Penske 8780 shock

[GSpeedR]

Dave, do you have idea of how the in-damper inerter is configured? I wonder if the resultant inerter forces are in series with the 'normal' damper forces or with the gas spring force(?).

In a good damper (with a correctly sized & loaded reservoir) the gas spring force of a non-through rod damper should constant, and equal to the charge pressure multiplied by the rod area. It acts exactly like preloading a spring, and has (or should have) no impact on the dynamic behaviour on the suspension.

An inerter acts (normally) directly across the damper. It generates a force proportional to its "equivalent mass" times the acceleration on one end of the damper relative to the other. In a similar way, a (coil over) spring generates a change in force proportional to the displacement of one end of the damper relative to the other, and an ideal damper generates a change in force proportional to the velocity of one end of the damper relative to the other.

Hence a suspension transfer function, comprising spring stiffness K, inerter equivalent mass M, and a damper strength C, can be written

,

where is the frequency, and is the square root of -1.

This demonstrates that an inerter acts to reduce the dynamic spring rate at low frequencies. It has a number of other characteristics, useful or otherwise, & interesting things start to happen when the dynamic spring stiffness becomes less than zero...

I hope this helps.

That is a nice summation. Thanks. I recall from Malcolm Smith's Cambridge paper that the tested configuration was an inerter in series with a damper/spring, though that was a single DOF system. So parallel arrangement must be better for a motorsport application, it seems.

So in the undamped case, dynamic stiffness has a zero at in the admittance transfer function, is this typically matched to problematic frequencies as one would tune a mass damper, or is that too small-minded? I imagine these elements can be quite violent if not properly tuned.

[DaveW]

If I remember correctly, Malcolm Smith's reason for re-inventing the inerter was to gain access to the considerable knowledge base of passive electrical filters. I say re-invented because, whilst he devised the term "inerter" (I believe), he was not the first to use the mechanism. The first reference I have found was in US2471857, and the first reference to its use in a suspension was in US02856179, in 1958.

Most sane people contemplate its adoption as a "3rd" element in a suspension, inserted between rockers, either stand alone or in combination with "heave" dampers & springs. The equivalent circut diagram becomes quite complicated in that application, involving springs and dampers in series & parallel. However, it turns out (surprisingly, perhaps) that it also can work as an element inserted in parallel with the main suspension elements, as observed by Pierce89.

Importantly, perhaps, the "installation" compliance of a suspension becomes a significant element when inerters are used in suspension elements. That fact changes the single pole "lead" transfer function into a more complex two pole all pass filter (2 poles & 2 zeroes), and very similar to the transfer function of a damped engine mount. Lots of fun....

[GSpeedR]

I reference Malcolm Smith only because his paper(s) are widely accessible. I'm not even sure he devised the term 'inerter' as I have textbooks which call the accelerance transfer function:



the "inertance", though I understand Smith has worked with these elements long before his publications in 2000-2002.

There seems to be some pretty nasty amplification above the resonant frequency (negative dynamic stiffness) in both parallel arrangement and including the series installation stiffness. Lots of fun, indeed.

[DaveW]

Thanks for that. There isn't much that is new in the world....

I believe that some teams have (mis)used inerters to destabilise the hub modes as a way of generating heat in the front tyres. Effectively exchanging mechanical grip for "chemical" (temperature) grip. Probably causes problems with wets...

[joncho]

The damper can also be outfitted with Penskes 3000 Series Active System which expedites rig and track development. The 3000 system allows teams to electronically develop curves, save them to a data base, and easily upload them with the push of a few keys.

Will like to know more about this in the Indycar racecar link but nothing is in penske webpage

[DaveW]

Will like to know more about this in the Indycar racecar link but nothing is in penske webpage

This might help:
http://www.penskeshocks.com/PRESS_2009-10-01.php

The article is a little contradictory ("At the heart of the damper, are standard Penske pistons, shims, etc. to maintain passive/active correlation. However, the control valve allows for an infinite spectrum of damper curves within a wide damping range...", for example). At least on the face of it, it shouldn't be patentable (more than one manufacturer has been doing something similar for several years).

I see that Penske use the magic "DOE" acronym. That reminded me of a test where one customer announced that he planned to do a "DOE" optimisation. He carried out 96 runs (if I didn't lose count), ran his software & got nowhere. With the benefit of a view of a couple of his runs, I was able to rescue his day with four additional iterations - matrix search vs steepest descent. I'm too cynical, I know.

[joncho]

Thank you Dave but I dont understand whant the control valves are and where are
how if shock piston, shims, etc are standard the curve can vary?

and another thing is that the 3000 serie is different from 8780 damper/shock that has mass damper? I am right?

[DaveW]

... I dont understand whant the control valves are and where are
how if shock piston, shims, etc are standard the curve can vary?

Good questions. They need to be answered by Penske people, really.

... another thing is that the 3000 serie is different from 8780 damper/shock that has mass damper?

I imagine your are correct but, again, a Penske input would be welcome.

In general, the three levels of damper "build" are:
1. Fixed characteristics,
2. Switched characteristics,
3. Remotely switched characteristics.

In most cases, both for race & road applications, dampers operate at level 1, at least during a session. Road vehicles & F1 vehicles usually use non-adjustable builds, for reasons of cost for the former, and packaging & weight for the latter.

Many race dampers are at level 2, mainly for cost reasons, but also because damper settings are often changed between qualifying & race, and wet & dry running (something that F1 have regulated themselves away from).

Level 3 dampers are not unknown, but are not common (with the exception of "adaptive" road car dampers), largely because there is an overhead in packaging, cabling, power consumption, etc.

It is usually not made clear (or, perhaps, not widely understood) that most damper designs cannot be migrated from one level to another with absolutely no changes in damping characteristics. Thus, for example, road car R&H engineers will usually work only with level 1 dampers, claiming that level 2 dampers "are not helpful". Again, many R&H engineers privately state that cars fitted with "adaptive" dampers are seldom an improvement over the same car with well set-up level 1 dampers.

At the other end of the spectrum, one F1 team last year packed a couple of sets of level 1 dampers for a race weekend (set up by simulation), & no level 2 backups. They found their race compromised because the simulations were wrong (belying the belief that mechanical set-up is unimportant, & simulation can replace testing) ...

Incidentally, I can't speak for other designs, but Multimatic DSSV dampers can migrate between the three levels with no change in characteristics, and a Level 1 damper can usually be produced with no iterations starting from a free-form damper specification.

[hardingfv32]

This could be a similar type of test shock from JRi:

http://www.jrishocks.com/assets/CAS-VP%20TechSpecs.pdf

I have a PDF of an article from 2008 Racecar Eng/Stock Car Technology, "'Shock and Awe" that details its use. PM me with your Email.

Brian

[Belatti]

At the other end of the spectrum, one F1 team last year packed a couple of sets of level 1 dampers for a race weekend (set up by simulation), & no level 2 backups. They found their race compromised because the simulations were wrong (belying the belief that mechanical set-up is unimportant, & simulation can replace testing) ...

Unbelievable! I knew F1 teams used "level 1" dampers because of weight and packaging as you said, but I cannot believe that an F1 level engineer could be so obtuse. In a level where we have got 15 cars within 1 second in laptime, there should be no such things as "unimportant".

[DaveW]

Unbelievable!

Exactly my reaction to the news...

This could be a similar type of test shock from JRi

Your reference contains the statement:

Unlimited Damper Flexibility
> Each shock can reproduce 5 independent needle/jet adjuster configurations usually requiring 5 different passive shocks or totaling 20 passive shock matrices.
> Each shock in the system is individually adjustable in compression and rebound.
> Configurable ECU is capable of open and closed loop control algorithms with user defined damping curves.

I can accept the first two bullet points, although they might be at odds with the "accurate emulation" statement later in the reference. The third is a noble objective but, in my view, requires a very large increase in capability, & characteristics might be difficult to transfer to another design of damper.

In my experience it is not that easy to get two copies of the same damper to match accurately. "Accurate emulation" requires the port flow characteristics to be identical as well as gallery restrictions, leakages, etc.

I would very much like to rig test them on a vehicle.

p.s. I think the "safe" way of producing a damper range is to start with adjustable dampers, make them non-adjustable by removing the adjusting hardware, and make them remotely adjustable by modifying the adjustment mechanism. In all cases the internal flow controls are unchanged.

[GSpeedR]

This could be a similar type of test shock from JRi

Your reference contains the statement:

Unlimited Damper Flexibility
> Each shock can reproduce 5 independent needle/jet adjuster configurations usually requiring 5 different passive shocks or totaling 20 passive shock matrices.
> Each shock in the system is individually adjustable in compression and rebound.
> Configurable ECU is capable of open and closed loop control algorithms with user defined damping curves.

I can accept the first two bullet points, although they might be at odds with the "accurate emulation" statement later in the reference. The third is a noble objective but, in my view, requires a very large increase in capability, & characteristics might be difficult to transfer to another design of damper.

In my experience it is not that easy to get two copies of the same damper to match accurately. "Accurate emulation" requires the port flow characteristics to be identical as well as gallery restrictions, leakages, etc.

I would very much like to rig test them on a vehicle.

p.s. I think the "safe" way of producing a damper range is to start with adjustable dampers, make them non-adjustable by removing the adjusting hardware, and make them remotely adjustable by modifying the adjustment mechanism. In all cases the internal flow controls are unchanged.

Both the JRi and the Penske adaptive dampers use a similar architecture and I believe the same Moog DDV. The control flow path is in parallel with the traditional damper piston/shim flow path and thus the controllable range is dependent upon the characteristics of that main valve. An advantage of this configuration is that resultant damper curves (when in open-loop control) can more easily be represented by a traditional damper hardware (piston/shims/bleed). This is important particularly for race series with restrictive damper rules. However, since the DDV does not have full control it will be limited when high bleed/low stiffness shims and pistons are used on the main stack, particularly higher velocity characteristics. Ideally (from a control perspective) the main piston would be solid allowing the DD valve full control over the damping properties, but this obviously creates great difficulty when attempting to build finished product. So a shock engineer must have some experience in relating adaptive builds/settings to damper builds. As DaveW mentioned, the flow characteristics are not identical. This becomes less of an issue with low-hysteresis, well-balanced dampers, but makes things more difficult to emulate otherwise. I don't have any details of their closed-loop control developments (force control rather than valve position) and I'm not aware of any official releases.

Keep in mind that JRi is an acronym for Jeff Ryan Inc who is a former Penske engineer responsible for many of their past innovations.