When I got my current ND last year, the alignment was off. Like "handles like a drunk hippo" off. The local Miata specialist race shop did a fast road type alignment that had a bit less camber and a bit less caster than the FM recommended numbers. This improved handling massively.
However, I recently started driving the car after it sat for a couple of months thanks to winter and noticed that when I have to make minute directional corrections at speeds > 50mph, it almost feels like I'm pushing against a detent on the steering and have to use what feels like way too much force to correct, then correct the correction because of overcorrecting it the first time overcoming the steering forces.
I recently had an ND3 RF as a rental car (thanks, National!) and while I noticed that the steering had a pleasant heaviness at similar speeds, it didn't have the stiffness around the straight ahead position that my ND2 has and as a result, was much more pleasant to drive.
Does this vague description ring a bell with someone? Potential alignment issue?
My current plan was to take it back to the same alignment shop (who have always done good work for me) after getting new tires and have them align the car to the recommended FM numbers to see if it makes any difference.
The original NDs (2016-17 soft tops) had a dead spot in the steering, but the ND2 should have decent feel and no "stiction" in the center. I'd definitely like to check out the alignment first.
The ND3 does have a new steering rack, so it is going to feel a bit different. I haven't dug into it too much, it's been too long since I drove a stock ND2.
Does Mazda have a steering recalibration that needs to be done after alignment?
Alignment shouldn't cause what you describe. It sounds like a mechanical issue, although so many manufacturers stick steer-ahead stability routines in their electric power steering systems that I'd make sure it didnt need to be recalibrated first.
The way you describe it makes me wonder if you have a bad steering rack or an issue with the EPAS.
Maybe a worn EPAS coupler bushing?
I'm hoping not - the car only has 26k miles on it and I would hope that the steering components would last a bit longer than that.
Then again, this is me we're talking about...
"Stiction" like Keith described it is a pretty good way of putting it. Like, I have to overcome some resistance to make small corrections. Almost like the EPAS is holding the steering straight for me. Feels different compared to what I would think of as a dead spot.
No reset procedure I could find on a quick search, DDG actually brings up this thread as the first hit.
I wonder if you get the same feeling with the front wheels off the ground?
You could also compare with engine running and with it off to remove the assist from the equation. That should at least tell you if it's in the steering column (sticky u-joint?) or something related to load/speed.
In reply to No Time :
Hmm. Probably not a bad idea to throw it up on my QuickJack and see how it behaves.
I've driven an ND with a glitchy EPAS, you wouldn't describe it like this. It's very clear when the assist drops out.
In reply to Keith Tanner :
No dropping out. At least it's consistent in its behaviour.
Threw the ND onto my QuikJack this afternoon to try and dig a bit deeper. Here's what I've learned so far with the car in the air:
- Don't drive an ND right after you drove an Elise, especially when you're trying to diagnose steering issues.
- No discernible steering play at either the front wheels. Feels like I would expect a low mileage car to feel.
- With engine and ignition off, steering moves smoothly from lock to look. No binding/loosening going on. Again, feels like you'd expect a low mileage-ish car to feel. Because I was looking for it, there is the tiniest - and I mean tiniest - hint of some stiction/minimal play in the straight ahead position. I was surprised how heavy the steering felt though.
- With the engine running, obviously there is a lot less steering effort lock to lock. Again, feeling smooth lock to lock with no binding. The stiction/play around the straight ahead position was a bit more pronounced with the engine running and thus the EPAS on, but nowhere near what it feels like at speed. And when I'm talking about play, I'm not talking about 60s Caddy levels of play, just maybe a 1/4" movement (if that) at the wheel before there's a little resistance and the wheel are starting to turn.
I don't think the alignment should have much of an impact on the wheel movement with the wheels about a foot in the air. The tyres aren't that sticky...
That said, I could see how the small amount of stiction I noticed with the unloaded wheels at zero mph would be more noticeable when I'm driving at higher speeds when I expect the EPAS to dial down the assist level a bit.
When the EPAS (with over toiletrained controls programming) is miscalibrated, this can create a pull to one side because it's trying to force the steering to what it thinks is straight ahead. Very annoying.
You said the alignment was extremely wonky. The EPAS usually learns to a bad alignment over time and will need to be reset, usually this involves a steer ahead recalibration (connect scan tool, do whatever hokey pokey it wants with the steering and brakes, click done).
There's also a possibility that the alignment was wonky due to a major suspension impact either by collision or hitting some righteous suspension bending road features. There's a chance that this may have caused a bind in the power steering part of the rack, which on EPAS racks is generally part of the opposite side of the rack (left side on RHD, right side on LHD)
Will write more in a moment
In reply to Pete. (l33t FS) :
Thank you. My google-fu seems to be fairly weak, I don't seem to be able to find information about recalibrating the EPAS, although it intuitively makes sense that there should be a way to do so. Wouldn't be surprised if that's another thing that requires dealer level scantools. There might be some info on miata.net but they disabled my login because it's been so long. I probably should try to get that reinstated.
The car had a clean Carfax when I bought it and I couldn't find signs of major collisions when I went over it. Of course that doesn't mean that someone didn't take an agricultural line at a trackday or auto-X and shifted things around that way. Didn't find any signs of that other than the wheels showing evidence of a fair amount of "parking by ear".
Since I had the alignment done, it does track straight and goes around corners in the manner you would expect and in both directions. But that doesn't mean that there isn't something that's not quite lining up but doesn't show up when you're doing a normal alignment.
What you are describing is what happens when the steering box in a reciprocating ball set up goes bad or is adjusted to tight. This having a rack I would be looking at the rack as the problem.
Here is the EPS CM (electric power steering control module) auto calibration procedure:
12.If the EPS CM is replaced, perform the auto configuration using the following procedure.
(1)Switch the ignition ON (engine off or on) and wait for 1 min or more.
(2)Switch the ignition off and wait for 3 s or more.
(3)Switch the ignition ON (engine off or on) and complete the EPS CM auto configuration.
(4)Clear the DTC. (See CLEARING DTC [EPS CONTROL MODULE].)
(5)Switch the ignition off and wait for 3 s or more.
(6)Switch the ignition ON (engine off or on) and wait for 1 min or more.
(7)Using the M-MDS, verify that DTCs U2300:54/U2300:55/U2300:56 are not displayed.
IIRC you can pull EPS DTCs with the ForScan freeware, but you'll need a laptop as it's not a phone app.
This is a behavior that appeared while the car was just sitting, correct? Are you 100% sure it's the car and not a perceptual change with the driver? As noted, the ND3 has a new rack so it's not a direct comparison. Perhaps getting a moment behind the wheel of another ND2 might be a good check before we dive too deep into looking for something weird. This stiction that you're finding as you perform a test you've never performed before may be how the car improves straight line feel and stability, the earliest cars were a little loose in this regard.
i have a 22 BMW that has a similar feel. It is still under warranty so I have taken it to the dealership twice for them to say "we can't reproduce the problem" The issue is most prevelant after driving down a straight road for an extended period of time (ie if you drive around town it doesn't appear but jump on a highway for 20+ miles and it shows up) it goes away if you move the wheel more than 90 degrees to the left and right but the dealership swears they see no issue...
Keith Tanner said:
This is a behavior that appeared while the car was just sitting, correct?
Yes, sitting with some battery drain, just enough to cause some issues when starting the car.
Are you 100% sure it's the car and not a perceptual change with the driver?
I'm not, which is why I asked for opinions/feedback. As mentioned it doesn't help that I have a couple of other cars with known superb steering feel.
As noted, the ND3 has a new rack so it's not a direct comparison. Perhaps getting a moment behind the wheel of another ND2 might be a good check before we dive too deep into looking for something weird. This stiction that you're finding as you perform a test you've never performed before may be how the car improves straight line feel and stability, the earliest cars were a little loose in this regard.
MX5s are rather uncommon here, so I'm not sure how easy it is to get behind the wheel of another ND2. I think you might be onto something re the straight line feel and stability as it exhibits plenty of that.
Right now my non-plan is to await the results of the used oil analysis and as long as that doesn't come back as an annoying negative, I'll throw on a set of fresh tires and have another alignment done. The front tires are pretty EOL and I want to rule out that they're the problem.
I keep finding subtle modifications on this car, so it's still a voyage of discovery.
Does this car have any Ujoint in the column? This is classic Ujoint sticky in a Volvo 240 that has seen a underhood pressure wash as a used car.
In reply to TurnerX19 :
It's a little early for that, age wise, but good point.
A sticky U joint will have a bind every 90 or 180 degrees, all along the steering range.
Very common for people to bring their cars in for a "power steering problem" that only happens in a few discrete steering wheel positions. Usually on SUVs or other things with underhood steering shafts, as most newer transverse cars have the rack sealing right to the floor/firewall so the U joints are all indoors.
I found that when this happens on an E46, it also damages the rack from the side loads on the quill. 
The rack is cheaper than the steering shaft but only because the steering shaft is amazingly expensive, at least on an all wheel drive model.
If the alignment was off that far on your low mile car...what caused it? You didn't note anything that the shop reported as loose or worn out when it was aligned, and usually things don't go out of alignment without either something wearing out or a lot of force being put through parts. Makes me wonder if someone hit something and the rack or a ball joint or something has some damage.
Otherwise the only comparable experience I've had is my 924. It sat for a long time before I got it, and it had kind of a hanging/high friction steering feel. Ended up being a ball joint that had dried out and then wore a series of notches in the straight line direction when I started driving it again as the suspension cycled up and down going down the highway. The joint didn't have any play. I found it by accident when I had the struts off to replace the inserts and was moving the joints around when not connected to anything.
tfaubus
New Reader
3/13/25 3:31 p.m.
In reply to BoxheadTim :
Same here, also seems temperature related as it's not noticeable when cold outside, last year I got it up on my qj and did full travel left to right and back multiple times and it was resolved, until last week driving in the warm again, morning it doesn't happen but afternoon yes, 2019 40000 miles, club rf
gearheadE30 said:
If the alignment was off that far on your low mile car...what caused it? You didn't note anything that the shop reported as loose or worn out when it was aligned, and usually things don't go out of alignment without either something wearing out or a lot of force being put through parts. Makes me wonder if someone hit something and the rack or a ball joint or something has some damage.
Car has a clean history and nothing I've found so far indicates that it's been in any case of incident other than some marks on the wheels. Not even any marks on the OEM tires. I remember that my LE came from the factory in dire need of an alignment as well, although it wasn't as bad as this car was.
The local race shop that does a lot of work on various Mazdas did the alignment and from prior experience with them, they would've pointed out anything that was off that they could detect during the alignment.
tfaubus said:
In reply to BoxheadTim :
Same here, also seems temperature related as it's not noticeable when cold outside, last year I got it up on my qj and did full travel left to right and back multiple times and it was resolved, until last week driving in the warm again, morning it doesn't happen but afternoon yes, 2019 40000 miles, club rf
Interesting - in my case, my impression is that it's worse in colder temperatures. A lot of the recent driving has been in temperatures ranging from the 20s to high 30s.
Yeah, my comment about it being worse in colder temperatures was wishful drinking.
Now that civilians like me can read forums posts on miata.net again without being logged in, I found someone mentioning similar "sticky" issues that were caused by a lack of lubrication between the rubber grommet on the firewall and the steering shaft. Figuring that it couldn't hurt I added a bit of Sil-Glyde there and it seemed to help, but it didn't fix the issue completely. It did remove the noticeable "stick", though. I guess it's time for an alignment to see if that makes any difference.
Either way I'm getting the impression that what I'm dealing with now is the programming of the EPAS.
