A Tale of Two Condensers

[legendre]

Some folks hate spark problems. As for myself, I mostly just dislike and distrust condensers.. as I have yet to devise a foolproof bench test to determine their condition. So let’s see if we can make any headway.

I have here two condensers, both removed from a functional Johnson TD20 magneto. We’ll call the first condenser "white" and the second "green". I also have several instruments, a Heath V-7A VTVM, a Heath IT-28 Capacitor Checker and a generic uC-based component tester similar to eBay item #331535438736. Manuals for the Heath instruments are available, upon request.

First, the data from the component tester (averaged from five runs each):

Green – Value = 0.217uF / Vloss = 0.5% / ESR = 2.2 ohm
White – Value = 0.223uF / Vloss = 0.1% / ESR = 0.73 ohm

You’ll note that they are very similar in value (capacitance), with White slightly larger. However, Green has five times the Vloss (0.5% vs 0.1%) and about three times the ESR (equivalent series resistance – 2.2 ohm vs. 0.73 ohm). But since their values are very close, they should both establish the same time constant when used in an R-C circuit.

So that’s next. Using the Heath V-7A (because it’s an analog meter, with a very high upper resistance range – X1Meg ohm) I measured how long it took for the meter to charge each condenser to the point where the needle read "100" on the scale. On the X1Meg scale, this indicates 100 Meg ohm – not exactly "infinity" but a nice, high, somewhat arbitrarily-chosen value.

Here’s that data (averaged from 5 runs each):

Green – 18.8 s
White – 9.21 s

Quite a substantial difference, I’d say. Green takes twice as long to reach the same level of charge, compared with White. In other words, in terms of R-C timing constant behavior, Green is behaving like a 0.44uF cap, rather than its marked & measured value of 0.217uF. Since we know the values measure nearly identical, the only real possibility is that it must be leaking off some of that charge current..

(more to come..)

[legendre]

Now on to a new instrument.. this time, a Heath IT-28 Capacitor Checker. As of this late date, the IT-28 is pretty well obsolete, but it does have one very useful set of functions – a highly sensitive leakage test at actual DC operating voltages from 10V up through 600V.

Each condenser was, in turn, connected to the test terminals, the voltage was set to max. (600 VDC) and it was left to settle for five minutes at which point the leakage characteristics were observed. Rather than a typical analog meter movement, the IT-28 uses a green ‘Eye’ tube as an indicator: Eye open = no leakage / Eye closing or closed = leakage.

White – no leakage observed at any voltage.

Green – initially no leakage at 600V, but within a few seconds the eye began to close. By the time five minutes had elapsed, the eye was fully-closed. Then the voltage was progressively lowered one step at a time, taking note of any leakage. Significant leakage was found all the way down to 350V (fully closed) and showed moderate leakage one or two steps below (~200V range).

Conclusions? If White is to be accepted as a standard (it seems reasonable), then Green is most certainly defective. Interestingly, Green’s failure mode suggests that when in-use, it may give acceptable spark at cranking speeds – for instance, when checking for spark – yet cause a great deal of trouble once the engine has started and run for a minute or two. Bear in mind that Green was once part of a magneto that ran one of my TD20 just fine..

Again – and this is a salient point: Green was apparently working just fine in one of my TD20 motors, though my second TD20 never ran right with Green in the magneto. What is the source of this apparent discrepancy – was Green the culprit, or simply part of a larger issue?

Either way, Green’s a piece of crap, it was most recently in my troubled TD20 motor, and it’s not going back in there. (What if all that grief was caused by a condenser with a sneaky little failure?). In any case, I have new parts on the way.. so we’ll see what, if any effect, they have on the running quality of bogey TD20..

[legendre]

When I began this, I knew it would have a limited audience.. but the silence is deafening.. lol.

Did I exceed the geek-limit? 🙂

[debe]

No not realy. But Component testers are not good for testing Ign condencers as they dont use high voltage for the test. They will give the value & ESR, which is great if you are repairing switch mode powersuplies. I use component testers for that type of work. For Ignition type condencers I use either a Mercotronic or a Neon lamp & HV to test capacitors for leakage open circuit / shorts. For me the only test is under HV conditions as that is where a capacitor is likely to break down. Horses for courses.

[wedgie]

It’s interesting and good stuff to put out here, even if I don’t understand half of what you wrote. I’m sure there are lots of members that do know, so keep posting. Maybe one day all the pieces will come together in my little pea brain 🙄

It would be interesting to check several new condensers and see what(if any) difference can be observed between them.

[debe]

I think you just proved that the Heath IT-28 tester is the best tester that you used as it tests under voltage. The faulty capacitor eventualy showed its self, probably due to internal heating at that voltage.

[legendre]

quote debe:

I think you just proved that the Heath IT-28 tester is the best tester that you used as it tests under voltage. The faulty capacitor eventualy showed its self, probably due to internal heating at that voltage.

Perhaps so, but there are a couple of ‘buts’.

– You’ll notice that the component tester did find +significant+ differences between the parts, in the Vloss and ESR figures

– The IT-28 has, at least once, failed to find a truly, horribly faulty condenser – which showed no leakage, and was measured as close to its marked value as possible, using that feature of the IT-28.. (I didn’t have the component tester at that time).

I sure wish that I had the aforementioned condenser, so I could run it through the same tests – but that was years ago. It was from a Motobecane 50V moped, and unlike most, they are rated at 2.0uF (not 0.22uF).. and that’s the only time I’ve ever seen such a part used on any older ignition.. 2uF!

[chris-p]

I think you may be overthinking it a touch.

Hook it up to a tester, made for these condensors. They pass or fail. Easy Peasy. Even a dumby like me can test them.

[legendre]

quote Chris_P:

I think you may be overthinking it a touch.

Tell that to the engineers who designed the tester you’re about to suggest.. how do you think they determined the design parameters? 😉

quote :

Hook it up to a tester, made for these condensors.

I’ve used several different instruments, as you’ve read. Which others would you suggest – I’m all ears! I don’t own one of the Merc-o-tronic sets, or any other tester geared towards flywheel magneto systems. I’d sure like to have one at this point, though!

Anyone have a unit for sale? Doesn’t have to work perfectly, I can rebuild & repair as necessary.. in fact, that’s how I’d prefer it, as I’m going to go through it in any case.. 😉

quote :

They pass or fail.

Oh how I wish..

[legendre]

quote Wedgie:

It’s interesting and good stuff to put out here, even if I don’t understand half of what you wrote. I’m sure there are lots of members that do know, so keep posting. Maybe one day all the pieces will come together

If there’s something you don’t understand – then just ASK! There’s no shame in that, only the opportunity to learn.

This is primarily an educational forum, so don’t feel so shy about using it that way… we all need to start somewhere, and I tell you, this stuff just ain’t that complicated. I will do my best to patiently instruct anyone in the basics of the relevant electrical theory, as time and space permit.

If you want a head-start, look up "Ohms Law" and the "Power Formula" – once you understand these basics, the rest will be a cinch.

[Author]

Posts