+- STD (https://www.superturbodiesel.com/std)
+-- Forum: Tuning (https://www.superturbodiesel.com/std/forumdisplay.php?fid=6)
+--- Forum: Engine (https://www.superturbodiesel.com/std/forumdisplay.php?fid=7)
+--- Thread: Electronic variable vane control (/showthread.php?tid=207)
Pages:
1
2
RE: Electronic doodads - Gross Polluter - 03-24-2009
(03-23-2009, 06:45 PM)300SD81 The MPX4250GP from Freescale Semiconductor makes a great boost gauge too. Its rated to I think, 45psi, and outputs as a 0-5v analog voltage.
The MPX4250 reads 250kpa absolute so it only reads to 21 psig @ sea level. The MPXH4600 will read 40psig, however, it's only offered in a 8-SSOP where the MPX4250 is offered in a 6-SIP.
The -GP is a Gauge sensor, reads to 36ish psig.
We use the same sensors with MegaSquirt.
RE: Electronic doodads - Tymbrymi - 03-25-2009
(03-23-2009, 10:10 AM)300SD81 I've been working on electronic controls for months. I remember posting some stuff at various forums, but I'm getting really close to getting my electronic actuator working now. The electronic actuator that came with my turbo had its own computer on board, and I spent about 2 months trying to interface with it before I decided to scrap it and design my own, so now I'm bypassing all the electronics and directly controlling the motor with my board, using a throttle position sensor off another car to sense vane position. Should have a working prototype within a couple weeks. I already have sensor boards and a USB interface built, so after I get the motor control system done, it'll be time to program the actual controller system...
Sweet!! Keep up the good work! Will you be posting schematics/code? If you aren't that's cool, just wanted to ask.
(03-23-2009, 10:41 AM)GREASY_BEAST Yeah, so I was a little arrogant/foolish about the software... but its no space shuttle either. Basically it all comes from tach sensor input, EGT input, boost input, rack position input, fuel pedal input, and spits out electrons to the actuator controls.I wasn't trying to be an ass or anything, but I work with EE's that think software should ALWAYS take JUST an afternoon, and it gets tiring. So that was more of a knee jerk reaction
Correct, that is the basic premise, but the devil is in the details, and if we're controlling the IP I will be paying a lot of attention to the details!(03-23-2009, 10:41 AM)GREASY_BEAST The "actuator controls" consists of a relay or set of relays that can handle the current the actuator needs (or if you have electronic turbo controls, actuators plural).
Relays will not work. You need something that can do PWM signals at high current (2-3 amps I'm guessing), at a frequency of greater than 100Hz. Relays just don't cut it at that frequency. So power electronic circuits are required.
(03-23-2009, 10:41 AM)GREASY_BEAST It would be nice to also have ambient temperature input and engine temperature input as well to be able to do some cold start programming and additional engine protection. However, transforming the signals into a form that the board can interface with seems like the major challenge. Around May I'll have a 617 on an engine stand. I have a couple cruise servos kicking around. I'll have a board of some sort. Let the games begin.
IMO, sensors are easy. As 300SD81 and Gross Polluter have mentioned, there are already chips that will convert some of the tougher signals to something nice and easy for the microcontroller. I think the hardest part is the actuators and PID control of said actuators. Power electronics aren't *hard*, but they aren't trivial either.
(03-23-2009, 10:41 AM)GREASY_BEAST Oh, and BTW the MW IP for that engine is kaput so it will need new (larger) elements, thereby granting a means to test your theory about shadetree IP calibration
Let us know how it goes!
(03-23-2009, 06:45 PM)300SD81 If you plan on using the board for EGT, might I suggest the MAX6675 (http://www.maxim-ic.com/quick_view2.cfm/qv_pk/3149 ) I'm using it for my EGT probe and its alot more accurate than using the microcontroller's AD inputs.
A friend used this chip on his senior design project, and it worked perfectly the first time and was super accurate. A lot more accurate than the ADC inputs is an understatement!!!
(03-23-2009, 06:45 PM)300SD81 As for my turbo project, I just finalized the design for the vane actuator circuit. I'm using a dsPIC33 microcontroller and H-bridge chip to drive the VNT's actuator motor directly, bypassing the built in computer.
VNT Actuator Board: http://img15.imageshack.us/img15/2223/vntboard.jpg
Nice work! Is that Orcad Layout? You have way more patience than I do!!!! I'm a big fan of Eagle... Much easier to use! Well, for me at least. Everyone here at work uses Orcad layout.
(03-23-2009, 06:45 PM)300SD81 As for circuit boards, the classic laser printer toner transfer method has worked for me for years now. 10 mil traces and 0.5mm pitch parts are possible if your good.
If you want to get yours fabricated we use 4pcb at work (we've done 16 layer, 260*C polyimide, 1 square foot sized boards, etc through them (not all at the same time though!) My next small project will use BatchPCB, which is perfect for small boards. You just can't be in a hurry.
http://www.batchpcb.com/
http://www.4pcb.com/index.php?load=content&page_id=134 (student special)
http://www.4pcb.com/index.php?load=content&page_id=130 ($33 pcb)
The $33 boards can be as large as 60 sqin, which is pretty big!
RE: Electronic doodads - 300SD81 - 03-29-2009
Heres my finalized motor control board, http://img6.imageshack.us/img6/6163/50120720.png
I did end up changing the dsPIC into an 18F series, since I didn't really want to deal with logic translations between 3.3v and 5v. I'm pretty sure 12MIPS at 48MHz should be enough to read from a serial bus and position sensor and set the vane position. My actuator is driven by a motor with reducing gears directly to the vanes, so its pretty simple to control with a H-bridge.
I don't really make schematics when I'm laying stuff out, just tend to do the PCB from chip pinouts, so none are avaliable. I'll post software whenever I get the whole system built, as right now, its just bits of code here and there that I can't finish until I have the boards to debug it on. If anyone else plans on using a PIC microcontroller, I have about 5 years experience working in MPLAB C18.
RE: Electronic doodads - 300SD81 - 04-09-2009
Little update.. All the parts for the boards came in, but I can't make the boards until the next time I go home since my dumba** roommate f***ed up my laminator that I use to make the circuit boards...
I got the USB interface board to talk to the sensor board though, already a step ahead of trying to use CAN bus, where I could not get them to communicate at all... I2C is far easier to debug using my cheapo oscilloscope.
RE: Electronic doodads - oel_brenner - 04-20-2009
so my Holset HE351VE
will be here in a few days..
I guess I have three options for VNT control
1. fully mechanical system with diaphragm actuators
2. system like 300SD is working on.. problem is my turbo actuator motor is a stepper ( I think...conflicting info here) , so I can't use an H bridge to drive the motor ( maybe )..
need more research.. Im more of a software guy...
3. as my turbo is a CANbus unit, I could go whole hog with a CANbus pressure sensor, put it all together with a CAN embedded controller
but this option is clearly the most complex and probably outside of my capabilities.
RE: Electronic doodads - 300SD81 - 04-20-2009
a stepper is very simple to control with a microcontroller, actually easier than the one I have since you don't have to worry about a feedback sensor.
RE: Electronic doodads - ForcedInduction - 04-21-2009
It might be easiest to ditch the entire VGT circuit board and make your own to control the motor how you like.
RE: Electronic doodads - oel_brenner - 04-21-2009
(04-21-2009, 05:45 PM)ForcedInduction It might be easiest to ditch the entire VGT circuit board and make your own to control the motor how you like.
thats probably the plan...
fiddling around with CANbus and making it work with the original setup sounds like
fun.. but I think I have too much on my plate right now.
like install a 47 pound turbo into my 124 !
RE: Electronic doodads - oel_brenner - 04-29-2009
well I took apart the holset controller, and the motor is definitely a stepper
looks like a six wire unipolar unit.
now to figure out what micro-controller to use
get some sensors, and get busy..
RE: Electronic doodads - oel_brenner - 04-30-2009
(04-29-2009, 10:41 PM)oel_brenner well I took apart the holset controller, and the motor is definitely a stepper
looks like a six wire unipolar unit.
now to figure out what micro-controller to use
get some sensors, and get busy..
edit:
well it 'feels' like a stepper..
but according to this info I dug up:
The HE341VE does NOT use the same actuator system as the HE351VE. It is very different. The 341 uses a linear actuator rod and a position sensor for ECM control of the actuator. The 341 has an EXTERNAL actuator position sensor. The 351 doesn't have this at all. I don't think any of the 341 actuator information applies to the 351.
The HE351VE actuator control does not use a stepper motor at all. It's a 3-phase brushless DC motor. There is an ST microprocessor on the board that runs a Freescale motor control chip. The chip controls the 6 MOSFETs you see on the top of the board. You can't see these devices in your partial disassembly because they're underneath the board and you have to carefully desolder the motor pins and external wires in order to take the board off. This isn't easy. Another interesting note, the board is multilayer; it looks like there are at least 5 layers. Not fun to trace.
There are 3 hall effect sensors in a cage around the brushless motor. I haven't figured out what they're for yet but probably for determining relative motor movement. To figure out what the actuator position is.
The motor is geared very high against the actuator crosshaft. One of the gears in the gear set has a magnet on it. I think this may do some position encoding but I'm not sure yet; I don't see any detector that the magnet goes near. It may have been installed in the system but not implemented on board. Perhaps the prototypes did not work correctly.
A serial bus EEPROM is on board (also hidden on the backside IIRC) and may contain some usable program data.
and possibly verified here:
http://delphi.com/manufacturers/cv/powertrain/sensors-actuators/smart-remote/
RE: Electronic doodads - Gross Polluter - 06-05-2009
After years of faithful service from my STK-200 development kit, the ISP dongle took a dump on me. A new ISP costs more than the Arduino board so guess which one I bought?
Sounds like the others on this board that are using them are having fun with them.
RE: Electronic doodads - winmutt - 09-03-2009
Thanks to Tymbrini in this post here : http://www.superturbodiesel.com/std/electronic-actuator-t-614.html
We finally have some headway! I will have some working code samples this weekend (tonight maybe even? I am excited as shit on this). And with any luck maybe some moving vanes by Sun/Mon. Stay tuned!
RE: Electronic doodads - Tymbrymi - 09-03-2009
(09-03-2009, 01:27 PM)winmutt Thanks to Tymbrini in this post here : http://www.superturbodiesel.com/std/electronic-actuator-t-614.html
We finally have some headway! I will have some working code samples this weekend (tonight maybe even? I am excited as shit on this). And with any luck maybe some moving vanes by Sun/Mon. Stay tuned!
Guess I need to take a video of it working on the bench tonight, huh?
RE: Electronic doodads - winmutt - 09-03-2009
If you want sure. My plan is to have a simple map based on boost signal from the boost sensor used for the over boost protection circuit. I'll have to get a new roller switch for the linkage position sensor on the valve cover and pull that in as part of the map as well. Maybe I'll find some LED boards laying around so I can add pretty blinking lights.
RE: Electronic doodads - tomnik - 09-04-2009
So it's boost and pedal-position as input?
I installed a vac gauge (taking the signal to the vac actuator) and boost gauge into my VW TDI to learn about the vane control.
I could see that there is very low, nearly no boost when cruising. The vac is somewhere in the middle range (maybe to avoid surge).
When pressing the pedal the vac immediately jumps to max. so that boost comes up easily. Remaining the pedal somewhere in the mid range (coming from slow cruising) I have the impression that the unit just allows as much boost as stored in the map, way away from max. boost to get the car where the pedal wants it to go, like limited acceleration.
I wonder if boost and pedal are enough and if the pedal signal is an analog signal as 0-10 V or just a switch near WOT or even.
Tom
RE: Electronic variable vane control - winmutt - 09-04-2009
I have some prototype code for the arduino. I spent a few hours last night researching and there is more I need to understand about the different types of PWM. The arduino is capable of the following frequencies :
List of Possible Frequencies:
Quote:For pins 6 and 5 (OC0A and OC0B):There are a total of 6! pins that can produce PWM. I think it is time to buy an oscilliscope.
If TCCR0B = xxxxx001, frequency is 64kHz
If TCCR0B = xxxxx010, frequency is 8 kHz
If TCCR0B = xxxxx011, frequency is 1kHz (this is the default from the Diecimila bootloader)
If TCCR0B = xxxxx100, frequency is 250Hz
If TCCR0B = xxxxx101, frequency is 62.5 Hz
For pins 9, 10, 11 and 3 (OC1A, OC1B, OC2A, OC2B):
If TCCRnB = xxxxx001, frequency is 32kHz
If TCCRnB = xxxxx010, frequency is 4 kHz
If TCCRnB = xxxxx011, frequency is 500Hz (this is the default from the Diecimila bootloader)
If TCCRnB = xxxxx100, frequency is 125Hz
If TCCRnB = xxxxx101, frequency is 31.25 Hz
The processor is a 16mhz RISC processor (thats like 8x faster than my old kaypro2!) and can produce up to 20mips. Plenty of speed for our needs.
The one issue I have is that if I am unable to use the internal timer to generate the frequency then I have to bit bang (use code to manually generate the frequencies) which limits the amount of interrupts (input reads from boost or foot position or tach) in which case 2 boards are needed. There may also be a daughter board (shield) that can be purchased.
Here is a good explanation of what I am doing : http://arcfn.com/2009/07/secrets-of-arduino-pwm.html
RE: Electronic doodads - winmutt - 09-04-2009
(03-23-2009, 06:45 PM)300SD81 If you plan on using the board for EGT, might I suggest the MAX6675 (http://www.maxim-ic.com/quick_view2.cfm/qv_pk/3149 ) I'm using it for my EGT probe and its alot more accurate than using the microcontroller's AD inputs. Your board has a Atmel ATmega168, which I believe has a SPI interface to connect to the chip.
The MPX4250GP from Freescale Semiconductor makes a great boost gauge too. Its rated to I think, 45psi, and outputs as a 0-5v analog voltage.
I have 2 MAX6675ISA+ and 2 MPX5500DP samples in the mail.
This is like christmas in September.
RE: Electronic variable vane control - Tymbrymi - 09-04-2009
(09-04-2009, 08:27 AM)winmutt I have some prototype code for the arduino. I spent a few hours last night researching and there is more I need to understand about the different types of PWM. The arduino is capable of the following frequencies :
If I'm reading the datasheet properly, those are the max frequencies when using those specific clock sources. Look at page 120 of the datasheet, and use "Fast PWM mode". You'll use the 16Mhz clock frequency divided by 8 as the clock source to the timer (TCCR1B = xxxxx010). This determines the frequency of the "ticks" (500ns). Since it is a 16-bit counter, you can have up to 65,535 ticks in the period. 7.1ms / 500ns = 14,200 ticks, which will give you the 140Hz frequency. Use the fast PWM mode where you put "top" in ICR1 (WGM13:0 = 14), and load the 14,200 into ICR1. In order to change the duty cycle you will load a number into OCR1x between 0 and 14,200. The percentage of 14,200 you enter will be the duty factor.
Did that make any sense? The datasheet is still pretty confusing to me...
RE: Electronic variable vane control - winmutt - 09-04-2009
(09-04-2009, 01:33 PM)Tymbrymi(09-04-2009, 08:27 AM)winmutt I have some prototype code for the arduino. I spent a few hours last night researching and there is more I need to understand about the different types of PWM. The arduino is capable of the following frequencies :
If I'm reading the datasheet properly, those are the max frequencies when using those specific clock sources. Look at page 120 of the datasheet, and use "Fast PWM mode". You'll use the 16Mhz clock frequency divided by 8 as the clock source to the timer (TCCR1B = xxxxx010). This determines the frequency of the "ticks" (500ns). Since it is a 16-bit counter, you can have up to 65,535 ticks in the period. 7.1ms / 500ns = 14,200 ticks, which will give you the 140Hz frequency. Use the fast PWM mode where you put "top" in ICR1 (WGM13:0 = 14), and load the 14,200 into ICR1. In order to change the duty cycle you will load a number into OCR1x between 0 and 14,200. The percentage of 14,200 you enter will be the duty factor.
Did that make any sense? The datasheet is still pretty confusing to me...
The part I wasn't clear on was if it was possible to adjust the duty cycle at 140hz. From what I read 140hz could be obtained by picking the near lowest starting freq and then adjusting the duty cycle to hit the frequency, it just wasnt clear if you could adjust the duty cycle then. I will go back and reread, I may have been confusing one of interrupts variables with the one that controls duty cycle. In fact that may very well be it as analogWrite controls with cycle. I was a few beers into a festive mood
. I looked for scopes on CL, nothing cheap.....
RE: Electronic variable vane control - winmutt - 01-22-2011
So I am finally moving ahead on this. I dont know what to do about a display though. I finally found 2 GREAT places for lcd displays on the w124. One in the lower left hand of the window in the corner and the other in the flip down sunvisor in the middle. The other option which I am liking more and more is to use the USB jack to hook up to my droid and write a droid app. Its a bigger screen and can do audible alarms etc etc etc. They are also getting dirt cheap so that you can pick one up used for 40-80 bux. Using a droid would also free up more input pins on the arduino AND allow me to charge the phone off the arduino. Seems like a win win situation.
The best part? I present to you android meets arduino toolkit : http://www.amarino-toolkit.net/
RE: Electronic variable vane control - aaa - 01-22-2011
I will nuke my ashtray.
A droid? Heh, $40 color lcd, nice.
RE: Electronic variable vane control - winmutt - 01-22-2011
Well you dont even have to use a droid but the car mount for it is pretty decent. There is also the possibility to get it to not only act as a systems management console but possibly to play back mp3s over the usb to the arduino to the stereo... So much you can do with it.
RE: Electronic variable vane control - Spaceraver - 01-23-2011
You could always use the outside temp display placeholder for the lcd. I'm thinking a Nokia 6100 LCD fits perfectly in that slot, albeit at bit tall.
RE: Electronic variable vane control - winmutt - 01-23-2011
I thought about this as well. But one line display is not nearly as cool as something that can be interactive.
RE: Electronic variable vane control - winmutt - 01-24-2011
Poo. So it seems that using the USB on the android requires rooting the phone. There is a way to do reverse forwarding with a PC but nothing with the arduino. I posted up on the android-developer list to see if I can get a response.
RE: Electronic variable vane control - aaa - 01-24-2011
I'm sure G1s are dirt cheap as well, and they're easy to root. I think mine sold for $40... not sure if it has USB host support though.
RE: Electronic variable vane control - Spaceraver - 01-25-2011
(01-23-2011, 03:55 PM)winmutt I thought about this as well. But one line display is not nearly as cool as something that can be interactive.
Hmm.. Well I thought about hacking the instrument cluster to get it to fit. Provided I can make it look "stock". A 128x128 lcd that you can get to display colour is cool if it will work with the arduino.
And then you could theoretically add Olefejers complete package of boost and turbo rpm and display everything. If it gets too small, omit something Or cycle things every 3 seconds or so between a lot of things.
RE: Electronic variable vane control - winmutt - 01-25-2011
In the W124 their might be enough room to put a long one line display with buttons on each side where the seatbelt indicator is in the dome light. I will probably end up doing both a LCD and the droid. I ended up getting the blue tooth shield for the arduino, it was $20. The USB host shield would have cost the same. By doing both I can make either optional. One of the many reasons I like the droid is for tracking information. You can combine all the inputs from the arduino with GPS, accelerometer, compass, and tilt angle in the droid and map all the data out. Yes a g1 would do the trick to. Making anything look "stock" is going to be difficult at best. The seatbelt warning though, is almost perfect for it. I also understand that 95 and up have some kind of special header for the cell phone interface. I have yet to see any pics of it tho.
RE: Electronic variable vane control - 300D50 - 01-25-2011
I'm C&Ping this from my other post, in case you didn't see it.
Quote:I've got a small 2x16 lcd, it's ~ 11 point font, overall size of 2.5cmX4cm or so. I can get the part number off it if you want.
<snip>
Electronics Goldmine still sells the displays, here's the link: http://www.goldmine-elec-products.com/prodinfo.asp?number=G16081
Very very small, might end up being too small for easy readability in a car now that I think about it...
RE: Electronic variable vane control - winmutt - 01-25-2011
The problem is that it is parallel and I would want a serial one. Also that is too big. I found sparkfun.com to have lots of stuff, albeit expensive in some cases...
RE: Electronic variable vane control - MTUPower - 01-25-2011
If someone gets around to a unit that will control a VGT that fits a 617 let me know- I'm willing to be a test case.
RE: Electronic variable vane control - aaa - 01-25-2011
Wonder if there'll ever be any point to integrating this stuff into one board. Then you could toss in things like parallel support for less than it costs now.
RE: Electronic variable vane control - winmutt - 01-26-2011
I have a boost sensor however it does not have a nipple on it. I was thinking about gluing a tube to the sensor to make one but I dont think superglue or whatever will hold. Any suggestions?
RE: Electronic variable vane control - tomnik - 01-27-2011
Try to find metal-epoxy it can make higher temps.
Then create some kind of tube to get away from the heat...
Tom
(01-26-2011, 06:28 PM)winmutt I have a boost sensor however it does not have a nipple on it. I was thinking about gluing a tube to the sensor to make one but I dont think superglue or whatever will hold. Any suggestions?
RE: Electronic variable vane control - Syncro_G - 01-27-2011
(01-26-2011, 06:28 PM)winmutt I have a boost sensor however it does not have a nipple on it. I was thinking about gluing a tube to the sensor to make one but I dont think superglue or whatever will hold. Any suggestions?
Neither does my BD X monitor gauge sender. It just uses a short, tight-fit, thick-walled hose. That piece then converts to one of those barbed fittings for nylon or Teflon tubing. It seems fine but I don't like it. I tested it up to 30psi with a mightyvac and it held fine.
RE: Electronic variable vane control - 300D50 - 01-27-2011
Get a picture of it, so we can get a better idea of what there is to work with.
RE: Electronic variable vane control - aaa - 02-01-2011
Along the lines of OEM and olefejer's vacuum-electric setup, I've been eying the EGR transducers. Presumably they introduce varying levels of vacuum to their respective actuators, why not repurpose one for vane control?
RE: Electronic variable vane control - winmutt - 03-04-2011
(02-01-2011, 08:01 PM)aaa Along the lines of OEM and olefejer's vacuum-electric setup, I've been eying the EGR transducers. Presumably they introduce varying levels of vacuum to their respective actuators, why not repurpose one for vane control?
So the transducer is both a senor and an actuator?
Finally got my hands on an oe pressure sensor (or atleast its in the mail). After that I need one more daughter board for the EGT and I can move foreward with my project. I need to work up a display for the sensors. I am thinking a mix of numbers and gauge style in a tiled layout. Time to start reading code forums.
RE: Electronic variable vane control - 300D50 - 03-04-2011
MAX6675 is the best for ease of use in the EGT department, awesome chip. One cap, it does the rest.
RE: Electronic variable vane control - winmutt - 03-07-2011
(03-04-2011, 11:45 PM)300D50 MAX6675 is the best for ease of use in the EGT department, awesome chip. One cap, it does the rest.
I don't remember seeing a cap in any of the diagrams I saw?
VSTech finally pulled the OE boost sensor, should be here this week!
Time to get off my ass and get this all connected electrically and get a proof of concept done.
RE: Electronic variable vane control - 300D50 - 03-07-2011
You need a decoupling cap as close as possible to the chip, across the 5V rail. It is very clearly spelled out in the datasheet, but people assume it's not required...
.01uf ceramic iirc is what's required.