So I'm no good at this math thing.

My car was automatic from the factory, meaning it's supposed to have the 3.90 rear. As many of you already can tell from my signature, I converted the car so that it would have the gearbox it should've had from the factory.

I used an S5 manual tranny and whatnot, which for the N/A cars should be identical to the S4 N/A manual tranny ratio wise... at least according to Icemark's FAQ greatness.

Yes, I used the 3.90 speedo gear in the transmission so that my speedo would read correctly, but it simply is not.

When the speedometer is reading 60-ish...more like 62-63 I'm doing 50mph. When the speedo reads a bit over 80, I'm doing 70mph...so on and so forth. By 90 on the speedo I'm doing 75mph.

Basically, I'm trying to determine what's going on here. The only things I can think of are that the S5 N/A gearbox does indeed have different gear ratios than the S4 or maybe I don't really have a 3.90 rear...like if someone had the diff replaced and didn't know there was a difference.. It is an LSD for sure, I do know that much.

Now the question: Assuming the gearbox ratios are identical which I'm sure they are.. let's say I'm using the 3.90 speedo gear (which I am) but the diff was really a 4.10... mathematically speaking is this variance in the speedometer reading accurate?


I'm so confused.

 

If your car was an auto from the factory, I don't think it would have had an LSD too. The spec sheets say all S4 autos had 3.9 rear ends, and I've never heard of a 3.9 LSD. Only 4.1 and 4.3. You should have had a 3.9 open diff stock, so unless it has been swapped, that's still what you have.

I also know that even if you have a 4.1 rear end and a 3.9 speedo gear in the transmission, it won't mess up your speed readings by 10 mph. That's a rediculous amount of error. I ran the exact setup I just mentioned in my vert (3.9 speedo gear & a 4.1 LSD rear end), and never noticed any error at all. My speed was always dead on when I'd drive past those radar traps. If anything, I wouldn't expect more than about a 1-2 mph difference or so at higher speeds, at the worst.

 

You are incorrect. The automatic GXL cars got a limited slip 3.9, at least in the Series4 cars.

Also, as speed builds the amount of error grows as well.

Can I get someone's input who knows how to spell 'ridiculous', please? Where are my math whizzes at? I know you're out there!

 

Oh God, a typo. That's really the end of the world, now isn't it? Congratulations on being an ******* for no reason. You might try searching if you can't handle someone giving you a reply.

I've read conflicting information on the auto rear ends. I've seen where Icemark said they had LSDs, but never anything that said they were 3.9 and not 4.1s. At this point, it doesn't even matter. You said you switched the damn speedo gears.

And of course the error will grow as speed increases, but a 10 mph difference isn't going to happen at 60 or 70 mph. You've got to be incredibly naive to think that kind of difference is going to happen from a 0.2 change in final drive ratio. Plug in your gearing/tire size here: http://www.catherineandken.co.uk/sti/tyres.html and compare speeds with a 3.9 and a 4.1.

Do you know what it comes to? It's about a 3 mph difference at 60 mph. You'd need to be going about 180 mph before the error grows to 10 mph. Clearly you have a mechanical or calibration problem.

 

Sarcasm.. Go look it up.

Icemark is correct. I believe I found out first via a Mazda advertisement, which I have plenty of (Hardcopy!) that the automatic GXL received a limited slip differential as well. Also, there's an LSD tag on my diff and the simple spin the wheels test further set it in stone.

Define this 'mechanical' problem. So far as I know the speedometer is a set of gears..

 

Sarcasm or not, you come off as a douche. You might work on that.

I also happened to come across this thread: https://www.rx7club.com/2nd-generation-specific-1986-1992-17/mechanically-correct-speedometer-challenged-804222/, where you asked the same question 6 months ago. You said in that thread "...when I do 73 on my speedo, it takes me (slightly vague) a minute from mile marker to mile marker. This would imply that I'm really doing 60mph when my speedo says 73mph. And while doing 80mph (according to my speedo) I still get passed like a mother ****** on I-85." If that's still the way you're "measuring" your exact speed, then that may very well be your problem. Guessing what speed others cars are doing on the freeway isn't very scientific. Timing between mile markers is more so, but that also depends on whether you're using a stop watch, looking at your clock, or counting in your head.

A much better way to decide if your speedometer is accurate or not is to use the same speed calculator I posted before. Using these specs:

3000, 4000 & 5000 RPM
205/60/15 tires
1st gear: 3.473
2nd gear: 2.002
3rd gear: 1.336
4th gear: 1.000
5th gear: 0.697
Final drive: 3.909

your speeds at 3000 RPM would be 16, 28, 42, 56 & 80 MPH. At 4000 RPM, it would be 21, 37, 56, 75 & 107 MPH. And at 5000 RPM, they would be 27, 46, 70, 93 & 134 MPH. Recording your speedometer readings at these RPMs, for say 1st through 3rd gear, would provide a highly accurate comparison. The tach signal is completely CAS based, so it has no relationship with the speedometer, speedo cable, speedo gear or transmission.

If these tests reflect the same discrepacies that you had noted before, then a mechanical possiblity would be with the speedometer itself. At the transmission, it is a set of gears that shouldn't be able to change. But on the speedometer end, the cable plugs in with a sort of pin. I don't think there's even a tab on it, so it's just friction that keeps it in place and rotating the speedometer at the correct rate.

 

I've got GPS that has proven itself accurate to within 1mph without fail.