Interesting story from Africa: My friend is a professional photographer who does several thousand-mile long road trips through the remote African wilderness. Wherever he goes, he gets cash offers on his tricked out SUV dating from 1998. The reason is that this was the last year before the SUV became loaded with all kinds of computer chips. If the pre-1998 SUV model breaks down in the middle of the Sahara, all you need is a screwdriver and a wrench to fix it. If you have a recent model however, good luck to you. You're going to have to wait for someone to FedEx a new controller board to you in the middle of nowhere. Most likely you'll have to abandon your SUV entirely.
If it really matters, you can't rely on a car that requires a computer to start or to drive. You want something purely mechanical.
2) Your preparation -- access to spares, tools, materials
3) Your knowledge and resourcefulness
There is a lot of critical mechanical parts on a car that are unrepairable, even at a car shop. Most of critical mechanical stuff on a car can only be fixed with a spare. Even changing your tire is just an act of using a spare. Ever broke anything within engine or transmission? Good luck fixing it with basic tools in wilderness.
When car mechanic says he's repaired something, there is good probability he means he has replaced the part that was faulty.
If you go across Africa, take spares for what is needed and tools to diagnose / repair. If you worry your computer may break, carry a spare. Carry a multimeter, some spare tools and materials to fix common electrical issues.
Electronic spares and tools do not take that much space and if you are worried about electronics there is no excuse not to take them.
People complain they can't repair electronics on a car but that's only because they never applied themselves the same way they applied to learn and fix mechanics.
My father specializes in repairing car / truck / bus etc. electronics. He uses simple tools and his experience to repair things.
He has been flown to from Poland to South Africa to fix a large crane that Volvo could not fix themselves. All with basic tools.
There is nothing inherently more difficult about fixing electronics on a car.
---
As a side note, I would prefer older car to drive over Africa but not because of lack of electronics but rather because of more margins, higher clearances and resistance to wear tear.
As to repairing in the wild they are easier to repair because they are SIMPLER, and the parts are more accessible (as in easier to reach them), not because they don't have electronics.
I am not expert mechanic therefore I will have better chance repairing simple car.
I am not expert electronics engineer therefore I will have better chance repairing car with simple electronics. Does that make sense?
I have a 2015 pickup, in the US, and its the speed up changes with the electronics that worries me. My touch-screen controls most features of the vehicle. (after some nice jeep hacks, I learned it controls much more than I thought) but they only used that model for a year or two.
The Axle, drive line, brakes, etc, have been used for at least a decade on my model of pickup, and Cummins engine parts are the same across trucks, generators, marine engines, and easy to find my size.
I just worry about the 'cost' of finding a working computer module in 2030, for my 2015 pickup, if it fails. At $2500 now, I'm not going to buy a spare.
older diesels are valuable because they came before a lot of emissions regulations and will run on dirty fuel, desirable in Africa, this could be the issue the OP is actually describing.
very very tangentially related, but I think interesting.
US Sherman tanks were not very good tanks. But there were a lot of them, and they were crewed by kids who grew up on farms. Those kids likely spend a fair amount of time keeping the tractor running.
Two big things going for them. 1, a ready supply of spare parts, from other tanks that didn't make it. 2, some understanding of the system, allowing people to macgyver solutions out of bubblegum and bailing wire.
As to your side note, I agree. I'd want something dirt simple, partly because, even though I might be able to pull off an amazing feat of engineering, I'd rather have the solution be obvious.
Since we're wishing, I'd also like everyone in a 1000 mile radius to have the same car, so I can find a dead one to strip if I need parts. (but that's not realistic given the implicit remoteness of the situation described).
The main person's opinion[1] has been debunked by multiple authors. It was a very good tank and fulfilled the American objectives. Particularly when you take into account the US Army's belief "tanks shouldn’t take on other tanks". The later gun upgrade even took care the of late war tank-on-tank battles.
I would add that better is the enemy of the good. It is better to have a good tank in large numbers than a better one but so expensive that you can only have limited numbers of it. Germans had better tanks but it did not help because they were expensive to produce and were heavily outnumbered.
The Germans also committed the criminal sin of tinkering on the production line. Consistency is the foundation of logistics. Its amazing how some lessons from WWII might benefit software developers in the present.
I tried to repair some electronics on a modern car and it was impossible for me to buy a new SIM card holder of the same sort. The dealer would happily take £3k off me for a new unit.
I’ve now got a salvaged waiting to go in
I have a car where a mechanical component with a electronic controller board was broken. When you swap the mechanical component you also end up swapping the electronics. I had to send the car to the dealership after being rejected by two independent mechanics because they are the only ones with the authorized electronic repair tool.
Most of the faults can be diagnosed without the computer, if you know how.
Diagnostics is nice because it gives you readout in one place, but the output is what the computer sees, not what actually happens.
Also, where is your licensed computer to diagnose mechanic problems? There isn't any? What a shame... how can you fix a mechanical problem without car telling you which part broke...
The computer diagnostics is a response to mechanics having no knowledge of electronics (also because it kinda make sense if you already put computer in).
This is just straight untrue. You can throw parts at the car while taking educated guesses, but to cost and time effectively fix the car you absolutely need the diagnostic software.
Crank position sensors, cam position sensors, EGR, Fuel system pressure, per-cylinder electrical failures, etc, etc., etc.. can all be guessed at by a shade-tree mechanic. But good luck fixing it quickly and without buying a bunch of parts. My 2010 Mercedes van require parts to be electronically "mated" to eachother, so even if I know the right part, I need the Mercedes software to tell the computer to hook up with the new part.
I've been a "shade-tree" mechanic for most of my life. I would say '98 is roughly correct when full fuel injection became pretty standard. Once you have that, you need a half dozen or so other sensors to make the system work. And you need precise fuel pressure. It just gets more complicated from there.
All these things have known resistances and waveforms. They can all be diagnosed with a $5 multimeter and $50 scope. Fuel injection became standard closer to 88. It's not complicated. The systems are pretty tolerant and only need a few sensor to operate (albeit not perfectly). The precision required from various sensors and things is far less than what internet commenters will lead you to believe. Though if someone is gonna build something finicky of course it will be the Germans (requiring software resets when changing basic things is but one example).
I too am a shade-tree mechanic. It's worth noting that early electronic fuel injection didn't require any external software or device. Rather, you would use a model-specific technique to put the vehicles computer into diagnostic mode, and it would blink codes at you from the check engine light, or perhaps show the code on an LCD that was normally for another purpose. This was late 80s and early 90s stuff using OBD1, the earlier diagnostic standard.
On my Toyotas, I would need to jumper a couple pins in a connector under the hood with a paper clip, labeled T1 and E1, I recall. On my BMW, I had to turn the key to ON and then stomp the gas pedal 5 times in a few seconds. After counting the blinks, the meaning of the codes could be found in manuals. IMHO this was a great system and every bit as useful as what you get with basic OBD2 scanners, just without the scanner investment. I needed no tools to be told my catalytic converter or primary o2 sensor failed, etc.
> "but to cost and time effectively fix the car you absolutely need the diagnostic software."
> "etc, etc., etc.. can all be guessed at by a shade-tree mechanic. But good luck fixing it quickly and without buying a bunch of parts."
That was exactly my point. Car CAN be fixed without the software.
It just takes more time (and knowledge).
If you are an owner of a repair shop then that is probably deal breaker. You can't hire actual engineers to fix cars, the cost would be prohibitive.
But nothing is stopping you to fix the car (except if manufacturer like John Deere makes it their mission to stop you from repairing it).
If electronics is broken your mission is to find the broken part and replace it, not reprogram it, which is basically the same as with mechanical faults.
Eg. Sometimes you can fix the issue, but you still need a computer to reset the "error light" on your dashboard.
In general, the direction is clear. While you can still have some manufacturers that aren't requiring it now. It will be in the future.
Source: a family member has a "non-brand" garage shop with specialization in one brand and now he can work better on cars, because he bought the computer when an official shop went bankrupt ( that was ~2 years ago).
Last time I saw him ( 1 week ago), there was a big similarity to Jhon Deere and the Ukranian developers that have unofficial software to work on the machine)
I don't need anyone to tell me which mechanical part broke because I already know that. It doesn't change that I still need the licensed computer to do a simple mechanical part swap because there are electronics embedded into the mechanical part.
Actually, anyone can get an OBDII scanner. The trick is figuring out what the code means, and hope you bought the service manual and have a space to remove half the bloody car based on the part that died, because (if you have say, a 2000 Chevy Venture with a cracked oil pan, you have to pull the frigging engine out of the car to get to and replace it) auto manufacturers build for automated assembly, not service and easy access..
When a bearing fails, have fun remaking ball bearings with a hammer...
Some mechanical parts are just like electronic ones - you need a massive factory to produce them.
You might argue that ball bearing balls are standard parts so easy to get anywhere, but the same is true of electronic components - a 10uf capacitor is probably within 3 feet of you right now.
All that's missing are the skills to be able to diagnose issues down to a single component rather than saying "the control board is bad, we need to order a new one".
And often times the repair shops themselves indulge in "replace part after part" marathons and see if it fixes the issue. Both for mechanical and electronic/sensor problems.
Saw the same with a cost estimate to repair a DeLonghi automatic coffee machine (quote was to replace both broken heater elements, the broken pump and renew seals for 270€) when in fact none of that was broken but the firmware had a bug in one of its stats counters (overrun integer brew counter) that was fixed by a fw upgrade.
1998 isn't even pre-OBD2, that's a computer controlled engine if gas, to say the least.
The spirit of your comment has truth to it, but it's unclear what exactly is so desirable about your friend's SUV in Africa. In your own words it's "tricked out", how do you know it's not just his modifications that have everyone making offers?
In choosing vehicles for such purposes, the criteria is more nuanced than "no computers". There are factors like fuel type, compression ratio (lower the better, runs fine on shit gas), non-interference design (in case timing belt fails, non-fatal), fewer moving parts in general (so you probably want to avoid an entire decade of non-computer, post-smog vehicles, which tend to have miles of vacuum lines and dozens of solenoids to leak and fail).
The 90s are kind of a sweet spot for computer-controlled post-smog simplicity, where vehicles likely still have throttle cables and no ABS. But you have to go much further back for no computers at all and no insane pre-computer emissions controls, basically pre-75.
That was the case in Russia villages. Old Russian cars were used everywhere and one could fix anything with a wrench and hammer. Those people were reluctant to adopt new cars. I think that it's changed now, as local mechanics got more experience with electronics and stuff. Most of electronics issues actually could be fixed, you can replace capacitors, etc. But of course not in the middle of nowhere.
That said, old Toyota cars are still treasury, because they just don't fail. But they're quite costly.
People in deep rural China few decades ago used to have diesel motorcycles because diesel, or some semblance of it was the only fuel available to other agricultural machinery.
Now these times are long past, and few Chinese today even remember that.
Everyone chimed in that electronics are so easy to repair(ofc its HN). Whoever was offering money for the truck obviously knew how to repair machinery but not electronica. So they prefer what they know to repair, lets bash their lack of knowledge?
We could also try to, hm, make electronics easier to repair by providing an fn manual and not being so anti-repair as an industry. But how are we ever going to be as rich as Apple and Tesla if we started selling actual products and not just onboarding to a crappy service?
This place needs a name change, VC news - definitely more representative of its readership.
One thing is for sure: cars of today are not at all collectable and poor investments. No one will know how to or care to reverse engineer some long obsolete proprietary CAN/lin protocol 50 years from now. The myriad of bespoke modules, electromechanical bits, computers, and other frail plastic bits will be near impossible to repair without an engineering degree and dozens/hundreds of hours in reversing, designing, fabrication. Absurdity. The bottom line is these vehicles are environmentally irresponsible.
I bought a 1961 Mack B61 (year and model #'s coincidental) to restore (slowly) and its a 100% mechanical turbo diesel (ENDT673B). It so dumb that the "ignition" key is just a two position on-off switch which enables the electrics allowing the starter to run when the start button is pushed. To stop the engine, you pull a stop knob which operates a lever on the injection pump pushing the fuel rack to the full-off position. I once let it sit two years without touching it; started up with fresh batteries after checking oil and coolant. That truck will still run and be serviceable in the next 50 years. Maybe parts will be hard to find, but most can be fab'd in a simple machine shop by a single person.
> The bottom line is these vehicles are environmentally irresponsible.
If you’re right - which I don’t think you are - we’re going from a world in which maybe 2% of vehicles are collectible to one in which none are. Not really much of a difference from an environmental perspective, especially when you consider the large efficiency and emissions benefits of computer controlled engines.
> especially when you consider the large efficiency and emissions benefits of computer controlled engines.
Is that enough to offset the entire environmental cost of the cars short life span from manufacturing to the scrap heap? Today it feels as if cars are not meant to last much past 5-10 years.
Just think of that Tesla tablet screen who's flash memory failed after a bug in their software wore it out. Will you still own that Tesla in 10-15-20 years? Of course not. It will be impossible to service by then. The same will apply to VW, GM, Toyota, Nissan, BMW, MB, etc. They're as disposable as smart phones because they are smart phones.
We can still have that old school stupidity if we:
A. Start buying practical cars with as little gadgets as possible (e.g. the base model). Luxury and convenience is how they lure you into walled gardens. Those extra tidbits increase the likelihood of electronic failure decreasing the longevity of the vehicle and increase operating cost. They fully know humans are lazy and exploit that trait. Don't fall for it.
B. Force manufactures to implement 100% open protocols at all level of interconnect. e.g. I should be able to hook into and diagnose the the CAN/LIN/etc network between my cars switch module and the heated/illuminated/dimming/motorized side view mirrors or tail light module. I should also be able to fabricate or buy a 3rd party device that can replace the functionality should the OEM gadget fail (e.g. tail lamp module).
C. decouple car function from ancillary function, e.g. removing or replacing the infotainment/navigation/radio from the vehicle should not interfere with the cars ability to function. Likewise the HVAC has no business displaying its status on the radio. These should all be discreet systems with the option to interconnect via above mentioned protocols.
The biggest problem in new cars is DRM (software) and custom parts (pcbs are cheap to customize). As such every
Make and model and someone year has parts that are unique or software with expensive lisences. Mechanical parts especially old ones were more likely to be common and therefore higher availability.
HN is funny sometimes. You guys seem to either want a battered down Toyota that you can fix with a screw driver or the absolute latest AI-powered Tesla 666.
It's a mix of software engineers who want to run the latest bleeding-edge framework and the operations engineers who have to make it work in production.
Hum... I'm currently a system admin and have spare parts and a toolkit in my vehicle. I will have to think on that when I go back to being a developer.
I mean, I want both. Or rather, there's a cutoff point where I am okay with one, or the other. Right now the state of things are in a worst of both worlds scenario. I think cars are:
Too expensive and complex now in the name of "safety". Most additions to required car safety features beyond airbags has been minimal in actual gains compared to the added cost and long term repair complexity induced.
I think entry level cars should be able to be much cheaper and barebones. The recent requirement for backup cameras is one example of too much for too little. I just want a car that gets me from A to B reliably. I don't even care if it has good radio/speakers, power windows, etc.
On the flipside, I hate driving in general. I'll happily pay for a self-driving (or nearly self driving, at least) car with all the gizmos needed to make it work. However I will only "buy" such a car (in the sense of legal ownership) if it has a lifetime warranty or reasonable trade-in guarantees. Otherwise it'd be lease or long term rent only as the repair and upkeep costs on so much proprietary tech would be insane past X years, if its even possible.
Cars are a tool, not a fashion / status symbol for me.
Well, he described me. I also doubted between an iPhone 12 mini and a PinePhone (the FairPhone was also in the mix). Both (or all 3) appealing for very different reasons.
Would anyone grief a car mechanic for owning a miata, a truck, and probably half a dozen vehicles in total? One form of being a professional involves being an enthusiast.
HN is (sometimes rightly) perceived as a hive mind with regards to the consensus that emerges on certain topics. It's a classier country club than Reddit, but it's still one.
Quite the opposite, I rip cyclocomputers off of every bicycle that I buy.
I can't be the only software dev that dislikes gadgets. I work with computers all day but don't particularly like them and don't want more of them in my house.
With that said I'm not against computers in ICE vehicles, the emissions improvements and yes reliability improvements they've brought are amazing.
Well, to be honest, this https://a-ride.com/ has caught my nephew and nieces attention as of late. Way to expensive for them but I guess bicycles get boring too.
Automakers should make cars with less chips. Touchscreen controls, licenses to accept, firmwares to update, tracking to spy on you and Internet-attached locks to crack are abominable. I want a car as analog as possible without loosing much of the engine efficiency.
For me only the engine is essential. I don't care about LEDs, AC, ABS, ESP or whatever. All I need is moving wheels, steering wheel (not necessary with amplifier), brakes (don't care about ABS). Heater at winter. Everything else is optional and can break. All I care about is to get to home alive when I need to ride 300 km without a single car around me. Whether my cars works or not might be live or death.
Though I don't think that chips are root of all evil. You can easily have analog systems failing. I think that engine chips just should be very reliable, may be even duplicated. We have vehicles on Mars, surely we could get reliable vehicles on Earth.
And, of course, I don't care about reliability of those second-class systems. They're not essential and car works without them. They could be made with chip or whatever, it does not really matter.
Going back to carburetors, vacuum lines, and choke valves? I can't even imagine what that would do to air quality around us. If you remember cars in the 1970s, it was awful.
I've never seen an ECU you couldn't tune with a laptop. It's a hell of a lot easier to dial in a modern engine than it was when you had to try to find the best combination of jets, needles, and springs.
You usually need more than the electronic equivalent of a screwdriver though, right? Maybe I'm mistaken. I was under the impression you would typically need something like a Cobb port.
If you can access them, I certainly imagine the efi profiles are much more tuneable.
Yes, it's usually a fair amount of work, on most modern ECUs there is some form of signature based protection which will require an exploit chain to bypass, and then the need to write either software patches to enable introspection into the parameters necessary to tune, or to reverse engineer the protections on the factory calibration protocols if they were left in production builds. This is all possible with the electronic equivalent of a screwdriver, a Raspberry Pi, but requires a lot of reverse-engineering effort and understanding.
Sure, but I suspect reducing unnecessary automation will lead to using noticeably fewer chips.
People have different preferences, but I think many, like me, would prefer a simple heater on/off and a linear stronger/weaker selection over a full climate control that auto-selects whether to run a heater or an AC and how much. My 2c.
Sure, but digital electronics are programmable and it's also possible to use programmable analog stuff for many things.
So it's certainly possible to use a few standard electronic parts, with 2nd,3rd sources to implement most of the electronics, and be easily protected from supply disruptions.
On the contrary, my first car still had an analog speedometer. It really sucked when it failed because the failure mode was that it turned decidedly non-linear past about 50 mph indicated. The first time I drove it after the failure I hadn't driven it in probably a couple months and I was driving it at night without any traffic around. I wound up getting pulled over on the highway and the indicated speed was off by 15 mph.
At least with a digital speedometer almost every failure mode is going to be obvious and not some subtle accuracy degradation. Part of the supposed rationale for police almost always not enforcing the speed limit until a good margin over is because of the historic inaccuracies of speedometers. Nowadays assuming someone hasn't modified some gear ratio or fitted the wrong size tires you can more or less be assured that the speedometer will be quite accurate.
On my manual transmission car you can pretty reasonably guess your current speed based on the gear you’re in and on the engine sound, at least I can do that from 0 to approximately 80-90 km/h (after which everything becomes too noisy, I need the speedometer). Failing that one can always follow the flow of traffic in order to remain on the safe side of the law.
Automakers should make cars that are still usable when non-critical components fail. Chips aren't the problem; too many points of failure is the problem.
In an engine, there critical parts are the rpm sensor and fuel supply / ignition. Failure of one of them means an engine not running.
Now on top of that you have emissions regulations which force you to run an engine with 100% functional emission systems. To ensure it, the automakers are limiting the engine performance to force the user to repair, which is a major source of annoyance for the user. Moreover regulations require ABS and ESP on every vehicle.
So even if automakers would do such a car, it would not pass the homologation phase due to these regulations which can only be achieved with chips.
Africa needs an African carmaker, or import from Russia or other non regulated markets
Custom electric conversions seem great (see e.g. Zero EV's MX-5 on YouTube). Not analog, but not "smart" either. No internet-connected computers, only the necessary microcontrollers.
Thesis: automakers were already making more cars far above and beyond demand before 2020, they used the pandemic to cut their part orders and now have repercussions from that, but in reality, that is what they want. They need a lack of supply and for current stock to be used up before they will be profitable.
Absolutely not. Automotive factories are only profitable when run at high utilization. If ICs are short, then factories cannot run at high utilization.
I am strugling to understand the nature of the global chip shortage.
Is there a single driver?
Or is it just a confluence of events: The US-China tech war (preventing expansion of Chinese factories, causing stock-piling from companies fearful of sanctions), factory shutdowns due to the pandemic, unforseen demand in certain areas (work from home equipment, TVs, GPUs for cryptomining), random accidents like this one.
One could fear there to be massive overcapacity in a short time (1-2 years). As mainland Chinese alternative capacity comes online and when some of the weirder demand peters out (GPUs for mining).
My intuition is also that there is less pressure in the "let's buy the latest phone game" than there used to be. (For example I am using a 3 year old iPhone X and have no inkling to buy a new one).
Stocks are not a single commodity, but stocks in thousands of companies over many exchanges, and there exist indexes and ETFs to try to aggregate them. The same can be done with slightly different silicon fabrication facilities.
For example, a company who needs their chips manufacturered at Facility A or B, but cannot use facility C would buy a futures contract letting them do that.
I believe OP means that certain types of IC's (for example memory chips) require pretty different technology and equipment than other chips (eg. processors). That means the production capacity isn't fungible.
Isn't that true of most things, one can't farm pigs with an oil rig, those facilities really only make one thing. Certain facilities could be retooled/repurposed and this is maybe a bit easier in agriculture for certain things (plant soy vs corn) but you can't just plant an almond orchard overnight.
Yeah but I think that is the point. In their view the commoditization of something means how easy is it to retool to make something similar. Fabs can be retooled, after all its usually all the same machinery, but the process knowledge is a huge barrier.
IMHO some of it is climate change. TSMC is only running with trucked in water due to drought conditions in Taiwan, and multiple fabs in Texas had to go down with their power outages and fab restarts are not fast.
What's with Japanese chip fabs catching fire? It's very odd that Renesas was producing some chips for AKM at that plant after the AKM plant caught fire. From earlier this year: "Renesas steps up and lends capacity to AKM after the fire"
That's a good answer for "why would a fab catch fire?", not as good an answer for "why would two fabs in Japan that were producing some of the same chips catch fire?". But as they say, twice is coincidence.
I wonder when in-vehicle computers come in standard sizes and with standard ports. like 4x gb eth, 8x can, 3x hdmi, 2x usb something along such lines.
just like you buy another system from another vendor for your data centre, if your usual supplier happens to have production issues.
this would also simplify computer power performance updates, to dunno go from level 2 to level 3 autonomy (assuming sensors are there already).
the vehicle part of the vehicle easily lasts 10-15 years. such standardized vehicle computer infrastructure would not only make automotive vendors more independent from board and soc fluctuations, it would keep the car fresh for a longer period of time, and allow for performance software after-market sales.
There are 2 cleanrooms. The 2nd story one is undamaged but needs the 1st story one to be repaired (I assume it's a structural thing?). So I'm curious how, from 1 picture, your comment can be informed?
Also from the article, the fire damaged 5% of the first cleanroom and 11 machines. Maybe they're able to get the other 95% back up and running without having the damaged 600sqft fully functional?
> There are 2 cleanrooms. The 2nd story one is undamaged but needs the 1st story one to be repaired (I assume it's a structural thing?).
I don't believe they have multiple clean floors.
Fabs are usually done with space below the clean floor being where machinery for supply of liquids/gasses is, and space above being used for air handling equipment.
Some have a storage floor, but that's usually it, especially for a smaller fab.
Sub fabs are not always about liquids/gases etc. Some fabs, especially bigger DUV fabs, use the sub fab to run their lasers so that scanners can be arranged on the track in a sane way.
edit: btw, almost all fabs I have seen have multiple clean floors. Even the sub fabs doing only plumbing work would maintain a clean environment. I'm guessing because thats cheaper than maintaining the interface which goes from clean to unclean environments.
I work in a tiny R&D company, mostly building prototypes, and currently, _every single project_ we have that involves hardware is in some sort affected by the general semi-shortage + Mouser (being a Texan company) delivery delays.
We weren't able to buy the _150_ micro-controllers we needed for a tiny series production, I'm currently waiting for a literal handful of crimp terminals to be shipped, the 220Vto12V AC/DC converter I normally use is unobtainium.
No PS5. No graphics cards. No Vaccines. 6 month wait for bicycles. The biggest LSD producer in Europe got caught, so no LSD either.
It's like in communism, but you still have to pay rent.
If it really matters, you can't rely on a car that requires a computer to start or to drive. You want something purely mechanical.