Author Topic: EEVblog #1409 - The HUGE Trap of Inductor Back EMF  (Read 6015 times)

0 Members and 1 Guest are viewing this topic.

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« on: July 25, 2021, 01:52:22 pm »
A practical demonstration of Lenz's law and back EMF in an inductive relay coil and how to solve it using a Freewheeling/Flywheel/Flyback/Snubber/Clamp diode. Also the downsides of clamping diodes, and switch arcing supression.
This is a follow-on from the DC Transients Fundamentals video:
Also a look at an AMAZING potential phenomenon you probably haven't seen before!
Actually, two rather cool things you probably haven't seen before.
Along with transistor ratings, transistor storage current, and Collector-Emitter breakdown voltage, there is a lot to unpack in this video.

00:00 - Recap of Relays, Inductors, Faraday & Lenz's Laws
02:30 - Relay Back EMF Explained
07:09 - The Flywheel analogy of Inductors
08:30 - Relay circuit demonstration
12:35 - 700V Back EMF!
14:43 - BJT Transistor Storage Time
17:03 - Back EMF Diode clamp demonstrated
19:06 - An AMAZING demonstration!
24:43 - Trap for young players
25:23 - DOWNSIDES of Back EMF Diodes
28:38 - BONUS cool effect of Back EMF diode DEMONSTRATED

 
The following users thanked this post: SeanB, thm_w

Offline sandalcandal

  • Supporter
  • ****
  • Posts: 641
  • Country: au
  • MOAR POWA!
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #1 on: July 25, 2021, 02:35:48 pm »
Nice work editing. Pretty slick without any wasted time; just under the 30 minute mark.

The extras are interesting too, never thought the spark gap type oscillation possible with a relay coil and a BJT. I wonder how much EMI it actually kicks out and if it's comparable to the other switching EMI in such a system.
Disclosure: Involved in electric vehicle and energy storage system technologies
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #2 on: July 25, 2021, 07:43:10 pm »
The oscillation may radiate quite a bit - it is quite some energy, not just from the coil, but also new energy when the transistor turn on again. It is kind of lucky to turn off after some 1ms.


In the old days before diodes got cheap, they had RC elements in parallel to relays. There where ready made RC components allready in the 1920s for this.

 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #3 on: July 25, 2021, 08:13:49 pm »
Relay contact arcing is a big problem for noobs, "muh why did my Arduino/Raspberry Pi crash?". The radiated EMI generally causes havoc, especially WiFi mains remote stuff ESP8266 sitting an inch away.
Back in the day we used snubbers, when there was room and the cost was minor. I always put them in but the parts are huge and frowned upon. You don't see them in appliances like ovens, dishwashers, furnace, A/C etc. they just let the relays arc and have a short life.

I find any back-EMF diode will experience maximum whatever the inductor current was, and whatever the supply voltage is.
After all, you can't get more current from an inductor than what was flowing in the first place, unless you are Nikola Tesla  ;)   If the relay coil current was say 100mA then even a 1N4148 is adequate.
But the back-EMF diode significantly extends the time for the B-field to decay (relay contacts to open) so a poor choice when you need speed.
Automotive will also just put say a 680R resistor across the coil, to pass the reverse-battery requirement yet have tame spikes.

edit: the relay coil-to-contact capacitance is another path for contact's arcing EMI to get to VDD on your MCU.
« Last Edit: July 25, 2021, 08:15:27 pm by floobydust »
 

Offline JustMeHere

  • Frequent Contributor
  • **
  • Posts: 724
  • Country: us
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #4 on: July 26, 2021, 06:39:01 am »
Love this series.  This is what brought me to EEVblog in the first place.

One on different snubbers please.  In particular I'm interested in using a relay to switch a 240v pump motor.
 

Online Berni

  • Super Contributor
  • ***
  • Posts: 4922
  • Country: si
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #5 on: July 26, 2021, 11:02:15 am »
Yeah those relay coils can really kick back.

We used to use the self oscillating relay trick to make a simple prank "tazer" out of a relay and a battery. Most relays have normally closed contacts so they can be made to oscillate and the amount of voltage they produce can make for quite a zap. You could probably boost it up to even more zapping power if you added a fluorescent lamp ballast in series (That's the other battery powered zapping trick, sometimes its used as a physics demonstration)
 

Online nali

  • Frequent Contributor
  • **
  • Posts: 656
  • Country: gb
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #6 on: July 26, 2021, 11:30:53 am »
Yeah those relay coils can really kick back.

We used to use the self oscillating relay trick to make a simple prank "tazer" out of a relay and a battery. Most relays have normally closed contacts so they can be made to oscillate and the amount of voltage they produce can make for quite a zap. You could probably boost it up to even more zapping power if you added a fluorescent lamp ballast in series (That's the other battery powered zapping trick, sometimes its used as a physics demonstration)

Lol, that was one of my first "projects" back in the mid-70s when I was a kid ;D I even wrote to I think it was Everyday Electronics and got my "idea" published which was quite a thing then for a small town 13 year old  8)
 

Offline Refrigerator

  • Super Contributor
  • ***
  • Posts: 1540
  • Country: lt
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #7 on: July 26, 2021, 04:29:15 pm »
Guys what's your opinion on this: (pic included)
Keep your EMF down and your BJT out of saturation.  :-+
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Online Berni

  • Super Contributor
  • ***
  • Posts: 4922
  • Country: si
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #8 on: July 26, 2021, 05:27:44 pm »
Guys what's your opinion on this: (pic included)
Keep your EMF down and your BJT out of saturation.  :-+

Never seen that before but it looks like a pretty neat idea. The main benefit i see is that you don't need access to the other relay terminal, so it can be applied to just generic open collector outputs.

I don't really see how it keeps the transistor out of saturation tho. The forward drop of the zenner is about the same as the transistors forward drop down the base, then the transistors saturation voltage is additionally added to the diode, so it would never really be able to divert away enough current to properly turn the transistor off. Likely needs a schottky diode in parallel to do that. Then again don't think BJT saturation is that much of an issue, relays are pretty slow anyway so its not like you are getting a substantial turn off speed boost from it.
 

Offline Refrigerator

  • Super Contributor
  • ***
  • Posts: 1540
  • Country: lt
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #9 on: July 26, 2021, 08:19:00 pm »
Guys what's your opinion on this: (pic included)
Keep your EMF down and your BJT out of saturation.  :-+

I don't really see how it keeps the transistor out of saturation tho. The forward drop of the zenner is about the same as the transistors forward drop down the base, then the transistors saturation voltage is additionally added to the diode, so it would never really be able to divert away enough current to properly turn the transistor off. Likely needs a schottky diode in parallel to do that. Then again don't think BJT saturation is that much of an issue, relays are pretty slow anyway so its not like you are getting a substantial turn off speed boost from it.
Looks like a bit of an oversight from my side but this would work with darlington transistors.  :)
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26751
  • Country: nl
    • NCT Developments
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #10 on: July 26, 2021, 09:15:17 pm »
Guys what's your opinion on this: (pic included)
Keep your EMF down and your BJT out of saturation.  :-+
I always put the diode in parallel with the transistor; not the relay coil. Just like the parasitic diode inside a MOSFET.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline nfmax

  • Super Contributor
  • ***
  • Posts: 1556
  • Country: gb
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #11 on: July 26, 2021, 09:20:03 pm »
If you drive the relay coil with an emitter follower, no separate diode is needed (think about it), although the transistor has to pass the coil current as it ramps down to zero, with the full supply voltage across it. But no more than the supply voltage.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 6923
  • Country: ca
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #12 on: July 26, 2021, 09:20:53 pm »
Guys what's your opinion on this: (pic included)
It's typically a zener diode from C-B to limit back-EMF voltage. This is standard practice in automotive ignition coil drivers BJT and IGBT around 400V zener.
« Last Edit: July 26, 2021, 09:22:51 pm by floobydust »
 

Offline jesuscf

  • Frequent Contributor
  • **
  • Posts: 499
  • Country: ca
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #13 on: July 26, 2021, 10:56:09 pm »
This phenomena may be easier to understand if we look at the current/voltage relationship in an inductor:

 \$v=L\frac{di}{dt}\$

In Dave's circuit explanation \$\frac{di}{dt}\$ is negative (the final current is zero), hence the negative voltage across the inductor.  But this also happens when the switching device closes up.  Not much of a problem for BJTs, but you can see the problem happen when a relay is switching and inductive load: both huge negative and positive voltage spikes.
« Last Edit: July 26, 2021, 11:06:55 pm by jesuscf »
Homer: Kids, there's three ways to do things; the right way, the wrong way and the Max Power way!
Bart: Isn't that the wrong way?
Homer: Yeah, but faster!
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14072
  • Country: de
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #14 on: July 27, 2021, 08:57:11 am »
I always put the diode in parallel with the transistor; not the relay coil. Just like the parasitic diode inside a MOSFET.
[/quote]

With just a single transistor the diode would not help with back EMF. It is only in a bridge circuit, that the diode parallel to the upper transistor/FET would prevent the voltage for the lower transistor get too high.
 

Offline DenzilPenberthy

  • Frequent Contributor
  • **
  • Posts: 408
  • Country: gb
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #15 on: July 27, 2021, 10:18:38 am »
After all, you can't get more current from an inductor than what was flowing in the first place, unless you are Nikola Tesla  ;) 

yep, there's always a way :) Especially if you have Cold War budget money :)

https://en.wikipedia.org/wiki/Explosively_pumped_flux_compression_generator
 
The following users thanked this post: Berni

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26751
  • Country: nl
    • NCT Developments
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #16 on: July 27, 2021, 10:46:26 am »
I always put the diode in parallel with the transistor; not the relay coil. Just like the parasitic diode inside a MOSFET.

With just a single transistor the diode would not help with back EMF. It is only in a bridge circuit, that the diode parallel to the upper transistor/FET would prevent the voltage for the lower transistor get too high.
No, but the diode will limit the voltage spike and keeps the transistor from breaking. Because MOSFETs have an internal diode (in some cases even with reverse breakdown / energy specified) you don't need an external diode at all to drive inductive loads like relays, motors and solenoids. Ofcourse you need to check the specs.
« Last Edit: July 27, 2021, 10:52:40 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Berni

  • Super Contributor
  • ***
  • Posts: 4922
  • Country: si
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #17 on: July 27, 2021, 11:06:26 am »
I always put the diode in parallel with the transistor; not the relay coil. Just like the parasitic diode inside a MOSFET.

With just a single transistor the diode would not help with back EMF. It is only in a bridge circuit, that the diode parallel to the upper transistor/FET would prevent the voltage for the lower transistor get too high.
No, but the diode will limit the voltage spike and keeps the transistor from breaking. Because MOSFETs have an internal diode (in some cases even with reverse breakdown / energy specified) you don't need an external diode at all to drive inductive loads like relays, motors and solenoids. Ofcourse you need to check the specs.

Yeah the trick is that the parasitic body diodes in MOSFETs actually act like zenners, so they will safely reverse breakdown once above the mosfets voltage spec and start diverting away the current. Since this diode is formed along the whole channel of the mosfet means that if the mosfet is designed to handle a lot of current, then so is this parasitic diode. It's not like the built in diode in IGBTs where the diode is actually a extra separate component on the die.

If you want to use the forward conducting mode of a regular diode, then indeed they only work in bridge configurations because they divert the inductive kickback into the supply rail as soon as it rises above it.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: EEVblog #1409 - The HUGE Trap of Inductor Back EMF
« Reply #18 on: July 27, 2021, 10:03:11 pm »
Yeah the trick is that the parasitic body diodes in MOSFETs actually act like zenners, so they will safely reverse breakdown once above the mosfets voltage spec and start diverting away the current. Since this diode is formed along the whole channel of the mosfet means that if the mosfet is designed to handle a lot of current, then so is this parasitic diode. It's not like the built in diode in IGBTs where the diode is actually a extra separate component on the die.

If you want to use the forward conducting mode of a regular diode, then indeed they only work in bridge configurations because they divert the inductive kickback into the supply rail as soon as it rises above it.

FYI -- AFAIK, the "zener" is a wear mechanism and not to be relied upon for repetitive service.  Something about hot carriers lodging in the gate, eventually causing breakdown?

MOSFETs of this design (for which there are basically two types, because avalanche robustness trades off with reverse recovery speed) are rated for single-pulse avalanche, partly as a metric of die size and specsmanship, but also partly because, it's exactly that, one time and that's it.  Ever.  A device can only take a few (10s?) of hits like that, AFAIK.  When you do see repetitive ratings, it's considerably lower, corresponding to the larger number of pulses tested (you'll have to look up the quality data or appnotes to see what reliability level they're testing to -- e.g. thousands/millions of cycles..).

So, it might be ~okay~ for solenoid drivers and stuff, depends.  It's definitely not usable like to handle leakage inductance in switching converters.

(The easy solution is clamping the pulse with a proper zener/TVS, either directly, or D-G so that majority current flow carries it instead of as avalanche breakdown.  The latter is what's used in integrated protected MOSFETs.)

BTW, note carefully the current and energy (inductance) used in those tests.  The current used, is generally about as high as the device can handle.  Then the inductance is cranked up to meet Tj(max) after the pulse.  It's not necessarily the case that more current can be handled even at far less energy!

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: chickenHeadKnob, thm_w, sandalcandal


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf