As much as I wish it was, I'm afraid the dyno these tests were performed on does not belong to me, no. Yes, the VAG engines do indeed keep all valves closed on deactivated cylinders, but this is not what decreases pumping losses. Freeing up airflow through the intake system reducing pumping losses, get rid of the horrendously inefficient butterfly, and you do exactly that. Why else do you think the highest performing N/A engines such as the current breed of 2.4L F1 engines do not use a butterfly, but instead a barrel throttle? Pumping losses again - the lower pV loop becomes smaller, albiet very small and the majority of flow control through a butterly is in the ~0-50%, but not to be ignored in those sorts of applications, too small for road yes, however.
Think about it another way, systems like Fiat's new Multiair (okay, not all that new as we're now in 2013) - open the throttle wide open, then to allow a say 50% cylinder fill, open the valve fully, then snap it shut fully half as early as normal. That way, you've created the 50% cylinder fill without throttling the air flow - you're filling with as little restriction as possible, then closed. The engines still have throttles to allow you to still drive the car if the actuators that control oil volume flow to the valve actuators fail and leaves you with 'normal' cam operation.
If you performed steady state dyno tests with a multiair system, where you snap the valves shut where you normally would, then 80%, 60%, 40% etc - you wouldn't see the same increase in the lower pV loop as you would if you restricted cylinder flow with a throttle. People seem to get confused with variable valve lift, and the new systems such as Multiair - those systems allow maximum airflow, then snap the valves shut early.
There will ALWAYS be pumping losses, so saying a throttle eliminates them is not strictly true - it only eliminates the losses 'eliminateable' for a given engine design. The very fact you're pulling air through the gaps in an engine bay, into a manifold, through a filter, through the ports etc always creates some sort of pumping loss, that's inevitable - get rid of those, you no longer have an engine.
However, I'm afraid you've still misunderstood - even when staring dyno data directly in the face *sigh*. VAG and other manufacturers keep valves closed for a number of other reasons; intake / exhaust ports have restrictions of their own, and to keep the deactivated cylinders hot. Think what happens to thermal losses through the activated bore walls if neighboured by a cool deactivated cylinder. I think you'll find if you look a number of engines, there are many which still deactivate cylinders with 'traditional' camshafts, as I've already described. Clever valve trickery simply increases the potential for gains.
The below quote from my earlier post explains exactly why compressing air in a cylinder does not contribute to pumping losses, pumping losses should not be called pumping losses at all, but rather 'intake flow losses.' The below explains how you get back all (minus an almost insignificant amount) of the energy you put in to compressing the air in the first place.
Talking in terms of theory however, yes - you're right - closing the throttle does reduce some kind of loss, the losses caused by having to compress a full cylinder charge, but not pumping losses. Even then, the reduction in loss is so small - as with these 'compression losses', most of the work done by compressing the air is returned as the piston is pushed back down the bore by said pressure - only the energy converted into heat by compressing the air that is then conducted through the bore walls and head is actually lost.... A very small amount indeed! Thus, completely outweighed by the huge increase in pumping losses with a closed throttle.
I'm afraid there is a vast amount of incorrect information floating about the internet, and thus it may be easy to believe the almost everybody on forums around the world has it wrong, but I'm afraid in this case - they haven't. Pumping losses are caused by intake restrictions. Please, using the provided dyno data, explain what else is causing the lower pV loop to decrease in size with an opening throttle? Also, please explain why BSFC decreases with an opening throttle, and IMEP increases with an opening throttle with respect to percentages of upper and lower pV loops?
I suppose you could always fire Honda an email and kindly explain to them that their website is wrong, and they haven't understood pumping losses with so many Championship wins to their name...
http://world.honda.com/automobile-techn ... CM/detail/
"Pumping loss is the air resistance incurred when the pistons pump intake and exhaust gases through the cylinder. It is particularly high during intake with the throttle closed."
Note the wording,
through the cylinder, not
in the cylinder - if you allow the air to pumped
through the cylinder more easily, pumping losses go down. The only reason we're not discussing exhaust systems here too if they're almost always fixed, conduct the same tests on a dyno - but with a throttle in the exhaust (clearly, please don't attempt this with a brass throttle butterfly) - and you can create pumping losses by throttling the exhaust.
Or these lot
http://www.mechadyne-int.com/vva-refere ... -si-engine
Or Fiat, who you are sure have wasted a decade developing their multiair system
http://www.fiatpowertrain.com/FPTPressR ... GENEVA.pdf
Not sure BMW would be happy about having got all their information wrong in a published book
http://books.google.co.uk/books?id=AZKF ... mw&f=false ,
http://www.bmw.com/com/en/insights/tech ... ronic.html
I'm not trying to be an arse, I'm simply trying to make you see it from a different point of view - and realise that, unfortunately - you have it wrong. I think I've now run out of different ways to try and explain it.