@twenty8 Here's a tip for you. I took the picture link in your reply and went to Bing. You can go to Google too. Click on the Image Search part, then paste in the picture link. It will do a lookup of the same and similar pictures. This is commonly called a reverse image search when discussing checking for scams. Anyway, I got this result. https://www.burnsco.co.nz/ac***ulator-tank-2lt-pr-wave
Thanks @RodStRace , I know how to image search but wanted the direct link. I shouldn't be so lazy....
https://www.globalwatersolutions.com/products/pressure-tanks/ Also possibly useful: https://www.globalwatersolutions.com/products/pressure-tanks/shock-arresting/ Stainless versions are unfortunately only in the larger sizes: 8l/2 gallon and up. The smallest is 8" in diameter and 12" tall, and should be sufficient for pretty much any vehicle likely to be considered on any part of the HAMB. There are also a few much smaller stainless options in the shock-arresting range. Their models intended for indirect solar water heating, right down to 2l/0.5 gallon, are approved for ≤50% propylene glycol: perfect for the purpose at hand.
In daydreaming around the diy Hydragas I fear I've hitherto rather hand-waved the matter of damping. I've specified no more than "a damping valve block", though given that the bag, valve block, and ac***ulator are separate components that is perhaps not unreasonable. The issue is that the fluid displacement per any given distance of travel is going to be substantially greater than with conventional telescopic hydraulic dampers. A monotube damper with a 46mm piston will displace 42cc of shock oil for every inch of travel. That is towards the upper end of the range of typical damper displacements: only some severe off-road dampers will displace more. Getting similar figures for air springs isn't straightforward. At best an effective instant piston area can be deduced from the fluid pressure at a given load at ride height. Thus, to take a fairly realistic example, a bag which carries 800lbs (3556N) at 100psi will displace 131cc of aquaeous glycol in the inch of travel centred around ride height. That is over three times as much as the aforementioned damper. While this means that available rebuild pistons and shim stacks would be inadequate, unless we can somehow incorporate three sets of each, and still be unsure how they would perform over time with aquaeous glycol rather than mineral oil, there are a few advantages to the situation. Large fluid displacements mean large valve hardware, which is likely to be very durable and allow more leeway as regards precision. Moreover, as the medium is under pressure as a necessary part of the suspension's operation, there is no need for external pressurisation to suppress cavitation. I'm thinking needle valves, two fixed and two spring-loaded, might be easier to arrange in a way which allows for external adjustment than shim stacks, but I'd have to sketch something up and see. Edit: Working back, the bag represents an equivalent piston diameter of about 82mm, or 3¼". Manufacturers of dampers for off-road racing trucks, like King, Fox, etc., make dampers in diameters larger than that, and moreover sell parts like pistons and shim stacks separately. They are not cheap but not prohibitive either. The simplest arrangement would probably be a chamber holding, say, a 3" damper piston on a short stem, fitting snugly into the chamber's bore. Because the piston doesn't move relative to the chamber, its wear band serves solely as a seal and does not wear at all. Here is a 3" King piston: This also means the entire damping block is sprung m***, which is another advantage. I get the impression that the hitherto dark art of shim stacking has been blown wide open over the past decade or two. The rise of suspension on bicycles probably played more than a minor part in that. The suitability of an aquaeous medium remains a question, but the corrosion-inhibiting charactersitics of the glycol is cause to hypothesise that it'll be OK. Distilled water might be a good idea if scale is a concern.
It's always possible to simplify things. It's part of the design process. While enjoying my habitual Wednesday afternoon coffee in the centre of Cape Town, I noticed a work crew from the City installing steel bollards along the pedestrian mall where I was sitting. The top of each bollard comprised a hemispherical steel dome, which I measured roughly to be 90mm (about 3½") in diameter. There were small holes in various places to let molten zinc drain during the hot-dip galvanizing process, and I was able to stick the nib of a pen in one, and then measure, to get an idea of the dome's wall thickness. It looked like 3.5mm. So, it seemed likely that 90Ø x 3.5 steel hollow domes were available from local steel merchants. I Googled it when I got home, because I hate doing that kind of thing on my phone. And indeed. All this because I still had the damping valve block on my mind:
That immediately had me wondering how to plumb 1" lines to the airbags. Most bags used in automotive aftermarket applications aren't rebuildable, and have ⅛" or ¼" air fittings, which are woefully inadequate for the purpose at hand. Cutting a hole in the end plate of an ***embled bag probably isn't a good idea. How close could you get to getting all the swarf out? I couldn't find an alternative way to recrimp crimped bags. That seems to be a non-starter. After a lot of going around in circles it became clear that the only way to do this is using rebuildable bags with bolted bead rings. It illustrated differences in website culture. I came to dread Chinese websites, because the technical information they contain is typically a handful of random specifications plus lots of shipping information and bulk order prices. At the other extreme are Australian websites. I love how forthcoming they are with hard technical nitty-gritty! Boss Air Suspension manufacture a very small range of rebuildable airbags, no more than a dozen basic airbag products, but they even give dimensioned drawings of the stock end caps on their site, so you have a basis on which to design an alternative configuration. And, Boss seem to have a solid reputation, as far as I could see. But why stop there? If we're fabricating end plates, why not have the damping valving communicate directly with the bag? A little bit of wrestling around the order in which to weld things together produced this: Bleeding this lot is a consideration. Air traps are to be avoided. To that end the upper ***embly has resin cast inside it, at an angle to lead air up to the connection to the ac***ulator. The pipe elbow there has a bung welded to it to receive a bleed valve. I took the interconnection fitting from the lower end cap, in the interest of keeping the total as short as possible. The plumbing needs some refining. I'm sure it can be streamlined a lot, e.g. by going all-AN except for the 1" female NPT fitting to the ac***ulator.
Have you ballparked operating/max pressures? The resin sounds like an issue waiting to happen. High pressures and temp fluctuations will work on the bond. Ask anyone who has chilled a bearing or gear to fit on a shaft. You could include a capture ring in the top, then cast the resin above it, but still seems like a patch. I'd just have the reservoir with the main fitting on top, and the pipe/hose be the high point with either a bleeder or fill at the top. No reason to have a gl*** or bottle of liquid sideways, unless you are wetting the cork! If this is going to be like the Mclaren with bladders in the reservoirs, it makes sense to have the air over the fluid. If/when the bladder fails with air below it would immediately allow air under to get into the rest of the system. Lower center of gravity, too.
My calculations have been as per airbag suspension, i.e. ±80-120psi. The pressure ratings for the bags are what they are regardless of the media inside them. Keeping air out of the liquid is perhaps not that critical in this instance, as the system is all about pneumatic elasticity anyway. The odd bubble or two probably wouldn't make that much of a difference. Perhaps I'm overthinking it. The only place we'd seriously not want air is in the interconnection, which I've got at the bottom of the system. Coming from my background in architecture I saw the resin being cast around top-hat-shaped lugs spot-welded to the underside of the top plate of the damping valve chamber, the way steel components cast into concrete might be. Sometimes the crudest solutions are the most elegant: if a roof has a small leak upon completion, the best solution might be an empty tin on the ceiling directly under the leak. Any water that ends up there will evaporate long before it gets close to filling the tin. The resin is probably easier than trying to shape the steel enclosure to taper to the outlet, but simply living with a bit of air trapped in a top corner might be easier still. The ac***ulator here is a diaphragm type rather than a bladder type. Bladder ac***ulators need to be installed vertically lest liquid get trapped under the bladder. Diaphragm ac***ulators are fine horizontally. Horizontal is probably a worst-case scenario, but there might be packaging constraints which require it. I wonder what would precipitate diaphragm/bladder failures? The system functions on static liquid and air pressures being equal, so the diaphragm or bladder isn't holding pressure. And the inside of the ac***ulator never sees UV. Plasticizer loss into either medium?
There was a time I'd just sit down at the drawing board and sketch this out on paper. I tried it now and found myself sitting there like someone who has never learned to draw. 3D modelling has actually become easier for me than drawing. I was trying to figure out what the absolute simplest interconnection would be. Quarter-elliptics allow for a very simple arrangement, especially if they are used to locate the axle. I can see that becoming a braking issue, though. The thing is to have whatever we use for damping rock with the quarter-elliptics, so the damping doesn't fight the interconnection. I've included options for friction dampers, lever-arm hydraulic dampers, and telescopic dampers here:
A couple observations and questions: Since the quarter elliptic springs are solid at each end, they locate the axles within their arc. I would ***ume they are fairly light, since you want the fluid to ***ume some of the 'push' of the suspension, not just the damping. But if the spring mount pivots and the hyd is acting at the bottom, the fluid is acting through the spring so they must be the stiffest. Balancing act here. The spring frame mounts pivot and are held at ride height by the fluid acting on the mounts, right? Or is this a simple straight rod linkage front to back in place of the hyd pipe in the figure as proof of concept? I get that this is the simplest fluid acting on the mount, but would want to include a raising rate linkage that could be adjusted for tuning on the real thing. This could probably also keep the main hyd line from being the scrub point. Braking and drive axle twist can be controlled by a triangulated linkage, so long as it doesn't bind in expected range of motion. Careful positioning could even use bushings to impart anti-sway and/or additional resistance at furthest reaches of normal travel. I wonder how much damping is going to be necessary with the hyd acting as part of the suspension support? Valving would be critical, between small bumps and a 'roller coaster' hill/dip combo. Renault and Mclaren use the hyd system for the damping, without a separate shock. Having the friction, lever or tube shock would allow tuning of the prototype, but once numbers are gathered it should prove redundant.
Yeah, I rewrote this a few times as it dawned on me what you were doing. Sorry if it seems like fractured thought. The shock on the side of the frame is like a steering damper on the linkage. At least in this simplified mechanical only system it strikes me that it's overly complex and added weight for what amounts to a bit of a different spring rate at each wheel while also imparting forces to a wheel that is not experiencing any change in surface. It puts me back into thinking of the different speeds causing disruption at frequencies. Great at anticipating a change at the rear at a very specific speed, not great at others and bad at another exact speed. Yes, hyd can be valved for this. I'm still interested, but don't see the leap forward by adding it all. Front tire drops into a pothole. Front spring pushes down, pivot swings, linkage pushes back, rear spring swings up some. Front and rear of car drops some instead of just the front. I guess this helps the ch***is stay more level longitudinally rather than seesawing and twisting, but how much are you gaining?
I might be off target here, but I think the front-to-rear connecting rods will have far too much compliance to do the job. The bending action of the rods may negate the transfer of the movement of the suspension from one end to the other.
I believe it's a demo drawing to show the system and how it interconnects, not a concept. He's wanting to do it with hydraulics, not a mechanical link bar. It's the dumbed down version so non-engineers can grasp it.
https://journals.sagepub.com/doi/10.1177/09544070241245473?icid=int.sj-abstract.citing-articles.2 Interesting read, looks a little complex but achievable. Dan
Are you guys aware of the FRICS system (front/rear interconnected suspension) banned in Formula 1 in 2014? Here is some good information. https://www.racecar-engineering.com/tech-explained/f1-2014-explained-what-is-frics/
Interestingly, I had some contact some years ago with local guys who were developing something very similar, though they were coming from an off-road background. Their website is down and their Facebook page has been inactive for about a decade. I was trying to convince them to go for the US performance aftermarket, but they were looking at selling their patent to semi-trailer manufacturers, where I suspect the idea went to die. Before then I did some promotional 3D modelling for them: Their system relied on sliding hydraulic seals which they hadn't managed to get reliable, which is why I've been leery of sliding seals in my subsequent thinking. I got the impression that the mad scientist guy was too ready to hand-wave the sealing issue, while the moneybags guy was very well aware of the problem.
There might be situations where that might be an issue, e.g. where one or more tyres is in the air. Otherwise the rods would be in greater or lesser tension rather than actual compression? It mainly damps the transverse-leaf pitch spring, but it'll add a bit of warp damping too. The thing most people miss is the role of the car's structure in all this. I'm seeing the interconnection as making the complexity required to get the car's structure rigid in torsion superfluous, which is why I believe that it'd be particularly useful in a hot rod application. It isn't just adding interconnection; it's interconnection instead of a roll cage or a space frame or an engineered unibody etc. Like a lot of small hatchbacks of its era, my DD was engineered for a distinct rearward bias in roll stiffness to counteract the understeer imparted by the front-wheel drive and the forward weight bias. Mine sits a bit lower than stock and is a bit more stiffly sprung, which amplifies all the effects of that setup. So, entering a right-hand turn, the heavy engine and transaxle want to lean left, and the relatively roll-soft front suspension lets it. The relatively roll-stiff rear suspension tries to counteract that, but it can only do so by turning the entire unibody into a big torsion bar. Chucking the car into that right-hander, the unibody creaks and groans, and the front end's roll moment gets imparted onto the left rear wheel, less the cause of all the creaking and groaning i.e. the torsion-bar action. The rear suspension has to be more roll-stiff than it would have needed to be if not for the torsional elasticity of the unibody. That right-hander is a specific curve I take fairly often, and it has a pronounced bank to it. As the front wheels strike the increasing banking the left front wheel is momentarily more heavily loaded, so the ch***is goes a bit more understeery for a moment, until the rear wheels catch up to the banking. The opposite happens at the end of the curve: there is a momentary oversteery twitch. You get a feel for this, and late-apexing becomes second nature, it's still a lot of fun and you don't lose it coming out of the curve (as I had more than once in my wayward youth) but the car is more nervous than it needs to be. And the price of all that fun is my wife cursing the suspension every time we need to transport anything fragile. The advantage of interconnection here would be that the car's structure would no longer be turned into a big torsion bar. Instead of roll moments being transferred between one end of the car and the other by structural torsion, they get distributed according to the leverages in the interconnection. They self-balance according to whatever bias is designed in, regardless of any deformation of the car's structure or the contours of the road. With interconnection, my DD's creaky-groany, barely adequate unibody would be grossly overdesigned. With interconnection, a Citroën 2CV's lightweight ladder frame is entirely equal to the task — as would be a T- or A-based frame. Your only remaining worry would be driveline torque, for which an X-member or K-member is adequate.
@twenty8 good overview of it! @Sharpone haven't read it yet, but will. @Ned Ludd Okay, let's say you can shave 10 pounds off the frame due to less need for rigidity, and add 10 pounds of suspension bits. I'd imagine that there is a point of adding too much lightness to the frame, since it still supports all drive train, humans, cargo and the rest, along with resisting the suspension loads. We won't even throw in crash resistance. With modern computational ****ysis, structures can be light and strong, especially with composites and modern manufacturing. https://www.metal-am.com/articles/m...redefining-next-generation-car-manufacturing/ I'll mention that a T frame was designed to flex as part of it's ability to handle uneven terrain, just like the 2CV was designed for this. Useful for unimproved terrain, but of diminishing returns for road going, unless the infrastructure crumbles even more! I'm not opposed to the idea, and having F1 usage along with the other production cars shows it's viable. I hope my questioning comes across as curious but not sold, as opposed to just hating it.
Excellent stuff guys, a very good mental exercise for me, the more I learn the less I know. Thanks Dan
Skimmed through the article (MATHS) and found it interesting. Tidbits like the sizing differences F and R, mean that optimizing operation would require more unique parts, not a single subset of parts at each corner. I'd guess differences in weight, center of gravity and of course steering and drive functions would change this too. The curves shown do support better dynamics, just like how anyone who has ridden in a Citroen describes. As a mental exercise, it is entertaining, but then so is 4 wheel steering. I think it could be done, but then I think about Legit Street Cars and his efforts in getting the MB hyd suspension sorted. Maybe it's just a typical American's desire for common + simple + cheap, but I am not a fan of over-engineered stuff that relies on everything working properly. See tesla's electric inside door handles in the current news.
I have no doubt that a futuristic Gigeresque lattice could achieve 15000+lb.ft/° within a 3" structural depth, but that's not the context I'm thinking in. In fact I'm not thinking in any kind of volume manufacturing context, but in an alternative to it. None of these ruminations are about revolutionizing the motor industry, but rather about working without it, supposing a hypothetical opportunity to do so. Hence looking at past applications of interconnection with a view to hot rod/homemade applications. What I'd like to see (and I live in hope) is hot rodders getting their teeth into this stuff and working it into a kind of hip-bone-connected-to-the-thigh-bone type of tech, which can be applied to a frame set out with lines on the garage floor, which is moreover adaptable to a wide range of body configurations without having to redesign everything every time. More than a small part of me is all for designing clever suspension instead of roads! It's about proportion. We know that early Fords were useful vehicles, having ample structural beam strength to carry a range of different bodies, including ones meant to carry payloads, while having so little frame torsional rigidity that it can be regarded as zero for any practical purposes. That can certainly be improved. Simply boxing the rails improves torsional rigidity by orders of magnitude, but that is still far from sufficient for the present purposes. I'm still more than an order of magnitude short. The design of the cl***ic early Ford suspension is probably over half a century older than the road-racing rule-of-thumb which states that adding structural torsional rigidity ceases to make sense at ten times the overall roll stiffness of the car's suspension. Yet early Ford suspension satisfies this rule in an unexpected way. We've seen that the frame's torsional rigidity was negligible. One tenth of that is for all intents and purposes zero. Early Ford suspension has zero roll stiffness at rest, because the car hangs by the spring shackles like a sling or hammock. The system self-corrects in roll because the shackle instant centre shifts in the direction of the lean, causing the car to rock back to vertical. Roll stiffness increases gradually with roll over the first 3½° or so, because the loading on the spring becomes slightly asymmetric, whereafter roll proceeds conventionally but on the basis of that diminished spring loading. That, rather than any supposed designed-in frame flexibility, is why early Fords could accommodate up to about 7½" of surface non-planarity in everyday use without doors sticking shut or flying open. Between the suspension design, customary hot-riveted construction, and Ford's predilections for saving material costs and having machinery do more and workers less, there was no need to design flexibility in specifically. The rails were open C-channels because Henry would rather have them come out of the presses pretty much complete than rely on a worker who could get uppity to weld them together. This becomes inadequate as soon as we use something other than the shackles alone to locate the axles laterally, which is probable, which is why it is good that boxing rails is a thing. And after that is where hot rod handling diverges: either dynamic expectations get lowered or a lot of steel gets welded to the top of the rails. I want to avert both. The aforementioned rule-of-thumb flies out the window when it comes to interconnection. The 2CV certainly doesn't safisfy it. Its frame rails are very narrow open C-channels tied together with a thin-wall stressed-skin integral belly pan. That makes sense as the main suspension loads are translated to fore-aft forces through the volute springs, and the belly pan resists lozenging of the frame with less material than an X-brace etc. It handles rough terrain by having some 10" of suspension travel. Nothing about the frame suggests anything intended to accommodate or resist torsion, because it doesn't have to. We're OK. You're usually good for intelligent comment.
Thanks for that. I like a good joke or reference, and called out BS, but am trying to keep it to share and learn here. That's interesting about the Ford and rigidity. Had not thought of it like that, but you are right. I understand this is a stream of thought and you are hoping to bounce this off others. It also makes sense to hope that this could be thought out to be a DIY setup carried out by rodders. They are a crafty bunch who want to expand the parameters and be different, while still fitting in a narrow subset of automotive hobby. This also applies to other car hobbyists. I'll start this with focusing it down some. In this forum, that's 4 wheels (let's not do trikes!). It's also a perimeter frame, so it supports the body/drivetrain and has suspension locating points at all 4 corners. A backbone like vintage Lotus is not really what HAMB is about. A case could be made that these rodders want big power, so drivetrain torque must be considered in the rigidity of the frame. Boxing C channel is a good idea, but so long as the corners are strong enough to support the suspension loads, this is not a major concern. It is usually either drive it until it breaks and weld more bracing, or overbuilt. Your mechanical drawing and most stuff here is based on solid axles front and rear. Let's not muddy it up with leading or trailing arms, A arms, or other more complex wheel locating apparatus. I'd also say that rods in general tend to have pretty limited travel, mainly for the low stance, which is opposite of the long travel off-road stuff you mention. If it works for 4-8 inches of travel, it can be 'stepped up' for longer travel by mods to the axle locating components and sizing up the hydraulic system. Go for the small basic design first. I'd also venture that while there are a number of very competent, fully equipped people here, the normal way for rodding is to use something that is already made and adapt it. Home built can include a full machine shop and fabrication from bulk material, but it's more likely and typical to take ready made components from other fields and adapt them in new ways. Once the concept is made and proven, it grows to the point where small manufacturers will offer kits. Ron Aguirre's vette hyd suspension experiment is now the lowrider suspension industry. That goes from budget street stuff to jumping compe***ion wild. https://kustomrama.com/wiki/Ron_Aguirre's_1956_Chevrolet_Corvette_-_The_X-Sonic A typical car weight is 2000-4000 pounds. That's 500-1000 per corner static, with say a 10% variable for weight bias. Here the low CG and limited travel hot rod parameters also help simplify, since the corner weights aren't going to have as great a change as a 4X4 with long travel. I'm picturing monster trucks for example.
Ways to Skin a Cat Department: if anything has come out of this thread, it's that there are potentially hundreds of ways to do this: coils, torsion bars, leaf springs, mechanical, hydraulic, pneumatic, etc. etc. etc. Here is another, a bit of a development of the idea I had over 30 years ago, and where this whole obsession began:
A lot of this thread has been in the form of the history of my own creative process. I fully appreciate that a probable majority of you guys don't give a **** about my creative process, and arguably rightly so. Indulge me nevertheless. Having eureka moments doesn't make me a genius; it just makes me a human being. Everyone has eureka moments under the right conditions. A whole managerial culture has sprung up in Silicon Valley in an attempt to create those right conditions by giving coders ping-pong tables and latte. And the managerial types fully understand that eureka moments work at a fairly low success rate. For every eureka moment which leads to a workable solution, ten or more are duds. It's obvious then that we'd want more eureka moments, and not fewer. So they set up brain-storming workshops, and tell their ideas people to come up with ten eureka moments, under strict instruction not to develop them further than bare brainwaves. But my dear little managerial tinpots, eureka moments don't work that way! Getting to the eureka moment requires hyperfocus, and that means chasing one idea to the exclusion of the rest of the world. And it means thinking concept and application at the same time, even if the application isn't completely sorted in the process. In reality, what comes out of these workshops is dozens of ******** notions just to make up the manager's quota, and then the ideas people go away and hyperfocus from scratch. This one is the result of taking a day, while my wife and I were on holiday outside Herm**** eleven years ago, staring into the middle distance. The key is the yellow things, torsionally stiff large-diameter hollow shafts able to rotate about their axes, linking the bell crank pivots: I subsequently devised a number of configurations, but the thing which kept eluding me was why I couldn't do the obvious thing and have the yellow things act on the pitch springs without losing pitch control entirely. Last Sunday it came to me, in church of all places. I don't know if it was the general atmosphere of serenity or Fr. Mcebisi's homily threatening to turn into a full-blown sermon, as it does, but all of a sudden I saw it: The missing link was in this case literally a link: that lateral link at the middle of the wheelbase. It constrains the roll-control interconnection to allow only free warp, and not pitch, so the pitch springs can deal with the pitch. And it creates the situation where the pitch springs and the roll springs act in series, like on the 2CV but not the Hydrolastic/Hydragas or the Packard, which has advantages for determination of spring rates and spring travel. But best of all, I've got the pitch springs carrying all the weight of the car, and the roll springs none. This is the opposite of all the extant production systems. It allows carte blanche as to spring rates etc. Now to work it into something practical.
Thinking more about the diy Hydrolastic approach: sorry, I can't help it! The 2-litre (½-gallon) ac***ulators were proving hard to package in small-car applications, which was what got me dreaming back in the direction of mechanical interconnection for a while. But the chances of that being even harder to package in any given application are great. Mechanical is probably the simplest, but the system would have to inform the entire ch***is design, as for instance in Post 41. Accommodating any pre-existing architecture will complicate things: which is potentially not quite the case with the water-glycol-based approach. It'd be nice to be able to use the little 0.3-litre (0.08-gallon) shock-arresting ac***ulators, but they lack the capacity to accommodate full-travel fluid displacement. They'd be perfectly adequate if they were the only things dealing with roll motions, like the small lower springs shown in Post 58. The smaller ac***ulator can easily be integrated into the displacer ***embly. Here I've gone to a machined aluminium upper case, in which the damping piston is held by means of a large circlip. I wanted to avoid the language of billet, but it makes accommodating the fluid fitting on the side a lot easier. A coat of paint will hide a mul***ude of sins. So, the idea is to get the displacer ***embly to deal pretty much only with roll by introducing something which does the same job as the yellow tubular doohickeys in Post 58. The principle is this: the main road springs are very stiff, stiff enough to give whatever roll stiffness we want, but there are far softer springs acting in series with them as long as the combined spring motion on the left matches the combined spring motion on the right. Insofar as left-side inputs match right-side inputs, the inputs are p***ed on to the soft springs. Insofar as left-side inputs oppose right-side inputs, no inputs are p***ed on, and only the stiff springs are in play. Here's a way to do that with aquaeous glycol in bags: Two axial bags are connected by rods to match each other's travel. The overall motion is resisted by any of a number of means — I've shown a rising-rate combination of coil springs, a primary rising-rate coil with a short secondary coil co-axially inside it, held in place by an anti-rattle spring. There are legion other ways to do this, including an airbag, or even a coil spring inside an airbag. The bags are damped by the same kind of pistons and damping valves as the displacer units', but at a far softer rate. The ***embly measures about 22½" x 5½" x 5½", but it can be made taller/wider but shorter, depending on the packaging requirements of the application. Here, then, is the entire system: Not shown is any means of resisting pitch, because I haven't been clever enough this time round to figure out a simple enough hydraulic equivalent to the purple bell cranks and cross-link in Post 58. But some kind of arrangement comprising a steel spring pushing the axle in question up counterbalanced by an airbag pushing the axle down would provide a way to do self-levelling. In fact the real-world situation I had in the back of my mind with this one is a '90s Japanese 1-tonne pickup truck (what Americans inexplicably refer to as a "mini truck") which has torsion bar front suspension. Subs***uting very slender bars, which are unstressed at ride orientation/height, plus a single airbag or interconnected pair of airbags at the rear, would achieve the same end.
The Jalopy Journal
