Exactly.
I think I still have some bits hanging around somewhere. 
The local supply bypass capacitor needs to connect from the coil supply to the FET source and the diode directly from the coil supply to the FET drain. This constrains the switching transient within the triangle of those three components.
Of course.
I believe you stated, not explained. Can you put some numbers to your claim?
Well, whatever a “typical Arduino user” is, designs that crash the microcontroller when a relay is (de)activated, are a moderately common topic here.
My statement was to satisfy more advanced folks who know the suppression can be accomplished with a resistor and capacitor and/or combinations of diode, zener, resistor, capacitor. I also thought it was diversionary to mention the dynamic changes of a relay, motor etc resulting from the use of a diode for suppression.
Regarding your insistence the wire leads are of major concern to a hobbyist. Can you put some numbers to your assertation?
see wire transient voltage
Also can you suggest why in this type of application transmission line response applies?
So you know, I only responded to your obfuscation of the simple need to address the "flyback" voltage resulting from turning off an inductive element.
I've designed enough products that have met automotive, military and circuits carrying classified information to know what is really important. BTW I never used a feed through filter, they are a non technical persons way out.
So if you are such and authority on EMC then go look at the typical Bulk Current Injection EMC test and suggest how one might pass this test with a circuit packaged in a plastic housing.
I'm done with this topic. In fact I'm embarrassed I helped drag it this far..... 
Well, we use 0.1 µF for logic circuits where the commutation currents are in the order of mA while the typical relay operates at 90 mA, so this would imply at least 10 µF but 100 or 220 µF intuitively sounds appropriate. 
no numeric examples? What about the resonant frequency of the 100 to 220 (presumably) aluminum electrolytics?
David Jones has recently put up a good video at EEVBlog on some pitfalls of omitting a free-wheel diode (including unintended RF transmission risk):
https://www.youtube.com/watch?v=hReCPMIcLHg
I have avoided discussing angels as per #50.
Everyone here agrees a diode (or somethig else to dissipate the energy) is needed. Your explanation why it is needed is nice. However most of this discussion is about where to place this diode. I think it was already exhausted by reply #5. The fact a guy claims NASA wants diode at each inductor provides very little insight into this until why they want it is explained.
Seriously though.....
I think the "discussion" phase of this thread has been beat beyond death, so I will not try to explain why I believe the diode must be at the source.
The only way this will be resolved (if that is possible) is to build a circuit and measure the results. Alternately a simulation which included various load to switch impedances might might be enough to convince those who believe one way or the other.
John 
Perry, sent this on this thread as the new flyback sticky post is locked.
it would be informative to use the data sheet schematic of a TIP120 or some such to show the diode.
and then put that into the schematic you have on the flyback post.
since it is not across the coil, it is not going to achieve the desired results.
since it does come up from time to time, it seems a worthy addition to that post.
Hi Dave,
I'm not at all clear what you are asking me to do, the TIP120 is a Darlington transistor as far as know, I don't see the relevance.
The core purpose of the tutorial was that I had noticed that people were being told they needed a flyback diode but often there was little or no justification, sometimes it was just something to be taken as a matter of faith from the clever electronics people who new best. I wanted to show that there is a very good reason for having the diode. So my point was you might tell someone they need the diode in some particular situation and point to the tutorial to explain why the diode is needed. I deliberately kept away from specific situations because I thought once I start on things like 'here is a relay with a transistor, put the diode across the relay coil', or 'here is a motor of type x, put the diode across the motor', or 'here is a motor driver note the diodes here in the driver they work like this...', then I'd expand the tutorial way beyond what my original intention was. My intention is entirely summarised by the title: 'Flyback diodes and why you need them', not 'flyback diodes and how to use them'.
If you fancy writing part 2 of the tutorial to show particular situations and how to connect the diodes then I'd be happy to link to it at the end of my tutorial. I think of particular interest is the diodes in a motor driver because at first glance the don't seem to be across the motor.
Im thinking how to use.
To to clarify the use was my point.
Once in awhile there are discussions that you need a diode . And chip so-and-so has a diode
I definitely want to stay away from 'how'. Are you interested in writing that tutorial?
Thanks for the offer.
Alas every time I try to write a tutorial.
I try to dispel myths. (Some think an opto is required to operate a relay ) i try to encompass all the subject and discuss wrong information (some think the diode in a FET is the same as accross the inductor)
I wind up with a textbook.
First you need to decide exactly what the tutorial is about, in this case I guess how to use flyback diodes in various scenarios. Then put together a first draft, most likely it will contain too much material and not be well written. Then go for a beer (or whatever it is you like to drink) and sleep for a while. Then re-write it, removing and improving as you go. More beer, more sleep, more re-writing. Once you think it's perfect get someone to review it, if they are any good they will point out that it's far from perfect. Re-write again based on their feedback. And repeat.
If you decide to do this start by sending me a PM and put your first draft in the PM so I can see it. I won't comment until you are ready. The advantage of a PM is you can see it as it will look when finished and you can invite anyone of your choice to review it by inviting them to see the PM.
If you really don't want to do it, that's OK with me.
Continuing the discussion from Flyback diodes and why you need them:
The use of diode in any circuit is down to the design, however if a manufacturer shows something that looks like a diode in its component diagram, it does not mean that is strictly the case.
Take what looks like a diode across a MOSFET,the automatic assumption that it is just a normal diode is of course erroneous in this case whereas it is a clamping diode designed for the MOSFET.
The best way is to ignore things that are shown on a component diagrams other than as information for the component and design the circuit as though the component is a black box to perform a function.
As far as flyback/ free wheeling diodes are concerned, then as far as basic usefor DIY projects, the article written is more than ample for this purpose.
I like how you say this here.
Because really as a whole look at a circuit minus its discrete components.
The entire circuit in its entirety is a capacitor, inductor, resistor, etc.
From a trace on circuit board to the insulation of a wire, all parts work but also exhibit undesirable behaviors.
Theres a LOT we still dont understand about electrons in physics.
This post altogether reminds me of Lenzs law