On / off power steering?

[JohnLear]

I've recently removed the power steer belt on my daily driver Honda Accord, and very much like the change in steering feel this creates, yet I do miss the PS in parking lots (and my wife would simply not be able to park the car if she had to drive it, and might have some difficulty steering it in tighter corners). It occurred to me that what I want is a PS system that could be easily switched on / off at driver discretion.

Has anyone ever tried, or heard of anyone, fitting an AC style clutch to the PS pump pulley? Such a clutch might be actuated by the driver simply pressing a button (on steering wheel maybe), so assist only exists when it's really needed, and not when it isn't (which is most of the time in my case).

Of course PS robs power, and removing the PS belt has resulted in a noticable 'butt dyno' measured improvement to the cars performance in both full and light throttle driving (more than placebo I think), and better engine response when 'blipping' for downshifts.

While I haven't yet carefully measured the economy, it does appear that the fuel guage is dropping significantly more slowly during my routine driving, i.e. the guage is dropping to / around X by Thursaday rather than by Wedensday (hardly an exact measurement I know...).

It would seem a win / win / win to me if the PS could be switched on / off at will, but are there any reasons why:
a) It seems to never (or rarely) to have been done?
b) Why it would be a bad idea?

I would think that such a system might appeal to quite a few people, especially more 'sports' oriented drivers. I know that quite a few people do disable their PS...

Regards,
John

 

[GregLocock]

I've never seen that done, but it sounds like a good idea.

I've driven some astonishingly bad speed sensitive PAS systems in the past that more or less replicate that effect, they were most disconcerting to drive at around the switchpoint speed.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[JohnLear]

Damn Greg,
If you say it's good idea then it most likely is. I hope I haven't just carelessly tossed a potential money making concept into the public domain...

It does seem such a ridiculously simple solution to the problem, I'm sure someone must have tried it somewhere sometime.

Regards,
John.

 

[GregLocock]

I don't think it would be acceptable for OEMs, but if you think it is worth a try you can red flag the thread and hope that no-one has cached it. I have seen roughly similar proposals, but nothing as straightforward.



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[NormPeterson]

It might be worth considering some sort of valve (solenoid-operated along with the pump clutch?) to change the fluid circuit such that in the unassisted case the human is not trying to pump fluid back through the pump and turn it.


Norm

 

[cibachrome]

This routinely done in the name of EVO, droop flow, MSVA, etc. to name a few. power systems. The belt spins the pump but no fluid is moving. The main problem is all the law suits that have been filed and already settled because of the safety issue it creates. You now have an unpressurized system and fluid blockage ("pump catch") will inhibit your ability to make a sudden lane change ("Moose/Elk Test"). The extra compliance now in the system because there is no longer any boost flow, is a major lane keeping problem. This on-center mainly understeering phenom: drives most people nuts after a few miles.

Its not that much power loss to just spin the pulley, but is a few hp to move the fluid. With "open" hydraulic systems, the oil is always moving so that when you need the power assist in emergency conditions its ready when the valve diverts the flow.

The real answer to you question is that the trial lawyers had a field day with this one. Even if the switch was automatic, someone has already claimed that it didn't work properly. Been there, done that, time to move on. Pump up your tires if you want to save fuel and empty out the boot, trunk and fill the car with paying passengers...

 

[Tmoose]

I have a magnetic fan clutch pulley off a Peugeot 404. On and off via a single carbon brush contact running on a slip ring. hmmmm.

 

[swall]

Wouldn't electric power steering be able to do this?

 

[GregLocock]

Sure. But that has its own drawbacks, not least of which is the very high referred inertia and friction from the electric motor.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[cibachrome]

EPS does do this. However comma, the bandwidth of such systems is lower than hydraulic assist so you give up some feel and capability. 42v or 48v systems have more authority but those electricals are not popular, especially by EMS workers. Yes there are multiple forms of EPS (column mounted, rack mounted, dual pinion, and EHPS with an electric pump on demand, but each has limitations: cost, serviceability, reliability, bandwidth and packaging. Even a return to manual steering has some appeal to me with high tire pressures, high steer ratios, low understeer, manual transmission (keep the sled moving, boys & girls), negative caster... You know, my golf cart is a good example of the KISS steering performance criteria. You should note that even ATVs are now appearing with EPS for reasons not just for steering work.

 

[JohnLear]

Greg,
It wasn't really a serious lament, even if it were a new idea I'd be surprised if it would be worth patenting(?), or even patentable, or rather, that a patent would be defensible(?). What is the idea really, taking a clutch commonly used to control a pump... and adapting it to control a pump...

I could be wrong, but in any case a patent is only a good as the resources one has to defend it, meaning any patent of mine wouldn't be a strong one...

Norm said; "It might be worth considering some sort of valve (solenoid-operated along with the pump clutch?) to change the fluid circuit such that in the unassisted case the human is not trying to pump fluid back through the pump and turn it."

cibachrome said; "you now have an unpressurized system and fluid blockage ("pump catch") will inhibit your ability to make a sudden lane change ("Moose/Elk Test").

In this case, if I jack the front wheels off the ground and then as fast as I can turn the steering wheel one way then the other, I can feel some almost imperceptible reverse pumping resistance, certainly not enough to affect emergency swerve manouvres. I can't see why the affect might be different with the wheels on the ground.

At least with my Accord, the steering action seems to suffer less from resistances related to hydraulic pressures with the PS disabled than with it active, i.e. with assist active it's not that difficult to 'beat' the assist when turning one way then suddenly turning the other way, this causing an added resistance that seems purely related to forcing the steering back the other way against the pressure still trying to turn the steering in the original direction.

Of greater affect on emergency swerves would be the simple increase in steering 'weight' caused by having zero assist, which isn't really a substantial increase in more open corners or lane changing. It is a substantial increase at walking speeds and in some faster corners where the steering wheel needs to be turned past 90°.

In this case I'm sure this is related to forces acting around the steering axes (most particularly trail, which on my car is significantly greater than stock) not to 'reverse pumping'. The steering does become fairly heavy to turn past 90° in faster corners, but returns to centre easily (forcefully) and quickly.

I know that many people who disable their PS 'loop' the hydraulic lines at the rack housing to avoid the possibility of what you're suggesting, but at least with this car it seems not to be necessary.

cibachrome; "The extra compliance now in the system because there is no longer any boost flow, is a major lane keeping problem. This on-center mainly understeering phenom: drives most people nuts after a few miles."

What compliance is this? Are you referring to the lash associated with opening / closing the rotary valve ports in the spool assembly? There is some very slight lash in my steering that may be caused by this, but I haven't noticed any increased tendency to tramline etc with the PS belt removed.

What is more noticable is the 'feedback' at the wheel, i.e. I can feel undulations and bumps at the steering wheel more easily, the steering feeling more 'alive' than with PS active. This might be disconcerting for many drivers, but I think a lot of others will like it.

Any understeering tendency seems caused simply by increasing steering weight in some corners, i.e. not really understeer per se but lack of muscle, which would be an issue for some drivers but not others. This is the nice thing about having tha abilty to switch the PS on / off, i.e. if the steering weight is onerous for driver X then they can turn the PS on, but driver Y can switch it off (and on in the car park).

I have to say that I've yet to drive a PS system that I actually liked on the open road, and have often thought 'this would be such a nice car if only the steering were more communicative'. I'm sure I'm not the only one...

 

[cibachrome]

In God we tust, all other bring data. For your "emergency" swerving maneuver, what was the steer input velocity? 100, 500, 1000, 1500 degrees per second? Your opinion is fine but not good enough when Grandma's lawyer wants $20,000,000 from you for her nursing room care (she only gets 1/2, remember?

The compliance of the toque sensor in the valve is a soft spot in the gear which is an understeer effect in general, mucho moreso when the boost feedback is not supplied. This is an easily and commonly measured nonlinear quantity. The effect, even when full boost is supplied, rolls out at .1 to .2 lateral g in most cars. Yes even Honda's. Yes, even in manual steering cars because the I-shaft has a nonlinear compliance element similar in magnitudes to the torque sensor's. This is not lash, it has a finite stiffness, measureable and quantifiable.

I suppose its necessary to inform you that not all people who drive motor vehicle have the necessary levels of perception capable of choosing the often subtle changes in vehicle motion variables AND their derivatives (per angle and per g). That is why we make measurements.

As I recall, early Hondas did have a clutched steering pump which kicked on at about 20 mph or so. This system was dropped as many people spun off the expressway/motorway entrances when accelerating up to speed when the clutch kick off and the car undesteered from steer angle and torque levels jumping and also pulling the wheel from drivers hands. (Not your capable hands, I'm sure !) This is extreme version of droop flow, which drops oil flow with engine rpm. Since motors run at low rpm now at highway speed to save gas, the ability to differentiate the necessary flow rates vs. speed is too difficult to manage.

Like I wrote earlier, Its all measureable with torque, angle, speed and yawrate instruments.

A patent search should show you a huge number of "Variable Effort Steering Effort" proposals, including your A/C clutch idea. There are some cool implementations of this in the NASCAR racing field. The rule is for an engine driven belt drive pump, but who checks for wires? No intermediate speed worries, just full throttle or pit lane parking maneuverability. They use a VERY large diameter T-bar in the valve so the unboosted compliance is low.

But I've digressed... make some measurements, then report back.

 

[JohnLear]

cibachrome,
I'm not suggesting that no compliance exists, just asking a question as I was under the impression that compliance from the T bar was rather slight. If it isn’t then I’ve picked up some misinformation somewhere. I’m not an expert on power steering systems, I possess enough information to be dangerous, maybe….

I’m merely stating my experience with this particular car as I see and interpret it. Sorry no data, only what I can feel through the steering wheel, which to my satisfaction tells me that there is less hydraulic resistance in my steering with the PS belt off than there is with it on (which may or may not translate to other cars). Is this a rigorously conducted measurement that would stand up in a court of law, of course not.

I'm amazed anything ever gets sold to anybody for fear of Grandmas lawyer, but appreciate your point. I’m not intending set up shop flogging this, but I might make one up for myself. Gran’s Lawyer will need to chase a different ambulance.

'Lash' is the wrong word, apologies for being careless in word choice. I can detect very slight lash in my steering, which is obviously wear and not the torsion bar twisting. I understand that compliances in the steering system will result in more steering wheel angle for X force passing through the system, but my understanding of understeer is that of something that occurs at the contact patches, not including compliances between the contact patches and the steering wheel. By a definition that includes steering wheel angle as a definer of understeer, then a car with a more direct steering ratio would ipso facto understeer less than an otherwise identical car with a less direct ratio??? What am I misunderstanding?

It's interesting to hear that Honda has already tried an automatically engaging / dis-engaging clutched PS pump, which I can imagine being disconcerting to dangerous when it decides to turn on or off at inopportune moments (just asking for trouble, it should have been obvious, I’d already considered this and even I’d dismissed it as not a good idea…). Much better to be either on or off at driver discretion.

My driving hands aren’t all that expert, but I was local kart club champion a couple of times, and 2nd in the NSW state dirt track titles years ago, but nothing particularly impressive. I think if the power assist decided to switch itself off half way around a corner, then more capable drivers than I am might find it very easy to lose control…

 

[cibachrome]

Vehicle understeer (a gradient in units of deg/g) includes all lateral g dependent components: tire lateral stiffness, tire aligning stiffness, camber stiffnesses, geometric contributions driven by lateral g (roll steer and camber), deflection reactions (steer and camber) and any other moment rections (tractive force, aero, etc). The tire contributions is not the biggest one unless you have a very large front to rear weight bias. The steering system is almost always the biggest contributor. This is because the tierod loads are reacted in the control arms, steering gear mounts, power assist reaction and steering shaft torsional flexure. This is usually managed, taking into account handwheel vibration tolerance, durability and wear concerns, and vehicle payload changes. Your Honda is probably about 2.5 to 3.0 deg/g, for linear range understeer, (from measurements) more so with the belt off. Nonlinear understeer is probably no different with the belt off because compliance effects are swamped by the load transfer distribution setup in your car. The degrees in the understeer metric are "reference steer angle". This is an imaginary angle theoretically at the spindles by normalizing the handwheel angle by the overall steer ratio gradient. (The gradient word here is most often missed by even the most knowledegable handling types. In some cases, the "steer ratio" is good enough, however in many steering systems, especially ones with Cardan joints, the "overall ratio" metric is incorrect.

Since the ratio normalizes the angle in the definition, it theoretically doesn't affect the understeer description. However comma, it can. Because the steering loads (internal and external) are often higher with the faster gearbox, you can add understeer with a "quicker" steering box. This frequently leads to the most maddening of conclusions that a "faster" box produces a "slower" car. Since the understeer is in the steering gain function, understeer is designed in to slow the steering response down to make the vehicle more manageable at very high speed. So, a 14:1 gearbox in a 4 deg/g car "feels" better than a 18:1 gearbox in a 1.5 deg/g car (just some rough numbers). The 14:1 job is more fun to park, and is comfortable to drive while going 80 mph down I-96, while the 18:1 job is not as easily maneuvered in town and takes some concentration at 80 mph. This assumes that the designers have also adjusted the effort levels for a constant steering workload level.

The final ingredient in the vehicle understeer recipe is the rigid body moments contributed by the tire aligning torques. Since tires resist slipping they want to go back to a relaxed state. These 4 moments act on the whole car and thus add understeer. In fact, there are very few if any "neutral" steer cars. Perhaps at some very controlled payload, but none out on the road. I'm sure your car feels "neutral" but as measured it won't be. Perhaps you've decided this based on the CHANGE in understeer trim you sence is fairly constant

 

[cibachrome]

Vehicle understeer (a gradient in units of deg/g) includes all lateral g dependent components: tire lateral stiffness, tire aligning stiffness, camber stiffnesses, geometric contributions driven by lateral g (roll steer and camber), deflection reactions (steer and camber) and any other moment rections (tractive force, aero, etc). The tire contribution is not the biggest one unless you have a very large front to rear weight bias. The steering system is almost always the biggest contributor. This is because the tierod loads are reacted in the control arms, steering gear mounts, power assist reaction and steering shaft torsional flexure. This is usually managed, taking into account handwheel vibration tolerance, durability and wear concerns, and vehicle payload changes. Your Honda is probably about 2.5 to 3.0 deg/g, for linear range understeer, (from measurements) more so with the belt off. Nonlinear understeer is probably no different with the belt off because compliance effects are swamped by the load transfer distribution setup in your car. The degrees in the understeer metric are "reference steer angle". This is an imaginary angle theoretically at the spindles by normalizing the handwheel angle by the overall steer ratio gradient. (The gradient word here is most often missed by even the most knowledegable handling types. In some cases, the "steer ratio" is good enough, however in many steering systems, especially ones with Cardan joints, the "overall ratio" metric is incorrect.

Since the ratio normalizes the angle in the definition, it theoretically doesn't affect the understeer description. However comma, it can. Because the steering loads (internal and external) are often higher with the faster gearbox, you can add understeer with a "quicker" steering box. This frequently leads to the most maddening of conclusions that a "faster" box produces a "slower" car. Since the understeer is in the steering gain function, understeer is designed in to slow the steering gain characteristic down to make the vehicle more manageable at very high speeds. So, a 14:1 gearbox in a 4 deg/g car "feels" better than a 18:1 gearbox in a 1.5 deg/g car (just some rough numbers). The 14:1 job is more fun to park, and is comfortable to drive while going 80 mph down I-96, while the 18:1 job is not as easily maneuvered in town and takes some concentration at 80 mph. This assumes that the designers have also adjusted the effort levels for a constant steering workload level.

The final ingredient in the vehicle understeer recipe is the rigid body moments contributed by the tire aligning torques. Since tires resist slipping they want to go back to a relaxed state. These 4 moments act on the whole car and thus add understeer. In fact, there are very few if any "neutral" steer cars. Perhaps at some very controlled payload, but none out on the road. I'm sure your car feels "neutral" but as measured it won't be. Perhaps you've decided this based on the CHANGE in understeer trim you sence is fairly constant, but that's not likely a zero understeer condition. Keep in mind that a true neutral steer vehicle has first order transient response traits and this is judged to be "sluggish". A bit of understeer loosens the system up a bit to quicken it more to your liking. And finally, an oversteer vehicle isn't necessarily unstable in open loop control. Its only unstable when the wheelbase to speed (squared) ratio is less than the amount of oversteer. There's a 9.8 and a 57.3 in that formulae, but you should get the jist of it.

OOPS, one more: Because tires aand steering systems are getting "better", getting enough understeer in the car to be safe and comfortable is more and more being done with wider rims on the rear and different sizes (larger on the rear). The downside is the issue of a spare tire and rotation of tires for wear optimization. Even run-flats don't solve this problem because you're limited in speed and range on the bum tire. But, it saves on trunk space and actually is a loser from a weight savings (and thus fuel economy) standpoint.

 

[patprimmer]

So the short story is don't turn off the PS pump?

I have to agree that compliance in the torsion bar in the steering box can cause a discerning reaction delay leading to oscillating over corrections.
Also a not working power steering system is a lot heavier than a properly designed manual steering box.

I had an EA Falcon S pack multi point 5 speed. It had a habit of throwing power steering pump belts that was bad enough that I always carried a spare or two and some custom made tools to it the belt on the side of the road. It always threw the belt when the tail was out and I was power sliding through a curve on a dirt road. Probably sudden high load on the system and a slight over rev as the wheels spun. I was taught to drive in old heavy car with lots of steering kick back, so I learnt not to wrap the thumbs around the wheel. Also I have quite high upper body and arm strength so n loss of control ever resulted, but the car was considerably faster on these winding dirt roads with the power steering working, simply because you can steer faster and get to opposite lock a lot quicker when required.

A real world example being an Aussie (I think you said earlier), you might remember the Channel 9 Camero driven by Colin Bond and Kevin Bartlett.

It was KBs car and it had power steering. When practising for Bathurst one year, CB complained it was slowing them down. They decided to do a few hot laps each, first with the PS operating, then without. Both drivers consistently did faster laps with the PS operating. Needless to say, they raced it that way.





Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[cibachrome]

If you accept the increase in efforts from the unboosted gear, get a spare valve from a scrap yard and have it welded up to eliminate the t-bar action. Then, either drain the oil or plumb it back on itself to reduce the back pressure. Or even stuff a manual gear in it. Chances are this gear won't be 17:1 but more likely 23:1. The gain will be lower but the car will be greener.

 

[JohnLear]

Sorry to ignore my thread for so long, lack of spare headspace recently…

----------------

cibachrome,
I very much appreciate the time and effort you’ve taken to go into this. I can’t pretend I have a handle on everything you’ve said, some of it supports my existing understandings, but at least some is a bit confusing (and I’m afraid some is over my head).

So one of the things you’re saying is that(?); when a ‘sporty’ variant of a road car is factory fitted with a ‘faster’ steering ratio, then this doesn’t ipso facto increase understeer, but, that the handling boffins may deliberately increase inherent understeer (or do you mean steering response?) by whatever means in order that ‘the average driver’ won’t be scared witless at highway speeds by highly reactive steering, while still being ‘thrilled’ by the steering ‘directness’ at slower speeds?

So, in such a case we may find a ‘sporting’ variant fitted with a ‘faster’ steering ratio that when pushed may actually be a worse offender in the excessive understeer stakes than is the base model shopping trolley?

That seems easy enough even for me to understand and I can see the rationale behind it, but am I understanding this part of your comment correctly?

With my Accord I used to feel that the steering ratio was too low geared, but after all the changes I’ve made to the car the steering feels a lot ‘faster’, even though the steering ratio is unchanged (and still on stock sized tyres on the stock wheels). The car still understeers, but a lot less than it used to, and is far more responsive at all speeds.

------------

I am still having trouble understanding how the “steering system is almost always the biggest contributor” to understeer, by which I don’t mean to suggest you’re wrong, just having trouble grasping how it is so.

I was under the impression that it is the front / rear distribution of dynamic weight transfer that is the largest factor affecting understeer / oversteer characteristics (all else being equal such as contact patch presentation to the road etc. etc). You say (at least I think you’re saying) that it isn’t, so I’m now trying to deal with a significant tilting of the Earth’s axis…

If the distribution of dynamic weight transfer (whether through geometric and / or elastic vectors) isn’t very important to front / rear grip bias, why can changing the roll couple (geometric and / or elastic) affect understeer / oversteer so strongly…?

You start to lose me when you speak of “linear” and “non-linear” understeer, and deflections that occur within the steering mechanism contributing to understeer (e.g. torsional flexure in the steering column shaft etc). As best as I can interpret it, it seems like you’re talking about a difference between ‘real’ understeer’ and a ‘perceived’ understeer, though I’m not at all sure I am getting the correct gist of your comments.

My conception of understeer doesn’t include deflections between the driver and the wheel rims, only the relative front vs rear slip angles at the tyres. In this case deflections (and changes in effective leverages as caused by single cardan joints or unequally angled double cardan joints, or steering arm arcs etc) within the steering may well mean the steering wheel is turned more for a given corner taken at X speed / lateral acceleration, but in my private universe this doesn’t equate to an understeer, just a vagueness, or rather, a slightly erratically ‘elastic’ steering ratio (not really a good thing, but hard to avoid entirely).

I keep feeling I’m thinking of something slightly different to what you’re thinking of, but then maybe my conception is just too simplistic…

At any rate, with my PAS disabled I haven’t noticed any change in the handling balance of the car, though of course this doesn’t mean that none has occurred, I may just not be sensitive enough to pick it. Something that made a far greater difference was the deletion of the front ARB (decreasing understeer, though turn-in steering response also lessened, requiring an increase in front damper stiffness and stiffer damper rubbers to ‘re-sharpen’ the turn-in).

-------------

I know that a neutral handling car is something of a phantom that doesn’t usually exist, or is very difficult to drive when it does, and that a car that feels ‘neutral’ would generally just have less understeer than is usual with most cars.

My understanding is that if you get too close to an (on average) neutral set up then the car also gets too close to repeatedly stepping over the line into abrupt oversteer. Better to have a slight understeering tendency that can be provoked into a more easily controlled neutrality or oversteer, than a fundamentally neutral tendency that readily trips into a difficult to control oversteer (possibly fluctuating too easily between understeer and oversteer). If that’s not too simplistic…

----------------

Pat, I remember the Bartlett Camaro at Bathurst, on it’s roof (7 out of 10 for that crash…). Years ago I had a brief drive in a “GT” Falcon (XC from memory) that had had the PAS disabled, and it was hard work to the point of being almost un-drivable (and I have good upper body strength, maybe the 13” steering wheel didn’t help…).

It’s not so bad with my Accord, I can easily accept the increased effort required to steer it with the PAS disabled (even with more than double the stock caster), but my wife won’t I’m sure.

As far as the “short story” goes (i.e. “don’t turn off the PS pump?”); it seems that (other than a suggested increase in understeering tendency that I’m not sharp enough to understand and can’t feel) the only answer to my question about other not so obvious downsides to disabling the PAS is cibachrome’s comment re torsion bar non linearity being excacerbated by the lack of assist.

Since this was suggested as a potential issue, in some circumstances I suspect I can occasionally detect (psychosomatically?) some slight ‘woolliness’ that might (or might not) be attributable. If so then it doesn’t appear to be a significant problem (I still far prefer the PAS disabled), but it would be interesting to try a welded up assembly as per cibachrome’s suggestion.

Of course this is only an option if the PAS belt remains off permanently, which it might not, or at least it’s good to leave the option open (and I may or may not get around to fitting an AC clutch on the pump).

Any rough idea of what might be a typical range of input shaft rotation / T bar twist before the connection between the input shaft and the pinion becomes ‘solid’ in the direction of steer?

 

[patprimmer]

It depends on the kick back weight. If you put more effort into the steering wheel you get a bigger twist in the torsion bar so the spool valve flows more fluid or more power assist to help it catch up with your input. When steering very fast, you can put so much twist into the shaft that the power boost does not assist enough to keep up once you require more than valve wide open fluid flow. That' about when the belt typically threw off. At least he fact that I was ahead of the assist reduced the impact as the belt threw.

I guess he short answer is it can twist enough to fully open the port in the spool valve. If you can get a steering box to disassemble and inspect, you will see.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[JohnLear]

Thanks Pat,
I've disassembled a fair few steering boxes (from Hillman Imp to Bugatti to Mini to Lloyd-Hartnet to...), but never a PAS system (at least not the spool valve). I was thinking there would be a point past which spool valve assembly became 'solid' to further movement in that direction, as a failsafe should the T bar fail (?).

Pulling apart a PAS rack would tell me what I want to know, but asking was a lazy and cheaper alternative...