Apparently the Rotax starter draws up to 300A during the first second. So all Rotax installations need to have a battery that has a low enough internal resistance to be able to supply this current with less than 12V voltage drop (assuming a 12V installation).
Therefore, I fail to see how an alternator, that is able to supply several tens of amps is even able to produce voltage spikes more than a few volts. That is, unless the cabling topology is completely braindead.
@Peter: I repeat and confirm: the C that Jesse mentioned, and is indeed a standard requirement in any Rotax91x installation, has nothing to do with suppressing transients. If that were its prime function, it would never be of the electrolytic type. A low ESR might allow it to help there, yes, but I never saw low ESR as a specification.
The C is there for complete unrelated reasons, as explained. It might be a good idea to bypass it, though, with a parallel capacitor with good HF/transient handling, ceramic or so. Though I never heard of such a reommendation, it can never hurt.
@Tom: 300 amps seems a bit much, even if it is only about the very first second; I always saw figures of 80-100 amps. Curious to see your source. But it doesn’t really matter: even if the maximum starter current were only 50 amps, your point is perfectly valid.
Interesting about 300A. That is about the worst possible with Lycos and the Skytec starters especially the 12V ones… as a number of owners have found out via welded relays 
In my aircraft there is a thick wire running to the starter only. The wire powering the rest of the airframe comes off near the battery. But this topology will differ by airframe.
Peter wrote:
Filling it with grease would be neat but I see no easy way of doing it. One would need to construct a gadget which goes over the end, allows the centre wire to pass through, and has a grease nipple on it.
Such devices exist. My Dad lent me one for doing the Auster’s trim cables.
Jan_Olieslagers wrote:
Curious to see your source
That was I think a 912 installation manual in google cache which was no longer at the original location.
But even if it’s only 100 Amps this doesn’t change my point.
tomjnx wrote:
If anything, long thin wires should dampen fast rising transients.
True, this is why long thin wires to the capacitor are useless (also for reducing rimple).
tomjnx wrote:
How do you know? Or is this conjecture? Switch mode power supplies usually generate quite some transients on their own, furthermore they typically have a huge input voltage range, and thus quite some headroom for transients and “dirt” coming their way from the outside…
True, sure it could be a conjecture. Have had lots of issues with several switch mode avionics products, same fault on several Rotax PM generator powered aircraft. I haven’t seen this fault on Lycoming / Continental powered aircraft, neither on Rotax aircraft with an external alternator. Oscilloscope view reveals output with lots of transients. at the bus. These don’t occure when switching from internal alternator to external alternator. Peter wrote:
Peter wrote:
It probably depends on where the capacitor is. If it is at the start of the long wire then you get a poor transient response.
This is why IMHO the capacitor is placed incorrect / incorrect wiring.
Peter wrote:
Also I am sure some voltage regulators have marginal stability. It’s a crude system – controlling the voltage by varying the field current
True, tough this system is as Jan also wrote with permanent magnets. The AC output is going in to the regulator, which rectifies and regulates. Quite poorly, there are lots of issues with this regulator.
Jan_Olieslagers wrote:
It might be a good idea to bypass it, though, with a parallel capacitor with good HF/transient handling, ceramic or so. Though I never heard of such a reommendation, it can never hurt.
That wouldn’t be much of help, due to the long thin wiring. Your basic idea of using multiple capacitors parallel is always a good idea.
Jan_Olieslagers wrote:
But it IS queer to see those accessories on the business side of the firewall, in my craft they are on the much more protected – and less hot – cockpit side.
We don’t see what the top cowling look like. It may very well be a duct there guiding fresh air to that component. I doubt it is needed though, lots of air comes through the front opening of the cowling, and the engine doesn’t give off much heat being water cooled. It’s certainly higher than ambient, but not that much. It also depends how the heat exchanger is mounted.
You really do want to mount as much as possible on the business side, especially things that may eventually heat up.
No I don’t.
Fair enough, but the idea of a firewall is to keep the fire on the outside, not to lock it on the inside. There is no right or wrong here. A Sonex with the ULPower engine crashed because he mounted the FADEC module on the engine side of the firewall without ensuring proper cooling. In my opinion as much as possible should be on the outside, and cooling issues should be handled there. For a Rotax with a heat exchanger, I don’t see what cooling issues there should be unless you for some reason force the exhaust air into the engine compartment.
And back to Peter’s mystery aircraft: all of the cabling is red – can it be fire-resistant? I only know of tefzel but that is usually white-coloured?
Not that fire-resistant cabling is a requirement for me, but if this installation is so exemplary than it definitely ought to have that.
