Vložit
- čas přidán 2. 07. 2024
- The MOSFET transistor has all but replaced BJTs (Bipolar Junction Transistors) in Arduino-style circuits which require a decent switching device to control motors, relays, LEDs and just about everything else.
PCBWay 6th Anniversary Carnival Coming Soon! With Larger Coupons and Gift Giveaways: www.pcbway.com
See my GitHub:
github.com/RalphBacon/BB4-MOS...
Not only are they cheap to buy, they are also easy to use - but just watch the static handling requirements! In other words, do NOT touch the pins with your hands.
Let's have a look at how they fit into the Arduino World in this video.
00:00 Welcome to Bacon Bytes
00:13 Intro: what is a MOSFET
01:06 PCBWay SMT PCBs - ideal for beginners
02:55 Resume MOSFET video
03:13 All about the Si4599 Dual MOSFET module - great for experiments
04:13 Whiteboard All About MOSFETs. First, N-Channel.
09:30 P-Channel MOSFETs
10:38 Keyword Summary (for Datasheet interpretation)
13:18 Summary
PRODUCTS
10pcs Si4599 dual N and P Channel 40V MOSFET Expansion Board Module (about $1 each + shipping)
www.banggood.com/10pcs-CJMCU-...
10pcs 2N7000 N-Channel Transistor Fast Switch MOSFET TO-92 Just $1.37 for 10! + Shipping
www.banggood.com/10pcs-2N7000... ?p=FQ040729393382015118&
INFO
Si4599 Data Sheet
www.vishay.com/docs/68971/si4...
IRL3705N Logic Level (TTL) Gate Power MOSFET Data Sheet www.infineon.com/dgdl/irl3705...
How a MOSFET is constructed
www.electronicsforu.com/resou...
List of all my videos
(Special thanks to Michael Kurt Vogel for compiling this)
bit.ly/CZcamsVideoList-RalphB...
If you like this video please give it a thumbs up, share it and if you're not already subscribed please consider doing so and joining me on my Arduinite journey
My channel, GitHub and blog are here:
------------------------------------------------------------------
/ ralphbacon
ralphbacon.blog
github.com/RalphBacon
------------------------------------------------------------------ - Věda a technologie
Finally I found it... a simply good mosfet video. Packed all important information and no confusion in 15 minutes. Thank you gentleman
Glad it helped!
Cool video, MOSFET's are fun to play with. Love the way you bring additional components into the world of Arduino. Thanks for your time putting together the videos
Glad you like them!
Hi Ralph - have been following you for a couple of years now and have really enjoyed your work. CZcams's recent changes re ads have made me seriously reconsider whether CZcams is worthwhile. My personal policy is to immediately unsubscribe from any channel where the author permits mid-roll ads. I have no difficulty with pre-roll and post-roll ads. I do mostly watch the former because I think the author deserves some recompense and I sometimes watch the latter for the same reason, providing they are not repeats. I also don't mind paid promotions such as you do for PCBWay etc but I feel the need to vote with my feet on the subject of mid-roll ads. If it eventually becomes the case that I no longer have any channels I'm subscribed to, so be it. It'll be an extremely cold decade in hell before I'll pay Google so much as a single cent to avoid ads. I note that some authors are also sufficiently irked by the changes and the appalling positioning of the mid-roll ads that that are running parallel streams on lbry.tv. You might want to consider that too. CZcams is probably "too big to fail" but even they might "get the message" if libry.tv starts to eat into their viewership.
A timely post, Phil, as I was having this discussion with a friend of mine a couple of days back. Mid-rolls ads are, by their very nature, intrusive and despite CZcams's assurances that they are placed in appropriate places (not mid-sentence) that is not always true. And overlays are one thing, a full-blown, TV-style ad is another.
Additionally, just how much extra revenue does the CZcamsr make by allowing mid-roll ads? They are not saying so it's guesswork. I'd be surprised if it was more than $1-2 per video on my channel. I'm coming to the conclusion that they are not worth the hassle, but CZcams have automatically included them now in all existing videos that are over 8 minutes long. I can kill them for all future videos, however. I'm definitely considering this.
Incidentally, I don't think viewers have to watch the ad for it to be counted. That's what the delay is for. When I watch CZcams I always skip the ads as soon as I can, unless it's something I'm interested in. So don't feel obliged to watch them any longer than CZcams makes you watch them on my channel, anyway.
In less then 15 minutes the amount of answers I got is AMAZING!
thank you again for another great video.
You're very welcome!
Pedantry alert: 0.01 ohms is 10 milliohms but what is 9 milliohms between friends? When I returned to maker world (about 4 years ago) , I was stunned by how cheap MOSFETs were. Back in the 1970s and 80s, they were relatively exoctic devices with price tags to match. The sensitivity of MOS devices to static charge is a low obstacle which is easily justified by the efficiency they offer. BJTs can be used at 3V but are much happier at higher voltages. Great video, Ralph.
Yes, I should have said 20 and 60 milliohms, oh well. It changes nothing. As I said though, some power MOSFETs are £1 each. EACH! Others just 10c! Choose wisely!
Just to add to the pedantry.. The difference is... a lot of money! :-) Well, not a lot, but a MOSFET with low-single-digit milliohm on resistance is quite a bit more expensive than one with 10's of milliohms. Last I checked, about $7-10+ US vs. $2-3.
"BJTs....are much happier at higher voltages"? Doesn't sound right to me.
Very good and comprehensive explanation!
One addition to the different Vgs one might find in a MOSFET datasheet:
-> Vgs is just the voltage from gate to source (with the red lead of the multimeter on the gate and the black one on the source for an N-Type)
-> The Absolute Maximum Vgs is the maximum voltage you should ever subject the MOSFET to under any conditions (stay well away). Usually the actual breakdown is quite a bit different.
-> Vgs(th) is the voltage at which the MOSFET will 'turn on' to some degree (usually not very well turned on). For example, it might be 2V minimum and 4V maximum for a drain current of 0.25mA at Tj = 25°C (the die itself is at 25°C).. That means that if you want your 20A MOSFET to really turn on fully (not just conducting 250uA) you need a lot more voltage than 4V to be sure about it, but if your Vgs is well under about 2V you can be pretty sure it's well turned off (at least around room temperature).
Cited from here: electronics.stackexchange.com/questions/157065/explain-in-laymans-terms-vgs-and-vgsth-of-mosfets
Very useful information, Raimund, thanks for sharing that. I will definitely be looking at that link to improve my understanding of the quirkiness of MOSFET specifications!
Love your videos, the number of times I come google something and your channel turns up is priceless!! Hope you are well.
I'm glad to see that Google is my best friend too!
Thanks for showing clearly how MOSFETS apply to the Arduino aspects of power switching and supply. Really enjoy your fundamental series on the important components with the basic overview.
You are welcome!
Thanks Ralph! Great video. I really like these fundamentals videos that you are creating.
Glad you like them!
Really gleaning a lot from your efforts, Thanks Ralph!
Great to hear!
Thank you for another very understandable video Ralph. I was unlucky in that my last year at school was the analogue one so I got released into a world of chips with all their related jargon without knowing what the hell it was all about. Yes I am that old. Thanks to your videos I am now beginning to get my head around it all. Could have done with this instruction 38 years ago but will certainly keep my mind active for years to come while I release the magic smoke more often than light a LED
I've heard of analog... I think.
Excellent video. Learbing how to use a mosfet and the best explanation I have come across.
Glad it was helpful!
Excellent series. Well done Ralph.
Glad you like them!
Easy to understand explanations. Worth watching. Great learning value with many practical pointers.
Glad you think so! Thanks, Paul.
Another great video. Thanks, Ralph.
Glad you liked it!
Excellent video ... keep up the great work!
Thank you very much!
Great video Ralph keep up the good work.
Thanks a million!
Good video and series, Ralph.
I suggest people look at the Vgs curves in the datasheet for their specific MOSFET when using them with 3.3V logic such as ESP8266 or ESP32. The Drain-Source resistance can be higher than the stated RDSon when driving the gate with a low voltage. This decreases their efficiency and increases the amount of heat that needs to be dissipated.
Higher, yes, Andrew, but not by much according to the graphs I've looked at. It doesn't suddenly shoot up to 10 ohms for example. If the VGS(th) is one or two volts (as per my snippet from a real datasheet) then at 3v3 it should be fully on. But that's what datasheets are for and I totally agree that makes must read them and not use any old MOSFET blindly..
Hi again, Ralph.
Sounds like a plan ! I will surely use these small wonders on my projects in the future.
Thanks for sharing the main difference between N and P channel MOSFETs. N channel is what I plan to use for sure.
Have a great day, my Friend !
Glad you liked it Daniel.
The Arduino has become such a ubiquitous tool now, and knowing how to expand their capability is very useful. Thanks, Ralph.
Glad you found it useful.
Well this is a keeper. Always used BJT as I sort of know them, these seem really super and thanks for the anti-static warning.
Thanks
Yes, I've killed several just by touching them.
I've not had a problem with regard to that, somehow. Lucky, I guess?
More great information Ralph. Thanks
Glad it was helpful!
Thanks again Ralph, absolutely invaluable information, now adding a ~200 ohm resistor to the 2N7000 gate of my temperature controlled attiny based fan controller, to be used in an kinds of fan cooled equipmenet. Startet with my DPS3050 where i got tired of the constant fan noise, even without load..
Sounds great!
Started binge watching your videos. Being a complete beginner in Arduino
Welcome aboard! Take your time and do try and actually _do_ some of the stuff I show, that's how you'll learn more quickly.
Finally a video on why we'd use a MOSFET over a BJT, thanks Ralph
Glad you found it useful, Chris.
another great instructive video :-)
Glad you enjoyed it!
Very interesting Ralph, for real protection we used to drive the mosfet from an Opto-Isolators.
Very safe, or you can use a purpose designed MOSFET driver, of course. I've never used one but I'm assuming they provide protection somewhere too.
I always add a pulldown resistor on N-Channel MOSFET gates. It's just good practice to make sure it stays off until you want it on.
Good practice, I agree. Stops all the floating about nonsense.
Got to watch your RDS(on) as it's not always as low as expected at the VGS threshold. Quite a few need a VGS of 10V's to get the low value even though they turn on at
A TLL (Logic Level) device will switch on fully at 5v. Whilst it may still have an even lower RDS(on) at 10V that is not what is important for Arduino users. We could also say that the RDS(on) goes back UP if your draw significant current (10s of Amps) but hey, at this level who's bothered?
Lots of modern mcu are 3v3 now, So that has been my problem also, or vdrop on 5v can still be a problem also.
@@tablatronix some mcu pins can be used for charge pumps to obtain suitable gate voltage.
Very informative video
Glad you think so!
Very useful, thanks
Glad it was helpful!
Fantastic, thanks for sharing :-)
My pleasure!
Thanks Ralph
You're welcome, Norm.
Ohh -now I learned something new.
Now I don't need to do the basic reseach - thank you very much. (so much easier to learn when you take the time to explain it).
Neat trick with the P-channel fet as protection for power supply.
P channel can be used as protection but D and S is swapped because internal diode.
@@LimbaZero Did Ralph show the PNP the wrong way around when he was showing how to use it as reverse polarity protection?
Glad it was helpful! The P-Channel MOSFET has its SOURCE as the +ve and the DRAIN as the output to the load. The PNP transistor was also shown correctly (emitter is positive wrt the collector).
@@RalphBacon I only ask because I saw this hackaday post hackaday.com/2011/12/06/reverse-voltage-protection-with-a-p-fet/
See my replies above, it was drawn back to front. Drat.
I hadnt realised you already covered ... so I also got a number of 2N7000 which hopefully will act as switches to the inputs of a Arduino and being powered from PCF8574's via a 180 ohm resistor... which I am using on Model railway for point control via a Mega and PCA9685 to run SG90 servos.. again another Great video
Yes, they are a reasonable, general purpose MOSFETs, the 2N7000, a higher than normal RDS(on) of 5Ω, 200mA current capability and a VGS(th) of 3V. A good BJT substitute.
Thank you!
You're welcome!
thanks for the video
You're welcome
Great vid Ralph, but there was one thing missed that I think is quite important. To turn the FET off properly you need to give it a ground reference to pull that capacitive charge off the gate if your microcontroller doesn't automatically provide the path.
Yes, Jeff, indeed. I usually use a 10K to 100K and I even had this videoed but it got snipped onto the cutting room floor. I'm going to do another video on this entire subject, it's too big to be done in 15 minutes!
Had a bit of trouble seeing detail on this one Ralph. Cast to my new 65" telly sorted that!
Wow! I'm on the big screen!
Thanks for another great Bit of Bacon. And, of course, something learned. Reverse voltage protection. I still have (and use) many BJTs, but any power critical (high current or battery driven) implemenrations now feature MOSFETs.
That reverse voltage protection had D and S wrong way. S should be on load site because mosfets have parasitic diode between D and S.
No, a P-channel MOSFET has the Source as the +ve input and the Drain as the output. You're thinking of an N-channel (which would not work in this circuit) which is the other way around. The body diode on the P-Channel MOSFET conducts from Drain to Source (so no current passes through it in the normal course of events). Look at my circuits for my Smart Phone Charger, I use both N and P channel MOSFETs in that circuit.
@@RalphBacon Mosfets have also diode between D and S. So if you connect negative voltage to Source (P Channel) then it will go through to Drain and there to load.
Edit: For normal switching operation S to Vcc with P-Channel and S to GND with N-Channel
Edit2: if you little bit google reverse voltage protection with p channel fet. Then you notice all schematics are D to Vcc and S to load. If it's other way it won't protect because of body diode.
I looked at my video and I see I drew it backwards for polarity protection. Damn. Well, when you draw the same symbol several hundred times you end up drawing it on auto pilot. I might have to cut that bit of the video out. Whilst it would not stop a circuit working it won't protect it either. The actual (correct) circuit is still sitting here on my workbench in the breadboard. So annoying.
Thank you.
You're welcome!
As usual, good little tuto😊👍🏻I put a 220 ohm resistor (instead of your 180…), like you said : just in case….
Yes, that value is just fine, it's not critical.
BJT's is more robust when you use coils, (when the coil backfire when disconnected). Don't ask why I know this....
Thanks for sharing 👍😀.
Did you put a protection diode across the coil, Asger? Like you would for a relay coil? That should protect BJT and MOSFET.
Great video, I would have liked to see where the load was for P channel!
It was supposed to be connected to the point marked OUT (as in, 5v OUT).
About that reverse polarity protection. Don't forget that parasite diode between D and S. With reverse voltage protection with P-Channel usually Source is in load side. Parasitic diode will pass enough current from D -> S to make enough big Vgs to turn mosfet on. If mosfet is connected as your picture then parasitic diode will pass current to load if input voltage is reversed.
Edit:
Parasitic diode is always included in mosfet symbol
Drat, I drew it the "normal" way to switch power on/off. Hmm. Tricky to fix now.
While a gate resistor may protect the driver when something goes wrong with the MOSFET (but even then if there's only something between 0 and 5V on the MOSFET), it's not the reason that it is used.
The purpose is to limit the current to a value the driver can deliver (40mA in case of an AVR/Arduino). The gate behaves like a capacitor and is like a short when switching. The resistor also reduces ringing that can happen because of the inductance of the gate trace and the capacitance of the gate. That ringing can lead to slow turn on and a lot higher losses.
Several 100kHz to MHz directly driven by an atmega is also very - let's call it - ambitious. Frequencies like that definitely need some external drive circuitry if you want your MOSFET not to be in the not-on-not-off state most of the time.
If we assume the MOSFET gate is, perhaps, 1nF it should charge in about 1.25µS if we use (for simplicity) a 250Ω gate resistor and assume a capacitor charges (not fully, but fully enough) in 5 time constants. Enough to generate a 0.8MHz signal.
In the Arduino world we tend to switch on relays, LEDs and the such like and that's fine for that. Generating pulses at several 100kHz is a different matter so it's a pity I suggested this, as although it can most certainly be done (it "works for me" on my workbench without blowing anything up - yet).
I haven't done it recently but I'm pretty sure you can get 1Mhz out of an Arduino? The PWM frequency is much lower by default but by changing the parameters we can get some incredible frequencies. In a much earlier video I adjusted the registers and got 15kHz to speed modulate a fan. In fact, it might have been higher now that I think about it, as I can hear 15kHz as a terrible high pitched whining.
Anyway, anyone watching this video (as a beginner) is unlikely to be creating PWM frequencies of several 100KHz without a driver of sorts (eg a single, fast BJT, how ironic).
very well explained, but I would like to know how to setup an Arduino to a control P channel as the pins from Arduino are +
To turn a P-channel MOSFET on, just bring the Gate to Ground. You can do this via an OUTPUT GPIO and writing LOW to that pin.
A slightly easier way is to connect an N-Channel MOSFET's Drain to the Gate of that P-channel, with its Source to GND and the Gate held LOW via a 10K resistor to GND - then you can write HIGH to the Gate of the N-Channel MOSFET to control the P-Channel one, much easier to get working.
Nice video, as usual. One word of caution: I've bought two batches of (power) MOSFETs from "The Far East" as you so euphemistically referred to it, and both have failed in the gate threshold spec. That V(gsth) threshold voltage is likely one that often results in rejects and a steady supply of dodgy parts to the unsuspecting hobby market. A nasty trap for young players!
Thanks for the info!
It is true that we hobbyists are probably the main market for sub-spec devices. Normally it won't matter, but when it does it's best to buy from a reputable source (at a considerably higher price, of course!).
10:30 0.06 Ohms is not 6 milli Ohms, milli is 10 to the minus 3 en.wikipedia.org/wiki/Metric_prefix 0.06 its 6 centi Ohms #SI lol, thanks for the video, explained well.
Two/Twenty, Six/Sixty. What's a few milliohm between friends?
Or 60 milli oHms
@@RalphBacon not much, but at least you know your viewers are alert. I'm really sympathetic to the problems of making a video and trying to hold so many things in your head and talk at the same time. I have no problem with corrections from viewers, with no foul to you. ;-)
Keeps us alert at least.
My viewers are like Raptors with bionic vision. Nothing gets past them. Keeps me well on my toes. What a pity that CZcams does not let me edit my video, not even text overlays any more.
@@RalphBacon What's an order of magnitude difference between friends.
Nice video series, thanks! I would note that BJT transistors can be easier to use with low-voltage circuits, say with 3.3V Arduinos (I like the Arduino Pro Mini). You can easily feed any common-or-garden BJT a few mA and get it into saturation at low voltages, but at 3.3V it takes a pretty special MOSFET to turn on fully.
Mosfet drivers are your friend and relatively cheap these days even for the optically isolated ones.
MOSFETs that switch fully on at 3v3 are becoming more common. www.infineon.com/dgdl/irl3705n.pdf?fileId=5546d462533600a40153565f29f42532
@@RalphBacon They are in the last few years but I'm a hoarder and following an involvement with low voltage lighting systems a few years ago I still have a collection of mosfets that need 10v to hit saturation. I think mosfet drivers still have a place in terms of gate capacitance which can really help with thermal performance.
You can always use a MOSFET driver to switch on those 10V Gates. But cheaper to replace in some circumstances I think?
@@RalphBacon It's getting that way but you can't ask a Yorkshireman to throw something away! There is still a small advantage in switching speed and it's amazing how much those nanoseconds at high resistance can add up in terms of component temperature. Again some of the newer mosfets have amazingly low gate capacitance so that's becoming less of an issue as well. I recently repaired a lighting controller for a friend (read replaced an SMD fuse) and wished the mosfets in it had a part no, It's running 24v/4A a channel @ 1kHz through a through a sot-23 package with the gate driven directly from the unmarked 3v3 micro and they don't even get finger warm. It's witchcraft I tell you !
Good after noon sr
Nice to hear from you Yogesh!
At 8 mins into the video you say that only a tiny current flows to charge the gate capacitance. Without the gate resistor, the current can be 100mA or more depending the the actual gate capacitance and the switching speed. The 180 ohm gate resistor will limit the current to about 30mA. It is always a good idea to have a gate resistor to protect your microcontroller from current spikes.
There's two schools of thought here: use a gate resistor, or don't.
As a risk-averse type of person I probably would but might not. Gate current flow is [usually] miniscule; it is the voltage that determines how the MOSFET works.
I'm pretty sure that 100mA might flow down the gate but for a pico second to charge the capacitor. Putting a resistor in the gate path actually slows down the charge (and hence the switching time).
Looking at P channel mosfet datasheets is fun and crazy. Negative current ratings.
Well, I'm not sure we'd call them _negative_ currents but if current is flowing from -ve to GND then I guess we must put a minus in front of the values.
Thanks Ralph thanks for the video, I believe P-Channel MOSFETs are capable of switching higher voltages so can be used like a relay , where as PNP Transistors cannot (would always be on) I have seen them used a lot with Arduino for working between 3.3v and 5v. I have not used them much as I worry static will kill my projects too easily, and hard to test once soldered to a circuit, perhaps I should have a go.
Well, there are high voltage BJTs as there are MOSFETs, The trick is to select the right component for the job. Static can be an issue but wearing an earthing wristband can help. And not shuffling your feet on the carpet whilst touching them.
@@RalphBacon sorry i meant switching higher voltages using a 5v arduino, this does not work with PNP transistors regardless of their rating. But does with P-Channel Mosfets.
I've not used many PNP BJTs in the Arduino World, Andrew. But something tells me it must work because we can run a bridge circuit for a motor this way (2 x NPN, 2 x PNP) - not that we would want to with MOSFETs now available.
@@RalphBacon yes you can do it with a combination of PNP and NPN but If you try to switch 12 volts off with five volts on a PNP T you will still be -7 volts and as a result still on. Apparently it's not a mosfet limitation. electronics.stackexchange.com/questions/248796/pnp-high-side-transistor-switching-with-microcontroller
Hi Ralph shouldn't you be using a resistor from Gate to Source to ensure Mosfet switches off or is this always nesecarry
Ditto here. I believe it's to drain off the charge on that gate-source capacitance, to aid turning off the MOSFET. Might not be an issue in the case of lower speed switching.
If your microcontroller has a push-pull output (like Arduino), I’m guessing that’s enough to turn the MOSFET off.
@@markday3145 yes mark, when the port switches low it actively pulls the gate low faster than a resistor would, so there's no need. However, I'm nervous about leaving a module lying around with the gate open. A megohm resistor from gate to ground will provide protection from static, so it's not a bad idea to include it in a detachable module.
Yes, a 10-100K resistor from Gate to GND is always a good idea (see my circuit to see that I do this) but not absolutely required. That information and more is still on the cutting room floor. I had to stop somewhere!
Side note: A lot of the "5 volt MOSFETS" you'll find will also work just fine at 3.3 volts. So you ESP32 people are good to go, too! (Check the voltage curves to be sure!)
True, Chris, but be very careful. The manufacturers will quote RDS(on) for 4V5 and 10V usually (even for TTL level devices) but to what extent the MOSFET has switched on at 3V3 must be checked (VGS(TH) will give an indication, that's all, as you say graph is best). The last thing you need is a partially switched on MOSFET (unless you intend toasting marshmallows over it).
@@RalphBacon What's wrong with toasting marshmallows? lol...
Beneficial side effect, possibly?
Useful info, thanks.
Been having so many problems with motor shields not outputting the amps that I need and went back to basics with a MOSFET. Turns out my output pins do not reach 5v on my UNO. I bought some IRL540N Fets and they would not turn on fully until I applied VCC 5v to it. Even though the data sheet says 1c-2v VGS. I ended up testing an old IRLB3036 and it worked so much better from the output pin. Data sheet shows 1c-2.5v VGS so I don't understand what's going on. I still need to pump slightly more voltage into it to get it fully on or maybe PC USB power just isn't outputting enough current to my Arduino? I've been at this for weeks now.
If your GPIO pins are not outputting the full 5v (well, let's say 4.95v) are you sure the power supply you are using is up to the task? Measure the input voltage on the 5v pin and compare it to a GPIO pin that is set to OUTPUT. Should be pretty much the same.
Also, check that the MOSFETs you are using are TTL-compatible. This means the Gate will switch FULLY on with a 5v input voltage. You can check the datasheet for the device and look at the Gth (gate threshold). Lots of them are really quite low, 1-2v these days. Although that only shows when the device will START to conduct it won't need a lot more to fully switch on.
@@RalphBacon Hi, thanks for replying. It measured different things under load but was around 1v when I measured under load. I tried several different pins. The power pack is 5v 3 amp and the VCC is 5v. I think the 540N isn't up to the task. I swapped for an old R~IRLB3036 and it worked much better. Not quite perfect, but increasing the voltage to compensate got me where I needed to be. Not sure why it didn't work as the VGS is only 0.5 v difference on the datasheet
What is the significance of the source to drain resistance? Would a higher resistance cause more heat? Why should you care when resistance is relatively very low anyway? Excellent video!
Yes, heat is your enemy. Or the MOSFET's anyway. If the RDS(on) figure is relatively high (eg 10 ohms) then it will dissipate significantly more heat than a device with a resistance of 0.01Ω. That's how some of the "power" MOSFETs are able to pass 50A without melting!
Ralph S Bacon Thank you for your reply. I really enjoyed your video. Great work!!
7:00 Actually, all pins in an Arduino are capable of doing PWM output. It is just that for a few pins the Arduino itself takes care of it, while for the other ones you need to manually do the PWM programming, the raises and lowering of voltages, etc.
Ah yes, you mean bit bashing. The very thing that might be time-critical and therefore worthy of a bit of "compiler optimisation". Woo hoo! A good example.
good video for its purpose, but...
- the current needed to switch a MOSFET during the transition is not negligible at all. If you want to switch quickly you should at least supply 1 A ("one amp") which would charge up a standard power MOSFET gate (1 nF of capacitance) within 10 ns
- limiting the gate current by 180 Ω (which is a good idea!) limits the durrent to 30 mA and the maximum switching frequency to below 1 MHz (good enough for most), using 1 kΩ which many recommend already brings you down to slow transitions and a max frequency of just about 100 kHz
- as pointed out by others, the threshold voltage is not a sharp transition between OFF and ON - always look at the corresponding datasheet which will give you curves for ID over VD for different VGS
Would you say that a gate resistor would still be necessary in order to provide protection to the Arduino, in the manner of limiting the current draw? I would think that an infrequent large current draw of a nS might not damage the arduino, but sustained and frequent transitions might degrade the part over time.
It's probably overkill but I tend to use a mosfet driver for anything switching at more than a few kHz especially as I'm mainly driving 3.3v at
Like a BJT the VGS(th) on/off is not a cliff edge, admittedly, but we're talking Arduino World here, not GHz frequency shifting. Keep the 180Ω resistor in to protect the Arduino!
@@newburypi the protection also goes in the opposite direction. A not-too-uncommon failure mode of a MOSFET is a partial or complete short between drain and gate. A resistor between the Arduino and the gate limits the current which can flow back into the Arduino and through its protection diodes. Here, however, 180 Ω might be too small to make a difference...
The video was very helpful to me for describing the terms needed and I have been struggling to find good candidates for use with 3.3V logic devices. So... I am a little confused. The two examples you list IRL3705N and Si4599DY show VGS of 16V and 20V respectively. Not the 1-2V you describe in the video as what we need. It's probably wrong, but my current understanding is that Mosfet's, if the gate is not at these voltages, the drain resistance is very high and thus waste a lot of energy in heat vs driving the motor. Do you have any Mosfets that show VGS in the 1-3V range?
The terminology is particularly confusing.
The 16V/20V is the VGS which is the *maximum* voltage allowed between Gate and Source.
The VGS(TH) is the *threshold* voltage at which the Gate will start to turn the MOSFET on. I know, it's done to confuse us.
A number of MOSFETs start to turn on at 1V-2V, some require 3V or even better 4.5V to ensure they are fully on - as the resistance is indeed high before that point and they get hot real quick, just as you say! The only way to know is to look at the DataSheet for the RDS(on) graph at particular Gate voltages. As an example:
IRL3705N will have an RDS(on) of 0.018Ω with a Gate voltage of 4v5 but will start to turn on at a VGS(TH) of just 1V (min) - 2V (max). Always assume the higher (worst) figure as you don't know what you might have bought.
Have a look at my video #195 on Reverse Voltage Protection (czcams.com/video/eS4_6XQi74c/video.html ) as that has more details too.
So, I should look for a TTL N-Channel Mosfet. What is a good general purpose one that would also work for ESP32 3.3v? I'm looking on Aliexpress and they have too many options.
For 3v3 use you need better than TTL. The only way to find one is to use the filtering on something like RS Components to narrow the search. That's how I found the IRL3705N but it's very chunky. You can always use MOSFET drivers from the ESP8266 that will guarantee your MOSFET to switch on fully, but that's a different topic. Learn more here: www.microchip.com/design-centers/power-management/motor-drivers/mosfet-drivers#:~:text=A%20MOSFET%20driver%20is%20a,(IGBT)%20or%20power%20MOSFET.
As succinct as a well-crafted Haiku!
UNO to MOSFET
Cooler than a B J T
Switching nirvana!
Thank you Ralph
You bet, Gord! You can't beat a good MOSFET as an efficient switch.
Would would have added a word or two about using a 555 as a gate driver. For pennies 555 gives you: logic level shifting so that you can get the full R ds on, will source and sink an order of magnitude more current than you micro-controller important if you have high switching frequencies.
True, but not really something I'd add as part of this beginners' video. As it is, so much information ended up on the cutting room floor.
Hi Ralph, nice explanations, bit of an "oop's" on units but the P Channel stuff was well explained,it defiantly can confuse at first.....edited cos spelt ya name wrong !
That OK, Sandy, I get called lots of things. Hmm, did I say 2 milliohms instead of 20 milliohms? Oh well, not important in this context.
@@RalphBacon Ha!
Love the introduction to 4599 MOSFET module.
Do you happen to know if the super spacious board they are on is to dissipate heat or is it just that big? I was looking to put one of these on an 8 pin DIL carriage to save space.
And then DOH! - I was building an assembly which used these to create a simple H bridge motor control. I use logic to keep the N channel MOSFETs off whilst I decide if the P channel MOSFETs are on or off. I am aware of the 'slow to switch off / fast to switch on' perils of 'shoot through' so in my project I allow at least 1ms of N channel MOSFET off each side of any directional change of the P channel MOSFETs whilst under MCU control.
My MCU has an 'enable' pin that gates an enable into each N channel MOSFET and just in case this looses the plot, there is a safety 10M resistor to ground the gate of the N channel MOSFET driving my bridge so if in doubt, it's disabled (My bridge is running at 8v so I need additional 'level shifter' MOSFETs hooked up to the MCU).
However I have no such pull down on the 'direction' N channel MOSFET as in final circuit this will be driven by MCU pin and the direction is one or the other - was keen not to waste power.
Was testing the build and thought I'd test the H bridge before soldering the MCUs into the circuit. The 'enable' MOSFET worked a treat but as soon as I tried to change the input voltage to the 'direction' MOSFET - a bad thing happened. I guess as I'm using very high resistive MOSFET switching loads (to save power) the 'direction' MOSFET started switching at mains frequency and with the 'enable' MOSFET 'on, the circuit started making a strange whining noise and then the magic blue smoke started escaping from the modules.
Luckily I have a few spare but I'll have to rethink a partial assembly test.
Keep those CZcams videos coming, they're great.
I think I know the root cause of my 'blue smoke' issue - I used some CMOS logic hung over my 8v power rail to drive the N channel MOSFETs in my bridge - to act as an enable for the bridge. I also used the CMOS to generate an inverted signal so one of my P channel MOSFETs would always be on whilst the other was off - to reverse the H bridge drive direction.
I was using a couple of N channel MOSFETs hung on my MCU power rail with very high resistive load (10M) up to the 8v rail of the H Bridge.
When I was testing the unit, I had a digital AVO measuring the voltage at the drain of my MCU switched N channel MOSFET - that was pulling enough current to slug the voltage down from 8v when the MOSFET was off and this was evidently enough lower than the 8v rail of the H bridge P channel MOSFET to never turn this off - hence the 'shoot through' and resulting 'magic blue smoke'.
Luckily I had a spare gate in my CMOS (CD4001) to act as a buffer between my MCU switched N channel MOSFET and the H bridge P channel MOSFET.
I have to add a new set of those 4599 MOSFETs (fully charged up with magic blue smoke intact) and give it another test.
I'm not sure why I never encountered this whilst at the solderless breadboard stage (dumb luck). Ho Hum.
You should be careful with those Si4599 modules, they are currently unavailable anywhere (what chip shortages?). I use an alternative, the AO4606 which is a drop in replacement (well, so far, so good).
@@RalphBacon Thanks Ralph - wow! - I just purchased 10 of these puppies form AE at four pence a piece (plus less than two quid postage) - that's insane - I rummaged through the bin to find the old si4599 pcb - I'll re-use these. (I have 6 x si4599 on order from AE at about a quid each but couldn't resist the AO4606 price point)
BTW - 'Chip Shortage' - on another post I was banging on about the unavailability of (genuine) ATtiny402 devices until Aug 2023 - my reply post disappeared - but just to say - if you don't need automotive / industrial grade devices (temperature rating), you can get these direct from the manufacturer and if you select the correct packaging, you don't even get charged the part reel cutting charge (specifically ATtiny402-SSN) my devices turned up in a couple of weeks any buying ten of them worked out at about a quid each. - Result!
Are you planning any 'How to' videos on these? - would love to get some tips.
I think better accronym is VGS, Vth reminds to voltage dividers :)
We have to be careful here, Josip. VGS and VGS (th) are two very different parameters (done to confuse us I'm sure). Perhaps that is why some manufacturers are starting to use Vth to differentiate?
When I ran your video yesterday I'm sure it had something about using a P mosfet for reverse polarity protection. I made a mental note to make some notes on this as it seems a very useful thing to do, but when I look at your video today its nowhere to be seen. Was I hallucinating?
Yes, you were under the influence. However, I shall be talking about this easy-to-use circuit in quite a bit more detail in a forthcoming video so you will not lose out. Stay tuned!
What exactly is the application for a dual P/N mosfet like the Si4599?
I use the arrangement of a P-Channel and an N-Channel almost exclusively as a mechanism for switching a device on and off (under program control).
See videos #123 Auto Shutdown (czcams.com/video/g1rbIG2BO0U/video.html ) #173 ATTiny85 Auto On Off Shutdown (czcams.com/video/S2y1oAVmxdA/video.html ). There are a couple of others too.
Spookily I've just retrofitted this exact device to a project of mine as I was fed up having to switch it off manually - all done via the Arduino now. And I'm using this in my new Bin Lid Monitor circuit too!
The body diode are also important
In what way, Dave, given the topic of this video? In what way would a _beginner_ use that information?
If you invert the drain source polarity, the mosfet conduct no matter the gate condition. If you are a beginner sound like that the mosfet is broken
OK, that's a good reason. I tried so hard not to have to mention that dratted body diode. Just confuses beginners. sigh.
in an N Channel Mosfet, can the gate have a higher voltage than the drain?
Hmm, the datasheet should provide the _absolute maximum ratings_ that the device can withstand.
However, think of the situation where you are switching a lower voltage (eg 3v3) but using a higher 5v (or even 10v) gate voltage to do so. That works just fine.
The only parameter worth watching is the max vgs (voltage between Gate and Source [not drain]) which is usually 20V or thereabouts but could be lower or higher for any particular device. But that's not your concern in this question.
I can't find anything on a MOSFET datasheet here to indicate the max voltage between Gate and Drain, though.
Don’t we need a resistor on the gate tied to ground? In order to switch it off?
You would if the arduino port didn't actively drive it low. So if you want to play tricks with switching the port to input, high impedance, then yep, put a resistor to ground. It's not a bad idea to do it anyway for antistatic reasons.
In all my testing I've never needed one - but I do tend to put a 10K from the gate to ground anyway. It would protect the MOSFET, as Ray says, if the Arduino pin is high impedance although I suspect the circuit board at that stage is protected anyway.
I used mosfets in an Arduino pcb print before, but I learned something new. Does anybody know where the specific value of 180 ohm for that resistor comes from? In case I want to use it on a 3,3V pin, for example. I assume it is the result of pin voltage devided by max current for the pin?
Correct, Jason. It limits the current from the pin to the maximum allowed - well, strictly speaking 40mA is allowed but we try and limit to 20mA for the 328P. Other micros will have different limits. Note that the resistor also limits the switching time for the MOSFET (as the Gate current is restricted by the resistor, hence it cannot charge as quickly) so if you are doing PWM on the pin bear this in mind.
I thought VGS(th) was the voltage needed to *start* turning on, and the voltage to fully turn it on (and get that really low RDS(on)) is higher. Look at the graph in the data sheet. Alas, it seems like even many logic level MOSFETs don’t fully turn on at 3.3V.
Correct, Mark. Many MOSFETs don't, but the snippet I showed from a IRL3705N turned a MOSFET on at between 1v (min) to 2v (max) so would be a good choice. See the datasheet: www.infineon.com/dgdl/irl3705n.pdf?fileId=5546d462533600a40153565f29f42532
Love the vid, but why is common to see dual channel MOSFETs used in the Arduino world?
I wouldn't say it _is_ that common to see a dual MOSFET in the Arduino world (not a P-channel and N-channel together) but of course bridge circuits for motors use multiple MOSFETs. However, that Si4599 could be deployed just for the P-channel MOSFET in the package, it's that useful.
@@RalphBacon so why would they be used at all, think iam still a bit confuddlled
In order to switch on a P-Channel MOSFET is it quite common to use another N-Channel MOSFET - hence this dual MOSFET package (to bring the Gate to GND). In motor control (bridge) you need two N-Channel and two P-channel MOSFETs too. Either individual components or in a dual package.
You have the same datasheet twice in the description :)
That's to ensure you read it. Or maybe I made a mistake. Now which one is it? Thanks for letting me know. Now corrected!
@@RalphBacon Ah the intentional read test :D
Aah! The basics. Just my speed.
And once again, no coding. I'm spoiling you.
Hello, I don't understand your resistance values 0.02 Ohm = 2 mOhm and 0.06 Ohm = 6 mOhm ?
Two/Twenty, Six/Sixty. What's a few milliohm between friends?
Ralph, I have a video request/suggestion for you. Could you do a video on how to use the ESP32 with WIRED Ethernet? I've looked around and there's virtually nothing on this subject.
There are some ESP32 shields, John, that offer wired ethernet, although they are not cheap (£20) but do offer to power your ESP via the same cable (PoE). Google search for ESP32 Ethernet Shield showed me:
www.mouser.co.uk/ProductDetail/Olimex-Ltd/ESP32-POE?qs=unwgFEO1A6tUQVMxdOBsBw==
The main reason that they are scarce is because everyone wants cable-free WiFi not a long Ethernet cable, regardless of whether that powers your device or not.
It's likely because it's not as simple as you'd hope using its internal ethernet support. This blog post that's referred to often might be useful if you wish to try: sautter.com/blog/ethernet-on-esp32-using-lan8720/
Another source for info, closed issues on arduino-esp32 GitHub repository: github.com/espressif/arduino-esp32/issues?q=is%3Aissue+ethernet+is%3Aclosed
Isn't Vth the absolute minimum for any current to pass? IE, the maximum off voltage instead of the minimum on voltage?
Well, it's described in Wikipedia like this:
The threshold voltage, commonly abbreviated as Vth, of a field-effect transistor (FET) is the minimum gate-to-source voltage VGS (th) that is needed to create a conducting path between the source and drain terminals.
I interpret that as being the lowest voltage we can apply to the gate to start to turn it on (N-channel MOSFET).
Hang on, isn't that what we're both saying? I'm now confused!
@@RalphBacon almost. At Vth it will start conducting but with high resistance (so it is not fully off)
From what I read, above Vth, the resistance goes down proportionally to (Vgs-Vth)² assuming Vds is high enough
Quite so. You wouldn't want to use the MOSFET at (or just above) minimum Vth, it would get very hot, very quickly. In fact, we need to ensure it's fully switched on to get the RDSon resistance to the value in the spec sheet. At which point it should run amazingly cool even with large current flows. Which it does, according to my experiments so far, like my intelligent phone charger.
@@RalphBacon precisely, that's what I mean with "maximum off voltage" in contrast to "minimum on voltage"
I don´t know a lot about electronics, but noticed, that people use many gate-driver-designs... Is a LL-mosfet also in need of those drivers, or does the fact, that it conducts at 1,3v, lets it work finely even with solely an arduino? I´m thinking thereby of course on the switch-on/off-speed of the mosfet. In other words, turning on already at 1,3 out of 5 available volts, lets it turn on faster (shorter turning-on time), cause at just 1,3v the arduino has more mA left to give, and thus shorten the turning-on-time (????????????)
In my case, I want to drive cleanly a mosfet at 5-30kHz, 1-2Α (so, not long delays while turning on/off), and looking for any easy way to do it, without having to study electronics first... ;-(
Thanks á priori for every answer...
A TTL-level MOSFET will switch cleanly but the propagation delay can only be found by (a) reading the data sheet (to see the delay in switching on) and (b) by trying it out using a 'scope. Best not to use any kind of gate resistor (slows down the gate capacitor charge time) even though some people seem to think they are necessary (I don't, but what do I know?)
@@RalphBacon Ok, thanks, so, my "very basic" assumptions were right... My perspective was to avoid artificially prolonging the data-sheet-delay by wrong-handling, and not to shorten that delay even further, so, thanks. The "trick" with the no-gateresistor, I´ve learned about it recently, too, after I have posted my question (above) to you. Still, the knowledge about the gate-resistor didn´t mean, that my question didn´t need any answer, so, your answer was bitterly needed, still... So, a big "Thanks" again...!!!
As a number of people have pointed out, your reverse polarity protection cricuit is wrong. See this application note from Texas Instruments for example: www.ti.com/lit/an/slva139/slva139.pdf
Yes, I drew it in the standard configuration to just turn power on/off. Just shows you shouldn't do things on auto-pilot.
I've never understood why BJTs use "Collector" and "Emitter" and FETs use "Drain" and "Source". Aren't they analogous? Why the difference?
They are only analogous in the sense that current flows from the collector to the emitter (in an NPN BJT) and from the drain to the source (in an N-Channel MOSFET) but, as you might imagine, it's more complex than that.
Here's some information: www.power-and-beyond.com/whats-the-difference-between-mosfet-and-bjt-a-909006/
20mOhm... @ 5:05
Two/Twenty, Six/Sixty. What's a few milliohm between friends?
0.02 ohms is 20mOhm not 2 :)
Yes, true. I spoke fake milliohms.
You forgot the gate pull resistor to prevent false triggering.
On what? All MOSFETs have the required pull-up/down resistors!
@@RalphBacon ....what? MOSFETs contain no resistors.
MOSFETs require the gate to never float, otherwise they will trigger randomly. A microcontroller will float its outputs until told otherwise.
This comment appears in a raft of comments about my latest project that used MOSFETs!!! Anyway, the gate resistors are (usually) necessary and I never omit them in Real Life but this is about the theory really. There is another video on MOSFET reverse protection, that (hopefully) has the resistor in place (or doesn't need one).
@@RalphBacon This has nothing to do with the series resistor. MOSFETs need a pull resistor from gate to source to prevent false triggering. Look it up.
Yes, I understand that. All my N-channel MOSFETs have a 10K resistor from Gate to Ground. All my P-channel MOSFETs have a 100K resistor from Source to Gate. You must have missed my videos where all this is clearly described.
Ralph, are you aware of this site: everycircuit.com/ I haven't tried it, but seems pretty cool.
I haven't seen it or tried it but I watched the demo and was intrigued. I will investigate further, thanks for the link.
0.02 Omh = 20 milli Ohm
Yes, indeed.
Isn't 0.02 ohms 20 milliohms?
Yes, my brain is stupid.
So a few years ago I uses a IRLZ44NPBF to switch a 5v fan using a 3.3v logic pin on a Raspberry Pi 4b. I'm brushing up on how to do this again on another Raspberry Pi. When I did this before I looked at the VGS(TH) and it was Min. 1v and Max 2v so i thought that meant the max voltage I could apply to the gate was 2v. So I added two resistors to make a voltage divider on the 3.3v Raspberry Pi GPIO pin (150K & 100K | 3.3v to 1.98v). Now that I'm looking at this stuff again I think that voltage divider was completely unnecessary since the VGS for this MOSFET is +/- 16V. Have I got this correct now?
Yes, you don't need to limit the gate voltage, it usually is fine up to about 10v. In fact, it works better the higher you go (up to the max, usually 10v) giving a lower "on" resistance.
The S in MOSFET stands for Substrate, not Semiconductor.
That's not my understanding. en.m.wikipedia.org/wiki/MOSFET
@@RalphBacon baffled at how many get this wrong. If you think about it, the word semiconductor doesn't fit because the semiconductor material isn't metal oxide, it's silicone or germanium. The Substrate, the material that insulates the gate from the silicone, is the metal oxide part of the FET.
No, the substrate or body of the MOSFET transistor is made of p-type silicon formed from a single crystal. The metal oxide (just a few atoms thick) is grown on TOP of the substrate.
Just sayin'. More here: bit.ly/38KMqxP
@@RalphBacon the metal oxide is not a semiconductor, since it doesn't conduct at all.
"short videos" - 17:35 Min
Short videos of 15 minutes, excluding my standard introduction, the sponsorship message and my goodbye message. How short do you want them, Frank?
@@RalphBacon i like the short version - 15 min + intro, sponsor & goodbye is ok
0.02 ohms = 20 milli ohms, not 2 ! Same with 0.06 ohms...
Yes, others have alerted me to this.
Actually, 0.02 Ohms equals 20 mOhms, not 2… (just saying).
Let me think, multiply by 1000, divide by 0,001, take away the number you first thought of... YES, you're right. Drat!
05:00 eh? 0.02Ω = 20mΩ
Put a note on the video and delete this message.
Yes, I realise I said TWO instead of TWENTY. And SIX instead of SIXTY. I'm not sure it changes anything (much).
MOSFETs: Resistance is futile!
Groan.
I'm sorry but they are too many mistakes in this video.
- there are big capacitances between the gate and the source and drain and you have to charge/discharge them so there's a BIG gate current during switching. When the n mosfet turns on the voltage at the drain drops very fast (going from Vcc to almost 0) so the capacitance between the gate and the drain that whats to keep a constant voltage between the drain and the gate makes the voltage at the gate drops (Miller effect), turning the mosfet back off. If you don't want oscillations you need to provide a good current to the gate.
- the gate resistor is there for a good reason: the output stage of the mcu is not designed to handle the currents needed for a clean switching. The good way is to use a special gate driving IC between the mcu and the gate. If switching speed is not that important the resistor is ok.
- there's a body diode in your p mosfet that will let current flow in the opposite direction. Your protection doesn't work! The drain and source are in the wrong way !!!!!
This has been covered by other viewers too but turning on a MOSFET without a gate resistor certainly seems to work without blowing up the Arduino. Perhaps because the gate charges in about 1µS (I'm using 1nF gate capacitance as a rule of thumb, although the datasheet will specify). Perhaps the port pin would eventually fail? But putting the resistor in place protects the Arduino and slows down the charge of the gate - but for the purposes of this overview I don't think that really matters if all we are switching on is an LED or relay or something like that. The MOSFET really requires a resistor from gate to ground too (I use 10-100K) to ensure it really switches off and protects it from floating about and reduces ringing but I also did not include that in the final cut. Took too long to explain every nuance.
Using a driver of sorts (even if it's just a single BJT) for high frequency switching is certainly advantageous but then again how much detail do I go into for beginners?
Yes, the reverse protection MOSFET is the wrong way round. I drew it on auto-pilot. I might cut that bit out yet,
@@RalphBacon if the beginners are your primary audience then giving proper informations is even more critical. If you switch a small mosfet onece per day Ilmaybe the mcu output will suffer but will survive long enough. If you switch a large one and do some PWM for exemple, then it will shorten the life of your mcu.
You are basically shorting the output for a brief time everytime you change the output. The gate capacitance is a bit misleading because of the Miller effect that amplifies it. Arround the threshold the gate's voltage will stabilize while you pump charges into it.
As Albert Einstein said, “Everything should be made as simple as possible, but no simpler.” Unfortunately in this video I tried to over-simplify and shot myself in the foot. I've removed the reverse polarity protection video piece as I could not let it stand. I'll probably do an entire video on it which will give be a chance to fill in the blanks (most of which are on the cutting room floor) on MOSFETs.
Milli = 1/1000 -> 6mOhm = 0,006Ohm
Yes, 60 milliohms not 6. Not sure it makes a huge difference here, though.