DIY rear ARB...

[johnl]

The Alfa 147 rear ARB is very weak. With one end disconnected it's very easy to move that end up / down by hand to a substantial degree with little effort. In my experience this is quite common with most FWD cars, it will encourage understeer and dampen down the steering / handling response. Disconnecting it resulted in no perceivable difference in how the car steered or handled (something I've found on other FWD cars as well). Especially as the previous owner had fitted a stiffer GTA front ARB, I wanted a much stiffer rear ARB, but didn't want to spend hundreds of $ on an aftermarket bar.

So, I took the Alfa bar to the Wreckers looking to compare and find another ARB from another car that might be adapted with the least modification. I came up with a front ARB from a Holden Rodeo. It's wide enough between chassis mounts, the ends are more or less the same distance apart, and it's 20mm thick (as opposed to about 14mm I think). However, it does need modification to fit it, and new custom brackets needed to fabricated. The brackets attach to the chassis via the four subframe bolts, and extends backward, which places the lateral part of the ARB a further back in the chassis (compared to the stock ARB location). This useful because the longitudinal arms are a bit longer than the stock Alfa ARB

Now, the bar itself needed to be changed in shape. Laying the bar on the ground with the longitudinal arms pointing vertically, it was necessary to heat and bend the bar at the junctions between the lateral and longitudinal parts so that the longitudinal arms were now horizontal. This is hard to describe in exact detail, and needs to be fiddled with until all is as it needs to be (keeping in mind that there needs to be adequate clearance between the ARB and the suspension arms so that the ARB doesn't contact the suspension as the suspension deflects). The Rodeo links (between ARB ends and strut mounts) just happened to be the correct length and fitted perfectly (and are steel, not plastic, so should be stiff enough with the stiffer ARB).

I was worried that a 20mm ARB might be a bit too much, with excessive harshness, but it's just quite firm, and not a problem. Understeer is significantly lessened, steering and handling response quite improved, and body roll substantially lessened.

Regards,
John.

[Colin Edwards]

Hi John,

You may find heating the bar up with oxy(?) to a cherry or bright red and letting it cool naturally has actually annealed the metal and made it softer!
I'd take the bar to a spring manufacturer and ask they heat treat / re-temper the bar.

Most FWD cars will benefit greatly with increased rear roll stiffness.

Colin

[johnl]

Colin,
the heating will have caused a loss of temper in the steel, assuming the steel was heat treated in the first place...

I have discussed this (on-line) with a some auto engineers (one of whom used to design suspensions for Lotus, and now Ford Australia) and been told that most ARBs while made from a good quality steel are not actually heat treated. There is (I'm told) no need since the deflection in the bar is typically well below the elastic limit of the un-hardened steel. There may be exceptions to this where the suspension is designed / expected to have a lot of travel, like off road applications.

This is not the first time I've done this. My previous car was a CB7 Accord, on which I swapped the rear ARB for a thicker and modified (originally front) ARB from a TR Magna. This bar required substantial reshaping with the blow torch and some brute force, and was on the car for at least 5 years with no issues at all (still on the car and working fine for it's new owner).

As to 'softening the steel, well yes the heat will reduce the hardness of the of any treated steel, but won't materially affect it's stiffness (i.e. hardness and stiffness are not the same thing, and stiffness doesn't rely upon the hardness). With any spring steel (or any steel other than some fairly exotic and rare exceptions) the modulus of elasticity (stiffness) is damn near identical (even for mild steel). The difference between two steels of differing grade and / or differing heat treatment is that the grade and / or treatment will affect the elastic limit, with virtually no affect on stiffness below the elastic limit. So, a steel with a low elastic limit will be just as stiff as a steel with a high elastic limit, so long as deflection doesn't exceed the elastic limit of the steel with the lower elastic limit (which will result in permanent deformation). I know this sounds implausible and goes against what most people seem to assume is the case, yet it is true. 

I totally agree that most FWD cars will benefit from substantially increased rear roll stiffness, or else I wouldn't be going to all this trouble!

Regards,
John.

[Colin Edwards]

Hi John,

In the late 80's I was involved in developing a range of aftermarket anti-roll bars and other suspension components to suit Toyotas.  Back then I was in the Celica and Toyota Owners Car Clubs.  We provided a range of cars to Lovells Springs.  I'm no metallurgist but I have worked with a few and all specified quenching and tempering bars (and springs) after forming / forging. 

A quick look at my old notes confirms the use of 4130 steel and quenching / tempering to 38-42RC.  The bars and springs were then shot peened prior to painting.  Seem to recall the Lovells product was very well respected.

One of the metallurgists I worked with professionally was adamant bars should be quenched and tempered especially rear bars.  He went on at great length about austenitizing and pre and post microstructure.  Given the most popular "bolt on mod" back then was to install a rear bar were none previously existed, the relatively thick aftermarket bar now made a considerable contribution to rear roll stiffness and as such was relatively heavily loaded under hard cornering. 

Going from a 14mm dia bar to a 20mm dia bar is considerable increase in stiffness.  I can't remember the formula but a doubling of diameter results in a 16 x increase in torsional stiffness.  Given your increasing your dia x 25%, the new bar certainly will have an impact!

I guess I'm just a tad concerned when springs and bars are set upon by an oxy torch.  I've seen a few failures and these motivated a number of the members of car clubs years ago to approach Lovells, Selby's, K Mac and Henderson's!

Colin

[johnl]

Colin,
I certainly wouldn't even contemplate heating a coil spring, for all the reasons you suggest. I would also agree that hardening / tempering of any spring (coil, torsion bar, whatever) is indeed 'best practice'. I would expect any metallurgist to specify it. Shot peening also wouldn't be a bad idea.

A manufacturer will spruik every good point their product may have, and may well exaggerate importance of some. I recall K-Mac advertising their ARBs as having forged eyes on their ARB ends, and stating that this was superior because the forged eyes would be stronger than flattened ends with drilled eyes, which is true. However, the added strength simply isn't required because the loadings at the bar ends are not great, and drilled eyes are quite good enough.

With a typical anti-roll bar heat treatment is desirable, but IMO (from advice given to me by qualified people and my practical experience) not necessarily imperative. If a given manufacturer doesn't heat treat an ARB (and as I said I'm told many don't) it will be for reasons of cost, though they would need to be sure the bar was still fit for purpose. I haven't re heat treated the bar (if it ever was) due to the difficulty of finding anyone locally who I would trust to do it (that, and evading added cost...).

A lot will depend on just how far the ARB is actually going to deflect in usage. A thinner ARB can be expected to deflect fairly substantially on occasion, so may possibly require heat treatment, but a substantially thicker ARB is by definition going to be a lot stiffer and thus will deflect substantially less, so may well not approach the elastic limit of the un-heat treated or 'softened' annealed steel. No issues so far, and none when I did more or less the same thing with the Magna bar for the Accord.

Yes, the stiffness difference between a 14mm ARB and a 20mm ARB is HUGE. When I bought the bar the wrecking yard bloke said he thought it was a 17mm bar, and I took this at face value and didn't check. Half way through modifying it, it occurred to me that it did look a little thick for 17mm, so I checked it and it was 20mm. I was then worried that it might be way too stiff, but in fact it's about as stiff as I was hoping for (stiff enough to do what I wanted it to do, not so stiff that it rattles my teeth, even on bumpy roads).

In this application the Rodeo ARB stiffness is somewhat lessened by having more bends in its middle section (than the Alfa bar), resulting in at least somewhat more bar length in the section of bar subject to torsion, which will reduce stiffness to some degree (relative to a shorter torsioned section of the same OD).

Also, the longitudinal arms are a bit longer, giving the suspension somewhat greater leverage over the middle section, somewhat reducing stiffness as well. This was serendipitously useful because for clearance reasons the modified Rodeo bar is mounted a little further rearward (than the Alfa bar), and because the 'lever arms' are longer their ends align nicely with the attachment points on the strut bodies (avoiding any need to orient the links at some angle off the vertical, or cut and shut the bar ends).

Regards,
John.

[Andrew Bose]

Hi Gents,
I assume the reason you wanted a stiffer ARB on the rear was to compensate for the thicker one added to the front by a previous owner. I have a standard 147 twin spark with Koni yellows Ti springs (a tad lower and a tad stiffer) and all I did to the rear to stiffen it up was to change to harder ARB bushes which did reduce understeer and generally sharpen the steering up. The standard front dampers seem far too hard and used to kick back when you need more controlled high speed damping (damper stroke not road speed). The Konis improved the car comfort grip and handling but the rear ARB bushes made it feel balanced like an Alfa should feel and the traction control light hardly ever comes on. I believe the balance of the car makes it smoother and more enjoyable to live with, $25.00 for the bushes and about an hours work had this set up fitted. Remember you car is not new and any wear in the bushes will make it appear/feel too soft but with new bushes it will be a fairly good compromise for the road. I also run "track/road" tyres which are far grippier than a road only sports tyre.

If you need any detail on fitting these type of bushes send me a message and I will call as it is a little hard and long winded to explain in writing.

One thing I was told to remember by a suspension specialist, was the suspension is to keep the tyre on the road especially under bumps/braking/steering, feeling stiff or soft makes no difference as long as you have grip.

Good luck with all the mods and make only one adjustment at a time to see how effective each one is.

Also have a play with tyre pressures it can make a huge difference to the balance of your car.

Andrew

[johnl]

Andrew,
I do feel that increasing front ARB stiffness (front roll stiffness) without also increasing the stiffness of the rear ARB is a mistake. This was done to my car by the previous owner, I think because the front ARB bushes were shagged, and his Alfa mechanic told him the only way to fix the bushes was to replace the whole ARB (sell him a whole bar...), and he might as well get the stiffer GTA bar. I think all it achieved was a reduction in roll motion (subjectively a 'good thing' for any car with 'sporty' pretensions), improve steering / handling responsiveness (also good), but at the expense of increased understeer shortly after turn-in and as cornering forces increase. It would also increase the tendency to lose traction at the inside front wheel when exiting corners under power (one of the main reasons to set up sporty FWD cars with higher rear than front roll stiffness).

With the 20mm rear ARB my cars' handling is now better balanced with reduced understeer and further improved turn-in and response. The amount of steering input and lock required for any given corner is also noticeably reduced. The car still fundamentally understeers, but pretty much all cars do, it's something that is needed even with racing cars. Cars that fundamentally want to oversteer are difficult to drive and very tiring (the novelty soon wears off, been there). What most people would interpret as a neutral handling car is really one that just understeers less than most do. That said, a car that is closer to being truly neutral in it's basic handling characteristic is more likely to cross the line into oversteer more easily and often, and catch the driver out, which is why most car manufacturers deliberately do a quite a few different things all aimed at promoting understeer, most by quite a lot.

The 'new' rear ex Rodeo ARB is still on the old Rodeo rubber bushes (which are in reasonable condition). I may change them for poly items at some stage ($50 for poly Rodeo poly bushes!), it would probably sharpen up the handling response etc just a bit more. The stiffness increase with a standard rear ARB using poly bushes will be minimal compared to fitting a much thicker ARB, even retaining rubber bushes. Your aftermarket springs may have increased rear roll stiffness relative to front roll stiffness if their stiffness was increased at the rear by a greater % than the front spring stiffness was increased. This would be somewhat similar to increasing the rear ARB stiffness more than the front ARB stiffness, i.e. increase rear roll stiffness more than front roll stiffness for a reduction in understeer / roll etc (but tends to make rear ride quality a little harsher than increasing rear ARB stiffness). When I'm shopping for aftermarket springs (for a FWD car) one of the things I'm looking for is that the % stiffness increase for the rear springs is greater than the % increase for the front springs (though this can be hard to ascertain because a lot of spring makers don't supply actual rates).

I'm surprised you find the standard rate front dampers to be too stiff. I find them to be a bit strange. In some circumstances (kinds of road roughness / bump) they seem about right. For undulations and controlling body roll on smoother surfaces they are OK (though I would prefer stiffer), but sharp bumps really upset them, even small sharp bumps. It seems the dampers are reasonably firm for relatively slow shaft velocities, but underdamped for rapid shaft velocities...

I would both agree and disagree with your "suspension specialist". He (I assume) is right insofar as feel doesn't matter (much) if the only target is maximum cornering speed. But this is road car not a race car. I don't really care all that much what sort of high G force numbers it might produce, but I do care that it be entertaining and fun to drive. For this, feel is very important...

He is also right that we need to keep the contact patches in good contact with the ground (with some exceptions to this related to handling balance). Other than how they are affecting chassis feel (unpleasant feeling), this is my problem with the stock front dampers, they are allowing the front contact patches to lose good contact with the road anytime the bumps get a bit too much for them...

Regards,
John.

[Andrew Bose]

Hi John,

I agree the previous owner was the culprit for upsetting the balance of your car in my opinion he should have put both front and rear bars to work more effectively. I was trying to suggest a method to fine tune to your preferences.

The springs on my car are genuine Alfa 147Ti springs which are a tiny bit harder and lower but not much difference at all.

I probably didn't explain myself well, the symptoms of not coping well with potholes you describe is exactly what I meant by feel too hard and need adjustment to their high speed damping. they cope with reasonably well with undulations but potholes are interesting/frightening mid corner depending on speed and/or run off area.

If you are ever in Melbourne PM me and take my one for a drive and see what you think.

Regards
Andrew

[johnl]

Andrew,
I didn't read you carefully enough, I thought you had some aftermarket springs fitted. My car is the 'TI' version, but still sits a long way off the ground (can't imagine it being higher with non TI springs...). I have no complaints about these Alfa springs (other than the ride height is surprisingly high for a 'sporty car), they are stiff enough for me and the roads I drive on.

Even with the stiffer GTA front bar and stock rear bar (made from coat hanger wire I think) my car didn't understeer nearly as much as most other cars I've driven. It still had more body roll than I liked (much improved with the much stiffer Rodeo ARB). I really couldn't pick a significant handling / roll difference with the stock rear ARB removed. That said, it's much better with the much stiffer rear ARB.

Potholes aren't my biggest issue, though there are plenty around here I don't hit them all that often. A car with fairly direct and 'pointy' steering / handling helps to avoid them (that's me weaving all over the road...). The biggest problem is with patched up surfaces. There are plenty of roads in my area where the road is basically patched areas overlapping patched areas overlapping patched areas, with little to none of the original surface still in existence (often no big bumps, just roughness). This is often the case in corners where the surface cops more of a pounding. And this is the bitumen roads, the dirt ones are often worse, but can be much better depending on date of last grading and traffic load...

It's a pity because otherwise a lot of my local roads would be a lot of fun, but as it is you can't reasonably 'attack' most of the corners because the front end of the Alfa is so discombobulated by the less than wonderful surface. I'm sure it's the high speed damping inadequacy of the stock front dampers, and maybe the overall rebound isn't quite stiff enough. It's on these surfaces that the front dampers lose the plot, not so much on bigger bumps but smaller irregularities. My old Accord fitted with aftermarket springs and aging leaky Koni 'Sports' coped much much better (and better again when the Konis were still fresh).

As I said, the rear TRW 'stock rate' dampers are surprisingly good compared to the front ones. I suppose the difference will be that the front dampers are having to deal with significantly higher mass inertia (unsprung weight of all the bits and pieces). Much bigger front brakes, the suspension 'upright', and the the drive shaft must all add a lot more unsprung mass than at the rear end, and the damper just can't cope with controlling it...

Regards,
John.

[johnl]

I've embarked on an experiment. My suspicion is that the rubber 'spring' attaching the damper rod to the chassis may be so soft that the damper cannot control the unsprung and sprung masses for a relatively short rod stroke length each side of the momentary ride height. This might be what is causing my cars poor front suspension control on roughish surfaces (?).

This isn't the case at the well behaved rear end, where the 'hat' that prevents the rear strut from dropping out of position also appears to prevent significant lowering of the damper rod (in the rubber of it's mount) when the suspension droops in rebound. So, I'm making some 'hats' for the front dampers too. I've made only one so far, but can't test drive anything yet because Australia Post seems to have lost my replacement fuel injector...

The 'hat' consists of a thick washer that sits under the rod nut, welded to a short vertical tube that protrudes above the hole in the aluminium casting (that the rod / spring mount is attached to). I've welded a 50mm OD circular disc to the top of this tube, with a central hole to allow the nut to be fitted and tightened. Between the lower face of the disc and the upper surface of the casting I've inserted a 9mm thick hard rubber pad (also 50mm OD with a central hole through which the tube is inserted). When the nut is tightened it preloads the rubber in the stock mount in a manner that stiffens the rubber in bump loading. The 'hat' virtually eliminates movement in rebound.

I'll report on how this goes, but as yet the car is still stationary due to the dead injector. But already, just pushing down on and pulling up on the fender the damper rod only moves a very small amount, way less than without the 'hat'. 'Jiggling' the chassis up and down the suspension is noticeably stiffer on the side to which the 'hat' has been fitted, compared to the side which is still 'unhatted'.

Regards,
John.

[johnl]

The injector finally arrived, so at last I can now drive the car (runs fine, thanks for asking).

The newly made front 'damper rod hats' do seem worthwhile. The front end is noticeably better controlled and feels somewhat firmer over coarse surfaces, the handling is a bit more responsive to steering input,  the front end is noticeably less 'floaty', the handling is generally more confidence inspiring, body roll seems less on corner entry (though much the same by mid corner, not that roll motion is all that much due to the GTA front arb and the 20mm rear arb).

It's all good and I think quite worthwhile to have made and fitted these 'hats', but it isn't 'transformational'. There is still some front 'damper judder' on course surfaces, but it's definitely less than it was. I think this may be because whilst the 'hats' dramatically lessen poorly controlled rebound flexure in the stock rod mounting rubber (i.e. the damper rod attachment), they don't do as much to effectively stiffen the mounting in bump motion. The effect of the 'hats' is that the damper rod is nearly 'solidly' mounted with regards to rebound force / motion, but only somewhat stiffened with bump force / motion.

The rubber mounts' bump stiffness is now more than it was without the 'hats', but only because when the rod nut is tightened it pre-loads the rubber 'in bump' (i.e. pulls up on the mount rubber because there is a 4 or maybe 5mm gap between the bottom of the 'hat' and the crush tube in the stock mount), vertically compressing the rubber in the mount by up to maybe 5mm. So, the mount is a bit stiffer in bump for the first 5mm or so of rod movement, but that's about it. I suspect normal loads can still compress the rubber significantly more than this (don't know for sure since I can't see it happen when driving the car), and if so then this is in effect still bump compression of an undamped elastomer 'spring' (that connects the damper rod to the chassis).

This may still be allowing or at least contributing to the remaining 'judder' in the damper mount (?). I have an idea to eliminate this by fitting some kind of plate above the 'hat'. The plate would be attached to the chassis using two or maybe three of the existing strut mount studs, and would bear against the top of the 'hat' with a hard rubber pad between the top of the 'hat' and the underside of the plate. This would largely eliminate vertical movement of the 'hat' in bump, and so eliminate unwanted and undamped vertical rod motion. At least that's the theory...

Regards,
John.

[johnl]

Just in case anyone is actually interested in this (maybe not judging by responses...), I did make the 'hat straps' for the top of the front suspension towers, and they are good, adding to the improvement created by the 'hats'.

Each of these 'straps' is made from a 55x25x6 steel plate, with a 25x3 'tongue' welded to each end (55mm to cover the diameter of the 'hats'). The 'tongues' are bent downward (next to the welds) and each tongue end bent so they become horizontal (parallel to the tower top) and drilled with 10mm holes so they can be held in place by two of the tower nuts (the two outermost ones). The thicker part of the straps (the 6mm plates) sit on top of the 'hats' I had previously made, with a thin bit of rubber (cut from a bicycle inner tube) sandwiched between. The 'hats' are directly attached to the damper rod (by the rod nut), and more or less eliminate rebound motion (of the damper rod in the soft stock rubber mount). The 'straps' are directly attached to the tower top, and largely eliminate compression of the mount rubber in 'bump'. I'd post some pics if were a digital native...

I was concerned that this all might excessively stiffen the manner in which the damper rod is mounted, specifically the lateral angularity of the rod as the suspension moves. Anyway, disconnecting the 'strut' at the bottom, I could still easily move the strut body in and out (laterally), changing its' angle at the strut mount with no significant resistance. I could see the 'hat' laterally move under the strap by about 1mm as I substantially moved the strut body in and out, there is some 'give' between the top of the 'hat' and the underside of the strap that permits the parts to move relative to each other without causing any partial 'bind', so I'm confident there will be no issues with this.

The damper rod now cannot significantly deflect the rubber bush in either bump or rebound (as it normally does), pretty much the entirety of the dampers' action is immediately 'felt' between the wheel and the chassis, instead of some of the initial damping action being 'lost' due to elastic deformation of the soft elastomer spring that attaches the the damper rod to the chassis (i.e. the damper rod mount at the top of the 'strut'). This 'lost' damping action normally causes the damping action to be at best somewhat non linear, with the action initially being very weak, then abruptly increasing after maybe a cm or so of suspension motion (either in bump or rebound). IMO any non linearity in the damper action is not likely to be a 'good thing', contributing to poor roll motion control and suspension motion control, etc. "Softness' also slows the speed with which weight transfer occurs, which is a big part of why stiffly suspended cars feel 'sportier' than softly suspended cars, and soft damper mounting would be a significant part of this.

The overall result (of both the 'hats' and the 'straps') is that the steering response and 'positivity' has been improved, body roll is initially less (at turn in and early corner), the car is more stable in corners. This improved corner stability is both directional and 'vertical', with the car not as readily 'wobbling' on the diagonal axis. Small irregularities in the road surface can now be more readily felt both through the steering and the chassis, and the suspension is noticeably noisier over rougher surfaces, so there is at least some price to pay with regard to harshness, but it's not a big price (especially as the suspension now seems to 'crash / bang' less heavily on the downside of bigger bumps and into potholes etc., which I assume must be due the wheel not 'free-dropping' in rebound quite so readily...?). There is noticeably less 'floatiness' with larger bumps / undulations in the road (which will I think be also be related to the improved corner stability, i.e. on the 'diagonal axis').

However, the stock front dampers (fairly new TRW units) are still (IMO) somewhat too soft, so not all front 'damper judder' has been eliminated. It's a lot better, but corrugations on dirt roads can still cause 'judder' depending on the size and  'frequency' of the corrugations. Having said that, the judder when it occurs is less, and occurs less often. I'd still like to get stiffer front dampers, not necessarily stiffer with 'slow' inputs (as you get with body roll and undulations etc) but stiffer with 'fast' inputs (as you get with 'sharp' bumps etc). Not sure if B4 Bilsteins (twin tube 'stock replacement') would be what I need, or if I'd have to spend substantially more on the B6 'sports' mono-tubes...

Regards,
John.

[johnl]

Earlier in this thread I used the word 'transformational' with reference to none of the changes I've made to the damper rod mounting quite being so, even if they were generally significant improvements. Having said that, in a certain circumstance I would use transformational to describe the effects, this circumstance being in colder ambient temperature. On cold nights the cars' steering and handling is close to brilliant, with almost none of the 'damper judder' and softness etc. that spoils things whenever the temperature significantly rises above chilly.

When it's cold outside (and so far it seems the colder the better); the cars steering and handling is generally much more responsive, turn-in is really good, body roll is less, directional stability in corners is better, understeer is less. It's just a lot more fun to drive. This temperature related effect did exist before I stiffened the damper rod mounting, but seems more pronounced after the mount stiffening. The car can be driven over coarse surfaces with much more confidence because the front wheels stay in contact with the road instead of bouncing off one bump to the next to the next to the next. The ride quality on rougher surfaces is also far better (though a bit harsher on smoother surfaces). The dampers are in control of the suspension, rather than than the road being in control of the suspension. All this wonderfullness deteriorates as the temperature rises, so summer may not be much fun with these TRW dampers...

This affect of ambient temperature on steering and handling is something I have often noticed in some degree with many other cars, but with this 147 it is very noticeable. My assumption is that the damper fluid must be significantly thicker when cooler, enough so that the damping becomes appreciably 'stiffer' (and conversely, thins out when warmer, with a resulting diminution of damping effect). Maybe this could be a reflection of the quality of the (cheap) TRW dampers? Perhaps the fluid used isn't all that great, affected by temperature more so than the fluid used in more expensive dampers? Maybe the fluid is thicker (thicker hydraulic fluids do change viscosity more with temperature change than do thinner fluids), perhaps to account for less precise valving in the cheaper damper?

Regards,
John.

[Cool Jesus]

I feel like I deserve a diploma after reading this ! 
Awesome info, thumbs up guys. 

[Colin Edwards]

Hi John,
Does the cold temperature handling / ride /damping deteriorate during the drive?  Given the damper fluid will heat up pretty quickly,  if the improved control is due to higher viscosity the improvements should dissappear just as quickly.

Maybe its time to replace iffy TRW's with something more sophisticated and with less non linear stiction.

Colin