I have a 2017 Tesla that I bought earlier this year with MCU1 and Enhanced Autopilot. I was going to get the MCU upgraded to MCU2 but saw that the price to upgrade with AP 2.0 is $2,000, as opposed to 1500 if you already have AP3.0. So as I was thinking of getting FSD I was thinking of getting FSD first for 5k (which includes upgrade to AP 3.0) and then spend 1.5k for the MCU Upgrade. Still on the fence though, alternatively I could just spend 2k for MCU upgrade and then use the monthly $100 Autopilot subscription.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Do not repeat the mistake I and many other MCU1 car owners have done. You will not get what you pay for.
Today, you will not get any real value for FSD purchase over what you already have with EAP. When/If Tesla actually ships* navigate on city streets, you might consider subscription to test if it is really worth it. (* as in you pay money and they give you the feature you paid for, no waiting, no games)
If you are committed to purchasing FSD and MCU2, I would wait a bit to see if Tesla will be forced to give MCU2 without extra cost to FSD buyers based on their promise of "we'll give you any HW needed for FSD" is not kept today.
One more thing: Purchasing MCU2 now that MCU3 is already shipping is like purchasing 5 years old computer for its original full price. If they start selling MCU3 upgrades (ands stop developing for MCU2), there might be regrets...
Interesting feedback for sure that I will consider.
Hmm is MCU3 shipping though? Is there a link for that? I know that the new model S / X are going with ICE (Tesla Model S, Model X Refreshes Replace MCU With ICE Used By Model 3, Y) instead of MCU2. However as that is landscape versus vertical, I wasn't sure that they were going to offer that upgrade back to pre-2021 model S vehicles. It would be great if there was some information on that to help me make an informed decision.
MCU3 is shipping on 2021 Model S'. However, it's a completely different screen and dashboard. Even Model 3/Y doesn't use MCU3. The idea that they will ever support MCU3 as a retrofit to a 2012-2021 Model S/X is really unlikely. The MCU2 retrofit because by design, MCU2 had to fit into the exact same space/interior as MCU1. MCU3 was never designed with that in mind.
GlmnAlyAirCar
Active Member
It's pretty safe to assume you will never get _early access_ FSD Beta as that is a test program that operates completely at Tesla's discretion. However, yes you are owed FSD when it is finally released to the public (with or without the Beta tag). At that point they will have to come up with some remedy.
What is the MCU2 Hardware actually based on?
from everything I was able to gather the MCU2 is based on the Intel Gordon Peak automotive platform. You can read on several sites that it uses an Intel E3800 cpu (which is not very specific because this is a series and not a cpu) since the Gordon Peak datasheet mentions this CPU. You can read on other sites that it uses an intel E8000 cpu, but I think this is complete horseshit or a typo. I found [1] and [2] as evidence of what is actually used: an Atom A3950, which seems to be a variant of the Atom E3950 (the A3950 cannot be found in Intels "ark" product database). While it might be possible that Tesla has used different CPUs in the MCU2 over the years, I don't think so since the A3950 was introduced by Intel in Q4/2016 (see intel ark link) and is the right one for the use conditions, including "Automotive" and "Industrial Extended Temp". I found a datasheet addendum which lists specifications for the E3950 and the A3950 and the A3950 seems to have a slightly lower TDP of 9.5W (A3950) instead of the 12W TDP (E3950).
How old is this MCU2 cpu?
the A3950/E3950 is a 4-core 1.6GHz base (2.0GHz boost) Intels Apollo Lake series CPU based on the Goldmont architecture which was unveiled April 2016 and supposedly released Q3/2016. Its siblings are commonly used in NAS-Servers, and the cpu architecture was reasonably powerful for an Atom based cpu, but it's not really a high performance CPU (like all cpus derived from atom cores). The refreshed "Goldmont Plus" architecture which debuted a little over a year later in Q4/2017 was a lot more powerful at the same power usage (see here for an appropriate comparison) but never made it into cars. With an Intel introduction date of Q3/2016 the A3950 CPU in the MCU2 it is close to 4 years old and it was never really powerful to begin with. That it still performs pretty well is an indication of how crappy the CPU in the MCU1 nVidia design was
In other words, MCU2 is quite outdated.
I would not be surprised if MCU3 in the latest Model S would have 5 X the TDP of MCU2 and thus being unsuitable as an retrofit in old cars. That said, there is no reason why all new cars should not switch over to MCU3. Having less displays wont be a problem.
And MCU2 being super old, maybe it would be more cost effective to replace A3950 with something more modern while keeling the same power consumption. They might call it MCU2.1 and get couple of more years out of it. Or even be able to sell an upgrade to MCU2 folks.
For practical purposes, MCU2 works fine for the foreseeable future. What I am worried of is that when software team can increasingly design for MCU3 (or MCU2.1), they'll become less likely to optimize or even ship new features for MCU2. This already happened with MCU1.
Why is everyone so sure you are "owed" FSD when it's in beta?
They said you had all HW for FSD. They could just say they are working on it, and FSD on MCU1 will be available in 2 years when FSD moves out of beta.
The fact that we're all oh, this is "early release" so it's OK that they limit it, but BETA.... well, everyone is OWED BETA???
Good luck. History here says Tesla will have no issue telling you that you have to pay for HW, or just wait and wait and wait for a future SW update. Just like they did on MCU1 when it would hang and reboot. "Just wait for a SW fix that has no current planed date of release, but no, you cannot have MCU2 even though that fixes the issue,"
If we would have "paid" for FSD, it would fair to speculate if we are "owed" Tesla delivering it.
But we actually paid for FSD (no "" around it) with actual money.
MCU1 was in about 150k cars. MCU2 is already in 1M Model 3/Y cars. MCU2 is going to have to be supported for a while.
I'm playing devlis advocate here.
You paid for FSD, not FSD beta.
Tesla is charging $1K to upgrade cars "with all HW needed for FSD" in order to subscribe to FSD. Historically, Tesla doesn't care what you paid for, and they are well within their history to not give betas out just because you paid for it. You paid for FSD, not FSD beta, and FSD isn't anywhere close.
I really doubt they'll give you MCU2 without a fight. Not saying you shouldn't take on that fight, you should.
GlmnAlyAirCar
Active Member
Because anything that Tesla ever released as "beta" has never gotten out of beta. AP1 is still in Beta. Summon from 2015 is still beta. For all practical purposes, beta is the general release.
I think you're confusing the MCU with the screen. The MCU is really just a small circuit board attached to the back of the display. I doubt the board is going to change size by enough to matter, given that the push has always been to make boards smaller, not larger, for cost reasons.
And there's no reason that the MCU shouldn't be able to drive a different size screen. LVDS is LVDS. It just would need to be configured to lay out the contents of the screen differently for the vertical orientation.
It's not outdated so much as it's a toy. The Atom CPUs are designed for low-end embedded stuff. They've almost invariably gotten stomped into the ground by ARM offerings even from the same time period. *And* it was a two-year old chip when they started shipping it. *And* that was three years ago.
Imagine using something with the total CPU power of an iPhone 6s, and the single-core performance of an iPhone 5s, to do everything an iPhone does while simultaneously dumping compressed video to a flash drive and various other car management tasks.
Quite possible, given that they are reportedly using a standalone AMD GPU (why!?!). But they could easily design a version without that monster GPU that would be more than good enough without exceeding the whopping 12W TDP of the E3950. For comparison, the current MacBook Air has an integrated CPU/CPU package whose TDP is only 10W.
That's why I'm not willing to spend thousands of dollars on MCU2. By the time they made it available, it was already an out-of-date design.
Looks like, if Tesla would be the one deciding if they deliver the purchase that was paid for, they would never deliver. Delivery costs money when promises were done hastily without a real plan. And now fixing things might become a distraction to their laser focus on growth.
Now that Tesla is not proactively fixing the problem they created, their policy will instead be decided by the cost of lawsuits, arbitration, customer support. Those costs push the decision in the organization to someone practical who just chooses the least expensive path. Thus it is important that every FSD buyer actually demands delivery of what they paid for.
It would be great if Tesla would be more thoughtful on preserving their brand image and just do the right thing. That should come from Elon directly. I think he actually would agree, but just does not have the bandwidth. Their brand image already went from "great car, great service" to "great car, horrible service" due to being run over by increased demand. Now they might end up losing the trust of a large share of their early adopters who increasingly consider Tesla is being borderline fraudulent or at least just not caring about customer. Personally, I think that they still have the best product out there and the demand generated by that blindsights them from looking at their brand with long term in mind.
What should they do? Hire and empower someone very senior from Apple to own the brand trust and customer experience.
You can say that, but Tesla replaces MCU1 screens when they do MCU2, and non of the SW for MCU3 is written for the landscape orientation of the MCU1/2 screen. This is as much about the software supporting the screen as the screen itself.
They changed so much in the MCU3 cars beyond the MCU, that they can't just drop them in. For MCU2, they already had to design it for exactly the same car to keep selling that car, so retrofits were easy.
Who in Tesla is going to go design HW, write SW, and create official retrofit support to shove MCU3 in the ~150k MCU1 cars, and then support that as a new configuration forever? It's not profitable, it won't happen. It only happened on MCU2 because it took almost no work
Durwood101
Member
Not to mention almost every Tesla on the road has MCU2........the days of Tesla upgrading that component on the vehicle are over. I doubt that would even happen at this point.......I was glad to have MCU2 upgraded when I did, at least I got a few more years out of my 2017
dbldwn02
Active Member
Agreed.
This is why there are and should be Alphas
List of things still in Beta
AP1
AP2 (Autosteer)
NoA
Traffic Light and Stop Sign Control
Windshield Wipers
Summon
Smart Summon
Trip Planner
And of course, FSD (Just as much Beta as the previous 9 items)
Things that aren't in Beta but probably should be:
Basic Navigation
Auto high beams
Dash Cam
Music player
Most of the time, they don't swap the screen. They charge an extra $500 if you want the screen replaced. Otherwise, they just swap the board.
Also, why do you think they have completely separate software for MCU3? They're running MCU2-equivalent functionality in the earlier ICE modules on Model 3, and roughly the same hardware supports portrait orientation on Model S and X. I can't imagine that they would suddenly split their entire code base and have to maintain separate code for Plaid cars just so they can drop support for portrait mode on MCU3. That would make way, way more work for them.
Arguably, there is no MCU3; it's really ICE version... 2? That module that contains both an MCU-like board and a self-driving board in the same package.
Either way, though, it's still just a single-board computer with a bunch of automotive Ethernet ports and CAN bus ports and an LVDS cable running up to the screen. As long as they have enough of the right connectors, the hardware can be super-generic. Sure, that means they probably won't have a port for the RADAR hardware, but that's no great loss. Everything else in the older vehicles should also be present in the newer vehicles.
And for the most part, as I understand it, Tesla's software is self-configuring, with a set of software features that are either enabled or disabled based on a configuration file, and then the software loads a given module or doesn't, and talks to a particular module over whichever CAN bus or automotive Ethernet bus it discovers that the module is attached to. And that software is written in high-level languages, so it should generally be a straight recompile for a new CPU.
So assuming they use a shared build system with a common source root for both platforms (and that seems likely unless they're insane), putting MCU3 in MCU1/2 cars should literally be a matter of recompiling for the right CPU architecture, configuring it for the set of specific modules that a given car should have in it, and installing it in the car, assuming they didn't change connectors, in which case they'd also have to build a trivial wiring harness to adapt from one style of connectors to the other.
Bear in mind that Tesla had to build a custom wiring harness for upgrading early MCU1 cars to MCU2, and that involves a much, much smaller number of vehicles than we're talking about here. So that likely isn't a showstopper. It's half a day for one engineer.
The only challenge is dealing with the screen orientation, and unless they drop support for MCU1 and MCU2 entirely, they're going to have to maintain that functionality for a long time to come. It almost certainly wouldn't be a huge amount of extra work to handle that in newer UI modules.
Speaking from a software engineering perspective, this shouldn't be a huge amount of work. I mean, it's not an engineer day, but I doubt it's an engineer year, either. Spread across 150k cars, the engineering isn't going to me more than a dollar or two per car. It's noise compared with the manufacturing cost of the hardware itself, even at such a small scale.
Similar threads
