Alternator rotor Inductive Kickback

TwoManyXS1Bs

BBQ Hunter
Top Contributor
Messages
14,670
Reaction score
15,273
Points
813
Location
Fredericksburg, Texas
Inductive Kickback of alternator rotors

Back in the `60-70s, as alternators replaced generators, and semiconductor electronics started appearing in automobiles, an electronics phenomen started causing problems for the automakers, Inductive Kickback. Essentially what happens is that current flow through an inductive device (like the alternator rotor) doesn't instantly stop when a circuit is switched off. The current will want to continue flowing like a freight train with bad brakes. What happens then is the now-disconnected positive supply line voltage will momentarily go below ground voltage, seen as a negative voltage spike on a scope. Inductive kickback spikes could be seen on some of those earlier cars on the order of hundreds of volts, but since they were point/coil ignitions with relay type regulators and vacuum-tube radios, nobody got hurt, except for the occasional backfire.

Enter the Yamaha XS-650, with it's large inductive rotor. The rotor is still trying to pull current when the ignition is shut off, and the brown ignition supply line is now isolated from the battery, and floating, subject to the inductive kickback of the rotor, and will spike to negative voltages. On the older relay type regulators, the other side of the rotor is grounded, and will drain there. On a semiconductor regulator, the green grounding line will spike 'high', stressing the regulator. With a points/coil system, if one of the points happens to be closed, most of that will be drained through a coil, possibly causing a backfire through a carburetor.

The early models, with point/coil ignitions and relay type regulators produce a lot of line noise, but with no semiconductor devices, were essentially immune to this and ran fine.

Fast forward to the later models with their solid-state regulators and ignitions, and you can see a potential problem. Typical nonprotected semiconductor devices of the period were damaged when exposed to inverted voltages of 5v or more. Well designed electronics included some sort of protection. Modern semiconductor devices typically have diode-protected inputs built-in.

The volume of posts about failed TCI's and regulators lead me to think that the aging, well-used, heat-cycled protection diodes in those devices appear to be failing, leading to the cascade of failures deeper inside the contraption. And the plethora of cost-reduced and quality challenged electronics available to the bike maintainer adds to this issue. Also, spikes could occur with bouncing brushes, and unusual methods of grounding the rotor's green line.

Standard electronics design involving inductive devices includes the use of a diode to shunt this inductive kickback back into the inductor, and are known as protection diodes, freewheel diodes, flywheel diodes, ...etc. This removes most of the negative spike from the supply line, and protects all devices on that line. View these attachments for typical implementations of Inductive Kickback Protection Diodes. For more info, Google "inductive kickback", "protection diode", "flywheel diode" and "freewheel diode".

freewheel.jpg flywheel.jpg 2-Methods.jpg

An excellent example of this kickback was observed by member fiveohindc, who inadvertently attached a neutral light "LED" backwards, noting a momentary glow when ignition was shut off:

http://www.xs650.com/forum/showthread.php?t=26974

How often do you hear, when the rider of a later model shuts off ignition, then later it fails to restart. This could be caused by this 'inductive kickback', and the damage could be permanent.
 
The experimental idea here is the addition of a suitable protection diode onto the XS650 rotor supply lines, the green and black lines, in order to remove this kickback from the brown ignition supply line, to help save any precious modern electronics.

Based on rough numbers of max 15v supply, 5 ohm rotor, 3 amp current, this diode would need, for a 100% safety factor, a minimum rating of 30v, 6 amp, 6 watt. NOT your typical 1N400x diode. What I plan to use is a 1000v, 6-amp, 0.7v forward-volt, 9mm diameter radial-lead rectifier diode, an 89¢ Digi-Key item.
The electrically ideal mount location would be at the rotor brush screw lugs, but due to heat issues, the better installation point would be at the rotor/regulator connector.

The connection is quite simple, shown in this modified schematic:

XS-Exp-Charging-Diode.jpg



I have no idea of the method of regulation of the rotor grounding line (green wire) in the current crop of semiconductor regulators, whether it's implemented as a simple linear transistor, or as a pulse-width modulated grounding transistor (which will produce spikes), or something else, or if various different methods are used by the different offerings. I also have no idea of what type(s) of protection are designed into these things, whether to protect the transistor, the rotor lines, or both. This is simply another layer of protection, applied at the source.

All the preceeding is theoretical, but standard practice. The values were computed for this application. I have NO late model to test this experiment, but have recently acquired a portable triggered/memory handheld scope for analysis. This is a well-known and managed phenomena, so I have confidence in it's implementation.

Currently, I see no conflicts in adding this protection diode, but openly welcome comment. Please do not implement this modification until it has been peer reviewed by the appropriate electronics experts here, and the diode device value is agreed upon, and potential conflicts are identified and resolved.

Adding this "protection diode" will NOT fix existing problems. If your electronics are fried, they're gone.
This will NOT give you anything extra, the only way you'll know it works is if your electronics don't fry as often, or at all.
This does NOT apply to Permanent Magnet Alternators (PMA) as they do not use an inductive 'rotor' or 'exciter'.
I believe that PamcoPete's electronic ignition system already has protection built-in. This could extend that protection.

I'm not ready at this time to do this experiment, but will post back here later when I get a round tuit...
 
Nice. Interesting stuff but I've never been particularly good at it, quite hard to digest some things. Inductance, reactance, I want to know, I don't want to know lol.

Been reading on inductance, still scratching my head a bit but I understand how the formula works relating to voltage, inductance and current change over time to produce the high voltage kickback with sudden current stop. That makes sense...

You know I saw a round tuit in a shop once when I was little... never seen one since.
 
I had not thought about spikes being a problem for those who run standard alternators however in correspondence I have had with Mark Whitebook (Probe ignitions) he said it was a good idea to fit a catch diode to the coil of the starter solenoid. This was also to reduce voltage spikes in the electrical system when the coil is de-energised.
I fit these diodes as a matter of course now.
 
It's been a very long time since I've used oscilloscopes, and those things were big/heavy and had to be wheeled about on welding carts.

Enter the new world of portable, handheld, pocketscopes, at 1/100th the price.

DSO-Nano-DSO201.jpg




Hooked this one up to the Brown (switched power) and ground, and got a trace of the brown wire's voltage swing when the ignition is turned off.

Kickback-Scope.jpg


Each vertical gridline is 5 volts.
Each horizontal gridline is 10 milliseconds (0.010 seconds).

The blue line is the voltage trace.
It starts at the left at 12 volts, then upon 'ignition off' it rapidly drops to MINUS 43.4 volts!

This negative voltage slowly dissipates thru other electrical devices attached to the brown wire supply, like the coils, neutral light, mechanical regulator solenoid, flasher relay.

After about 30 ms, the voltage rises to about -5v, which is the tolerable minimum for most unprotected semiconductor devices.
After about 100 ms, the voltage rises to about -1v, then takes about 1 second total to completely neutralize.

I hope y'all can see and understand this.
If necessary, I can snapshot and post the scope trace.
 
It is surprising how long it takes the back emf to dissipate. It will be interesting to compare after you have fitted some diodes.
 
Hey, Signal! Pretty serious spike, huh? I suspect that these early models may present the worst case scenario, as later models have more contraptions connected to the brown supply line, and run with headlights 'on', which would provide more sinking for that back emf. But, they also have a 'kill switch', which isolates the coils, and creates a new scenario.

Rectifier diodes come in a variety of sizes, mostly determined by the current rating. The 6-amp diodes are not tiny and will take a chunk of space.

Rectifier-pkg.jpg



I got a bag of (10) 6A10 rectifier diodes for a couple bucks, that's 20¢ apiece.
The 6A10 is rated for 6-amps, 1000 volts, stable up to 175°C (350°F).

6A10-Diodes.jpg
 
The diode can be hooked up at the rotor's brush block, the source of the kickback, but could be an annoyance during brush servicing.

The diode could be attached at the mechanical regulator, bridging the green and black (ground) wires. Then, it's out of sight, out of mind.

But, I decided to attach the diode at the regulator connector, on the harness side, bridging the green and black wire connectors. You can just see it below the scope probe.

Then, ran another scope trace, using the same scope settings to show a before/after comparison.

Kickback-Scope02.jpg


The negative voltage spike is practically gone.

To get a better look at the trace, and get better voltage values, I zoomed-in the scope settings.

Kickback-Scope03.jpg


Each vertical gridline is 0.2 volts.
Each horizontal gridline is 50 milliseconds (0.050 seconds).

The blue line is the voltage trace.
It starts at the left at 12 volts, then upon 'ignition off' it rapidly drops to -0.89 volts, clamped by the kickback diode.

This negative voltage slowly dissipates thru other electrical devices attached to the brown wire supply, like the coils, neutral light, mechanical regulator solenoid, flasher relay.

The voltage slowly rises, and after about 250 ms, it's neutralized.

Good 'nuff. Now I don't have to worry about blowing-out new-fangled electronic doo-dads...
 
Good 'nuff. Now I don't have to worry about blowing-out new-fangled electronic doo-dads...

Mp-3 players, gps, phone as a speedometer are these the doo-dahs you speak of?
Examples please because you are so far above my understanding 'm not sure what you are talkin bout.
 
Hey, WER! Yep, them kinda things, including fancy ignitions, USB power adapters, and more.

A couple of years ago I installed one of those tiny 2-wire digital voltmeters. It eventually blew-up. The 2-wire version doesn't have reverse-polarity protection, but the 3-wire version does. I installed the 3-wire version and it worked fine (so far). Lately, I've installed the Volt/Time/Oil-temp digital meter, claims to have reverse-polarity protection, but I'd rather avoid that risk.

Many 12v gadgets have reverse-polarity protection, so you don't toast your new toy if you accidentally hook it up backwards. But, that -43v spike shown earlier may be a bit too much for some of these things, or repetitive exposure (each time you switch ignition off) may eventually ruin the protection, then the gadget.

So, this is just a cheap fix to avoid those issues...
 
Thank you for posting the after results. A simple fix that could prevent many headaches. Good work.
 
I have been using IN4008 diodes mainly because I have some on hand.
Here is the information I received from Mark Whitebook.

"The diode must be rated to momentarily carry the solenoid's primary current, which is generally right around 3.5A to 4A. I use diodes from the 1N4001 to 1N4007 family (i.e, 1N4002, 1N4003, etc.) each increase in the last number of the part indicates a 100V higher "withstand" voltage - I have a bunch of 1N4004 types here, which withstand up to 400V applied in the reverse direction. All diodes in the family are rated for up to 30A surge current, which is many times what's required in this application. These types are extremely available, cheap, and rugged. A bigger diode, rated for higher average current, won't hurt, if you've got one around".
 
Very interesting 2M. I don't recall reading this thread when you 1st wrote it, must have missed it. So, I now have a "solid state" VR115 regulator installed. Do I need this diode now? Got any pics of how you mounted it in your regulator plug? Also, a link to the bagful you bought? Can we get them on eBay?

I'm not much up on electrical stuff but I assume from your pics that the "arrow" in the diode symbol points to the striped end of the diode? And from your wiring diagram, that striped end goes to the green wire?
 
Very interesting 2M. I don't recall reading this thread when you 1st wrote it, must have missed it.

Hey, 5twins. Completely understandable. Super-busy forum. And you chasing literally hundreds of carb and other issues.

So, I now have a "solid state" VR115 regulator installed. Do I need this diode now?

Need, no, not really. Those auto regulators are built to a standard, and should have a better protection diode, for the higher current demand auto armature. The folks that designed these were at the forefront of this "inductive kickback" issue. But, installing one gives you a second layer of protection.

In the electronics industry there's 4 grading levels:

Consumer grade - For the general public, usually indoor usage, toys
Commercial grade - For business-critical and harsher use
Automotive grade - Very harsh use, heat, vibration, moisture, wetness
Military grade - Withstands severe environment, atmospheric changes, and EMP

Got any pics of how you mounted it in your regulator plug? Also, a link to the bagful you bought? Can we get them on eBay?

Well, there's a pic, above, post#11, 1st pic, near the scope's probe, but not very clear. It simply bridges the black and green wires, at their blade connectors. I guess I should post a better pic.

Digikey, Mouser, eBay. Google or eBay search on "6a10 diode" will get a flood of results. I try to select the most common and economical stuff for these kinda things.

Bag of (10) 6A10 diodes, $4, on eBay: http://www.ebay.com/itm/281908022482
Cheaper if you want to wait on China post.

I'm not much up on electrical stuff but I assume from your pics that the "arrow" in the diode symbol points to the striped end of the diode? And from your wiring diagram, that striped end goes to the green wire?

Yes, the arrow points toward the stripe. And, I'm glad you brought this up. Very important that the stripe end goes to the (+) positive wire. That would be:

For the 70-79 systems: Stripe end goes to the green (regulated power) wire, other end to the black wire (ground)

For the 80-83 systems: Stripe end goes to the brown (power) wire, other end to the green (regulated grounding) wire

This same rule applies if adding a protection diode to the starter relay. Bridge across its coil wiring, with the stripe end on the (+) positive side...
 
So this spike is why PAMCO's tend to throw a spark at key off when the engine's not running? Been waiting for a spark with a just right crank position to shove the bike off the side stand to flop on it's side........ I think the standard crank stop is a bit before TDC so it's usually harmless. Any comments about using a diode on a PMA setup?
 
Would the 6A10 diodes work on the starter relay as well?
 
Back
Top