Hello from Milltucky Mi

ecologicalleds · Post by **ecologicalleds** » Tue Mar 17, 2015 10:02 pm

Hi everyone. I have built myself a nasty little 02 grand am with a hybrid 3400/3500 v6. I have been using the stock ecu but been tuning it with dhp power tuner and hpt tuner. I have hit a huge wall with this setup and believe I'm in need of rusefi. Gm programming sucks and very limited. I know this work in progress, but I would love to help if I can.

A little on my car. 98 montana van 3400 v6 block, 3.4 lumina 24v .04 dome pistons ( gets me over 16:1 compression), hand made copper head gaskets(to hold compression), head studs, fully ported and polished 3500 v6 heads(100hrs+ of labor time), ls6 valve springs, 1.9 ratio roller rockers, roller cam(close to stock because of radical rocker ratio), powder rods, steel crank, roller timing chain, 100psi oil pressure at idle, ported and polished 3500 lower intake and upper 3400 intake, 80mm throttle body, ls6 maf, stainless headers, 3" true exhaust, coil on plug(my design), bobcat oil cooler(keeps engine real cool and adds 2 QTS oil), 48lb/hr injectors, iridium plugs, I'm sure I missed a lot. I run 93 octane gas and frankenslut will push over 350 hp. I do not beleive in or use nitrous. Will not turbo it. Oh yeah, I put a 5 speed manual and lightened flywheel in it too!

Yes it is 16:1 compression, stil 375-380 psi at cranking with 20,000 miles on it!

It has one cam sensor, which I believe is 1x. Two crank sensors, one is 7x and other is 24x. I am no way a programmer, but have hacked around the ecu, could rusefi support this style of cam/crank sensors? I need individual injection, individual fired plugs, 8,000 rpm, and excellent fast knock detection. I don't have room for much error with this powerplant, but I believe this would be the cats ass, up and running.

What about using a 7" touch screen on the ecu? I played with arduino led controller with a color tft 7" on it. I don't know about the screen being fast, but would be neat for user interface.

Post by **AndreyB** » Tue Mar 17, 2015 10:32 pm

Welcome!

That's about 100hp per liter and a large engine, this car should fly!

ecologicalleds wrote: It has one cam sensor, which I believe is 1x. Two crank sensors, one is 7x and other is 24x. I am no way a programmer, but have hacked around the ecu, could rusefi support this style of cam/crank sensors? I need individual injection, individual fired plugs, 8,000 rpm, and excellent fast knock detection. I don't have room for much error with this powerplant, but I believe this would be the cats ass, up and running.

So far we did not even need 3 channel cam/crank signal - I'd say cam + the 24x crank should be good enough? If really needed 3rd channel could be added eventually. Knock detection is still work in progress

ecologicalleds wrote: What about using a 7" touch screen on the ecu? I played with arduino led controller with a color tft 7" on it. I don't know about the screen being fast, but would be neat for user interface.

with larger screens it mostly about having the memory to keep the picture. 640x480x4 bytes is about 1Mb of data, and our stm32f407 only has 196Kb. We have an option to use external RAM, but at the moment the shortage of software developers is becoming a bottle neck already, even without a screen.

ecologicalleds · Post by **ecologicalleds** » Tue Mar 17, 2015 11:05 pm

Yes, if frankenslut had wings, she would. Yes I aimed for the 100hp per liter. First gear is useless, lol, unless for showing off. She will smolder the 18" tires on demand through the gears. I learned a lot playing with the original ecu, but found the limits already, so I have to settle at a 350hp tune, I feel it can go alot further. Hacking gm code can only get u so far before u run into problems, maybe a full rewrite of gm code, but that's beyond me.

I was wondering about just using the two sensors. What needs to be done for the knock sense, programming and/or hardware? Sorry if asked before, but a lot of reading to catch up.

Post by **AndreyB** » Tue Mar 17, 2015 11:59 pm

ecologicalleds wrote:What needs to be done for the knock sense, programming and/or hardware? Sorry if asked before, but a lot of reading to catch up.

the hardware seems to be "90 %" ready (and I mean it in with irony, the last 10% are the hardest ones) and software is also "95%" ready (same irony here)

see http://rusefi.com/forum/viewtopic.php?f=5&t=778

kb1gtt · Post by **kb1gtt** » Wed Mar 18, 2015 4:14 am

Welcome along.

Have you seen the wiki? http://rusefi.com/wiki/index.php?title=Main_Page

Frankenso, might be a good match for Frankenslut http://rusefi.com/wiki/index.php?title=Manual:Hardware_Frankenso_board

My dream of a dash is something like this http://rusefi.com/forum/viewtopic.php?f=4&t=775

Yeah, knock has a clock starting issue. I think it can be fixed by changing a cap, but haven't gotten to test it yet. Life keeps getting in the way.

Oh speaking of one of those life issues, I think I have the same engine as you. The wife / kid wagon is a Buick Rendezvous, which has recently blown a head gasket. I hope your mod's include a change to green stuff or anything other than DEXCOOL. I'm a victim of that crap. Water pump started sucking small amounts of not noticeable air, which cause it to plug the block, which caused hot spots, which caused an unhappy wife. If I can access the harness, I might be able to get russian a log of the crank wheel that's in this thing. The engine compartment is very tight, not much room to access anything.

ecologicalleds · Post by **ecologicalleds** » Wed Mar 18, 2015 2:21 pm

I seen the hip9011 thread. I'm going to start buying what I need really soon, saving up now. I will try to match what you guys are doing. Frankenso is my thoughts for sure. I want to help and really love what has been done, and look forward to this adventure!

What year of ran-de-vu, lol, is it? Yes dexacool a big no-no! Horrible form arm room in engine compartment, changing back plugs suck, unless from underneath. Pleas, Before u pull heads, do a compression test! I have seen and worked on many that some one else diaged with blown head gaskets, but only bad lower intake gaskets!!! Intake leak will have almost exact symptoms of blown head gaskets(coolant in oil, overheating). 3100, 3400,3500,3900 v6 are really close in design, after 2005 or so they changed some stuff. Use real antifreeze when put back together, goto flush out dexacool. Thermostats have a high failure rate also.

Post by **AndreyB** » Wed Mar 18, 2015 2:37 pm

ecologicalleds wrote:I seen the hip9011 thread. I'm going to start buying what I need really soon

by the way there is a kit @ https://www.tindie.com/products/russian/frankenso-02-diy-kit-/ but you would still need to figure out 0805 componets separately.

ecologicalleds · Post by **ecologicalleds** » Wed Mar 18, 2015 2:42 pm

I use my android tablet for a dashboard right now, torque pro, built my own obd2 adapter. I love it, but there is a split second delay, gm ecu 40 pids/s is only so fast, especially when the car revs up like a crotch rocket. With rusefi i think there is huge possibilities with anroid linked dash. I think a com line would have to be used at or above 115,200 buad and that would take care of delays, plus not having to filter garbage out of the obd2 connector. The wife thinks its soo fun! I'm teaching her to drive it lately, all grins and burning rubber.

kb1gtt · Post by **kb1gtt** » Wed Mar 18, 2015 6:36 pm

I used the blue liquid color changing stuff to determine exhaust gases in coolant. I changed to green stuff some 8 months ago, and yes flushed the hell out of it. However the gum was already there. Do you know any stuff that will dissolve that gum? I've looked and tried my best, but seems GM no longer sells the stuff that actually works. I tend to think that any work like removing the manifold or heads will require engine removal, as it's just to tight in there. So even a manifold gasket is a huge issue I think.

ecologicalleds · Post by **ecologicalleds** » Thu Mar 19, 2015 3:21 am

That gummy crap is rough to get out. Gm stuff to clean out system is hard to find anymore even living near detroit. I do not ever trust that blue stuff, false readings from dexacol contamination(dexacol breaks down into a corrosive strong enough to eat plastic intake gaskets and react with chemicals). Ever mix dexacol and green coolant, that's a mess. Seen that a lot when dexacool first came out. Try good brake/electrical cleaner( the stuff that doesn't eat rubber), I've had to use that in extreme cases. Stinks like hell and definitely need to flush really well after. Yeah, the vans always make me cringe when they need work, but I've worked on much worse. Head job takes a good 8 hours with a car hoist. Intake job about 5 hours.

I haven't gotten to do much research/ reading on the hip9011 yet, was at the junk yard digging for parts. I'm going to check on 7" tft color for the rusefi, the one I had ran off a mega2560 and it has its own sd memory on it for the graphics and stuff. I think its a matter of making graphics and some programming to read the data if I remember correctly, but I am getting old and will have to dig out my notes(been a couple years).

kb1gtt · Post by **kb1gtt** » Thu Mar 19, 2015 10:10 am

About this motor, I have limited time, and I don't want to spend a bunch of $ on it to have it repaired, so I think it's basically a terminal situation for it. I'll still have it around for a while, so I can get some info from it which will potentially help stream line some of the development process.

About HIP knock, most of the software should be written, the issue russian was having was that the XTAL wasn't starting reliably. He has to physically touch the XTAL which is enough of ping to get it started. Once it was started he was able to program it via SPI commands and was able to debug it against ping signals sent from a laptop sound card. I believe the issue is a slight change in capacitance for the Cl cap, that's C166 and C167 on the Frankenso schematic. The HIP datasheet notes 20pF, which is what we originally used, but per this thread http://rusefi.com/forum/viewtopic.php?f=5&t=778&hilit=cl I have changed the schematic to call for 30pF. I need to first verify that I have a board with this problem, then I need to see if the 30pF will remove the issue. It's all a time issue, as noted above 5 hours of work on my car makes it a terminal issue.

ecologicalleds · Post by **ecologicalleds** » Fri Mar 20, 2015 1:30 am

I understand on the van. I have set up my wiring harness for easy access and will prolly build another harness for testing/setting up frankenso. My car has a funny way the ecu controls timing through an ignition module while cranking and under 550 rpm, gives me hell. When it takes over control of ignition module it uses pwm to control advance. I want that gone, so I'll have to make breakout connector for the 7x, 24x, and cam. I think the 3500/3900 use a 24x sensor in the same location at my 7x and no 24x under the front dampner. So I should be able to use just the 7x and cam signal reliably at high rpm? Would the other 24x on my engine be worth using for better timing resolution? Prolly a dumb question, but could a clock signal be sent to the hip from the main processor to prevent the crystal issue?

kb1gtt · Post by **kb1gtt** » Fri Mar 20, 2015 9:47 am

You can use just the 7X and CAM sensor, but if you can get to the 24X sensor that would be better. The issue with the 7X is that you can have a change between the crank pulses. This happens even more with light weight flywheels and high HP applications. AKA low inertia, and powerful combustion makes for variations. These variations will make it hard to predict exactly when to fire. This results in a larger +/- tolerance on your advance, and to ensure you don't detonate, you'll need to back off the timing to make sure you don't fire to early.

So yes the 7X will work, but it will mean you need to back off the timing more than you would if you use the 24X.

ecologicalleds · Post by **ecologicalleds** » Thu Mar 02, 2017 4:06 pm

Sorry for such a long delay.

My really nice msi gt70 laptop died(apparent main board fail) while flashing one of the five Arduino mega2560 on my 3000 watt led system(yes, you will get sun burnt working under them @ 40% power!). Even worse, when she went, it corrupted all three hard drives and my programs for the light system.

That was real fun, rewriting code for four led rails, one power supply, and the main 7" lcd touch screen controller, all from memory. But during that process I was able to get the led megas running at pwm of 62khz. that's including the radio communications, measuring 5 adc channels(128 times each), temperature sensor, three fets, isr routine, pwm fans 31khz, and some other neat functions. Which also turned me onto the stm32f4 drone controllers.

Also started cutting aluminum with my CNC router! HEHEHEHE ..... cant wait to tackle my aluminum v6 flywheel.

My poor car has been calling for me(I stare at her often). Now that I can dedicated time to her, she needs to breathe again. Stepping away from this for soo long, I'm very disappointed in myself.

You guys have came along quite well. I'm really digging the can bus and might use that to get a dash. My car has a one wire data link to the dash and would like to use my own homebrew style.
Junk yard, here I come, for another wiring harness to rework for the frankenslut.

Only thing holding me back it is which way to go on the frankenso.

Should I just buy the v4 frankenso diy kit(in my tindle shopping cart) or go with the v5 and have the pcb made?

I had a quote done for the frankenso 5 and would like to go that route, but their min order is 3 and cost over $370. I am not scared to do smd soldering, I like the challenge. Is there anyone wanting to buy or have an v5 board to sell?

A question for the gurus, I have my own maps from tuning the original pcm in my car, can those be used as a baselines? I'm unsure, is it either map or maf, or both that I can use?

I will be doing some catching up, my bad on the questions that have been answered before elsewhere. Thank you again for developing,doing this, and the time, all of you

Post by **AndreyB** » Thu Mar 02, 2017 4:53 pm

ecologicalleds wrote:Should I just buy the v4 frankenso diy kit(in my tindle shopping cart) or go with the v5 and have the pcb made?

I had a quote done for the frankenso 5 and would like to go that route, but their min order is 3 and cost over $370. I am not scared to do smd soldering, I like the challenge. Is there anyone wanting to buy or have an v5 board to sell?

The price for a four layer batch vs risk vs benefits is a big part of why we are still on v 0.4

An issue was recently discovered with Frankenso 0.5 power supply, it would be at least three weeks until Jared gets new version of power supply piece of the board.

If you want to take the risk you can order Frankenso 0.5 and I will buy a board from you, a safer approach is to order Frankenso 0.4 kit, the nicest thing you can do is to donate towards future Frankenso 0.5 run

ecologicalleds · Post by **ecologicalleds** » Thu Mar 02, 2017 5:12 pm

Thai you for the quick response.

I will go ahead and purchase the v4 kit now, and wait a lil bit on the v5.

I can donate some milling time and a carbide bit on my cnc router to cut some aluminum cases and or covers, would need some 6061 aluminum chunks/plates.

Most likely going to mill an aluminum ecu case eventually, unless there is a demand for a sealed enclosure.

kb1gtt · Post by **kb1gtt** » Thu Mar 02, 2017 7:52 pm

Cool to hear from you.

I recently got a bridgeport, and I'm working to CNC that puppy. So I'm a bit distracted.

I agree, go V0.4 for now. V0.5 is still being developed and I'm not sure if all the bits have been pushed to the repo at this point in time.

stefanst · Post by **stefanst** » Thu Mar 02, 2017 7:57 pm

kb1gtt wrote:[...]I recently got a bridgeport, and I'm working to CNC that puppy. So I'm a bit distracted. [...]

This post is useless without picture of Bridgeport!

kb1gtt · Post by **kb1gtt** » Thu Mar 02, 2017 10:08 pm

To avoid hijacking this fellow thread, I posted the Bridgeport photo here. http://rusefi.com/forum/viewtopic.php?f=13&t=1188

ecologicalleds · Post by **ecologicalleds** » Thu Mar 02, 2017 11:20 pm

thank you

I bought the v4 diy kit.

I found stepper board and some other misc stuff of ebay.

I'm ready for homework now.

Post by **AndreyB** » Thu Mar 02, 2017 11:29 pm

ecologicalleds wrote:I'm ready for homework now.

http://rusefi.com/
http://rusefi.com/wiki/index.php?title=Manual:Software:Simulator if you have a windows PC?

matt · Post by **matt** » Fri Mar 03, 2017 1:04 am

ecologicalleds wrote: I had a quote done for the frankenso 5 and would like to go that route, but their min order is 3 and cost over $370. I am not scared to do smd soldering, I like the challenge. Is there anyone wanting to buy or have an v5 board to sell?

last year i done my frankenso 0.2 is usd110 for 5 pcs. fabricate done in malaysia. for frankenso 0.5 i believe price is should be around that figure, unless malaysian
currency continuously drop.

Rhinoman · Post by **Rhinoman** » Fri Mar 03, 2017 8:36 am

ecologicalleds wrote:gm ecu 40 pids/s is only so fast

That's pretty quick for 'OBD2' if you follow the specification then something like 8.5 PIDs/sec is the maximum, I've done tests with a BT ELM and I was getting 4 PIDs/sec - its quite useless for any kind of tuning. RusEFI has a CANBus link that could run at 1Mb; 11-bit CAN has 54-bits per frame and allowing a maximum loading of 80% on the bus then you are going to be limited to around 14500 PIDs/sec, the limiting factor will be the ability of the software to stream the data. You still need to be sensible about how many PIDs you log simultaneously - I once had a vehicle manufacturer send me a log where they were logging over 400 parameters (using 28-bit CAN) and they were looking for a 10ms glitch!

ecologicalleds · Post by **ecologicalleds** » Fri Mar 03, 2017 1:23 pm

Yes, can bus is the way to go.

I had stripped down the data stream on the obd2 and still was horribly slow on the grand am for logging. I even disassembled the bin with idapro, and did some hacking(automatic trans to manual trans segment swap, fualt code removing, and more). Lol, I have a nice big pile of OEM ecus for the car. I'm done with the OEM computer system.

Post by **AndreyB** » Fri Mar 03, 2017 1:31 pm

ecologicalleds wrote:Yes, can bus is the way to go.

rusEfi has very little CAN-related logic. I did not find any open source implementation of the fancy higher-level protocols, also serial just works for me now.

I am also not aware of an open source tuning app to use with a fancy CAN protocol.

If we want to discuss this further please do that within existing CAN threads on the forum if possible. http://rusefi.com/forum/viewtopic.php?f=5&t=959

ecologicalleds · Post by **ecologicalleds** » Sat Mar 04, 2017 2:30 am

i did not like dealing with the oem datastream, fighting through the unknowns, staring at hex for hours, and finding the limits of the delco. no more of that for me.

i may have mispoke, 40 pids/s, meaning to say 40 parameters a sec(rpm,volts,temp,adv,o2,fuel level,fans, ect).

used the laptop or the tablet as my dash. thinking of using the original dash and the steppers out of that for instrumentation, maybe adding and lcd.

serial data is perfect for me.

work desk is almost cleared off, majority of the parts are on the way.

rusefi.com

Hello from Milltucky Mi

Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi

Re: Hello from Milltucky Mi