Resistors and MOSFETs and fire-prevention, oh my

[Brandon92]

@tlfong01
You can also measure how open or close the n channel is of the MOSFET. You simply need to measure the voltage across the drain and the source. And it should be a very small voltage across it. Because of the low RDS(on) value. So, you can also calculate what the real RDS(on) is at that time and compare that value to the one in the datasheet.

[tlfong01]

@tlfong01
You can also measure how open or close the n channel is of the MOSFET. You simply need to measure the voltage across the drain and the source. And it should be a very small voltage across it. Because of the low RDS(on) value. So, you can also calculate what the real RDS(on) is at that time and compare that value to the one in the datasheet.

Power MOSFET Channel Width Measurement

Yes, openness or closeness is a bit misleading, because "open" or "close" is usually either or. Of course I can say half open, slightly open etc. Channel "conductance" or “conduction” is bit too technical. So I just say "wide" or "narrow".

When Vgs is around 3V, channel is narrow, Rds(on) is big, so drop is big. I measured the voltage between Vcc and Drain and found it was 9V. That means Vds is 3V!

Looking back, I should measure Vds and not VccD, which means I directly place the two meter leads across the motor, and the kick back spark might kill my multi-meter.

So IRF540 should not be that useful when Vgs is between 2V to 4V.

I have just ordered 20 cheapy logical level (Vgs(th) 1.8V) Power MOSFETS (only CNY3.5 each), for testing. I guess this new toy would arrive tomorrow or day after tomorrow. Then I will test both the old the new toys together, ...

[tlfong01]

1. Your R2 (1k) has a to low value in comparison to your R1. So, you have made a voltage divider instead. And that is not what you want here.
2. So, you need a higher value of R2. I have used values of 50k as a pull down resistor at that point.
3. But, it need to be stronger that the internal pull up of the Rpi pin. So, that it will not turn on the MOSFET or when the GPIO pin is set as a input.
4. And that pull down resistor also ensures at start up that the MOSFET is turned off. And that it will not turn on by itself. That could let to a dangerous situations. So, at start up the MOSFET is in a defined state.

Power MOSFET Rin and Rgs Circuit Analysis
Thanks a lot for your advice. I agreed with all your 4 points above, though they are a bit advanced and professional for my casual Micky Mouse hobbyist exploring for fun projects. I need to do some calculations for the Rin and Rgs and let you know the results later.
I only drew the schematic, and too lazy to actually solder the resistors, not to mention the other mysterious capacitors suggested by the OP.

Power MOSFET Gate Charge Circuit Analysis Learning Notes

It appears that I need to do some hard homework to google around and read about the mysterious Vgs(th) and the capacitors mentioned by the OP.

So I googled and found the following application note by Toshiba and TI good for newbies. I think I need one day to read, digest, and perhaps to do some experiments.

Hopefully I can then do some MOSFET biasing (not sure if I can call that biasing, as in BJT) and determine the valves of Rin and Rgs.

https://penzu.com/p/23b3f240

Power MOSFET Gate Drive Circuit Application Note - Toshiba 2018jul26
https://toshiba.semicon-storage.com/inf ... ?did=59460

Fundamentals of MOSFET and IGBT Gate Driver Circuits - TI Application Report 2018
http://www.ti.com/lit/ml/slua618a/slua618a.pdf

MOSFET 和 IGBT 栅极驱动器电路的基本原理 - TI 应用报告 2017 (Chinese version of above TI Application Report
http://www.ti.com.cn/cn/lit/ml/zhca770/zhca770.pdf
...

[ptimlin]

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.

[Gavinmc42]

Mosfet transistors have a high gate capacitance, to turn them on and off fast, mosfet drivers chips are used.
These can source and sink 2-4amps, the GPIO pins cannot do this.
A high gate resistor will protect the Pi's GPIO but will charge and discharge the gate capacitance voltage changes slowly.
This puts the mosfet in a linear resistor mode during transition between on and off, there it must dissipate heat.

I have used Smart fets like the BSP75N with no problems.
Lots of automotive parts are very useful with Pi's, they are generally designed not to catch fire.
Fet datasheet can lie none are easy to read , just get a good part designed for the job to start with.

[tlfong01]

Fet datasheet can lie none are easy to read

All Semiconductor datasheets can lie none are easy to read [I mean, to/for me]

I have never heard of the phrase "can lie none". It took me a couple of seconds to guess that if it is "lie-none" or "can-lie-none", or both can be used. Of course I googled but found nothing. Very likely it is not Oxford English, not to mention Queen's English (my first English book was published in Malaysia, and the first lesson is "A man", "A pen", "A man and a pen". Of course that was last century, I have not used any pen for years, I only use a mouse. On second thought, "lie none" sounds like broken Hongkong/Singaporean English.

I am making a short comment now. But I got stuck in whether saying "easy to me" or "easy for me", and "I have never heard" or "I never heard". Or should I use the phrase "I mean" before or after "to me", and should I bracket or parenthesis the whole comment/remark?

So I need to exit rpi.org.forum now and enter english.stackexchange, ...

PS - Above scribbled rubbish is my this morning's technical English writing assignment, but not enough word count, need to add a long tail,

[tlfong01]

you have to be sure that the mosfet works on all conditions. Temperature is a big problem! Here in canada it goes down to -40 celsius , and even lower. If you have something outside it needs to work.

Working Conditions Disagreement

So the problem is that on one hand the professional wants to work in all severe outdoor conditions (perhaps except in Mars where there are dust storms). On the other hand the casual sloppy hobbyist just wants to play with a toy motor indoor, in the home sweet home.

A side problem is that the hobbyist seriously thinks that many engineering product specifications are lousy and misleading - see appendix below. (I do agree to serve Broadcom reps Polonium tea! )

Appendix - Rpi3B+ GPIO Voh/Vol no spec and I2C flat rate 100kHz no setting problem

RE: DOCUMENTATION BUGS: RPI2&3 GPIO ELECTRICAL SPECS Postby gwideman » 2018-Apr-02 Mon 10:48 pm
https://www.raspberrypi.org/forums/view ... h#p1296302
... What I don't understand then is how when something so apparently fundamental is missing, so many people have successfully used, and how so many have manufactured boards, that use the GPIO without actually having any problems. Like, millions of people....

RE: DOCUMENTATION BUGS: RPI2&3 GPIO ELECTRICAL SPECS Postby jamesh » 2018-Apr-03 Tue 12:14 am
https://www.raspberrypi.org/forums/view ... h#p1296369
Insurmountable wrote: ↑ 2018-Apr-02 Mon 11:09 pm
I completed my HND in electronic engineering in 1992, we had to produce a product manual and spec sheet or it was a fail, that was a large chunk of the course marks. And you had to calculate the specs using statistics none of this just add a bit extra for good luck BS.

Some of use haven't got years to fully test a product so specs and statistical analysis are important if you want to produce a reliable product. Just because something works fine in chilly England does not mean it will perform equally as well in the deserts of Dubai. How would you know ? By making sure you don't exceed the max ratings for components. What appears to be the max rating from bench testing a component will be a lot higher than what the actual max rating to get a lifetimes use out of it.

I went onto work in the aerospace and nuclear power industries, we would put Polonium in the Broadcom reps tea if they tried to sell us something with out scientifically calculated specs.

RE: RASPBERRY PI3 I2C BAUD RATE SETTING Postby tlfong01 » 2018-Aug-04 Sat 4:33 pm
https://www.raspberrypi.org/forums/view ... z#p1348886

Mars Rover Opportunity Is Dead After Record-Breaking 15 Years on Red Planet - Mike Wall 2018feb14
https://www.space.com/mars-rover-opport ... -dead.html

[Burngate]

Fet datasheet can lie none are easy to read

I have never heard of the phrase "can lie none".

Can I assist with the punctuation?

Fet datasheet can lie; none are easy to read

[Gavinmc42]

Can I assist with the punctuation?

Yes please, I failed English and so had to do Engineering.
My new Windows10 and Edge does not do good suggested? , er what's it called again?

[tlfong01]

Fet datasheet can lie none are easy to read.

I have never heard of the phrase "can lie none".

Can I assist with the punctuation?

Fet datasheet can lie; none are easy to read

Dear Mr Human

Thanks a lot. Elon Musk's DeepFaking GPT2 is not yet intelligent enough to help, I think.

Update 2019feb16hkt1724 - Elon Musk OpenAI DeepFaking GPT2

How OpenAI writes convincing news stories and works of fiction Unlisted - Guardian News Uploaded on Feb 14, 2019 60,325 views
https://www.youtube.com/watch?v=XMJ8VxgUzTc (youtube)

New AI fake text generator may be too dangerous to release, say creators - Guardian 2019feb14
https://www.theguardian.com/technology/ ... ws-fiction

It was trained on a dataset containing about 10m articles, selected by trawling the social news site Reddit for links with more than three votes. The vast collection of text weighed in at 40 GB, ...

By structuring the text that is input, it can perform tasks including translation and summarisation, and pass simple reading comprehension tests, ...
...

[danjperron]

So the problem is that on one hand the professional wants to work in all severe outdoor conditions (perhaps except in Mars where there are dust storms). On the other hand the casual sloppy hobbyist just wants to play with a toy motor indoor, in the home sweet home.

Well first this is the wrong mosfet when you have 3.3V for the gate.


I looked at your chart I see that it won't be able to switch the valve. If I look at my three irf540 only one was able to switch a small relay from gnd to 3.3V.

It is not because it works one that it will always work.

About the temperature. Well i was burned three times buying stuff from China. They works fine in summer but stop working in september when the temperature was below 20 C. Mind you all of the times was the oscillator crystal that just stop working. I had to change the crystal and they start to work again.

You are in a warm country. Here In Canada the temperature are extreme, its not mars but it will fluctuate from +35 C to -40 C . Right now in February one day is freezing rain and the next day is -25 C. We just got 35 cm of snow yesterday.

If you are using a part play with the border line specification then you are asking for troubles.

[tlfong01]

1. Well first this is the wrong mosfet when you have 3.3V for the gate.
2. i was burned three times buying stuff from China.
3. If you are using a part play with the border line specification then you are asking for troubles.

Trouble Maker vs Trouble Shooter

1. OK, so I have ordered the right mosfet, logic level input, Vgs(th) 1.8V, Rpi friendly.

2. From ShenZhen with love - Finger burning, cheapy CSD18511KCS ¥3.20 each x 25 = ¥80.

... https://penzu.com/p/acc88f4b

3. I will still ask for trouble, crossing the red line as usual!

Update 2019feb15hkt1633 - Express delivery tracking record
2019-02-15 16:07 place order

2019-02-15 14:38 goods dispatched, sfExpress to collect
2019-02-16 13:30 arrived, collected [23 hours]

...

[Gavinmc42]

It will turn on at 3V3 volts but it will not be saturated.
Figure 3 and 2 of the datasheet, show 25amps at -55C and perhaps 0.5V DS at 100amps, ie 50Watts of heat to dissipate.

RDS on against VGS, notice how high RDSon goes when VGS is under 4V?
I don't see the words "Logic level" in the data sheet?
Mosfets are transistors not logic level switches.
It will be better but not ideal.

[tlfong01]

1. It will turn on at 3V3 volts but it will not be saturated.
Figure 3 and 2 of the datasheet, show 25amps at -55C and perhaps 0.5V DS at 100amps, ie 50Watts of heat to dissipate.
RDS on against VGS, notice how high RDSon goes when VGS is under 4V?

2. I don't see the words "Logic level" in the data sheet? Mosfets are transistors not logic level switches. It will be better but not ideal.

The picture guy's Power MOSFET Hacking Diary

I am looking at the datasheet. I am a picture guy, cannot remember words, so I cut and paste pictures below.

https://penzu.com/p/acc88f4b

3. I remember I saw the verb "logical level" some where. Perhaps it is some sales rep trying to mislead the professionals. I will google later.

[tlfong01]

I remember I saw the verb "logical level" some where. Perhaps it is some sales rep trying to mislead the professionals. I will google later.

Logic Level MOSFETS

Logic Level MOSFET Selector Guide - NTE Electronics
http://www.nteinc.com/Web_pgs/LL_MOSFET.html

DESCRIPTION

The NTE series Logic Level MOSFETs are compatible with the 5-volt power-supply requirement of logic circuitry. These devices do not require an interface circuit between it and the CMOS logic driver; therefore, the extra cost of the interface circuit power supply is eliminated.

The chief physical structural difference between Logic Level and other MOSFETs, and the electrical reason for its difference in performance, is its gate insulation thickness, which has been reduced from 100nm industry standard to 50nm (500 angstroms), yet which retains the dynamic strength to handle the high voltage applied to power transistors.

Since the surface inversion of the MOS channel is determined by the gate-insulator voltage field, the halving of the gate-oxide thickness should be expected to have a major effect on the gate voltage required.

In fact, this reduction is the reason for voltage reduction from 10 volts (standard MOSFETs) to 5 volts (Logic Level MOSFETs).

[tlfong01]

1. It will turn on at 3V3 volts but it will not be saturated.
2. datasheet, show 0.5V DS at 100A
3. Mosfets are transistors not logic level switches, it will be better but not ideal.

Saturation comparison of BJT and FET

Well, I think the term saturation is misleading, what the we hobbyists bother is the power:

P = V * I = Vxy(sat) * I

Let us compare BJT NPN Power Darlington TIP120.

TIP120 NPN Plastic Medium-Power Silicon Transistors
https://www.onsemi.com/pub/Collateral/TIP120-D.PDF

Collector−Emitter Saturation Voltage Vce(sat)
= 2V at I = 3A
= 4V at I = 5A

So at 5A, power is 4 * 5 = 20W.

For FET (Fig 2, Saturation Characteristics of the datasheet summary below)

https://penzu.com/p/acc88f4b

Vds(sat)
= 0.1V at I = 25A

So at 25A, power is 0.1V * 25A = 2.5W only.

In other words, FET's non-sat is BJT's super-sat.

My hobbyist toy motor current is less than 1A anyway, so why worry?

My quick and dirty conclusion is that power MOSFETs are ideal logic level current switches for Rpi toy motor projects.

Actually, not just switches, but PWM toy car motor turbo overdrive speed controller, and I might remove the heat sink and let the device heat up, making the N channel wider Rds(on) smaller, channel wider, current flow larger, thus speed higher. In other words, an outside spec design.

Or are the datasheets misleading me?

All-Japan Model and Hobby Show Tokyo 2019
https://www.tamiyausa.com/blog/all-japa ... ept-29-30/

Tamiya Mini 4WD Pacific Coast Hobbies 2019
https://www.tamiyausa.com/blog/mini-4wd ... es-232019/

[ ミニ四駆 Tamiya Mini 4WD ] 2019 Are you ready?
https://www.youtube.com/watch?v=ZivmWMTy2Ek
...

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

I dream to become an Aji, but my IQ is only 95!

I am a huge fan of Norbert Wiener, the originator of Cybernetics. This old guy, 70 years ago，imagined An Age of Robots! .

In 1949, He Imagined an Age of Robots - John Markoff, NY Times, 2013
https://www.nytimes.com/2013/05/21/scie ... found.html

I'm not a robot - 110,312 views
https://www.youtube.com/watch?v=edaie-WxnX8

Jia's Ringtones - Please don't be sad - 302,994 views
https://www.youtube.com/watch?v=TlkNMi7QfTI
...

[tlfong01]

I have ordered the right mosfet, logic level input, Vgs(th) 1.8V, Rpi friendly CSD18511KCS ¥3.20 each x 25 = ¥80

WaterFall/Agile Rpi3 Based Mini 4WD Smart Vehicle Incremental/Continuous Testing/Continuous System-Integration/Test Driven Development Plan

https://penzu.com/p/acc88f4b

/ to continue, ...

[tlfong01]

IRF540N Test Circuit
IRF540N Test Circuit - With Rg and Rgs added
https://penzu.com/p/087c1ca8

Your R2 (1k) has a to low value in comparison to your R1. So, you have made a voltage divider instead. And that is not what you want here. So, you need a higher value of R2. I have used values of 50k as a pull down resistor at that point. But, it need to be stronger that the internal pull up of the Rpi pin. So, that it will not turn on the MOSFET or when the GPIO pin is set as a input.

And that pull down resistor also ensures at start up that the MOSFET is turned off. And that it will not turn on by itself. That could let to a dangerous situations. So, at start up the MOSFET is in a defined state.

Power MOSFET Pull down resistor Rgs Value

Thank you for your advice. I forgot I need to consider potential divider effect. So I think your recommendation of Rgs = 50k is good. Now the divider ratio is 470Ω / (470 + 50k) should be OK.

I also read the Electronics Tutorials' calculation of TTL pull/down resistor values. Their recommendation is that 10kΩ to 100kΩ is acceptable and 50kΩ is OK. But for pull down, the recommendation is 1kΩ.

https://penzu.com/p/acc88f4b (schematic)

Appendix - Pull up/down Resistors Calculation - Electronics Tutorials
https://www.electronics-tutorials.ws/lo ... istor.html

Calculating Pull-up Resistor Value

So if we assume a voltage drop of only one volt, (1.0V) across the resistor giving double the input voltage at 4 volts, a quick calculation would give us a single pull-up resistor value of 50kΩ.

Reducing the resistive value further, will produce a smaller voltage drop but increase the current. Then we can see that while there may be a maximum allowable resistive value, the resistance value for pull-up resistors is not usually that critical with resistance values ranging from between 10k to 100k ohms acceptable.

Single Gate Pull-down Resistor Value

Therefore if we assume a voltage drop of only 0.4 volts across the resistor, a quick calculation would give us a single pull-down resistor value of 1kΩ.
...

[tlfong01]

CSD18511KCS Preliminary Test Results

The penzu journal entry below summaries the test results.

My quick and dirty conclusion is the following.

1. Power MOSFET CSD18511KCS is fully compatible with Rpi 3V3 GPIO (direct drive).

2. No gate driver is necessary.

https://penzu.com/p/acc88f4b (Test results summary)

The test was very smooth and took only 15 minutes, much shorter than I expected. Now I can now take a long break, to watch the stupid Korean video "I am Not a Robot".

I am not a Robot episode 8
https://www.facebook.com/watch/?v=1181125228685085
...

[tlfong01]

@tlfong01
1. You can also measure how open or close the n channel is of the MOSFET.
2. You simply need to measure the voltage across the drain and the source. And it should be a very small voltage across it. Because of the low RDS(on) value.
3. So, you can also calculate what the real RDS(on) is at that time and
4. compare that value to the one in the datasheet.

Measuring Vds and Calculate the Rds(on)

Ah, your advice is inspiring.

It was stupid me to measure the voltage across the motor, because the Vcc should be dropping when the current increases. In other words, I can not calculate the Vds precisely.

One other thing is that measuring the voltage across the motor is risky, because the kickback current might pass my high class, expensive multi-meter and fry it.

This brings up another thing, that is the motor resistance is dynamic and changes in an nonlinear way, when heated up or changing speed. So I think I better use a resistor load instead of the inductive motor load.

So perhaps I should unearth some big power mud resistors from the under my bed junk boxes. They have been collecting dust for too long a time.

https://penzu.com/p/acc88f4b (10 Ohm, 2A resistor)

I also found that I could not bear the stupid Korean "I am not a robot" TV series for more that two episodes. I better come back to do the more exciting MOSFET testing right now, ...
...

[tlfong01]

1. simply need to measure the voltage across the drain and the source. And it should be a very small voltage across it. Because of the low RDS(on) value.
2. So, you can also calculate what the real RDS(on)
3. compare that value to the one in the datasheet.

Rds(on) Measurement

You guess is very correct. I varied Vin from 0V and found the device starts to conduct at round 1.64V (1mA), Vds 12V, and saturates at around 3.5V (1A), Vds 0.03V . So the Rds(on) is 1A * 0.03V = 0.03R = 30mR.

https://penzu.com/p/acc88f4b
...

[tlfong01]

Rds(on) Measurement

https://penzu.com/p/acc88f4b
...

[tlfong01]

1. measure voltage across drain and source.
2. it should be a very small voltage low RDS(on).
3. can calculate what the real RDS(on)
4. compare value to datasheet.

Twilight Zone Spec

Actually I am trying to create a twilight zone spec where Vgs = Rpi Voh = 2.4V ~ 3.2V. This Vgs range is what I hope Rpi can direct drive the MOSFET.

The datasheet does not include the twilight zone, so there is nothing to compare.

Anyway, I found Rds(on) in the twilight zone is of the order 30mΩ. Now I am going to zoom in, to get a better picture of what the hell is going on there, ...

https://penzu.com/p/acc88f4b (big picture of twilight zone)
...

[tlfong01]

I found Rds(on) in the twilight zone is of the order 30mΩ. Now I am going to zoom in, to get a better picture of what the hell is going on there, ...

Rds(on) vs Vgs (2.6V ~ 10V)

/ to continue, ...

https://penzu.com/p/acc88f4b (big picture of twilight zone)

...