Why Methanol can blow you up without a failsafe?
wish zeitronix will hurry up to finish their USER input with a "trigger when less than x volts" function......or even better/safer, trigger failsafe when voltage is above OR below x volts....
Funnest d@mn post Ive read in a long time
Im going with this very setup on my car.
I have no failsafe. I have my meth set to come on @ 12psi all the way up. When its shootin a LED light comes on the dash. If I dont see that light I dont press, I find out whats the problem.
bump for lots of good info and hopefully more. smoked mustang... great thread and topic! it just makes me all warm and fuzzy inside when everyone can get along and talk about tech. on cars without fighting!!!!!!!
Evolving Member
Joined: Dec 2006
Posts: 276
Likes: 0
It depends on how you measure the sensor and it depends on what you consider fast...
The most common way to use a "UEGO" sensor (Bosch or otherwise) is to measure how much current it takes for the pump and Nernst cell to maintain a balanced state. You can read a good article about both the underlying theory and the practical implications here.
The reaction in the sensor is quite fast, but from a control perspective it is still quite a bit like precision control of a DC motor. Close enough, in fact, that some variation of the same common solution - a PID loop, it typically applied. And, like most closed feedback loops, there is dampening and lag.
Also, like any PID loop, the balance between response time and resistance to oscillation depends a great deal on how the loop itself is fine tuned. A big part of Bosch's reputation for slowness was simply the 'tune' of the prepackaged control circuit Bosch offers. I don't have the spec handy, but I recall tao being very high, making both initial reaction and stabilization quite slow.
It is quite possible to use a 'tighter' tune with Bosch sensors, and many manufacturers do. A typical good controller has a tao on the order of 100 mS. So if a change occurs in the lamda, the controller should begin to react within a tenth of second and be providing a stable reading at best precision within 1/2 second (5 tao). There is really no evidence to suggest that other sensors, like NTK, can be 'tuned' appreciably tighter - though there are sensors which have more stable current response curves over time and better resistance to things like lead. So I guess it would be fair to say that the concern over inexpensive UEGO sensors used this way isn't speed, but accuracy over time.
The other way to use a UEGO sensor is to measure with a form of pulse width modulation. There is a myth that this method was invented as a cost cutting measure, but that is simply not true. The inventor was actually looking for better measurement speed. Half a second is no big deal on a static dyno or with an engine protection circuit, but with big engines on inertial dynos or actual tracks, the lag and dampening potentially matter. What he observed was that, like an ECU and a narrow band sensor, most wideband controllers are attempting to stabalize around a singularity, creating inherent oscillation which must be corrected for. By rearranging how the main components of the sensor are typically wired, he was able to bypass this all together.
Measurement speed is dramatically faster. Commercial implementations all achieve best precision about 200 times per second and the method could be theoretically faster still. There is another myth that this is a Bosch specific thing. That is also false. The measurement principle was originally implemented on multiple UEGO sensors, including NTK. It is still used with non Bosch sensors in some special cases, like were high resistance to lead is required. Commercially you see it almost exclusively with Bosch for basic cost/performance reasons.
The disadvantage (both real and perceived) that Bosch has in terms of accuracy is largely due to two factors, drift over time and inadvertantly exceeding the manufacturer's max bung temperature spec. As I noted in another thread, many manufacturers are aware of Bosch's max EGT specifications, but often miss the spec'ed max bung temperature, which is quite a bit lower.
What was discovered was that with pulse width modulation, the drift over time (which all UEGO sensors have to some degree, but which Bosch has a reputation for), is actually measurable (hence correctable). Also, due to a quirk of sorts in at least one Bosch sensor, bung temp also turned out to be monitorable. With no appreciable difference in terms of speed and accuracy, there was/is little incentive to support more expensive sensors.
BUT (sorry, I tend to run on about all things technical)...
My main points are:
* Any good UEGO based wideband controller (Bosch or otherwise) should be able to detect and respond to a dangerous meth injection situation in under .5 seconds.
* Even in situations where .5 seconds is too slow, UEGO sensors (Bosch or otherwise) can be operated much faster still.
* Regardless of the measurement principle used (current or PWM), Bosch's 'disadvantages' in terms of accuracy can be pretty much negated. For example, even if you are not using PWM, so you cannot correct for the drift over time, the sensors themselves are very low in cost and can be replaced regularly. Similiarly, virtually all Bosch installations will benefit (both in terms of accuracy and longevity) if care is taken not to exceed all the sensor's spec'ed temp limitations.
-jjf
Evolved Member
Joined: Oct 2002
Posts: 2,788
Likes: 1
Little bit off topic, but since we are talking abut UEGO sensors / controllers....
Why is that the Autronic Exhaust Gas Analyzer is over $1500, and this Zeitronix unit is like $200. Why such an enormous difference in price?
Why is that the Autronic Exhaust Gas Analyzer is over $1500, and this Zeitronix unit is like $200. Why such an enormous difference in price?
Evolving Member
Joined: Dec 2006
Posts: 276
Likes: 0
Looking at the top tier Autronics unit there are some obvious differences. For example, it is quite a bit faster (up to 10 Hz for true digital measurements), measures over a broader range, and can be free air calibrated. This last one is subject to a lot of myths. Companies that don't implement this often make a lot of noise that it isn't necessary. But the physics don't lie. With these devices lambda is a relative measurement, proportional to partial pressure.
If you tune your Evo to perfection in Santa Monica on a beautiful day, then head out past Lancaster to do some track runs in the high desert on a typical 90 degree day, you can see the difference first hand. Free air calibrate and tune to the same numbers in both locations, and engine performance will be quite close. Use fixed pressure and the numbers from Santa Monica will be too rich in the desert. Too rich, of course, in a performance tune means less power. Maybe only a few HP on an Evo, but in a race... And, of course, a guy who lives in Santa Monica might want to drive in Big Bear or Mammoth as well.
Two other differences stand out. The Autronics unit supports an assortment of sensors. This probably doesn't matter to most users, but it might matter in some special circumstances. Also, it appears that the two companies have somewhat different distribution and support infrastructures. It may not seem like a big thing, but once a company publishes a phone number, customers expect someone to answer it.
One could turn the question around and ask why Zeitronix is as expensive as it is. A couple of companies have now broken the $200 boundary for wideband measurement (say $80 less than Zeitronix), even while selling through more conventional channels and offering things like telephone support. But that modest gap is probably either economics of scale or channel building.
-jjf
Evolved Member
Joined: Oct 2002
Posts: 2,788
Likes: 1
Thanks very much for all that info. I'm having an Autronic SM4 installed and mapped in a few weeks. I got the SM4 rather than plug 'n play because of the additional external sensor inputs. I'd like to have wideband going into the SM4 but am trying to figure which one and how exactly.
Thanks again.
Thanks again.
Evolving Member
Joined: Dec 2006
Posts: 276
Likes: 0
Thanks very much for all that info. I'm having an Autronic SM4 installed and mapped in a few weeks. I got the SM4 rather than plug 'n play because of the additional external sensor inputs. I'd like to have wideband going into the SM4 but am trying to figure which one and how exactly.
Thanks again.
Thanks again.
It is pretty easy for us because our analog outputs are fully programmable. But I suspect that you could get a unit like the Zeitronix to work with a pair of external opamps and the right resistors (basically adjust the analog output to match with external circuitry). The one thing I'm not sure about is speed. Our users typically slow our analog outputs down to match the Autronic unit. I'm not sure how the SM4 would respond when the wideband measurements are quite a bit slower than expected.
Understandably, Autronic is quite unhelpful when it comes to you saving $1,300+ (or whatever) and not making a purchase from them - but the SMC, SM2, SM4, etc. are pretty popular, so most wideband providers should have some experience with them and be able to give you some guidance on integration.
Good Luck,
-jjf
Evolving Member
Joined: Dec 2006
Posts: 276
Likes: 0
I think you missed the point. It is the *Autronics* that has a 0-5 *input*. Zeitronix finally has a linear WB *output*, but the range would not match, so the output would need to be shifted and rescaled. In other words, it isn't enough that the voltage range matches, the meaning of each voltage is important as well.
The second issue is that the analog output may be too slow. Autronic, of course, designs for their own wideband. And the Zeitronic controller, which has no spec'ed speed, appears quite a bit slower in terms of settle time. This makes sense, and is quite likely a design choice. The last time I checked, Zeitronix was using CAT5 cable and RJ style connectors for connection to the Bosch sensor. This makes for a very small firewall hole, but there are some technical downsides as well.
Please note, I am *not* saying that the Zeitronix cannot work, I am only saying that the output would need to be adjusted to match and that speed *might* also be an issue. The best way to know for sure is to find users who have successfully connected the units together. This is an application we're happy to support, but I can't speak for other vendors and there are some technical hurdles.
As a side note, your disdain for our modular line has been repeatedly expressed. However, I am not sure why you would stress the point here. There are some cases where modularity comes out higher than a all-in-one box (that is part of the reason we offer both a modular line and a handheld series), but here would seem to be a case where modular 'wins' a price fight.
However, rather we are a little less or a little more should not be the initial concern. The first concern should be rather or not the unit is suitable for the specific user need at hand. I'd suggest you review the technical requirements for the exact situation cited before making a recommendation, let alone quoting a price.
-jjf
A few points, for what they are worth:
Regarding 'why':
To understand why things go boom, I think it is helpful to recall what meth injection is doing in the first place. As most of you undoubtedly already know, it really comes down to heat and detonation. Run lots of boost, you get extra heat, build up heat and the hydrocarbons in gasoline break down into auto-igniters and you get detonation.
Fundementally, I think of meth injection as a variation on water injection. We're looking to effectively increase resistance to detonation and have a cooler flame front. But, of course, meth has some potential advantages. For one thing, it has BTUs it can contribute. Also, not only is the flame front cooler, but potentially slower and more stable as well. One advantage I overlooked for a long time (until I was looking closely at someone's setup) is that, depending on where injection occurs, the vaporization can also provide a pretty significant level of intercooling.
Now, as SmokedMustang points out, the problem really starts when we put these advantages to work in the tune. He mentions a big one, we can run leaner - but it really isn't a matter of replacing gasonline with methanol, the lion's share of the energy is still coming from the gasoline, it is that we leaning closer to theoretical 'Best Power' than we could without the lower temps and extra resistance to detonation.
But there are other ways we can push performance. For example, we can push boost. We can also take advantage of the slower flame front with timing and so on. I'm bringing this up because if the tune is aggressive enough, the loss of the meth can leave the engine in a situation where there is too much heat and pressure for gasonline alone - regardless of how rich we take the mixture.
It is easy to think of the extra fuel we pour in for cooling as a 'coolant', but that can be missleading. 'Charge cooling', or cooling by evaporization does
occur, but if you look at the latent heats and quantities of fuel involved, it is miniscule, generally just a few degrees C. Most of the cooling with a richer mixture comes from changes to the flame front itself - and there is a very distinct curve to this effect.
I guess what I am taking a long time to say is that 'newbies' (to use SmokedMustang's phrase) is that what the first post meant by 'super lean' isn't necessarily an engine gasping for fuel. It is quite often an engine suddenly running just rich of stoich (pretty much the worst possible place to run an engine in terms of CHTs and probability of detonation) and .85 lambda (AFR of about 12.5:1 with gasoline) - .85-.86 lambda is a great place to generate loads of power, but is waaay too much heat and pressure at WOT for most engines, particular non-normally aspirated ones. In other words, setting the stage for extreme temps up top and detonation, all while the engine is still producing lots of power.
Regarding 'Detection':
Wideband controllers are a very good way to detect a sudden change in meth flow. The reason is that most are inherently lambda meters. To put it simply, the sensor 'finds' stoich and the controller measures the amount of offset. The stoichiometric ratio for methanol and gasoline are dramatically different. Gasoline is nominally 14.7:1, methanol is typically somewhere between 6/7:1. So if that methanol disappears and there is suddenly more O2 in the mix, a wideband controller can potentially detect the change very quickly.
Regarding 'Corrosion':
Methanol is definately corrosive to aluminum and can soften some rubbers and polymers, but perhaps a bigger problem is its affinity for water. It is truly a water magnet and, left in a standing pool, will gladly such enough water from the air around it to turn into a gelatonous goo.
Regarding 'Alarms':
True story, the first time I saw an XD-1 gauge start flashing in alarm mode, I thought it was the most annoying thing I had ever seen. And I was surprised to learn that the user was hooking up an audible alarm as well. When I mentioned this to Klaus and told him it seemed like overkill, he simply asked me "Ever hear of a perfectly working airplane landing gear up?" (we both have a background in aviation).
Suddenly the point was crystal clear. Most small airplanes with retractable gear have indicator lights and warning horns (typically linked to flaps or throttle). Yet a fair number of pilots get so busy and distracted that even though the lights are wrong and a horn is beeping, they don't realize that the gear is up until they hear metal on pavement instead of the squeek of tires.
My moral being - if you install an alarm capable gauge, or even a 'failsafe system', good for you. Just don't install it and forget about it. If you are pushing the edge and using something like meth injection to help you do it, try to train yourself to scan the relevant instruments as part of your driving (pilots use checklists and repetitive accronyms like "GUMPS" to help insure proper gear settings for landing). That way, alarms and etc. become a 'backup' to the primary 'failsafe' system.
-jjf
Regarding 'why':
To understand why things go boom, I think it is helpful to recall what meth injection is doing in the first place. As most of you undoubtedly already know, it really comes down to heat and detonation. Run lots of boost, you get extra heat, build up heat and the hydrocarbons in gasoline break down into auto-igniters and you get detonation.
Fundementally, I think of meth injection as a variation on water injection. We're looking to effectively increase resistance to detonation and have a cooler flame front. But, of course, meth has some potential advantages. For one thing, it has BTUs it can contribute. Also, not only is the flame front cooler, but potentially slower and more stable as well. One advantage I overlooked for a long time (until I was looking closely at someone's setup) is that, depending on where injection occurs, the vaporization can also provide a pretty significant level of intercooling.
Now, as SmokedMustang points out, the problem really starts when we put these advantages to work in the tune. He mentions a big one, we can run leaner - but it really isn't a matter of replacing gasonline with methanol, the lion's share of the energy is still coming from the gasoline, it is that we leaning closer to theoretical 'Best Power' than we could without the lower temps and extra resistance to detonation.
But there are other ways we can push performance. For example, we can push boost. We can also take advantage of the slower flame front with timing and so on. I'm bringing this up because if the tune is aggressive enough, the loss of the meth can leave the engine in a situation where there is too much heat and pressure for gasonline alone - regardless of how rich we take the mixture.
It is easy to think of the extra fuel we pour in for cooling as a 'coolant', but that can be missleading. 'Charge cooling', or cooling by evaporization does
occur, but if you look at the latent heats and quantities of fuel involved, it is miniscule, generally just a few degrees C. Most of the cooling with a richer mixture comes from changes to the flame front itself - and there is a very distinct curve to this effect.
I guess what I am taking a long time to say is that 'newbies' (to use SmokedMustang's phrase) is that what the first post meant by 'super lean' isn't necessarily an engine gasping for fuel. It is quite often an engine suddenly running just rich of stoich (pretty much the worst possible place to run an engine in terms of CHTs and probability of detonation) and .85 lambda (AFR of about 12.5:1 with gasoline) - .85-.86 lambda is a great place to generate loads of power, but is waaay too much heat and pressure at WOT for most engines, particular non-normally aspirated ones. In other words, setting the stage for extreme temps up top and detonation, all while the engine is still producing lots of power.
Regarding 'Detection':
Wideband controllers are a very good way to detect a sudden change in meth flow. The reason is that most are inherently lambda meters. To put it simply, the sensor 'finds' stoich and the controller measures the amount of offset. The stoichiometric ratio for methanol and gasoline are dramatically different. Gasoline is nominally 14.7:1, methanol is typically somewhere between 6/7:1. So if that methanol disappears and there is suddenly more O2 in the mix, a wideband controller can potentially detect the change very quickly.
Regarding 'Corrosion':
Methanol is definately corrosive to aluminum and can soften some rubbers and polymers, but perhaps a bigger problem is its affinity for water. It is truly a water magnet and, left in a standing pool, will gladly such enough water from the air around it to turn into a gelatonous goo.
Regarding 'Alarms':
True story, the first time I saw an XD-1 gauge start flashing in alarm mode, I thought it was the most annoying thing I had ever seen. And I was surprised to learn that the user was hooking up an audible alarm as well. When I mentioned this to Klaus and told him it seemed like overkill, he simply asked me "Ever hear of a perfectly working airplane landing gear up?" (we both have a background in aviation).
Suddenly the point was crystal clear. Most small airplanes with retractable gear have indicator lights and warning horns (typically linked to flaps or throttle). Yet a fair number of pilots get so busy and distracted that even though the lights are wrong and a horn is beeping, they don't realize that the gear is up until they hear metal on pavement instead of the squeek of tires.
My moral being - if you install an alarm capable gauge, or even a 'failsafe system', good for you. Just don't install it and forget about it. If you are pushing the edge and using something like meth injection to help you do it, try to train yourself to scan the relevant instruments as part of your driving (pilots use checklists and repetitive accronyms like "GUMPS" to help insure proper gear settings for landing). That way, alarms and etc. become a 'backup' to the primary 'failsafe' system.
-jjf
First off. This was an EXCELLANT read.
I agree with your points of not tuning more aggressively as the only option. I personally am tuned for pump gas on my 300ZX TT and just run more boost whenever I spray Methanol Injection.
Another safety option is the water injection kits by Snow Performance. I am a retailer of Snow Performance, so I am a little bias but they Snow Kit comes with an optional Safe Injection Module.
The Safe Injection can be wired to a relay or a lead / warning lamp. I know of people who have easily wired it into the power of an EBC and when the system detects low fluid, the user is back done to factory / base boost levels.
Besides having a safety feature trigger a map change, I think this is the next safest use of WI.
Once again , excellant read !!!
The second issue is that the analog output may be too slow. Autronic, of course, designs for their own wideband. And the Zeitronic controller, which has no spec'ed speed, appears quite a bit slower in terms of settle time. This makes sense, and is quite likely a design choice. The last time I checked, Zeitronix was using CAT5 cable and RJ style connectors for connection to the Bosch sensor. This makes for a very small firewall hole, but there are some technical downsides as well.-jjf
If there is no "spec released speed", then you do not have the proper information to even insinuate that speed is a factor, outside of a sales ploy to bash a competitor product.
For the record, for all those who do not know this new December site registrant jfitzpat, he works for Innovate and all responses are subject to bias to the innovate product line.
Evolving Member
Joined: Dec 2006
Posts: 276
Likes: 0
Similiarly, the posts regarding measurement principles and UEGO are quite sound. I see little reason for you to be displeased, since I spent most of them debunking myths about the sensors, and explained how the principles could be applied to any Bosch install, not just ours.
The concern about speed does not truly originate with me, it originates with Autronics. I broached it because the Autronics argument strikes me as technically sound. If the unit is assuming a certain response curve, different dampening and slew could confuse the unit. It also matches our own user experiences, where results were better when analog outputs were reprogrammed to match the Autronics native speed. STILL, I only broached it as a concern. And pointed out that other users who had successfully made the hookup would be a good source of information.
I thought that was the courteous thing to do. Particularly when the threads are technical in areas on which I have some expertise. And it should be clear that I have taken some pains to actively negate any 'sales advantage' that certain technical facts might incur. Still, my decision to curtail seems sound. As I pointed out to the moderators in our offline conversation, my long term involvement here would aggravate you no matter what herculean effort they and I put into sanitaizing and limiting my posts.
Case in point, I simply noted that you have confused analog input with analog output and need to rethink interconnectivity before selling a 'solution' for the specific problem at hand. Although this is indisputably true, you are chest thumping once again.
Try to relax, I am one of two product development engineers at a company that sells, quite literally, tens of thousands wideband controllers and sensors each year. Why I am here at all is easily verifiable. Some people here were trying to utitlize some open source freeware that I have personally contributed to the tuning community. One of them posted about their difficultiies, along with a link to a thread here, in our support forum.
I've already agreed, offline, to give code to Malibu Jack to contribute to his work, on which you appear to rely heavily. Similiarly, I've been putting together some additional open source snippets to help individual tuners here who have contacted me directly with specific needs. And, I'm continuing to contribute to Enginuity, which, while currently Subraru centric will, in time, help the Evo community as well.
And last, but not least, I've conceded that my involvement here can make things awkward for some people, particularly anyone who has played a little fast and loose when it comes to technicalities regarding our product line (or technicalities about certain subjects in general). Since we are not a sponsor, I agreed that this puts the moderators in an intolerable situation. So, if you can just contain youself a little, you win on all fronts - no technical information or corrections disseminated from me, and contributions to some of the free and nearly free offerings which you rely on when promoting your wares.
-jjf
P.S. You've been *so* pleasant to me, that the thought has occured to me that I, personally, should look into becoming a sponsor. If you think you dislike me now...
We do not use MalibuJack's Mitsulogger. We use Evoscan. Nevertheless, he does a lot of work to support the community. Please keep your input factual.
P.S.- It would be a good gesture to mention your relationship to innovate somewhere in your signature in all fairness to forum members just in case a few of the 15,000 members miss the post where you mention that important fact.
P.S.- It would be a good gesture to mention your relationship to innovate somewhere in your signature in all fairness to forum members just in case a few of the 15,000 members miss the post where you mention that important fact.
Last edited by TTP Engineering; Jan 7, 2007 at 08:26 PM.