MGU-K Seals

[hardingfv32]

Honda is the latest to have MGU-K seal issues, but it is said that this has been an issue with other manufactures in the past.

What type of seals are we talking about: oil or water shaft seals?

What is unique about the MGU-K's operation that makes this seal difficult to design?

Brian

[wuzak]

Who else had issues with MGUK seals?

[hardingfv32]

Ben Anderson:

"The seal problem is understood to be a key difficulty that affected other manufacturers when they first designed engines for F1's current V6 hybrid turbo formula, one for which AUTOSPORT sources say there is no 'quick fix'."

Brian

[Facts Only]

MGUK-K Seal is such a broad term, in my experience it could cover:

Sealing of the shaft running into the PU geartrain - Stopping the engine oil running out of the PU into the MGUK - Likely a rotary lip seal.
Sealing of the oil feed and return for the bearings - In and Out from of engine oil to the front and rear MGU-K bearings - Likely an O-Ring face seal (if fully integrated) or dash2 hose fittings.
Sealing the coolant flow for the MGUK cooling system - Could be oil or water - probably same type of seals as oil feed/return.
Sealing of the internal water jacket flows - if Honda are running "in stator cooling" they will need to seal the coolant flow inside the motor windings themselves.
Sealing of the windings - I doubt its this but they may be calling it sealing and not insulation

My guess is that its probably and internal integrated coolant seal that's gone, I think this because they were able to run in lower power mode, lower power means less coolant requirement which means they could restrict the flow and decrease the pressure in the system reducing the load on the seal.

[FW17]

Facts

Why is this being a problem on a 6th generation K-ERS system? Does the doubling of power output from 80 to 160 hp have a significant difference in architecture that previous experiences were irrelevant?

PS I am assuming a change in design every year since 2010

[autogyro]

Surely it would be better to build the KERS into the gearbox?
Cant do it with a lay shaft gearbox, ahh well.

[Facts Only]

Facts

Why is this being a problem on a 6th generation K-ERS system? Does the doubling of power output from 80 to 160 hp have a significant difference in architecture that previous experiences were irrelevant?

PS I am assuming a change in design every year since 2010

Well Honda bailed on F1 at the end of '08 just before KERS was introduced, we know they did the design work for the original KERS but never raced with it. Even if they have been ticking over with a design since then this is the 1st generation of KERS they have actually had to use in the real world. Whereas in that same time Renault, Merc and Ferrari have spent 6 years designing, developing, building and racing KERS units.

That is a huge disadvantage. This is why I have never been particularly confident in Honda making any sort of challenge with their PU, certainly for 2-3 years anyway. The amount of experience that has been gathered by the other 3 engine makers over the last 6 years is vast and money just cant buy that.... well not if you set-up in Japan and struggle to poach rivals staff. I know that they have a couple of big names to come up with the big ideas but you need the guys on the CAD boxes and CNC machines to be able to fire out designs for the nuts and bolts and bits and bobs and seals (quite literally) of the engine and get them made without these sort of niggles, and this comes with experience.

It may only be an MGUK seal but it has wasted a huge amount of their very limited test time, they may have the best PU out there but if they cant run at full power because a small seal doesn't work its all irrelevant.

I'm not trying to run Honda down, I'm impressed with the fact that they have made it to the first test and got running but I think the task is much bigger than people think.

[Richard]

From the Alonso Crash thread:

The seals they are referring to are most likely oil related seals. And McLaren have already stated there were no ERS anomolies.

The MGU-K is usually oil cooled with oil flowing both through and around the stator and through the central shaft of the rotor ..
Based on the 2009 MGU-K layout , I assume it to be the first front seal that stops the through rotor oil flow from escaping at the drive gear end. That seal probably has to cope with shaft speeds upto 30,000 rpm , and possibly at full engine oil pressure if they use that instead of hydraulic oil.

[PlatinumZealot]

A high speed oil cooled motor....
You what's interesting? The seal leakage won't stop the unit from working because the oil is non-conductive (I think) but imagine the huge windage losses afterward?

[R_Redding]

The seal leakage won't stop the unit from working because the oil is non-conductive

Although F1 engine oil is a complex cocktail ,fresh engine oil would be okay to use I'd of thought. I tested some new VW501 pumpe deuse oil with a 500v Impedance meter and got a "infinite" Ohms reading , and the same with new ISO32 hydraulic oil (hardly a definitive test though).

Oil is used to cool many electronic devices and electrical gear in direct contact with the windings, especially transformers ,but that will be using the correct oil for the application.

The MGUK is a highly thermally stressed component due to its size and power output (a discriminated 160hp in a package size similar to a 1hp AC Induction motor) so a fully functioning cooling system is a must ..

McLaren also stated on one MGUK seal failure that it damaged the engine aswell ... so it could be damaging the engine by leaking into places where the build up causes issues and has no scavenging. The Merc MGUK looks to be driven off of the crankshaft utilising the camshafts driving gears ... so a leak could fill those cavities... or the leak could be starving critical bearing surfaces of oil ... but who knows.

Rob

[riff_raff]

The rotor PM materials and the stator laminates and windings all require intensive oil cooling. The cooling oil flow is pressurized and passes into and out of the high-speed rotor shaft thru a pair of slip rings that are sealed with the type of piston ring seals used on turbochargers. Since the oil circuit is pressurized, even a small slip ring seal leak will dump large amounts of oil into the tiny gap between the rotor sleeve OD and the stator ID and create problems. With shaft speeds up to 30,000rpm, the bearings must also be kept from accumulating liquid oil, since this will produce huge viscous losses and heat in the bearings.

[Facts Only]

From the Alonso Crash thread:

The seals they are referring to are most likely oil related seals. And McLaren have already stated there were no ERS anomolies.

The MGU-K is usually oil cooled with oil flowing both through and around the stator and through the central shaft of the rotor ..
Based on the 2009 MGU-K layout , I assume it to be the first front seal that stops the through rotor oil flow from escaping at the drive gear end. That seal probably has to cope with shaft speeds upto 30,000 rpm , and possibly at full engine oil pressure if they use that instead of hydraulic oil.

http://i897.photobucket.com/albums/ac18 ... zmtiau.jpg

I know that not everyone is running a hollow cooled shaft on there MGU-K or MGU-H, it seems a lot of work for sealing. Perhaps this is why Honda is having a problem.

[tok-tokkie]

How do they machine that shaft? It appears to be closed both ends yet has a complex machined oil gallery inside it.
EDIT: With the bigger picture I can now see it is two parts pressed together. All is apparent.

[autogyro]

So much easier without a central rotating shaft.

[PlatinumZealot]

So much easier without a central rotating shaft.

Are you suggesting a disk shaped motor as a better option? That is a higher Rotational inertia.