Bloodhound LSR Run Update |
Вставка
- Опубліковано 27 вер 2024
- Today, 30th October, the team rolled Bloodhound LSR out onto the Hakskeenpan with the intention of running Profile Four: a max reheat start up to 400mph and then a parachute deployment to slow the car.
Bloodhound started perfectly and blasted off down the white line, however, when she crossed the area where an old farmers track had been graded flat, there was a very minor bump that caused an engine stop switch to engage for under 10 milliseconds - causing the EJ200 jet engine to safely shutdown and aborted the run.
The car was recovered back to the Tech Camp where Bloodhound’s System Engineers poured over the data and identified the issue. The data was verified by checking the video footage from the rear facing fin camera. This clearly showed the engine cutting out as the car crossed the darker coloured patch of desert. It was so quick that Andy Green couldn’t feel it from the drivers seat.
The Systems Team, lead by Joe Holdsworth, are confident that their software modifications will make the switch less sensitive and ensure the high speed test programme continues as planned.
---------
We’d love to hear from you! Please let us know what you think or any questions you’d like answering in the comments section below.
Don't forget to subscribe to our channel to never miss any of our exclusive videos about Bloodhound LSR: bit.ly/31C1RyT
www.bloodhoundl...
/ bloodhound_lsr
/ bloodhoundssc
/ bloodhound_lsr
/ admin
Your doing exactly what you should be doing. We all want Andy (and the team) to set another LSR but patience and safety is paramount. Well done to all !
Great report.
Easy on the music, hard on the raw data.
This is an amazing project.
Great to be kept in the loop, thanks. Excited to see her next run. Keep up the hard work and good luck.
Really interesting to hear this sort of information. I love how you guys are showing the good and the bad. Makes it so much more engaging and real.
They had problem, they looked at it, debugged it wrote a fix. Sorted.
No dramas, no need for exaggeration. So typically British. Way to go everyone.
This isn't a "typically British" way to solve an engineering problem btw.
@@swimcali Thats not what I wrote. My first sentence was about the problem solving and is typical of a high tech environment. However it was in my SECOND sentence where I was referring to "No dramas, no need for exaggeration" that was typically British.
And most interestingly they shared it all with us here on UA-cam. I am thrilled to watch this detailed coverage but the biggest benefit is the example it sets for young people. They get to see something of the process needed to accomplish a difficult task.
Too much drama from you.
@@swimcali Your point was exactly?
Incredible work all, been following this for a long time and one of your aerodynamcists was my uni lecturer! Good luck!
Great stuff! Many bugs to be fixed, but that's what these runs are all about (plus windspeed limits etc). Keep going!
Great to see such progress, as someone who was part of a team that very briefly held the Guinness World Record for a rocket powered model car!
Interesting, the power of software that gives so much data...can also be the source of an issue in the environment the car must navigate. Unforeseeable issues are the very reason for staged testing. Bravo team Bloodhound LSR!!
Best of luck to all at team Bloodhound!
And this is why extensive testing is required. Good work guys and girls.
fingers crossed for a good run tomorrow :)
Great stuff. Keeping Andy safe is important of course. With you guys all the way 👍😎
Great work. Such an impressive car and team
So cool. Thanks for the great info update, we know these little shakedown issues are all part of the process for building to success.
Today I read a story about a bloodhound who got out of her yard and ran a half marathon with the human runners. She finished seventh and got a medal! Seems like a good omen to me.
I look forward to these videos.
When this project was being developed (by the previous guys) we were promised `live data` from the desert for school students to study. Would be interesting.
Wow thats Crazy Good luck
these are all the things that are impossible to plan for and why testing and reacting quickly to unforeseen issues is so important...anyone working in design or engineering will understand this everyday...
Great video, many more like this please.
Second stage fail safe. Just hold for a few milliseconds - give it time to recover as 1st option
I like all of the updates. When is full speed run scheduled?
Next year. They might manage 500+ mph this year though
Well done guys
One they go over the sos no turning back good luck.
Is this going to continue building to the 1000mph record attempt, or is it a test program that will stop short of that?
This year is testing with the jet engine only. Next year they’ll be back with the rocket installed as well.
@@andylewis7360 I remember reading something like that ages ago but, with the stop/start funding problems, wasn't sure if they'd changed their plans to do it all in one hit, while they could afford it.
Without the rocket, isn't it still possible to slightly exceed the Thrust SSC record? Not sure if I imagined that bit.
tjp353 You’re right. I was working on the basis of the original plan, as I perceived it. But I’m sure the plan was worked out meticulously so probably best to stick with it.
I suppose bloodhound could beat Thrust SSC’s record, given a long enough track. We can hope, I suppose. She certainly seems to accelerate like a scalded cat! 😁
I pretty sure that the full record attempt will be next year, not least because Bloodhound’s 2019 exploits could attract a good deal more sponsorship support.
Not been fully following the progress, but is this the same track (with bumps) that the eventual record will be attempted on?.
Yes
No it's not.
They've got over 20 tracks mapped out. These are the preliminary test tracks for up to 600mph.
joe pineapples The surface is, yes. I think there’s a track going across the parallel tracks that Bloodhound runs on. If you watch, you can see a change to a darker colour at 0:53 in the video.
Ah, a software work around for a hardware issue. I spend half my life writing those. I've written at least two debounce routines in recent years that I can recall.
That did not work out well for Boeing.
Bob Smithies Because Boeing's software work around was badly designed, not documented externally, and the impacts or even existence of it on the control of the aircraft not in the pilot training material. I am sure Andy Green is aware of this one.
@@owensmith7530 - really ? Lets hope they put it in the manual.
10mS switch debounce? That does seem a tad short! At 1000mph you’d travel all of 4.5 Metres in that time. I think 30-40 is still a bit short but hey. These things need tweaking.
I'm not alone with button bouncing 🙈🤣😂🙈
simple fix then good work make some noise
Very sensitive. Fingers crossed.
Surprised your suffering with switch bounce.. somebody skimped on the switches!
Switches bounce, it's a simple matter of physics. The only switch that can never bounce is one so stiff you can never press it.
ACTUALLY.. @Owen Smith High end switches such as aviation switches don’t suffer with that problem. Just adding a 40-50 ms delay to the microcontroller isn’t a good fix as the bloodhound engineer had suggested, as the microcontroller is then occupied waiting out the delay! A high end switch or designing the switches with RC circuits in the first place would have been advantageous.
Robert Pitcher Depends how they've done the delay, the code doesn't have to busy wait. It could be an asynchronous timed event. Or it could read the timer on each switch transition and ignore any that are too soon. Or it could integrate and wait for a confirmed number of settled readings of the switch. There are quite a few different ways of doing it depending on your needs.
Unbelievable how many tiny little setbacks this project is having.
GladiusYTR welcome to running a prototype racing car. These are just teething problems and easily fixed
@@bloodhoundssc Totally agree. Keep grinding at it!
GladiusYTR It’s par for the course, mate. A volume car manufacturer will build dozens of prototype and pre-production vehicles and test them almost to destruction over 18 months to 2 years before the first model is sold.
@@andylewis7360 Yep. All true. It just puzzles me how sometimes small teething problems interrupt a whole day of running. I'm sure the Bloodhound team don't have all year to keep running the car and sort out niggle after niggle. Surely something like a steering wheel switch debounce filter should already be part of the control software and have some sensible calibration. Particularly given that NVH at 1000mph can't be a minimal factor. OEMs do a lot of prototyping, true, but small issues like these don't usually interrupt testing. Granted this is a LSR car and the risks involved are greater... Perhaps these things are normal in the context of breaking records and we're not used to getting the special insight we are seeing from the Bloodhound team. Either way, it's so freaking fascinating and awesome.
GladiusYTR It IS fascinating, isn’t it?! My guess is that they went for the minimum value they could set, on the basis that if it doesn’t cause a problem then they’d get the fastest possible reaction time. I’ve never used a debounce time of less than 50mS and 100mS in most normal circumstances. Given the fact that it’s a kill switch for an engine that takes tens of seconds to wind down, I don’t imagine that it’d make a whole load of difference but I suppose in an emergency, some 10’s of mS could make some difference, although I can’t imagine what. My main concern is that it’s an indication that they’re using aerospace, rather than automotive connectors. Auto connectors have got to be more tolerant to vibrations than something that’s fitted to an aeroplane.
Tesco switches
Literally "switch bounce". Hmmm. Wonder if the switch manufacturer envisaged their product doing 1000MPH across a desert in a vehicle with metal wheels? Ha! Go for it folks. Faster, harder, more!
Good point! They were altering suspension settings yesterday. Electrical control switches and connector contacts are usually rated for high G ratings. Usually tested by putting them on a big loudspeaker cone and sweeping the frequency up from 0-20KHz. It’s quite unusual to get a bounce in normal operating conditions without an intentional click. I reckon it was a connector contact. Extending the debounce period should cure it. I only hope they’re using automotive connectors and not aerospace ones. Planes don’t get many shock loads. Cars do all the time.
NO! Fix the hardware. It can get worse. Don`t mask the problem with software changes!
Exactly, hardware kills, software doesn't.
Stay safe guys, I don't want to be reading Andy Greens obituary any time soon.
@@hughcdavies you can put a hardware RC filter on the switch contacts or do exactly the same thing in software.
One you can do in the 'desert', the other you can't
bjofuruh The software delay was a bit on the short side. That’s all. They’ve just changed a counter. It’s no biggie.
Im guessing you don't really understand debouncing then?
You can actually do a better job of this in software than you can in hardware. An RC filter only does one thing. A software debounce can be as simple as a delay and re-read, or it can average, or take the median value, or after a change wait for a certain number of consistent readings to take as a real change, or any number of other methods. You can choose the best method for the job.
here, this is a good read for novice programmers - www.arduino.cc/en/tutorial/debounce
AMAZING INFO GENTS!!! Rooting for you guys here in Oregon!!!
Impressed with your attention to safety at every step along they way. Doing this on the ground is far more dangerous than in flight. Please be careful.
Thank you guys so much for the updates. Its been the best part of my day sense they started.
Inject this technical content directly into my veins
Those pesky milliseconds!!!
That's some force on the vehicle and the resulting movement of things that causes contacts to move.
Going to have to time after seeing the falling edge and check for a rising edge in the debounce period.
Is it practical to have that?
There's many digital inputs from switches, one can't implement that on all now?
I would be soldering caps across stuff lol.
That's concerning though, that vibration can cause contact movement and an unintended command takes place.
Luckily it shut down the engine, but this is still at low speeds.
Now I'd be worried about safety.
Would rather sort out the electrical problem as a cause, the software is doing what it should.
i would not change it, fix dat switch.
It's a safety circuit, can not have filters or delays on it, seriously.
Thanks for the update, D.
Good points! But, that said, there’s no way that Andy would be able to push a button for only 10mS. I’d be busy soldering caps too but I wonder if they’d have them in the desert. Time to call Farnell, maybe! 😁
Holdsworth this is the best Bloodhound info yet.... short and clear info.. good enuff to make you an HONORARY YANK !! (No problems we can't solve Houston)
my mum says drive it like your granny stole it!
🇬🇧👍🏼😊
Very interesting, thanks for the explanation.
That's insane! First time running this complex machine and all that went wrong was only a small wiring problem
Safety, safety, safety. 100%.
👍🇨🇦🇬🇧
Very cool.
Turn the switch on its side maybe? I imagine there's not nearly as much lateral G force as the is vertical to sorry about. :)
It's so cool to get these inside/meanwhile stories and hearing those solutions! When i heared that it was done by a bump i immediatly thought about adding some sort of a filter to the signal, that's just a bit like playing with the settings of a drone flight controller :D I dream of such a job.. Have a blast, team Bloodhound!
I'm glad this pursuit still exists and you are embracing the challenge. Keep up the great work !!
Option 3: Avoid the bump
Expensive testing 😎