-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
I have one car of each (S with AP1 and 3 with AP 2.5) so I think I have a decent perspective here.
Honestly I think they are very similar (at this point at least). I feel AP2.5 holds the lines stronger than AP1 and brakes/accelerates more smoothly with traffic than AP1. Where AP1 wins is that it can see cars in adjacent lanes, so if a car starts to cross the line into my lane and merge in front of me, the AP1 car reacts faster and slows down as it sees it coming before it's fully in my lane.
Honestly I think they are very similar (at this point at least). I feel AP2.5 holds the lines stronger than AP1 and brakes/accelerates more smoothly with traffic than AP1. Where AP1 wins is that it can see cars in adjacent lanes, so if a car starts to cross the line into my lane and merge in front of me, the AP1 car reacts faster and slows down as it sees it coming before it's fully in my lane.
Isn't there something about AP 2 and up that won't drive or engage in slower than 30km/h traffic ? While AP1 does engage even in standstill.
I don't think it's in any way "not smooth". Except it could very well have it's own acceleration mode, like AP could drive in less than sport, or maybe chill mode, while the car were in ludicrous mode.
I don't think it's in any way "not smooth". Except it could very well have it's own acceleration mode, like AP could drive in less than sport, or maybe chill mode, while the car were in ludicrous mode.
The limitations are exactly the same. AP1 AND AP2 require you to either be at 18mph to engage, OR be following or stopped behind a lead car at any speed.
Neither will engage if you are the lead car and going below 18mph.
gjunky
Trifecta: Solar and both cars are EVs
With creep turned off, I can engage AP1 from standstill as long as it sees the lines in front of the car. Works at any speed. I turn it on all the time below 10mph when traffic starts to back up.
HW1 cars only have a narrow beam radar/ camera. Yes they see cars in adjoining lanes WHEN THEY COME INTO THE BEAM. If someone cuts in really close to your car, it doesn’t see them until they are part way into your lane. This can cause unexpected evasions by the car with startling results.
Experience on my side: I have a brand new S75D (delivered in december) so HW2.5.
And due to an issue with the battery pack (see my other post on 12V battery failure), I had a loaner that was a S90D with the new front design but still AP1.
I can say that AP1 is clearly much slower (not smother) in changing lane and accelerating/slowing down when changing speed. I find personaly the calibration of the AP1 for those manoeuvre, I must admit, more natural and close to the speed a human driver would perform those actions.
The AP2 is much more authorative and dynamic on those. I would like to see this tammed down a little bit to come more in line with the AP1.
For the rest, the AP1 is giving a lot of confidence in the way it works (very smooth) but I experiences several time some break down or super strange behaviour which shows that it sees clearly less than the AP2. It also cannot sustain the same type of curve it seems.
The AP2 seems to have more potential from what I have seen and limited experienced so far but except for this difference of behavior, I don't think AP1 is still any superior to AP2 in capabilities.
Oh, and one aside benefit of HW2.5 is the additionnal number of camera (or ultrasonic sensors?). I thought there were not activated but some additionnal may be because for instance, the AP1 systematically failed to propose me to park my automatically in some conditions (like when entering in my garage and initiating the movement moving forward (before having to go back in a perpendicular slot) whereas the HW2,5 systematically work.
So the auto parking definitely works better in a broader range of situation with HW2.5, no doubt.
And due to an issue with the battery pack (see my other post on 12V battery failure), I had a loaner that was a S90D with the new front design but still AP1.
I can say that AP1 is clearly much slower (not smother) in changing lane and accelerating/slowing down when changing speed. I find personaly the calibration of the AP1 for those manoeuvre, I must admit, more natural and close to the speed a human driver would perform those actions.
The AP2 is much more authorative and dynamic on those. I would like to see this tammed down a little bit to come more in line with the AP1.
For the rest, the AP1 is giving a lot of confidence in the way it works (very smooth) but I experiences several time some break down or super strange behaviour which shows that it sees clearly less than the AP2. It also cannot sustain the same type of curve it seems.
The AP2 seems to have more potential from what I have seen and limited experienced so far but except for this difference of behavior, I don't think AP1 is still any superior to AP2 in capabilities.
Oh, and one aside benefit of HW2.5 is the additionnal number of camera (or ultrasonic sensors?). I thought there were not activated but some additionnal may be because for instance, the AP1 systematically failed to propose me to park my automatically in some conditions (like when entering in my garage and initiating the movement moving forward (before having to go back in a perpendicular slot) whereas the HW2,5 systematically work.
So the auto parking definitely works better in a broader range of situation with HW2.5, no doubt.
Going along with what others said, I think the biggest differences between AP1 and AP2 at this time are, in terms of AP performance and not cosmetic differences:
(1) AP1 is smoother. Not necessarily safer or more correct, but smoother. That instills a feeling of confidence in the driver. AP2 is much more willing to make rapid changes. This lets it react to sharper curves or sudden lane shifts but also gives it more of an erratic feel when lane lines are irregular or misdetected.
(2) AP1 is far more willing to trust the lead car compared to AP2. Even in areas with good lane markers, it sometimes will believe a car in front more than it believes the lane lines. As long as you're following a well-behaved driver, it's hard to tell whether AP1 is doing well, or just mimicking what the driver in front did a couple seconds ago.
(3) AP2 operates much more reliably at high speeds (e.g. 80-90mph) while those speeds seem to push AP1 near its limits. On curving highways, speeding to those ranges leads to rather dissatisfactory AP1 performance but AP2 often still performs acceptably.
(4) AP2 ultrasonics are slightly better than AP1. I would not go as far to say double the range (like what Tesla claims). I would say it's more like equal to 1.5x range in most conditions. This leads to marginally better detections of cars beside you, though neither AP1 nor AP2 makes very heavy use of this information right now.
Overall with .50.3 it's starting to turn into a toss-up whether AP1 or AP2 is "better" overall. I think both are great and useful in the same kinds of conditions right now. AP1 does better in some circumstances (especially crossing city intersections and hitting bumps). AP2 does better in a lot of other circumstances.
(1) AP1 is smoother. Not necessarily safer or more correct, but smoother. That instills a feeling of confidence in the driver. AP2 is much more willing to make rapid changes. This lets it react to sharper curves or sudden lane shifts but also gives it more of an erratic feel when lane lines are irregular or misdetected.
(2) AP1 is far more willing to trust the lead car compared to AP2. Even in areas with good lane markers, it sometimes will believe a car in front more than it believes the lane lines. As long as you're following a well-behaved driver, it's hard to tell whether AP1 is doing well, or just mimicking what the driver in front did a couple seconds ago.
(3) AP2 operates much more reliably at high speeds (e.g. 80-90mph) while those speeds seem to push AP1 near its limits. On curving highways, speeding to those ranges leads to rather dissatisfactory AP1 performance but AP2 often still performs acceptably.
(4) AP2 ultrasonics are slightly better than AP1. I would not go as far to say double the range (like what Tesla claims). I would say it's more like equal to 1.5x range in most conditions. This leads to marginally better detections of cars beside you, though neither AP1 nor AP2 makes very heavy use of this information right now.
Overall with .50.3 it's starting to turn into a toss-up whether AP1 or AP2 is "better" overall. I think both are great and useful in the same kinds of conditions right now. AP1 does better in some circumstances (especially crossing city intersections and hitting bumps). AP2 does better in a lot of other circumstances.
Detecting a parking spot for auto parking is handled by the ultrasonic sensors. The AP 2 and 2.5 ultrasonic sensors have a longer range than the AP 1 sensors, so it makes sense that they would be able to detect parking places more accurately.
arcus
Active Member
Having a loaner P85D with AP 1.0 for a couple of days I can attest to this as well. It definitely feels more refined (not even a hint of hesitation or bouncing from left to right). AP 2.0 still feels like an untamed stallion at times.
It happened to me on the first drive with AP 1.0 engaged - it decided to start following a van that was veering off the lane instead of lane markers.
Agree on the rest of @chillaban's assessments as well.
headcase
Member
Yeah, AP1 still feels more refined and confident than AP 2 / 2.5. Elon's boys still have some work to do.
I own an AP1 and had an AP2 loaner with 50.x software for a week. If you aren’t comparing the same stretch of road one can probably not tell the difference between the two. I can’t say one is better than the other, but they are definitely comparable at this point.
While AP2/2.5 may be catching up or finally equal to AP1, it's still not at a point where it's making most AP1 car owners want to upgrade. There's also a general lack of confidence that Tesla will be able to pull off FSD with the current AP 2.5 hardware. That's also being factored into upgrade decisions.
Basically, I feel like you're going to be fine with either AP1 or AP2/2.5 at the moment. AP2/2.5 will continue to improve but I don't think we're going to see any groundbreaking improvements until a future iteration of AP.
Basically, I feel like you're going to be fine with either AP1 or AP2/2.5 at the moment. AP2/2.5 will continue to improve but I don't think we're going to see any groundbreaking improvements until a future iteration of AP.
headcase
Member
From what I understand, AP1 reads speed limit signs whereas AP2.X uses the map data for speed limit information. I personally think signs are more reliable than the map data which seems to lack accuracy for those types of meta-data.
I also think that AP1 is likely at end of life, regardless of Elon's promise of upcoming on-ramp to off-ramp navigation for the hardware. I doubt we will see many, if any, upgrades for AP1 cars beyond what we have today.
I also think that AP1 is likely at end of life, regardless of Elon's promise of upcoming on-ramp to off-ramp navigation for the hardware. I doubt we will see many, if any, upgrades for AP1 cars beyond what we have today.
f-stop
Active Member
I own a 2015 70D with AP1. This week I had a new 100D service loaner for a day, on 2018.2 0eddd23. While I didn't do any freeway driving, I tried AP with the loaner a few places around town - on short stretch of divided road (80kph limit) and also in traffic on regular major street. To me AP2 especially at lower speeds or when first starting seemed to "wiggle" left and right a lot, I don't notice the same on AP1. Didn't really do enough AP driving with the loaner but that was my initial impression.
