Cheap VAFC, Boost Controller + (Turbo Timer)
12-24-2003, 03:29 PM
#11
Guest
Posts: n/a
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
i think id rather see a setting for 0-1k and then another for 1-3 or something like that (some people build a lot of boost before 3k especially on the highway)...also, you wont know what your settings are?...explain that more please...basically, minus the other options, you are building the smc
12-24-2003, 06:33 PM
#12
0.5 BAR
Thread Starter
Join Date: Jun 2003
Posts: 178
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
okay,
considering everything, I have decided to abandon the analog AFC. I am going to go digital, I should of went that route in the first place but only recently have I gotten the right microcontroller. tis a bit of a pitty though, there is alot more art in analog design than programming a Microcontroller. Anyways here are the new specs i am shooting for.
tunable across the entire rpm band to 10k by 100rpms and by every 1% of the TPS.
2x16 backlight LCD display.
vtec activation control
Digital electronic boost control
I am even going to do Retarding based on Map voltage. Yes retarding, I will even do retarding based on boost.
---- I will even throw in control of an extra injector
I might do dataloggin but the EEPROM on my microcontroller is tiny, so it will depend on how well the flash memory works.
Displays:
Current Boost
AFC settings @ current RPM
vtec engagement @ current RPM
timing retarding @ current RPM
Current RPM
Current AF ratio
extra injector on/off
I am putting all this ---- on chip because well the Micro controller runs at 20MHz and has 8 A/D and another 20+ digital input/output pins. It would be a pitty not to use the thing to its full potential.20 Mhz may not seem like much these days but consider that 10000 RPM only translates to 166.6667 hz, so the microcontroller is almost overkill for auto applications. If you guys are so crazy about tunability just for spite I might make it tunable by every 50th rpm using the EEPROM memory or use the flash memory and make it tunable for every RPM.
Anything else I can include? Now would this be worth atleast $150 bucks? honestly I do not wanna undertake this project for fun, it is quite a task but if the financial reward is there then I will definitely do it.
BTW how does the TPS come into play with typical AFCs, from what i read the TPS and MAP have practically a 1:1 relationship, does APEXI use the TPS for anything other than display?
considering everything, I have decided to abandon the analog AFC. I am going to go digital, I should of went that route in the first place but only recently have I gotten the right microcontroller. tis a bit of a pitty though, there is alot more art in analog design than programming a Microcontroller. Anyways here are the new specs i am shooting for.
tunable across the entire rpm band to 10k by 100rpms and by every 1% of the TPS.
2x16 backlight LCD display.
vtec activation control
Digital electronic boost control
I am even going to do Retarding based on Map voltage. Yes retarding, I will even do retarding based on boost.
---- I will even throw in control of an extra injector
I might do dataloggin but the EEPROM on my microcontroller is tiny, so it will depend on how well the flash memory works.
Displays:
Current Boost
AFC settings @ current RPM
vtec engagement @ current RPM
timing retarding @ current RPM
Current RPM
Current AF ratio
extra injector on/off
I am putting all this ---- on chip because well the Micro controller runs at 20MHz and has 8 A/D and another 20+ digital input/output pins. It would be a pitty not to use the thing to its full potential.20 Mhz may not seem like much these days but consider that 10000 RPM only translates to 166.6667 hz, so the microcontroller is almost overkill for auto applications. If you guys are so crazy about tunability just for spite I might make it tunable by every 50th rpm using the EEPROM memory or use the flash memory and make it tunable for every RPM.
Anything else I can include? Now would this be worth atleast $150 bucks? honestly I do not wanna undertake this project for fun, it is quite a task but if the financial reward is there then I will definitely do it.
BTW how does the TPS come into play with typical AFCs, from what i read the TPS and MAP have practically a 1:1 relationship, does APEXI use the TPS for anything other than display?
12-24-2003, 07:48 PM
#15
0.0 BAR
Join Date: Sep 2003
Posts: 0
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
maybe im retarded or something....whats the point of this when there are programs like turboedit and uberdata and such?
How much just for the EBC? Dual Setting EBC would be nice....especially for like $40 or something.
How much just for the EBC? Dual Setting EBC would be nice....especially for like $40 or something.
12-24-2003, 09:58 PM
#16
0.5 BAR
Thread Starter
Join Date: Jun 2003
Posts: 178
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
AFC are less hassle and arguably more straight forward, and allow you to do adjustments on the fly. Less of a learning curve and you do not have to muck around with a burner. Plus The rom editors are not compatable with all ODB systems. But overall I agree that a AFC alone is a bit less capable than a well programmed ECU. But if your did AFC and retarding together then AFC should approach the level of a well tuned ECU, but the ECu would still be better. As long as we are talking about what is the best of the best, that would be an upgraded ECU, old honda ECUS must have only a few K of memory, maybe a few thousand operating points are in memory, the rest they just linearly interpolated. Look at state of the art computers, we have cheap *** computers for dirt cheap with 256Megs Ram and CPUS that are in the Ghz. honda ECUS probally have a hard time breaking the Mhz barrier in comparision. Hondata and other systems are not much better, still into the Mhz range with only a few thousand data points in memory. Really my final goal is to use a commercial computer to run my car. But then you have to deal with the computer crashing which would kill the car, but I have a solution for that. Think about it, creating a standard hardware platform for Cars. I am not talking about OBD platform, it is not standard because I can not use my ECU in a toyota. Everyone can tune as they please through software. people can swap C libraries, algorithims, run commercial type mathematical tools to run their car. How bout using cubic spline instead of linear interpolation, how about not needing interpolation because you can store the entire fucken fuel map in ram down to 0.00001 rpm, and have a thousand more on your hard-drive. Haltech charges like 1700 USd for a shitty ECU, for 250 bucks I can get a AMD xp with 256 megs ram and a 30 gb hard-drive. what needs to be designed is a good D/A system and also some hardware/software needs to be designed to protect the car when the blue screen of death appears.
People are so content to pay a few hundred bucks for the extra 4k of EEPROM memory that the hondata system provides. If you charge people a resonable price for standardized hardware and leave it up to tuners to create algorithims and software then everyone should be happy. Honestly I do not call hacking bins and hex files really creating anything of momentum. In the end there is not standard ROM to hack or microcontroller to hack, there just is not enough of base support. Look at unix, see how much of a grass roots support there is for it, why can there not be such support for cars? there should be, even laptops are standardizing their designs. Yes using a 2 GHz AMD processor is super overkill for a car, but ---- would you rather be paying 200 bucks for a 2 GHZ cpu or 1750 for a 4 MHZ cpu? you pay 1750 not for the hardware but for the ---- inside (program), if you have an open platform you will end up getting software from free from other tuners. You pay for what the hardware is really worth.
Sorry I am just off on a tangent, but I am a bit pissed that people here are giving me ----, in the end I am trying to bring more value to the table through designing of an AFC. yes I am looking to make a few bucks, but in the end you pay less and get more.
40 bucks for a EBC is perfectly doable, if people were willing to pay $40 i would make a tonne of them. But you would have to buy your own boost soleniod. I could do the soleniod plus electronics for about 60-75 bucks, if you guys are interested I can do that, dual setting, triple setting, whatever setting. As long as people desire it i can pretty much do it. But just 1 guy saying that he wants it is not enough.
People are so content to pay a few hundred bucks for the extra 4k of EEPROM memory that the hondata system provides. If you charge people a resonable price for standardized hardware and leave it up to tuners to create algorithims and software then everyone should be happy. Honestly I do not call hacking bins and hex files really creating anything of momentum. In the end there is not standard ROM to hack or microcontroller to hack, there just is not enough of base support. Look at unix, see how much of a grass roots support there is for it, why can there not be such support for cars? there should be, even laptops are standardizing their designs. Yes using a 2 GHz AMD processor is super overkill for a car, but ---- would you rather be paying 200 bucks for a 2 GHZ cpu or 1750 for a 4 MHZ cpu? you pay 1750 not for the hardware but for the ---- inside (program), if you have an open platform you will end up getting software from free from other tuners. You pay for what the hardware is really worth.
Sorry I am just off on a tangent, but I am a bit pissed that people here are giving me ----, in the end I am trying to bring more value to the table through designing of an AFC. yes I am looking to make a few bucks, but in the end you pay less and get more.
40 bucks for a EBC is perfectly doable, if people were willing to pay $40 i would make a tonne of them. But you would have to buy your own boost soleniod. I could do the soleniod plus electronics for about 60-75 bucks, if you guys are interested I can do that, dual setting, triple setting, whatever setting. As long as people desire it i can pretty much do it. But just 1 guy saying that he wants it is not enough.
12-25-2003, 08:14 PM
#17
3.0 BAR
Join Date: Jan 2003
Posts: 7,082
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
I think that your hopes are too high. Tunability for every rpm and every % of the tps is overkill, we don't need to tune down to that precision. If you can get everything you plan on into a unit with a digital display for $150 and it all works and is easily tuned then it will be the greatest thing I've seen for the price.
I'm sure a few people would be interested in the ebc, but I don't know how many, you would just have to ask, wait, and see. If you need someone to test this out for you I would be willing to try, but keep in mind my car is non turbo right now.
I'm sure a few people would be interested in the ebc, but I don't know how many, you would just have to ask, wait, and see. If you need someone to test this out for you I would be willing to try, but keep in mind my car is non turbo right now.
12-25-2003, 08:26 PM
#18
Guest
Posts: n/a
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
i would be very interested in an ebc...would it be boost in stages or would it be truly adjustable?....for the 60-75 you mentioned, a group buy would work well imo...let me know and ill test one for you once i build my new manifold
12-25-2003, 10:07 PM
#19
1.5 BAR
Join Date: Aug 2003
Posts: 1,194
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
What family of microcontrollers are you contemplating on?
Converting digital to analog is easy using PWM, the hard part is filtering it to pure DC for sensor inputs, well it requires the right lowpass filter and an output buffer to keep the filter consistent.
What kind of idea did you have for the boost control? a trianglewave oscillator and a comparator or digital stuff?
I also thought about running my car off one of my old laptops and a parallel port interface, but I did not want to have to wait for it to boot up to be able to start the car, and the idea of depending on a hard drive in a vibrating thermocycling highly moist environment didn't sound like such a good idea.
I personally don't like piggyback units much, sucks to be limited by the oem ecu in any way, boost cut, rev cut, error correction and stuff, many even calculate the cranking pulsewidth entirely from coolant temperature (means your car is hell to start with huge injectors, needs tricks like triggering flood clear mode after cranking for a couple seconds), and don't forget that attenuating airflow/MAP signals advances up your spark timing too, requiring correction on that.
Piggyback units are good for "marginal" changes in my opinion, but they'll never be as good as standalone management systems.
Either way it's nice to see people who refuse to pay a load of cash for flashy junk from Apexi.
It's a boring fact that OEM ECUs are loaded with very lousy programs to make power, they're all about sacrificing some power for emissions, keeping the catalytic converter in the right temperature range, etc.
Converting digital to analog is easy using PWM, the hard part is filtering it to pure DC for sensor inputs, well it requires the right lowpass filter and an output buffer to keep the filter consistent.
What kind of idea did you have for the boost control? a trianglewave oscillator and a comparator or digital stuff?
I also thought about running my car off one of my old laptops and a parallel port interface, but I did not want to have to wait for it to boot up to be able to start the car, and the idea of depending on a hard drive in a vibrating thermocycling highly moist environment didn't sound like such a good idea.
I personally don't like piggyback units much, sucks to be limited by the oem ecu in any way, boost cut, rev cut, error correction and stuff, many even calculate the cranking pulsewidth entirely from coolant temperature (means your car is hell to start with huge injectors, needs tricks like triggering flood clear mode after cranking for a couple seconds), and don't forget that attenuating airflow/MAP signals advances up your spark timing too, requiring correction on that.
Piggyback units are good for "marginal" changes in my opinion, but they'll never be as good as standalone management systems.
Either way it's nice to see people who refuse to pay a load of cash for flashy junk from Apexi.
It's a boring fact that OEM ECUs are loaded with very lousy programs to make power, they're all about sacrificing some power for emissions, keeping the catalytic converter in the right temperature range, etc.
12-27-2003, 12:30 PM
#20
0.5 BAR
Thread Starter
Join Date: Jun 2003
Posts: 178
Re:Cheap VAFC, Boost Controller + (Turbo Timer)
okay what about this,
do 3 d mapping for the AFC:
Vmap_out=f(Vmap_in,RPM)
and do 3d mapping for timing based also on MAP and RPM
basically you remap the fuel and timing externally.
these setting would be super imposed onto the existing ECU fuel and timing maps, tunning a in 3d would be with respect to your existing fuel and ignition maps? basically all you would have to do is tap your MAP sensor, RPM sensor and the ignition output from the ecu. should work for any model ECU.
I know it is not the best approach because you can not take into account much of the stuff happing inside the ECU during closed loop, but during open loop when it really matters it should work like a charm.
well orginally the boost controller was done using a comparator so was the vtec activation, add some opamps and some caps with a few resistors plus some voltage regulation circuitry and bam you got it. I was gonna use Analog electornics to also do the AFC thing, but in the end it would of been a nightmare to look at. I would of need 8 tunable resistrs per 1000 rpm, plus 1 for vtec plus another one for EBC. It just would of not been an elegant approach, so I am now going digital, but I wanna build more than an AFC which is why I am goning to make a piggy back unit to remap the fuel and timing.
oh I have another question, using a normal af meter your readings swing wildly up and down, is it becuase during the compression stroke the af meter reads mostly 02 and during the exhasuast stroke the meter reads the "real" af of the mixture? if this is the case, reading the af at certain degrees after tdc should yeild much more coherent readings right? BTW what the ---- is "scramble boost" i read it as one of the features in the fancy EBC?
do 3 d mapping for the AFC:
Vmap_out=f(Vmap_in,RPM)
and do 3d mapping for timing based also on MAP and RPM
basically you remap the fuel and timing externally.
these setting would be super imposed onto the existing ECU fuel and timing maps, tunning a in 3d would be with respect to your existing fuel and ignition maps? basically all you would have to do is tap your MAP sensor, RPM sensor and the ignition output from the ecu. should work for any model ECU.
I know it is not the best approach because you can not take into account much of the stuff happing inside the ECU during closed loop, but during open loop when it really matters it should work like a charm.
well orginally the boost controller was done using a comparator so was the vtec activation, add some opamps and some caps with a few resistors plus some voltage regulation circuitry and bam you got it. I was gonna use Analog electornics to also do the AFC thing, but in the end it would of been a nightmare to look at. I would of need 8 tunable resistrs per 1000 rpm, plus 1 for vtec plus another one for EBC. It just would of not been an elegant approach, so I am now going digital, but I wanna build more than an AFC which is why I am goning to make a piggy back unit to remap the fuel and timing.
oh I have another question, using a normal af meter your readings swing wildly up and down, is it becuase during the compression stroke the af meter reads mostly 02 and during the exhasuast stroke the meter reads the "real" af of the mixture? if this is the case, reading the af at certain degrees after tdc should yeild much more coherent readings right? BTW what the ---- is "scramble boost" i read it as one of the features in the fancy EBC?
