I removed my uncharged AC compressor for better access to spark plugs. The K060365 Gates belt I used isn't the recommended belt but looked OK in terms of spacing + rib depth.
I'm pretty sure I routed it and tensioned it correctly.
The belt easily drifts forward and off the front lip of the power steering pulley by just turning the engine over without starting the car. See pic below.
Eyeballing the bottom block pulley, PS pulley + idler alignments, the rear lips all appear to line up properly. The idler pulley turns freely with no wiggle.

The belt coming off partially, wasn't an issue with the Mazda belt + 4 pulleys in the stock setup.
The Mazda alternator belt tracks OK.
I zip tied the loose AC lines and covered fittings with heavy tin foil.

Has an anyone else experienced anything like this ? I remember reading that running the car without the PS pump causes damage to a steering component, .. not sure which.

thanks


Belt riding on front lip of PS pulley

 

Flip the tensioner pulley around so the lip is facing outwards

 

Thanks, but if 6 ribs and the idler pulley lip can't hold this belt in place, I think either the belt or a wobley pulley is the likely problem.

I couldn't find any pulley wobble turning them by hand..
i'll watch the pulleys + belt while someone else turns it over, I guess.
If I flip the idler around, and the belt travels the other way, it could cause much bigger problems..
Does anyone know of a compatible aftermarket idler pulley with 2 sided lips is available ?

i'll try to call Gates to see if this belt looks like a possible fake. I don't think kevlar belts would stretch like this.

 

the ribs in a serpentine belt are SAE standard, so they are all the same, as are pullies and such.

the K060365 = 6 rib which is 36.5" long. if you multiply by 25.4 you get mm, which is a 6PK927. there is a fudge factor because some brands use the ID and some the OD... Mazda lists a 6PK920
there exist double sided belts too, its a 6Dpk920 or DK60365

if the pulleys are lined up, i would try a different belt. sometimes gates belts are like rubber bands, even at max adjustment they fly off or flip over

 

Your pulley appears to be on correctly but looks too close to the pump housing (may be pic deceit)....pulley edge clearance from pump housing should be about 9mm .

 

All good things to know .. thanks guys !

The pulley to bracket spacing is 9mm on mine.
The fact this belt so easily mounts the front pulley lip, aligned, or slightly misaligned, does suggest some rubber band belt qualities.

A Gates rep did contact me and offered an easy test as to wether a belt is true or not. Laying the belt on it's edges, on a smooth true surface like a granite or glass top, the belt edge should uniformly touch the surface along both edges and stand square to the surface.
The rep also suggested trying to move pulleys sideways + up and down by hand to see if there is any sideplay.
I didn't find any, but did sand and spray paint my idler pulley to discourage rust, so that's a possible factor. A friend is bringing over a different belt in a week, so I'll investigate further in the meantime.

 

I had a similar problem after my engine was replaced with another after 4 years use.
The problem was with the power steering pulley. It now was about 1/16" back too far. Thus one outer rib would be cut off.
Placing a washer between the pulley and pump bearing solved the problem.
Thus I question Mazda's standards with early 1992 made 13B-REW engines and later ones.
My original engine failed due to the second rotor oil cooling jet coming unscrewed from the e-shaft just a couple months past 4 years causing drastic drop in oil pressure.
Car is a March 1992 FD R1.

 

Thanks cewrx7r1.

Sounds like the easiest solution for me to try,
with my stretched belt as a guinea pig.
I didn't find any other potential causes, other than my cleaned up idler pulley.
If it works, I'll buy a better micro-V belt.