Capacitor degradation?

Thread Starter

Classictrial

Joined Aug 23, 2010
30
I am involved with old off-road motorcycles which are fitted with capacitor discharge ignition systems, and these seem to run an awful lot better after the CDI unit has been changed.

Is the capacitor itself likely to deteriorate over time, and if so how long before this is going to make any noticeable difference to running? Have little or no knowledge relating to these things, so any responses are very welcome.

Thanks

Chris
 

SgtWookie

Joined Jul 17, 2007
22,230
Hello Chris,

Aluminum electrolytic capacitors can and do fail over time; and are the most likely suspects in any given circuit that's more than 5 years old; particularly if the circuit is unused for extended periods of time.

The capacitors are basically made of two sheets of very thin aluminum foil that's rolled up, placed in a can-like package, and filled with an electrolyte fluid that is alkaline in pH.

Then the capacitors' dielectric is formed by very slowly charging the capacitor up to its' rated voltage via a current-limited source, until the "leakage current" through the capacitor falls to an acceptable level.

Automotive environments are particularly brutal; extreme temperatures, shock/vibration, corrosion, electrical noise all take their toll.

Even if an electrolytic capacitor is simply unused for an extended period of time, the dielectric will degrade; the longer it is not used, the worse the dielectric becomes. The capacitance is reduced, and the leakage rate increases. If the leakage rate becomes excessive, there will be enough power dissipation in the package to cause the electrolyte to boil, rupturing the package forcefully.

There are other failure modes; for example the package may have not been hermetically sealed or become corroded and the electrolyte simply dried/leaked out, or the internal connections to the "plates" may have corroded or separated.

Replacing electrolytic capacitors with fresh ones is a very common practice, and will help improve the reliability and performance of the circuit.

Buy fresh electrolytic capacitors from high-volume authorized distributors, like Mouser.com, Digikey.com, AvnetExpress.com, Newark.com, etc. Don't buy electrolytic caps from surplus vendors or auction sites unless you have the knowledge and equipment to test and re-form them, or you will likely experience failures within a short period of time. "Stale" electrolytic capacitors are not a bargain for a novice.
 

jpanhalt

Joined Jan 18, 2008
11,087
Do we know that the CDI capacitor is aluminum electrolytic? I have seen polyester/mylar capacitors used that are less than 1 uF and 450 to 600V.

It would be nice to know what the markings are on the capacitor being removed.

John
 

marshallf3

Joined Jul 26, 2010
2,358
Do we know that the CDI capacitor is aluminum electrolytic? I have seen polyester/mylar capacitors used that are less than 1 uF and 450 to 600V.

It would be nice to know what the markings are on the capacitor being removed.

John
Most CDI units use a non-polarized cap for storing the HV that is eventually used to pulse the ignition coil. They were rarely of any quality when new so replacing them with modern equivalents is probably a good idea. They also tended to use ones that could barely handle the required voltage to begin with.

I hate CDIs as they tend to pot the circuits wheree it's near impossible to get to the components and even potting compound has been known to break down over time. If I could easily get into mine I'd replace my output caps without hesitation.

Most large supply houses will have modern replacements, and if you can get them in a high enough voltage the caps they use for speaker crossovers are substantially better but rarely worth the extra $$. http://www.partsexpress.com
 

marshallf3

Joined Jul 26, 2010
2,358
"Stale" electrolytic capacitors are not a bargain for a novice.
Quite personally I don't think they're a bargain for anyone. I'm a "grab bag" addict, if I'm ordering from a place that offers grab bag assortments I'll usually buy one for the heck of it as I've never failed to get a decent amount of useable components from them just to add to my stock along with a lot of oddball stuff that's fun to figure out and/or play with. The electrolytics that come in these usually aren't worth keeping unless they're high value filter caps of a useable voltage, in which case I'll test them for value and leakage before keeping them. I still wouldn't use them in a mission-critical circuit but for normal power supply filtering they're usually fine.
 

SgtWookie

Joined Jul 17, 2007
22,230
John,
I don't know that they are aluminum electrolytic caps. I merely suspect that they are, as they have a distinct size and cost advantage over other types.

Marshallf3,
We're on the same page. It isn't difficult to check capacitance and leakage vs voltage rating, but it's a pretty good bet that the majority of hobbyists wouldn't have the equipment available to perform the tests, and re-forming the dielectric if necessary.

As you're likely aware, it's not difficult to re-form the dielectric of polarized aluminum electrolytic caps. I don't recall offhand how to re-form polarized electrolyics.

If anyone's interested, I'm attaching a .PDF document from a WWII-era polarized aluminum electrolytic capacitor re-forming instrument. It also has abbreviated tables of acceptable leakage rates vs voltage rating vs capacitance value. It's 4.5MB, so beware if you're on a slow connection.
 

Attachments

Thread Starter

Classictrial

Joined Aug 23, 2010
30
Thanks very much for the responses..........the CDI units are potted and very difficult to take apart, but I would imagine the component parts are very low quality, and as some units also include the HT coil then perhaps heat is also an issue?

I think perhaps that performance might be improved using a CDI with higher quality parts, and a separate ignition coil, and this is something I will have to try.
 

SgtWookie

Joined Jul 17, 2007
22,230
Potting compound is used to help provide physical support for the components on the board, which helps with the shock/vibration issues, and also protects the circuit against corrosion and fluid contamination.

You might try using very hot (boiling) water to soften the potting compound.

At a minimum, a circuit board being used in an automotive environment needs to be conformal coated. Conformal coating is a type of clear acrylic lacquer that protects the board/components against contamination and corrosion.

It's tough to remove.

Parts need to be secured to the board so they won't fail due to vibration/physical shock issues.

Modern automotive coils provide a lot more output than the old versions from the early 70's and before. You need to drive the primary side pretty hard though.
 

jpanhalt

Joined Jan 18, 2008
11,087
My experience is limited to hobby versions of CDI. I have never seen an aluminum electrolytic used for the charging (i.e., HV) capacitor, even on cheap ($25) Chinese imports.

John
 

marshallf3

Joined Jul 26, 2010
2,358
Due to the duration of the pulse and energy required it only made economical sense to use plain old 200V or higher regular caps.

Good luck on the potting compound, some is fairly easy to remove, most is a real pain. Denatured alcohol tends to help melt whatever they used in ours.

Most don't generate their own HV either, it comes directly from the magneto coil through a diode to the cap which is discharged across the coil by an SCR. The SCR is triggered by yet another small pickup coil in the magneto or mounted to the cam if it's a 4-stroke engine.
 

timrobbins

Joined Aug 29, 2009
318
If size is not a problem then most switchmodes or UPS will use a 0.1uF upwards AC X rated cap in the mains filtering section, which easily has a DC voltage rating above 400V (sometimes up to 750V, although many don't specifically have a DC rating) - they make for very easy to use replacements. (I have used up to 10uF blocks as replacements for old electro cans that used to have 500-600VDC ratings)
 

marshallf3

Joined Jul 26, 2010
2,358
It would sure beat fighting that nasty potting compound.

I would definitely go with the advanced circuit they illustrate, it's only a few cents worth more parts and it's important that the firing voltage of the SCR be somewhat compensated for temperature variations.
 

Thread Starter

Classictrial

Joined Aug 23, 2010
30
Seems to me that a CDI that doesnt work is probably scrap! Adapted PVL go kart CDI ignitions (based on systems originally intended for garden machinery!) are very popular for older off-road bikes, which seems very surprising as while they seem to have a start retard function, they do not have a proper linear advance curve, with max advance occurring at the same time as max torque.

The running of machines with these systems fitted is very much compromised by the ignition systems fitted, and curiously no one has looked at producing a proper digital inductive system, which seems to me likely to be far superior to crude garden machinery type CDI systems.
 

marshallf3

Joined Jul 26, 2010
2,358
Doing so would involve another million in develoment costs, many more millions in setting up to produce the parts required and in many cases major modifications to the engine so you'd hae a DC source to run the electronics off of.

There currently exist two ways they mechanically advance spark in the older systems. One is explained in that article I linked, the other is by a "swinging spring" method much like original automotive distributors employed.

Used to be you could buy a really nice dirt bike for less than $1,000, you can barely get a half decent used one for that nowadays and new ones can easily set you back $10K.

As with cars, televisions, toaster ovens or what have you they're going to continue building things as cheaply and simply as they can to maximize profits. If a manufacturer does design something outstanding it usually ends up with a selling price nobody wants to pay and they end up losing a ton of money.
 

marshallf3

Joined Jul 26, 2010
2,358
BTW: While still in the very final stages of restoration I ride a 32 year old Honda that at it's time was an engineering masterpiece and is still considered to be one of the best bikes they ever made.

You won't see many in the U.S., most were sold over in Europe and used by "couriers" that delivered things back and forth between companies. Only a few common parts need replacement at almost perfectly predictable intervals so it isn't unusual for one to run 250,000 miles before it finally gets retired.
 
yes the CDI's ive looked into dont have electrolytics, they have polyester types.

Potting compound is very hard to remove too, but there is some easier types. The silicone type is easy if you use a little heat from a heat gun. My friend found a white marble stone like potting mix in his that was still hard to remove but it was at least chippable. The black epoxy stuff is really hard to remove. Ive got a few more CDI's i will be un potting over the next month or so and to try a few different methods.


I have recently had a CDI x-ray scanned to see what its like on the inside.
You can see the x-rays here on my freinds site. There is a lot of info here that we have filed for our CDI explorations!

http://cid-407d7ef0965d3991.office.live.com/browse.aspx/Pure%20Enduro/CDI

I know someone who has made one of those silicon chip CDI kits last week. Dont think he has got it set up on his bike yet though.

I would like to get the circuit worked out for my CDI's which are the analog AC type and then see if there is a way to impove them but without going to a DC system and needing a battery on the bike.

The guy here explains a bit how you can alter the old CDI's a bit with different value capacitors.

http://www.geocities.com/a57ngel/moto/CDI-mod.html

I forgot to mention this before but the Suzuki CDIs ive seen inside that are made by Kokusan Denki Japan, use pretty much all Toshiba components, well at least for the semiconductors. Not 100% sure about the capacitors.
 
Last edited:

Thread Starter

Classictrial

Joined Aug 23, 2010
30
Doing so would involve another million in develoment costs, many more millions in setting up to produce the parts required and in many cases major modifications to the engine so you'd hae a DC source to run the electronics off of.

There currently exist two ways they mechanically advance spark in the older systems. One is explained in that article I linked, the other is by a "swinging spring" method much like original automotive distributors employed.

Used to be you could buy a really nice dirt bike for less than $1,000, you can barely get a half decent used one for that nowadays and new ones can easily set you back $10K.

As with cars, televisions, toaster ovens or what have you they're going to continue building things as cheaply and simply as they can to maximize profits. If a manufacturer does design something outstanding it usually ends up with a selling price nobody wants to pay and they end up losing a ton of money.
A uk manufacturer has already built a couple of prototype inductive ignition bike systems, which are digitally controlled, and can be powered for several hours by a very small lithium polymer battery pack.

The top line competition bikes using old British engines, that are popular here in the UK cost £10k+ to build, so continued use of the analog garden machinery CDI systems seems very surprising!

From the point of view of an older petrol burning motor, the inductive system is far better as spark duration is much longer, which means improved running with less efficient early engine designs, and due to lean burn requirements has also been universally adopted on all modern petrol engines.
 

marshallf3

Joined Jul 26, 2010
2,358
All depends on the manufacturer. Look at how long Mercedes used a mechanical diesel fuel injector pump that required absolutely no electronics at all to feed the correct cylinders at exactly the right timing. :)
 
Top