P85 modded to be a P85 w/Ludicrous

[Peteybabes]

Also, can't one just hold the steering wheel every few minutes? While one waits for an actual self-driving car?
/
Now, back to regular programming.

all i have to do is a quick tug every 30-90 seconds when the message pops up, seriously not a huge deal...lol

 

[Carl Reese]

It wasn't a big deal when the AP worked for 30-40mins without human intervention. Then the software was "updated" to make you touch the wheel every couple minutes. Then it was updated to remove the audible tone.
Personally I find it a a pain in the @ss when you get locked out of AP. I'm looking at the road not the dash cluster, when I'm driving. It's very easy to miss the glowing light.
When my team set our second EV record, I had early prototype to extend the time AP would stay engaged. I've got no problem proving this as we recorded the entire trip. Nobody ever asked how we accomplished 96.1% AP engagement back when AP1 was 4 days old.
Obviously Drivers Are Already Abusing Tesla’s Autopilot

 

[HankLloydRight]

Nobody ever asked how we accomplished 96.1% AP engagement back when AP1 was 4 days old.

I think you just answered it yourself...

It wasn't a big deal when the AP worked for 30-40mins without human intervention.

Also, by the time you get an actual patent (if you get one), the state of AP software in three or four years is going to be so far ahead of where it is now. The nags and human detection will be better, whatever 'full self driving' is will be out, and the market for your thing will be pretty small.

And the liability issues are very real. Your device intentionally bypasses built-in safety features in the car. When someone kills themselves and others when using it, I surely wouldn't want to be in your shoes. Good luck.

 

[TIppy]

Jason Hughes on Twitter

Unfortunately I'm an idiot and I left my laptop bag on the dyno with my GoPros and CAN data captures. Won't be able to get it until tomorrow or Friday.

553 HP @ wheels. 480 ft/lbs torque. 456kW @ pack (so 611 battery HP).

Runs were pretty consistent, aside from my start and stops at different speeds.

Your car is much more efficient than mine is. I get about 80 percent conversion of electrical power to motor shaft power. 456 kw is 611 electrical hp. 80 percent is 489 motor shaft hp. In addition on some tests I've done, I also loose about another 6 to 10 percent to the wheels. So 489 * 0.94 is 460 hp at the wheels, or even as little as 440 hp. How is it possible that you get 553 hp to the wheels? You're converting 90 percent of battery electrical power to hp at the wheels.

Is it possible this dyno is at altitude, and they are compensating for ice performance?

 

[TIppy]

A version 1 90 kwh battery didn't have any more power than an 85kwh battery. Some of them didn't even have any more capacity, not for long anyway. You want a version 3 battery, part no. beginning with 1088xxx, to have less droop and more power.

 

[TIppy]

From an earlier post in this thread, kwhftw said "p85+ put 395 to the wheels on the dyno."

The P85 was quoted as having 417 motor shaft horsepower. That's reasonable for the 1150 amp limit. So the efficiency from motor shaft horsepower to wheels is 395 / 417 = 0.947 which is close to what I've gotten on various tests.

 

[demundus]

@TIppy he's using a ludicrous pack fuse.

 

[TIppy]

@TIppy he's using a ludicrous pack fuse.

Yeah, so am I. I have a p90dl with a version 3 battery. The fuse has nothing to do with the efficiency of converting battery electrical horsepower to horsepower at the wheel. There is just no way he is getting 550 hp at the wheels.

 

[pnaecker]

You're missing the point entirely, being fixated on a single mechanism instead of the big picture. Show me a commercial example of a heatpipe being used as an intermediary heat spreader between a heat source and liquid cooling.

Hewlett Packard sells very expensive HPC computers called Apollo 8000's that do just this. (It is hard to find technical docs or photos online (it's a modestly proprietary technology), but I've got some of these, and they work just fine. It allows them to put 70kW+ in a single IT rack, which is a multiple of what is possible with air cooling. HP makes no secret of how they work - they show the open server trays at trade shows.)

The servers have a heat pipe (like in home overclocked coolers, but larger) that spreads the data from the chip to a "water wall". The other end of the heat pipe makes friction contact to the cold plate that is cooled by internal water traces inside the cold plate, thus greatly increasing the contact surface and heat transfer duration to the water.

I think the counterpoint is still valid, of course: you gotta get the heat off the IGBT, and therein lies a likely limit. But if it turns out that the first contact thermal resistance (silicon to heat pipe) is not the primary constraint, you can get the heat from the chip to a much, much larger water cooled surface quite readily with boiling heat transfer inside of a heat pipe.

 

[wk057]

There is just no way he is getting 550 hp at the wheels.

You're welcome to stop over for a demonstration ride.

Dyno wasn't compensating for anything, and values match pretty closely with data from CAN logs. You can also validate the numbers a bit from the graphs. For example, consistently the car was at about 320 ftlbs of torque at ~9000 RPM, so plug that into HP=(TorqueFtLbs*RPM)/5252 and you get 548 HP. And that's just an estimate eyeballing the graph.

The car has a 90 kWh pack with the Ludicrous upgrade (pyro fuse) which allows it to output up to 1520A. I'm pushing just over 1400A to the rear motor currently. P85 normally does 1050-1100A max. P85D does 1300A, PxxD does 1500A. Input power vs output power had me in the upper-80s% efficient, off the top of my head (not actually looking at the data right this second).

Can see the slow-refresh diag screen peak at 1417A in this one video I posted of a dyno pull.

I also recently street raced a P85D non-Ludicrous with this car from a stop to about 90 MPH. Was spur of the moment, so no logging equipment or max battery power or anything. P85D's traction helped at launch, but I was nearly two lengths ahead of him by the time we let off. Those extra 100 amps, plus not dragging around a second motor, make a huge difference after the initial launch. Pretty sure if this was taken out to a full 1/4 mile I'd have beaten the P85D by several tenths or more.

I've made a few more tweaks to the car since my last update here, also. I bumped the max current up a few more notches to hold high torque for a little longer as well as bumped max torque a little. I can get close to 100kW regen now with some pretty high regen torque when the pack can handle it... which has improved my overall efficiency significantly since I almost never use the brakes in this car anymore except to hold at a standstill. Recently did a 242 mile stretch on a single charge at ~288 Wh/mi average with an average speed of 72 MPH (max about 85 MPH).

Unfortunately I've damaged the rear halfshafts/CV joints when tinkering with the higher torque output, so, going to have to replace those... hopefully with something stronger.

 

[TIppy]

You're welcome to stop over for a demonstration ride.

Dyno wasn't compensating for anything, and values match pretty closely with data from CAN logs. You can also validate the numbers a bit from the graphs. For example, consistently the car was at about 320 ftlbs of torque at ~9000 RPM, so plug that into HP=(TorqueFtLbs*RPM)/5252 and you get 548 HP. And that's just an estimate eyeballing the graph.

The car has a 90 kWh pack with the Ludicrous upgrade (pyro fuse) which allows it to output up to 1520A. I'm pushing just over 1400A to the rear motor currently. P85 normally does 1050-1100A max. P85D does 1300A, PxxD does 1500A. Input power vs output power had me in the upper-80s% efficient, off the top of my head (not actually looking at the data right this second).

Can see the slow-refresh diag screen peak at 1417A in this one video I posted of a dyno pull.


I also recently street raced a P85D non-Ludicrous with this car from a stop to about 90 MPH. Was spur of the moment, so no logging equipment or max battery power or anything. P85D's traction helped at launch, but I was nearly two lengths ahead of him by the time we let off. Those extra 100 amps, plus not dragging around a second motor, make a huge difference after the initial launch. Pretty sure if this was taken out to a full 1/4 mile I'd have beaten the P85D by several tenths or more.

I've made a few more tweaks to the car since my last update here, also. I bumped the max current up a few more notches to hold high torque for a little longer as well as bumped max torque a little. I can get close to 100kW regen now with some pretty high regen torque when the pack can handle it... which has improved my overall efficiency significantly since I almost never use the brakes in this car anymore except to hold at a standstill. Recently did a 242 mile stretch on a single charge at ~288 Wh/mi average with an average speed of 72 MPH (max about 85 MPH).

Unfortunately I've damaged the rear halfshafts/CV joints when tinkering with the higher torque output, so, going to have to replace those... hopefully with something stronger.

1400 amps from a P90DLv3 pack is about 458 kW at 100% SOC. A P100DL puts out about 565kw and the motor shaft horsepower is 610HP.
That's 565 * 1.341 * 0.805 = 609.9HP. Your 458kW * 1.341 * 0.805 = 494HP. That's at the motor shaft. There is about another 5 or 6 percent to get to the wheel, so 494 * 0.95 = 469HP at the wheels. The service manual says that the rear P motor is good for 375 kW(502HP) at the motor shaft.

Using canbus, when I look at the power from the battery and calculate the mechanical power from the reported torque and rpm, I get about 82 percent efficiency. If you take into account that the rpm reported is the frequency of the stator field and that the rotor is rotating about 1 to 2 percent slower, that gives about 80.5 percent efficiency. Again this is at the motor shaft. So I just don't see how you can get 550HP at the wheels.

The dyno's are usually setup to compensate for local atmospheric condition, so that the results are for the stardard atmosphere of sea level and 59F. But of course, these things don't affect the power of electric motors.

 

[zwede]

The dyno's are usually setup to compensate for local atmospheric condition, so that the results are for the stardard atmosphere of sea level and 59F. But of course, these things don't affect the power of electric motors.

I don't have the energy to go through 16 pages of posts, but I think Jason dynoed on a DynoJet? The DynoJet SW will let you graph either SAE corrected data or uncorrected. As long as what was posted was the uncorrected graph there was no temp/humidity/pressure correction applied.

 

[TIppy]

I don't have the energy to go through 16 pages of posts, but I think Jason dynoed on a DynoJet? The DynoJet SW will let you graph either SAE corrected data or uncorrected. As long as what was posted was the uncorrected graph there was no temp/humidity/pressure correction applied.

It seems like dynos aren't particularly accurate. They are good for tuning, because the results are reproducible. I've read that mustangs are more accurate, but they read about 12 percent lower than dynojets.

 

[int32_t]

It seems like dynos aren't particularly accurate. They are good for tuning, because the results are reproducible. I've read that mustangs are more accurate, but they read about 12 percent lower than dynojets.

If I'm testing something, obviously I want the one that reads higher!

... which has improved my overall efficiency significantly since I almost never use the brakes in this car anymore except to hold at a standstill.

I'm curious, does upping the max torque limit reduce the lowest speed at which regen works effectively? Is that what you're saying here?

 

[zwede]

It seems like dynos aren't particularly accurate. They are good for tuning, because the results are reproducible. I've read that mustangs are more accurate, but they read about 12 percent lower than dynojets.

That's a misconception. DynoJets are perfectly accurate, as are Mustang dynos. They measure differently though which is why they don't come up with the same numbers.

DynoJet is an inertia dyno. The car accelerates a large drum filled with water. The dyno measures how fast the drum accelerates (how fast kinetic energy is added).

Mustang puts a load on the car and measures how much it needs to brake to keep the car at a certain rpm, or a slowly increasing rpm.

Both dynos are very repeatable. An ICE that is allowed to quickly increase rpm will make more power which is why the DynoJet sees higher numbers.

I don't know if an EV will measure differently on a Mustang and DynoJet.

 

[TIppy]

That's a misconception. DynoJets are perfectly accurate, as are Mustang dynos. They measure differently though which is why they don't come up with the same numbers.

DynoJet is an inertia dyno. The car accelerates a large drum filled with water. The dyno measures how fast the drum accelerates (how fast kinetic energy is added).

Mustang puts a load on the car and measures how much it needs to brake to keep the car at a certain rpm, or a slowly increasing rpm.

Both dynos are very repeatable. An ICE that is allowed to quickly increase rpm will make more power which is why the DynoJet sees higher numbers.

I don't know if an EV will measure differently on a Mustang and DynoJet.

When something is repeatable it is precise, not necessarily accurate. I agree they are precise which is why I said they are good for tuning. They let you see the difference in performance. If mustangs and dynojets are repeatable, but give you different numbers they are not accurate.
If you are tuning it's best to use the same shop, because even the same brand of dyno can vary from shop to shop.

If the dynojet lets you spin up your wheels faster than you could accelerate in the real world, then it doesn't accurately provide the torque and horsepower you'll have on the road or at the track. Supposedly, the mustang can simulate any road conditions by varying the eddy current load as the car accelerates. It can even pause the pull at any load and rpm.

 

[zwede]

If you are tuning it's best to use the same shop, because even the same brand of dyno can vary from shop to shop.

Years ago I was a member of a car club where this was heavily debated. Finally one member got tired of the endless discussions and dynoed his car at 3 different DynoJet shops the same day. All 3 showed the same numbers within 2-3 hp.

Of course it is tempting for a dyno operator to fudge the numbers and get a reputation as the shop to go to for the "best numbers", but at least in my area the dyno operators were honest.

Then there's WesTech over in California. That's where the car mags take their stuff to get "magazine horsepower". WesTech was caught with "miscalibrated" dynos reading 20% high. They claimed it was a mistake... It certainly explained all those magazine builds where they got 500 hp out of an old Chevy 350 out of the junkyard that they threw speed parts on.

 

[wk057]

I'm curious, does upping the max torque limit reduce the lowest speed at which regen works effectively? Is that what you're saying here?

Might be reading your question wrong, but let me clarify overall.

The torque limit normally is just over 100Nm of regen torque. I've been able to up this to about 200Nm. I've also popped the power cap on regen, letting it put in as much power as the BMS permits (basically supercharger power levels). At higher speeds the torque increase doesn't help much, but the power cap lift does significantly, since on a highway off-ramp I can see triple digit kW regen. The added torque means it can maintain higher power for longer and at slower speeds. I've actually had to tweak it over time because regen could lock the rear wheels at low speeds with high torque. Pretty much can get down to 4 MPH with just regen, and pretty quickly.

For example, if I let completely off of the throttle with regen set to max while going about 20 MPH, it feels like I mashed the brakes in the car pretty good, and I get flung into my seat belt, which locks in place during the deceleration event.

I'm actually going to be undoing quite a few of my more complex modifications temporarily, aside from the power increase modifications, because I've found a minor design flaw in my custom control hardware that I need to fix. Mainly for safety reasons, since the failure mode could cause unpredictable behavior.

 

[Peteybabes]

Then there's WesTech over in California. That's where the car mags take their stuff to get "magazine horsepower". WesTech was caught with "miscalibrated" dynos reading 20% high. They claimed it was a mistake... It certainly explained all those magazine builds where they got 500 hp out of an old Chevy 350 out of the junkyard that they threw speed parts on.

still happening...check out engine masters on motortrendondemand.com ...

 

[Peteybabes]

Might be reading your question wrong, but let me clarify overall.

The torque limit normally is just over 100Nm of regen torque. I've been able to up this to about 200Nm. I've also popped the power cap on regen, letting it put in as much power as the BMS permits (basically supercharger power levels). At higher speeds the torque increase doesn't help much, but the power cap lift does significantly, since on a highway off-ramp I can see triple digit kW regen. The added torque means it can maintain higher power for longer and at slower speeds. I've actually had to tweak it over time because regen could lock the rear wheels at low speeds with high torque. Pretty much can get down to 4 MPH with just regen, and pretty quickly.

For example, if I let completely off of the throttle with regen set to max while going about 20 MPH, it feels like I slammed the brakes in the car and I get flung into my seat belt, which locks in place during the braking event.

I'm actually going to be undoing quite a few of my more complex modifications temporarily, aside from the power increase modifications, because I've found a minor design flaw in my custom control hardware that I need to fix. Mainly for safety reasons, since the failure mode could cause unpredictable behavior.

when will your mods be publicly available?