I have a race tuned Nissan RB26 Skyline engine with a problem. The separate multi rib belt that drives the PAS pump form the new and stock crank damper turns over or comes off when the engine is given race type usage. I am 100% sure the pulleys are in alignment in all planes, having even borrowed a Gates laser pulley alignment gizmo. The drive set up is totally stock, the belt is new and genuine Nissan, and the same set up worked on my old engine with no issues. Now, what i am wondering is if the new Tractive sequential gearbox and triple plate clutch are allowing shock loads or rapid RPM changes to get a "wave" in the fairly long belt run, and it then flips over or comes off? I know Toyota run a hydraulic damper on the spring loaded tensioner on the multi rib drive belts on some manual models, but the same set up is un-damped on autos. These belts drive all the ancillaries, mine just drives the PAS pump. My stock Nissan set up doesn't use a pre loaded tensioner, the adjustment is manual and fixed. A similar, separate, but much shorter belt run drives the alternator with no issues. I have made an idler pulley to run on the back of the belt, on the none drive side, to possibly stabilise things, but to test it involves (expensive) track time. What do people reckon my chances of having this effect a cure are? Anyone into drive belt technology, it's frustrating something previously reliable and apparently simple is causing hassle? Thanks

Without a loaded tensioner, a ribbed belt works pretty much like a vee belt -- it grows with speed, forms ripples and standing waves, etc, and is then prone to turning over on the pulleys or jumping off altogether. Load oscillations could be a factor, but increased and/or prolonged rpm are generally more than sufficient.

Easy fix: somewhere along the slack side of the span, install a long bolt a few mm away from the back side of the belt. Not touching -- it only needs to make contact when the belt begins to stand up. If you want to be really fancy, install a length of tubing on the bolt (free turning) to serve as a bearing. Your back-idler pulley sounds like an elegant solution. I would be very surprised if it doesn't cure the problem.

Try a different brand of belt. It shouldnt make any difference but in my experience it can, Then the biggest thing is slow the pulleys down. I presume it comes off at high RPM and you probably do not need to turn the pump that quick. Bigger pump pulley or smaller crank pulley. A lot of P Steer pumps are adjustable for pressure, a bit less pressure may help.The standard set up is designed for low RPM street use and when you are using 3 times the RPM problems will arise, hence slow the pump down.
The sprung tensioner may help but is a patch not the problem.
I have never seen a rib belt flip, flick off yes,shred themselves yes, burn the ribs yes [with high RPM] That is a problem I used to have with V belts and a different brand solved the problem. Never worked out why though too many RPM was part of it.
A friend was having this problem with alternator belts [V] in the end he found leaving the belt very loose solved the problem. sounds wrong but it worked. And the powersteer pump is slowed down over 50%. Otherwise they boil the oil and have belt problems. And you may make a few extra HP too.
Personally I prefer V belts or gilmer belts on race engines but rib belts should work fine, just replace them very regularly.
While dyno time is also expensive, if the problem persists run it hard on the dyno and watch the belt. You may work it out that way.

If the belt is flipping this suggests that the problem is a torsional mode i.e. the belt is twisting along it's length. I agree with McGuire that constraining the belt on it's free side should be the first thing to try and the next thing would be more tension. Other thing to check is that your pas pump bracket is nice and stiff particularly if it's a home made effort and is not twisting under the steering load such that the pulleys are no longer aligned - get someone to turn the steering wheel whilst the car is stationary and see if the pulley axis changes.

Indeed. If you look at a ribbed drive belt, you can see it's really just some number of v belts in tandem on a common carcass (four-rib belt = four little v belts) and what makes any v belt flip over is torsional vibration. In effect the belt is standing off the pulleys and twisting. The pulleys needn't be out of alignment. Could be excessive load or load variation or increased speed/frequency. Even with perfect alignment, if the friction on both sides of the v were identical that would be a miracle, so the belt is always squirming to some extent. Too much and it flips over on the pulley and/or flings itself off altogether.

Some great answers, and so fast, thanks! The car is on the trailer and off to Oulton tomorrow to see if it's fixed with my pulley arrangement. I am developing the engine and box to eventually put in a brand new shell I found on Ebay in Japan that now has a nice multipoint cage in it. I want the gremlins ironed out before transferring stuff. Next on the list was going to be a small PAS cooler, the car is undriveable without PAS, as I found at Anglesey two weeks ago when the belt came off and I had "a little fright" ;) I like the idea of slowing the pump, and a bigger pump pulley would be far easier to have made than a different damper, as the PAS drive pulley is an integral part of the stock one. I have taken aboard the comment on bracket stiffness. The stock bracket is a forging, immensely stiff and rather heavy I doubt there's any deflection but I will try having someone heave the wheel whilst i look for pump movement in the morning. a good point, thank you. I am fairly confident the guide roller on the back of the idle side of the belt will help, I remember thinking it was quite a long belt run compared to the alternator belt.


Does altering the PAS pump pulley diameter affect actual belt speed and centrifugal force though? Wouldn't I need a smaller DRIVE pulley to alter that?


Thanks again. There are some photos of the engine showing the PAS pump arrangement on my web site, albeit with no belt fitted. I have fabricated a 3/8 inch thick steel plate to mount on the two gold irridited bolts holding the stock pump tensioner. On this metal arm I have fitted a new cam belt tensioner pulley of about 2.5 inch diameter that runs on the top side of the belt on its outer plain face. The pulley indents the belt run slightly.

http://www.chriswilson.tv/finished/finished.html

If you're using the stock pas bracket then that wont be the problem as it will have been designed with NVH in mind so should be plenty stiff enough. To get the belt speed down you are right in that the crank pulley diameter needs to reduce, this is probably not straight forward to do but if you reduce the pump speed you also get a reduction in parasitic loss and because you are not thrashing the hydraulic fluid round so fast you might not need the cooler either.

Hopefully the idler has fixed it, if not try a bit more belt tension.

Can't really say definitively from the photos as the parting lines are not visible, but the PAS bracket is not likely to be a forging, is more probably a casting... which is worthy of mention only because castings are known to crack, which can create a lot of problems with belt alignment until detected. Often a crack in an engine mount or accessory bracket (casting or stamping) can be totally invisible, spreading upon load but undetectable otherwise. You can't even see it until you unbolt the piece from the engine, when (surprise!) you discover it was in two pieces all along. Seen it many times. With late-model engines salvaged from collision jobs (their most common source) it's very common. And since the engine hasn't been used in that condition for very long, the customary tell marks -- a trail of rust along the crack due to fretting between the broken surfaces -- haven't had a chance to develop yet. So the crack may be more or less invisible.

I have no particular reason to believe that is the case here, just throwing it out there in the interest of information on the general topic. Mentioned in passing, I suppose.

After maybe 75 laps of Oulton yesterday, including an over rev through a missed down change the belt is still on, correct and in perfect condition, so the idler pulley running on the back of the belt on the none drive side has cured it Many thanks for the ideas and discussion! Great forum.

I see it's fixed but +1 for this /\/\ - the long bolt with a teflon tube is what I have used before, the belt is simply compressing on itself during rev fluctuations and flicking around the place.

Hmm I just remembered seeing this idea on a standard car but can't remember which.

...it put a rod through the block but the belt is perfect!!

No, that was last year at Anglesey..... http://www.gatesgarth.com/bang/bang.html

Ouch!

Agree, you got some great responses to this one. McGuire's suggestion is a good one.

It never occured to me (til this thread) that muti rib belts would be prone to torsional modes and "flipping" but it makes sense when you think about it, higher section MOI + lower torsional stiffness = lower natural frequency. Maybe even lower than the first bending mode in some cases - anyone have info on this? Anyway the idler (or bolt) at mid span will eliminate the first mode of either so - easy fix.

Interesting that the tight side (which is the same length in this application) does not suffer due to the higher tension and higher natural frequency. Could the belt be tensioned enough to move the slack-side natural frequency above the operating range?

Does the use of equal size pulleys make the problem worse by intruducing similar excitation frequencies at both ends?