Resistors and MOSFETs and fire-prevention, oh my

[Burngate]

Just with a quick look, there doesn't seem to be a path to discharge the gate - once it's been switched on via the opto isolator, it'll stay on.

[tlfong01]

Driving Big Motor Mabuchi RS-540SH 12VDC Motor
So now I have become a logic level Power MOSFET expert!
In order to build up my reputation, I decided to test the biggest motor in my junk box, the Mabuchi 12V 7,100rpm , 250mA (only!) motor.
I changed the 37GB motor by the big guy and switched on power, but motor not turning.
I removed the big motor and replaced it by the original small guy. But nothing happened. The solenoid and motor no longer responded. So now I am in big trouble!

Resettable PPTC Fuse vs Circuit Breakers With Manual Reset Button

I found the 1A fuse blown. So yesterday I went to the electronics parts flea market to get a PPTC fuse, to avoid wasting time changing fuse so often.

However the shop there does not sell PPTC, but recommended a black thing with a push button. I have never seen such a thing before, but it is 15 yuan only, worth taking a risk. I bought it and found it a good toy, because is hand push button resettable. I think this is better than pptc fuse for testing power mosfet, because I know if there is overload, while pptc automatically resets, so I won't know if there is an overload.

Another laughable thing is that the shop suggested me to consider a resettalbe thermal fuse. I refused and told them I don't want a fuse when temperature is over limit. But on second thought I think "thermal" means if there is over current, temperature rises and a bimetal strip bends and cuts the current. I knew this long time ago when hacking my rice cooker and saw such a thing, ...

Resettable PPTC (polymeric positive temperature coefficient) Fuse - Wikipedia
https://en.wikipedia.org/wiki/Resettable_fuse

Amazon Circuit Breakers with manual reset button
https://www.amazon.co.uk/Fityle-Thermal ... it+breaker[/url]

...

[tlfong01]

Resettable PPTC Fuse vs Circuit Breakers With Manual Reset Button
However the shop there does not sell PPTC, but recommended a black thing with a push button. I have never seen such a thing before, but it is 15 yuan only, worth taking a risk. I bought it and found it a good toy, because is hand push button resettable. I think this is better than pptc fuse for testing power mosfet, because I know if there is overload, while pptc automatically resets, so I won't know if there is an overload.

IRL540N Testing with 3A Circuit Breaker

Now I have started testing IRL540N.
...

[tlfong01]

... I failed English and so had to do Engineering ...

Oh my goodness. I failed English and so could not do engineering

Critical Thinking and Problem Solving Skills

I often hear that critical thinking and problem solving skills are also important. I am not sure how to develop these skills, perhaps I can try hard to:

..... Criticize others harder

..... Do more problem based learning (PBL)

Which Skills are Most Important on the Job and Which Skills are in Short Supply? - Monica Herk, CED 2015sep22
https://www.ced.org/blog/entry/which-sk ... n-short-su

Critical thinking (“the ability to exercise sound reasoning and analytical thinking to understand and solve work place problems”) and

Problem solving (“the ability to solve problems using one’s skills and knowledge in real-world settings to solve problems that have not been encountered before”)

stand out as the skills that are deemed essential to most jobs and are in relatively short supply.

What Skills Do Employers Say They Want in Online Job Postings? - Monica Herk 2015dec08
https://www.ced.org/blog/entry/what-ski ... b-postings
...

[tlfong01]

Just with a quick look, there doesn't seem to be a path to discharge the gate - once it's been switched on via the opto isolator, it'll stay on.

Gate Driver Design V2.0

Ah, many thanks for catching my careless error. So I have updated the schematic to V2.0.
...

[tlfong01]

Now I have started testing IRL540N.

IRL540N Test Results Bad

For IRL540N, Vgs = 3V, Ids = 0 (not conducting)

For CSD18511KCS Vgs = 3V, Ids = 1.77A (saturated)

I tested one more IRL540N, the results is as bad. Spec says at Vgs = 3V, IDS should be 15A!

This reminds me of Warren Beffett saying: "What you find is there's never just one cockroach in the kitchen when you start looking around,"

Now I have started looking around, and already found two cockroaches. My guess is that there are over 50 cockroaches in the kitchen!

Appendix - Comparing CSD18511KCS with IRL540N

Test Condition
Vcc= 10V, Rl = 5R (20W)

IRL 540N
Vgs Ids
3V 0.0A
4V 1.65A
5V 1.75A

CSD18511KCS
Vgs Ids
3V 1.77A
4V 1.77A
5V 1.78A
...

[tlfong01]

1. You need a load on the gate. If you touch the gate with the 3.3V. It will turn on.
2. Then you need to discharge the gate. ... simply use a gnd wire and touch the gate. It should turn off.

Power MOSFET Preliminary Testing - The Newbie Way

1. Use a current limiting resistor (10K) and super bright blue LED as the "status" load check. This test load is directly connected to the flyback diode (not to the terminals to connect to the real load (solenoid, motors etc). (See the status led/ flyback diode connection in the schematic below. Note that (1) the flyback diode is connected "before" the Tymiya female connector, (2) the status LED is connected "after" the flyback diode.) This is important, to make sure that if the flyback diode is accidentally disconnected, the "canary in a coal mine" won't sing. And the canary lives there forever, or until the coal mine is closed.

2. I start testing only with the LED as the fake load. No real load like solenoid or motor, which generates scary back EMF, EMI etc, will be entertained for now.

3. I connect a jumper wire directly to the gate (no initial current rushing in limiting resistor Rg 200Ω for now). I hold by hand the other end of the jumper wire and touch it first to 5V, then ground (discharging Cgs, not using discharging resistor Rgs 10kΩ to 50kΩ for now), then 3V, then ground. This is the first thing first test making sure if the mosfet can switch on the fake load at Vgs = 3V, 5V, and switch off at ground.

4. IF AND ONLY IF everything goes well with the fake load (LED) , then I will move on to test the real load (solenoid).
...

[tlfong01]

Which Skills are Most Important on the Job and Which Skills are in Short Supply? - M Herk, CED, 2015
https://www.ced.org/blog/entry/which-sk ... n-short-su
Critical thinking and Problem solving stand out as the skills that are deemed essential to most jobs and are in relatively short supply.

Power MOSFET IRL540N Cockroaches - A Case Study of Critical Thinking and Problem Solving

/ to continue, ...
...

[tlfong01]

Your analysis is flawed.

the rise time of the digital output is around 1.6ns

That's only true for an unloaded GPIO pin. Assuming the GPIO is ideal is nonsense, and leads to nonsense results like it supplying 2.8A to charge the 1350pf to 3.3V in 1.6nS. That just isn't going to happen !
The GPIO pin is a voltage source which has a maximum source current as well, so you can easily workout a realistic rise time (it will be much longer that 1.6nS).
If sufficient resistance is added between the GPIO pin and the gate, such that the peak current is less than the upper limit, then the GPIO pin can be modelled as a simple voltage source making rise/fall time calculations even easier as it's a simple RC circuit driven from a voltage source.

Power MOSFET Initial Gate Capacitance Current

Yes, I agree with @PeterO that @Brandon92's analysis is flawed. It is not necessary to consider any rise time. We can just treat it as a simple RC charging circuit.

The flaw is not considering Ri, the internal resistance of the voltage source (GPIO). In other words, the initial charging current Imax is at most 3V/Ri, and should be much, much less than 2.8A!
...

[tlfong01]

IRL540N Test Results Bad

Power Electronics Heavy Current Heat Problem

I also found the 5R 20W power ceramic cement resistor becomes very hot when passing only 1.8A. Cracks are formed at the base of the resistor. The touching connecting wire's plastic sleeve got burnt to charcoal. I worry that ifI don't pay any attention, sooner or later something would catch fire. So I am replacing the ceramic thing by a 100W metal aluminium cased resistor.
...

[tlfong01]

Bad IRF540N Still Can Pass 1.7A when Vgs = 5V

The cheap 1.2 yuan IRL540N, though not conducting at Vcc = 12V Vgs = 3V, can still pass 1.7A at Vgs = 5V. So it is still usable, with a gate driver such as optocoupler EL817C.
...

[tlfong01]

Bad IRF540N Still Can Pass 1.7A when Vgs = 5V
The cheap 1.2 yuan IRL540N, though not conducting at Vcc = 12V Vgs = 3V, can still pass 1.7A at Vgs = 5V. So it is still usable, with a gate driver such as optocoupler EL817C.
...

10 Bad IRL540N Guys Testing Results

I tested 10 bad guys and found them still useful, if Vgs > 4V.
...

[tlfong01]

Bad IRF540N Still Can Pass 1.7A when Vgs = 5V
The cheap 1.2 yuan IRL540N, though not conducting at Vcc = 12V Vgs = 3V, can still pass 1.7A at Vgs = 5V. So it is still usable, with a gate driver such as optocoupler EL817C.
...

The good and the bad IRL540Ns

The day before yesterday I unhappily found the cheapy 2.1 yuan IRL540N seems bad, not conducting at Vgs = 3V, I knew I was in trouble. I thought of two remedies:

(1) Check if I can upcycle the bad guys, using a gate driver to drive the bad guys at 7V to 12V.

(2) Order 25 genuine pieces at 3 to 4.5 yuans each, and check out it they indeed meet
the spec: conducting at Vgs = 3.0V.

The good guys arrived this evening. I tested one and found it good, with the following results.

Vgs Ids
3V 1.697A
4V 1.700A
5V 1.701A

In other words, the device seems saturated at Vgs = 3.0V, as specified. I noticed that the tube marked Mexico. So my quick conclusion is the Southern American country is better quality than the Asia mighty country.

Appendix - IRL540N Order

IRL540N new guinene quality assurance ¥3.50 each
https://detail.tmall.com/item.htm?_u=n3 ... 44071268

2019-02-22 11:37 ordered and paid
2019-02-22 20:31 good dispatched
2019-02-23 19:45 arrived Castle Pead Road SSP Collection Point (23 hours)

[tlfong01]

If the fet is getting hot then it is probably in the analog region.
A higher gate voltage turns it on more and so the resistance goes lower and hence heat goes down.

theitguyfromny » 2019-Feb-08 Fri 9:15 am
The FET was getting real hot before (too hot to touch), but not w/ the set up I have now, I don'think. I'll have to check again. I think it's because I was keeping the solenoid in its'open, non-default, position.
I have to believe there's ways to keep FETs cooler..but that's a different thread.

danjperron » 2019-Feb-08 Fri 9:39 am
If you read his last post it is not getting hot right now and this mosfet could drive up to 20A at 3V according to the speck.

Power MOSFET Finger Untouchable Hot

This morning I started testing 30 power MOSFETS:

.....(a) cheapy 1.2 yuan URL540N x 10,
.....(b) IRF540N (forgot price) x 10,
.....(c) IRL540N (high class, genuine, no fake guarantee! , 3.5 yuan each).

In the middle of the boring job, I fell asleep, and mistakenly inserted the chip bottom up, instead of top up into to KF301-3P screw terminal block connector (G, S swapped!) . The chip did not conduct at Vgs = 3, 4, or 5V. Then I corrected my careless mistake and try again. This time I had a pleasant surprise - the chip got saturated even at 3V!, also at 4, 5V. All other weak guys still do not conduct at Vgs = 3V. So the weak guy has become a good, guy!

/ to continue, ...

What Doesn't Kill You Makes You Stronger - Conan The Barbarian
https://www.youtube.com/watch?v=CfwXP7qj0E8
...

[tlfong01]

This morning I started testing 30 power MOSFETS:
.....(a) 1.2 yuan URL540N x 10,
.....(b) 3 yuan IRF540N x 10,
.....(c) 3.5 yuan IRL540N x 10.

3.5 yuan IRL540N x 10 Results Good - Rpi GPIO 3V3 signal OK

...

[tlfong01]

This morning I started testing 30 power MOSFETS:
.....(a) 1.2 yuan URL540N x 10,
.....(b) 3 yuan IRF540N x 10,
.....(c) 3.5 yuan IRL540N x 10.

3.5 yuan IRF540N x 10 Results OK - But Rpi GPIO 3V3 signal Not Compatible

...

[tlfong01]

This morning I started testing 30 power MOSFETS:
.....(a) 1.2 yuan URL540N x 10,
.....(b) 3 yuan IRF540N x 10,
.....(c) 3.5 yuan IRL540N x 10.

1.2 yuan IRL540N x 10 Results Bad - Rpi GPIO 3V3 signal Not Compatible, But still useful if Vgs >= 5V

...

[tlfong01]

Keyes IRF520 Power MOSFET Driver Module

I saw TaoBao selling fake IRF520 modules at ¥3 each . So I bought 10 for testing.

HobbyComponents IRF520 MOSFET Driver Module Model HCMODU0083 - £2.5
https://hobbycomponents.com/motor-drive ... ver-module

This little module is a breakout board for the IRF520 MOSFET transistor. The module is designed to switch heavy DC loads from a single digital pin of your microcontroller. Its main purpose is to provide a low cost way to drive a DC motor for robotics applications, but the module can be used to control most high current DC loads. Screw terminals are provided to interface to your load and external power source. An LED indicator provides a visual indication of when your load is being switched.

Schematic and Forum
http://forum.hobbycomponents.com/viewto ... =76&t=1872

TaoBao IRF520 MOS PWM Driver Module - ¥3
https://detail.tmall.com/item.htm?id=58 ... 3ad4cZbbmQ
...

[tlfong01]

CSD18511KCS Test Results

I checked out 10 pieces and found all Rpi 3V3 GPIO signals compatible.
...

[tlfong01]

HobbyComponents IRF520 MOSFET Driver Module Model HCMODU0083 - £2.5
https://hobbycomponents.com/motor-drive ... ver-module
This little module is a breakout board for the IRF520 MOSFET transistor. The module is designed to switch heavy DC loads from a single digital pin of your microcontroller. Its main purpose is to provide a low cost way to drive a DC motor for robotics applications, but the module can be used to control most high current DC loads. Screw terminals are provided to interface to your load and external power source. An LED indicator provides a visual indication of when your load is being switched.
Schematic and Forum
http://forum.hobbycomponents.com/viewto ... =76&t=1872

IRF540 (IRF520) Ids at Vgs = 3.2V, 4.0V, 5.0V

I compared datasheets of IRF520 and IRF540N and found them similar, except IRF540 has lower Rds(on). So I think IRF540N is better than IRF520 and therefore there is no point testing IRF520.

I tested five samples of IRF540N at Vgs = 4V, just to confirm that HobbyComponents IRF520 Module should work for Aruduino (Voh > 4V). The results is summarized below.

At Vgs < 3.0V, all 5 devices do not conduct.
At Vgs = 3.2V, unstable, unpredictable Ids 0mA ~ 50mA
At Vgs = 4.0V, all 5 devices Ids ~= 1.5A, slightly heat up
At Vgs = 5.0V, all 5 devices Ids ~= 1.5A, no noticeable heat up.

My quick and dirty conclusion is the IRF540N cannot reliably work with Rpi's 3V GPIO. On the other hand, no problem with Arduino's 4V GPIO.

So for Rpi, there are two solutions.

(1) Use IRF540N, but with gate driver 5V+

(2) Use IRL540N, direct drive at 3V, or use a gate driver to drive at 5V+.

Next is to check out if Rpi can indeed direct drive IRL540N with GPIO 3V signals.

[tlfong01]

OK I have to ask, is tlfong01 some sort of weird automated BOT designed to analyse posts and then make up almost related responses, but not quite correct because a BOT doesn't truly understand what it is reading?

But I am very serious when trying to point out that many electronic product specifications are misleading customers, and Vgs(th) is a case in point.

There is nothing misleading about it. It is your misinterpretation of the spec sheet that is misleading.

I am not a robot!
...

[tlfong01]

Next is to check out if Rpi can indeed direct drive IRL540N with GPIO 3V signals.

Rpi GPIO Pin 11 3.0V (Voh) Driving IRL540N Test

Update 2019feb26hkt1709

The circuit with the 14D471K (MOV) metal oxide varister seems problematic.

I found the 3.5 yaun IRL540N fried.

I removed the varister and replaced another IRL540N and found everything OK again.

Actually I know very little about the MOV. I only know it is for AC220V+. I thought it should also work for 12VDC.

So the lesson learnt is：

....."DON'T TRY ANYTHING YOU AREN'T SURE!"

Always remember the Murphy's Law:

....."Anything that can go wrong will go wrong."

...

[tlfong01]

Rpi GPIO Pin 11 3.0V (Voh) Driving IRL540N Test
The circuit with the 14D471K (MOV) metal oxide varister seems problematic.
I found the 3.5 yaun IRL540N fried.
I removed the varister and replaced another IRL540N and found everything OK again.

DSP Sig Gen 1Hz 50% duty cycle driving motor and solenoid, and 1kHz motor only

https://youtu.be/1cysjRcF40g (youtube)

No substantial ringing at Vin, and no flyback spikes noticeable. Next step is to use (a) Rpi GPIO upshifted 5V, and (b) GPIO 3V direct driving mosfet.
...

[tlfong01]

DSP Sig Gen 1Hz 50% duty cycle driving motor and solenoid, and 1kHz motor only

1kHz signal input to Power MOSFET IRL540N, Rg 470Ω, Rgs 4k7, Vcc 12V, Rload 5Ω 20W

Now I tested a 5 resistor 20W, and found everything OK.

...

[tlfong01]

DSP Sig Gen 1Hz 50% duty cycle driving motor and solenoid, and 1kHz motor only

1kHz signal input to Power MOSFET IRL540N, Rg 470Ω, Rgs 4k7, Vcc 12V, Rload 5Ω 20W
Now I tested a 5 resistor 20W, and found everything OK.

Power MOSFET IRL540N, Rg 470Ω, Rgs 4k7, Vcc 12V, Rload 5Ω 20W

The hardware setup selfie.
...