The I-90 Surge by Out of Spec Studios

[BS8899]

With respect to the Lucid in the I-90 Surge, are we focusing on

The whole point of discussion here is driving past 0% multiple times.

• I think we all agree that Team Lucid made some bad decisions. The out-of-charge/tow incident occurred when ~90% of the race was already covered. That said, putting the running out of charge/towing incident aside, I am wondering if there are some significant lessons RE: Lucid's charging/efficiency vis-a-vis:
  • before the Lucid ran out of charge and got towed, the Taycan was already significant ahead and would have won the race amongst the EV entries regardless, correct?
    • Furthermore, the Lucid wasn't exactly in second place either. It was in a cluster of 3-4 EVs behind the Taycan, and was that way for a long time before the tow incident. In other words, the Lucid wasn't a run-away 2nd place finisher right behind the Taycan.
    • Yet, the specs suggest the Taycan has significantly less range than the Lucid and lower peak charging intake, correct? Does that suggest that on long road trips with need for multiple charging stops, a flatter, sustained charge curve is more important than a short high peak charge rate?
    • all of the other EVs (Teslas, Hyundai, etc.) in the Lucid cluster have significant lower spec'ed efficiency and lower peak charge rates than the Lucid
    • Does it call into question:
      • is the Lucid's efficiency claims, when mapped to real-life freeway speed driving (avg of 66mph on the Taycan, including charging. Probably avg over 75mph-80mph while cruising), is no better than the other EVs?

 

[JB67]

@Shane_SLC thank you for doing that race. It has been interesting watching the progress of your team. I think you all get a bunch of credit being able to even walk after getting so little sleep!

In regards to your comment about having a podcast about the trip and topics to discuss:

1) Why does everyone think the BMS showed 0% previously and there being the “large-ish” buffer below that (ex. OoS range test and earlier in the race).
FYI after I saw you guys successfully use the buffer in the last video I thought for sure it was something that was always there. I now see why the “Don’t trust the BMS” is a good slogan!

2) Speculation as to why the car was not doing the auto lane changes (if that is the right term?) towards the end of the race. If that was prevalent throughout the drive, or just after lots of driving.

3) Availability of amenities at chargers (particularly bathrooms).

I know there was a question from Kyle C. about what is desired to be seen from OoS next year, and after watching all the race, and seeing how many comments were made about expected arrival charge vs. actual, a series where the vehicle drove the speed limit snd compared predicted state of charge vs. updated arrival SoS actual arrival would be very informative to see what system you could rely on the most. Probably not practical with the limited manpower OoS actually has though.

 

[Shane_SLC]

With respect to the Lucid in the I-90 Surge, are we focusing on

• I think we all agree that Team Lucid made some bad decisions. The out-of-charge/tow incident occurred when ~90% of the race was already covered. That said, putting the running out of charge/towing incident aside, I am wondering if there are some significant lessons RE: Lucid's charging/efficiency vis-a-vis:
  • before the Lucid ran out of charge and got towed, the Taycan was already significant ahead and would have won the race amongst the EV entries regardless, correct?
    • Furthermore, the Lucid wasn't exactly in second place either. It was in a cluster of 3-4 EVs behind the Taycan, and was that way for a long time before the tow incident. In other words, the Lucid wasn't a run-away 2nd place finisher right behind the Taycan.
    • Yet, the specs suggest the Taycan has significantly less range than the Lucid and lower peak charging intake, correct? Does that suggest that on long road trips with need for multiple charging stops, a flatter, sustained charge curve is more important than a short high peak charge rate?
    • all of the other EVs (Teslas, Hyundai, etc.) in the Lucid cluster have significant lower spec'ed efficiency and lower peak charge rates than the Lucid
    • Does it call into question:
      • is the Lucid's efficiency claims, when mapped to real-life freeway speed driving (avg of 66mph on the Taycan, including charging. Probably avg over 75mph-80mph while cruising), is no better than the other EVs?

I think there are certainly plenty of lessons to be learned for both the everyday novice EV driver to the hardcore nerds. Admittedly, a lot of results of the I-90 Surge don't have a lot of direct applications for most EV drivers, but there are certainly tips and tricks that can be useful.

• The Taycan had a pretty secure lead relatively early on into the race. Billings, Montana, (about 800 miles in) was the last time any of the other EV's saw them, IIRC. The team did a great job of optimizing charging stops throughout the trip and had very few issues with charging, to their credit.
• The Model 3 had a somewhat secure 2nd place for most of the trip, but towards the end things were very close. We were pushing really hard to stay as close to them as possible to possibly overtake them somewhere at the end or at least secure 3rd place.
• I'd say in this particular scenario (very long distance with no breaks) there is strong evidence to suggest that sustained high power charging wins out over range and efficiency. The Taycan was only 0.3 mi/kWh less efficient than the Lucid, and the maximum range had a noticeable effect only on the first leg of the trip when starting at 100% SoC. A scenario more true to how actual EV drivers travel (shorter distances, charging overnight) would show the Lucid in a significantly better light.
• I think what set the Model 3 apart was it's impressive efficiency and access to the Tesla charging network. The charging curve is far from class leading, but with the best charging options and efficiency out of the group (as well as great optimization from it's team) the weaker charger performance did not hold them back much. The other non-Tesla EVs in that group had strong and consistent charging curves as well as decent efficiency, which allowed them to keep us with us really well.
• I think that it's pretty well documented that Lucid's efficiency really suffers when travelling over 70mph. The same car we used can travel over 500 miles at 70 mph, but once we started travelling at 75-85mph our max range was more like 400 miles, at most. If you did a race where the speed limit was between 60-70mph the Lucid would no doubt do significantly better, but the other cars would as well, at least to some degree.

 

[Shane_SLC]

@Shane_SLC thank you for doing that race. It has been interesting watching the progress of your team. I think you all get a bunch of credit being able to even walk after getting so little sleep!

In regards to your comment about having a podcast about the trip and topics to discuss:

1) Why does everyone think the BMS showed 0% previously and there being the “large-ish” buffer below that (ex. OoS range test and earlier in the race).
FYI after I saw you guys successfully use the buffer in the last video I thought for sure it was something that was always there. I now see why the “Don’t trust the BMS” is a good slogan!

2) Speculation as to why the car was not doing the auto lane changes (if that is the right term?) towards the end of the race. If that was prevalent throughout the drive, or just after lots of driving.

3) Availability of amenities at chargers (particularly bathrooms).

I know there was a question from Kyle C. about what is desired to be seen from OoS next year, and after watching all the race, and seeing how many comments were made about expected arrival charge vs. actual, a series where the vehicle drove the speed limit snd compared predicted state of charge vs. updated arrival SoS actual arrival would be very informative to see what system you could rely on the most. Probably not practical with the limited manpower OoS actually has though.

Thank you for watching and following along! It was definitely an adventure I'll never forget. The Lucid deserves a lot of the credit for keeping us comfortable and refreshed throughout the trip.

1. OoS, at least, has shown multiple times that there is "normally" a usable buffer below 0% SoC, perhaps 1-4 kWh or up to 30 miles, depending on the specific scenario. I'm certainly of the opinion (especially now) that relying on that buffer is not something anyone should do routinely. However, we were in a race, and we were trying to eek out any advantage that we could. Obviously, in our case this was not the right move and ended up costing us dearly. It was a constant struggle for us trying to decide the proper amount of charging to do maintain speeds, as the built in navigation would make overly-optimistic calculations while charging, and we had multiple instances where we had a 10-15% SoC arrival go down to 0% very quickly.

2. I think this was just mostly user error. For auto lane change to work you have to hold the stalk down for half a second or so until the car acknowledges and begins the auto lane change. We did completely lose Highway Assist and Adaptive Cruise Control the first night through Montana and Wyoming, though.

3. I can't remember any particular charging stops that seemed devoid of any food or restroom options. The main issue was that most of the nearby locations were not open 24/7, so there were some pit stops at non-charging locations for food or restrooms by the other teams, and plenty of peeing behind bushes or dumpsters in the late night hours. Team Lucid didn't make a single extra stop other than to charge, thankfully.

I would love to see this kind of testing as well, but with so many different cars to test it'd be difficult. Plus you'd have to test each at different speeds to see how it might affect each one. They at least try to mention how the built-in navigation performs when they're doing range tests or races.

 

[Adnillien]

OoS, at least, has shown multiple times that there is "normally" a usable buffer below 0% SoC, perhaps 1-4 kWh or up to 30 miles, depending on the specific scenario. I'm certainly of the opinion (especially now) that relying on that buffer is not something anyone should do routinely.

The accuracy of the SOC indication accumulates errors when driving, charging, driving, charging... without stopping and letting the battery rest for multiple hours. The car determines SOC two ways. The most accurate is a temperature corrected voltage after the battery has been resting for several hours. Less accurate is counting electrons (Coulombs ) into and out of the battery. Coulomb counting has errors and those errors accumulate over multiple charging and driving cycles. The BMS will correct for these errors every time the battery rests for several hours. Without that rest there is no correction. Hence, near the end of a drive like yours, the indicated SOC will have more errors and you should give yourself just a bit more margin.

I do believe there is a 3 to 4 kWhr buffer below zero in the Lucid battery. I would not advocate trying to use it but I would have more confidence in being able to use it at the beginning of a trip rather than the end of a very long trip.

 

[tonybot]

The accuracy of the SOC indication accumulates errors when driving, charging, driving, charging... without stopping and letting the battery rest for multiple hours. The car determines SOC two ways. The most accurate is a temperature corrected voltage after the battery has been resting for several hours. Less accurate is counting electrons (Coulombs ) into and out of the battery. Coulomb counting has errors and those errors accumulate over multiple charging and driving cycles. The BMS will correct for these errors every time the battery rests for several hours. Without that rest there is no correction. Hence, near the end of a drive like yours, the indicated SOC will have more errors and you should give yourself just a bit more margin.

I do believe there is a 3 to 4 kWhr buffer below zero in the Lucid battery. I would not advocate trying to use it but I would have more confidence in being able to use it at the beginning of a trip rather than the end of a very long trip.

Very interesting and frankly I had no clue this is how it worked. I was thinking more like remaining KWh and real time efficiency based simple math. Really good to know!

Assuming other EVs experience the same challenge, do they have a better way to manage it through software etc or is this more or less a universal problem?

 

[Adnillien]

Assuming other EVs experience the same challenge, do they have a better way to manage it through software etc or is this more or less a universal problem?

All EVs have this problem but the Coulomb counting can be tuned to minimize the errors so some EVs are better than others.

 

[JB67]

The accuracy of the SOC indication accumulates errors when driving, charging, driving, charging... without stopping and letting the battery rest for multiple hours. The car determines SOC two ways. The most accurate is a temperature corrected voltage after the battery has been resting for several hours. Less accurate is counting electrons (Coulombs ) into and out of the battery. Coulomb counting has errors and those errors accumulate over multiple charging and driving cycles. The BMS will correct for these errors every time the battery rests for several hours. Without that rest there is no correction. Hence, near the end of a drive like yours, the indicated SOC will have more errors and you should give yourself just a bit more margin.

I do believe there is a 3 to 4 kWhr buffer below zero in the Lucid battery. I would not advocate trying to use it but I would have more confidence in being able to use it at the beginning of a trip rather than the end of a very long trip.

Ok, fair warning, now you have me curious @Adnillien as this explains the behavior that was seen with the F150 Lightning running out of charge earlier than expected in other OoS videos. Thanks for giving an explanation!

So in general vehicles need a “rest period” between frequent charges to calibrate the BMS. Do you happen to know a rough maximum error that the BMS can reach without a rest period? I completely understand how the vehicle could be designed with the assumption that there would be downtime as most people don’t drive constantly without sleeping!

You mentioned “temperature corrected voltage”: is there an “ideal temperature range”? (Ex. 32 degrees F to 1xx degrees F)

After that rest period, is there a nominal error?

And how do you know this? Do you engineer EVs?

I ask because I haven’t had an EV, but having fun learning about them (and all the “nuts and bolts” details) while I wait for one I can actually fit in that isn’t a sedan. My background is Physics and Aerospace Engineering thus why your explanation really peaked my interest!

 

[BS8899]

I think there are certainly plenty of lessons to be learned for both the everyday novice EV driver to the hardcore nerds. Admittedly, a lot of results of the I-90 Surge don't have a lot of direct applications for most EV drivers, but there are certainly tips and tricks that can be useful.

• The Taycan had a pretty secure lead relatively early on into the race. Billings, Montana, (about 800 miles in) was the last time any of the other EV's saw them, IIRC. The team did a great job of optimizing charging stops throughout the trip and had very few issues with charging, to their credit.
• The Model 3 had a somewhat secure 2nd place for most of the trip, but towards the end things were very close. We were pushing really hard to stay as close to them as possible to possibly overtake them somewhere at the end or at least secure 3rd place.
• I'd say in this particular scenario (very long distance with no breaks) there is strong evidence to suggest that sustained high power charging wins out over range and efficiency. The Taycan was only 0.3 mi/kWh less efficient than the Lucid, and the maximum range had a noticeable effect only on the first leg of the trip when starting at 100% SoC. A scenario more true to how actual EV drivers travel (shorter distances, charging overnight) would show the Lucid in a significantly better light.
• I think what set the Model 3 apart was it's impressive efficiency and access to the Tesla charging network. The charging curve is far from class leading, but with the best charging options and efficiency out of the group (as well as great optimization from it's team) the weaker charger performance did not hold them back much. The other non-Tesla EVs in that group had strong and consistent charging curves as well as decent efficiency, which allowed them to keep us with us really well.
• I think that it's pretty well documented that Lucid's efficiency really suffers when travelling over 70mph. The same car we used can travel over 500 miles at 70 mph, but once we started travelling at 75-85mph our max range was more like 400 miles, at most. If you did a race where the speed limit was between 60-70mph the Lucid would no doubt do significantly better, but the other cars would as well, at least to some degree.

Shane,

Thanks you for doing the I-90 Surge and thank you for sharing your insights.

We should not just focus on the Out-of-Charge and the Tow event in the I-90 Surge. As you captured herein, there is a lot of insights from this adventure that can we applied to better the Lucid platform.

I'd encourage you to reflect on your learning and share with this community. Learning from your experience and applying these learning to improve the Lucid platform is what we all wanted.

I hope Lucid take your deliberation seriously and relook at how they optimize their efficiency/range/speed /charging time matrix.

I routinely drive my Lucid AGT on (relatively) long trips (780 miles) between my two houses. I know that some of Lucid's efficiency/charge time claims are not realizable in real-life.

I hope Lucid also take a look at my other post RE: relooking at the charging curve optimization. In short, for road trips, a flatter, sustained (30m-40m) longer strong charging curve is (my opinion) much more beneficial than a short-mega burst of energy.

Again, thanks for your insight. Please share more to educate all of us.

 

[tonybot]

Shane,

Thanks you for doing the I-90 Surge and thank you for sharing your insights.

We should not just focus on the Out-of-Charge and the Tow event in the I-90 Surge. As you captured herein, there is a lot of insights from this adventure that can we applied to better the Lucid platform.

I'd encourage you to reflect on your learning and share with this community. Learning from your experience and applying these learning to improve the Lucid platform is what we all wanted.

I hope Lucid take your deliberation seriously and relook at how they optimize their efficiency/range/speed /charging time matrix.

I routinely drive my Lucid AGT on (relatively) long trips (780 miles) between my two houses. I know that some of Lucid's efficiency/charge time claims are not realizable in real-life.

I hope Lucid also take a look at my other post RE: relooking at the charging curve optimization. In short, for road trips, a flatter, sustained (30m-40m) longer strong charging curve is (my opinion) much more beneficial than a short-mega burst of energy.

Again, thanks for your insight. Please share more to educate all of us.

Totally agree! Thanks to Shane and team for this service to the community!! Do we know if Lucid engineering pay attention to such reviews and potentially factor in any lessons? It is possible that they are already aware of these insights but it not, it will be great incorporate data including telemetry and other data from this vehicle to validate and roadmap improvements. For the OGs here, is that something they tend to do based on past experience?

 

[illopp00]

@Shane_SLC thank you for doing that race. It has been interesting watching the progress of your team. I think you all get a bunch of credit being able to even walk after getting so little sleep!

In regards to your comment about having a podcast about the trip and topics to discuss:

1) Why does everyone think the BMS showed 0% previously and there being the “large-ish” buffer below that (ex. OoS range test and earlier in the race).
FYI after I saw you guys successfully use the buffer in the last video I thought for sure it was something that was always there. I now see why the “Don’t trust the BMS” is a good slogan!

2) Speculation as to why the car was not doing the auto lane changes (if that is the right term?) towards the end of the race. If that was prevalent throughout the drive, or just after lots of driving.

3) Availability of amenities at chargers (particularly bathrooms).

I know there was a question from Kyle C. about what is desired to be seen from OoS next year, and after watching all the race, and seeing how many comments were made about expected arrival charge vs. actual, a series where the vehicle drove the speed limit snd compared predicted state of charge vs. updated arrival SoS actual arrival would be very informative to see what system you could rely on the most. Probably not practical with the limited manpower OoS actually has though.

I wonder if #2 is just due to being in the 'middle of nowhere', any temporary glitch in the GPS and highway mapping disables HA. But overall my takeaways

1. Very entertaining series! But also quite boring in the sense that here are things that one might expect to happen that didn't:
   1. More than one team or instance of running entirely out of battery.
   2. Emergency situations from Low SOC AND chargers out of service/long waiting time.
   3. Flat tires, blowouts and any mechanical damages.
2. I know they are all new cars except the Old Model S, but there were absolutely no mechanical, electrical or software failures over the 3000 miles. Only minor glitches. All these cars are very very good at getting you from A to B.
3. To figure out the extra minutes, you have to be able to get the speed, efficiency right to basically eliminate an entire stop. All that faffing about coming off highway, parking, charger handshake, moving over chargers, plugging and unplugging. At highway speeds of 70mph+ you are covering 1.1+ miles per minute. So if each stop adds 5 minutes of non-charging, low speed or distance covering there is something to be said for staying and charging higher. At 100 Kw most reasonably, efficient cars are adding about 6.6 miles per minute of range, so basically you need 1 extra KW added to buffer past the extra stopping off the highway.

 

[Shane_SLC]

We just finished recording the podcast and it should hopefully be posted tomorrow. Unfortunately I was having internet issues for the first 30-40% so I wasn't able to contribute to the entire thing, but I think all the major discussion points were covered. I'll share the link here when it goes up and I'll follow up with anything I wasn't able to add afterwards. Thank you to everyone who has been following our adventure and contributing to the discussion!

 

[koflo]

I have already opted myself into a trip on I-90. If there are 2 others willing to give it a go, then we have a team. I already have an idea of the code to write to calculate maximal route efficiency, so the human part would be taken out of the equation almost entirely.

I don't own a GT 25', only a GT-P with 21" wheels. I plan on getting the 19", but that is 8+ months out. Obtaining a DE or GT 25' with 19" wheels would be a requirement for a 1-1 comparison.

Videoing is also something I am unfamiliar with.

I have a DE-P but have 21s. Would consider doing this for science or maybe assembling another team to test my taycan turbo j1.1. Both are on 21s though.

 

[Adnillien]

So in general vehicles need a “rest period” between frequent charges to calibrate the BMS. Do you happen to know a rough maximum error that the BMS can reach without a rest period? I completely understand how the vehicle could be designed with the assumption that there would be downtime as most people don’t drive constantly without sleeping!

The errors accumulate with charge/drive cycles so the error will be larger after 3000 miles than after 500 miles. For most normal driving, 500 to 700 miles in a day, the errors are small and less than a couple of percent SOC. After 3000 miles, the errors will be greater but the BMS system is still reading the cell voltage under load so there can be some correction. I do not know how far it can get off but for normal driving, you can ignore it.

You mentioned “temperature corrected voltage”: is there an “ideal temperature range”? (Ex. 32 degrees F to 1xx degrees F)

NMC batteries have maximum capacity in the 80-90 degree range and maximum lifetime in the 60 to 70 degree range. Battery cell voltage is a function of both temperature and SOC, the SOC and temperature curves are known for the cell chemistry. It is a matter of knowing the temperature to determine SOC, not that there is an ideal for determining SOC. Using the Home Assist API, you can see max and min cell temperature. There is usually several degrees difference between min and max. I believe this is measuring a group of cells not individual cells so even the temperature measurement is not perfect.

After that rest period, is there a nominal error?

SOC is always an estimation so see previous comment.

And how do you know this? Do you engineer EVs?

I do not engineer EVs, I have worked on semiconductor BMS components in the past and have done a fair amount of research on NMC batteries. With your physics background you would enjoy digging into this so please do and share here when you find something new. I am far from an expert.

 

[Unmutual]

Just one other thought on strategy in general: just because you CAN go 10 mph over the speed limit per the rules, given the efficiency hit on the Lucid in the particular, SHOULD you go 10 over the speed limit? I often go a little under the speed limit if I'm trying to maximize range, and since for reasons other have stated, it's the charging sessions that really kill you in this sort of a race, wouldn't minimizing the number of charging sessions be important?

 

[illopp00]

Just one other thought on strategy in general: just because you CAN go 10 mph over the speed limit per the rules, given the efficiency hit on the Lucid in the particular, SHOULD you go 10 over the speed limit? I often go a little under the speed limit if I'm trying to maximize range, and since for reasons other have stated, it's the charging sessions that really kill you in this sort of a race, wouldn't minimizing the number of charging sessions be important?

Turns out only on the first leg with battery at 100% would it make sense to maximize range by driving a bit slower than max. I did this whole math based on my real world data from my Air Touring on 19in. Seems optimal then to do charge stops that are 5-70%, eat the extra charge time and have the buffer to drive faster.

 

[Shane_SLC]

Here's the link to our discussions from last night. My sincere apologies for the connection issues on my side. I'll be recording my own video and sharing it later this week to make up for my lack of input. Please let me know what your thoughts on the podcast video are and if there's anything else I can bring up in my personal video.

 

[ATP2023]

Turns out only on the first leg with battery at 100% would it make sense to maximize range by driving a bit slower than max. I did this whole math based on my real world data from my Air Touring on 19in. Seems optimal then to do charge stops that are 5-70%, eat the extra charge time and have the buffer to drive faster.View attachment 24661

Now your talking like a race director of a Le mas team.

 

[illopp00]

Here's the link to our discussions from last night. My sincere apologies for the connection issues on my side. I'll be recording my own video and sharing it later this week to make up for my lack of input. Please let me know what your thoughts on the podcast video are and if there's anything else I can bring up in my personal video.

Just watched all the details, still have to give the team kudos for staying with it and completing the race. After all that discussion now I think its the somewhat hot take $50K question... Since you didn't really need the full pack and the power of the GT, I wonder if the Pure RWD might have done better with the extra efficiency? It charges pretty much exactly like the Tesla Model 3, has just about the same size pack 84Kwh vs 79.7 kwh. You would have missed the massage seats, but the rest of Lucid Air goodness would be there...

 

[momo3605]

Just watched all the details, still have to give the team kudos for staying with it and completing the race. After all that discussion now I think its the somewhat hot take $50K question... Since you didn't really need the full pack and the power of the GT, I wonder if the Pure RWD might have done better with the extra efficiency? It charges pretty much exactly like the Tesla Model 3, has just about the same size pack 84Kwh vs 79.7 kwh. You would have missed the massage seats, but the rest of Lucid Air goodness would be there...

Pure RWD charges pretty slow. It would not have fared well