you guys updated

20

Almost. mA, current, is actually determined by the load. The source of power can supply a limited amount of current. If you draw more current then available things will start to misbehave, if you pull less all is ok. Many CMOS based devices also have a surge rating (very short time) that they will supply more then the steady state rating. Turning on a MOSFET, which is a voltage controlled device you need to exceed its on threshold. The MOSFET also appears as a capacitor and only sources or sinks current during the switching time, on or off. When either state you may have a few piko amps or basicly nothing.

When the gate is connected to the microprocessor it will only draw current for a very short time, while it is switching. It is also good practice to place some relative high value resistance, 10K is popular, from the port pin to ground. This guarantees that the N-Channel MOSFET will be off during reset and set until the Arduino can gain control of its GPIOs.

1 Like
21

The problem was in bad battery connectors and thin wires. Once everything was moved to stripboard, soldered together and higher quality connectors were used on the battery, the problem was resolved.