British Columbia DSM
GVRD - BC DSM Club => Specs, Techs & Reviews => Topic started by: Richard Chwiendacz on September 01, 2006, 12:02:34 am
-
So, you're finally tired of hunting down boost leaks and dealing with that nonlinear, restrictive, easily overrun karmann vortex 1g airflow meter, huh?
Good, me too.
Thankfully there is an alternative. If you have an EPROM ECU, you can convert your ECU to speed density, using a MAP sensor to replace your 1g MAS. It may sound daunting, and expensive, but it really isn't when you think about the cost of an SAFC and upgrading to a 2g MAS, or MAFT + GM MAF.
Pros:
- No more clouds of black smoke and bogging from running rich when you vent your BOV
- Boost leaks no longer matter, as far as your fuel system is concerned. Your BOV can leak, your TB shaft seals can leak, your fuel injector seals can leak, your intercooler hoses can pop off even, and as far as drivability is concerned, the car will run great. Of course, you still want to avoid boost leaks, as they slow your turbo spool, limit your max boost/power, create more intake charge heat, and shorten your turbo's life. However, they will no longer throw your A/F ratios or timing off.
- No more intact tract restriction. Superior in this regard to both 1g MAS, 2g MAS, and MAFT setups.
Cons:
- Tuning is a pain, especially by yourself. Unlike the volume or mass airflow measuring methods used by our Karmann airflow meter or a hot-wire airflow meter, we need to know exactly how much air our engine can aspirate at any given rpm and intake manifold pressure. This means we have to make a nice big VE table, with 256 different load points, and this VE table will need to be retuned with every mod you make. Our stock air metering system measures the air entering the engine directly, and can compensate happily for changes in exhaust, cams, headwork, or anything else we do to it. Speed density-based air metering can't do that, and that means we have lots of tuning ahead of us.
A few things you should know:
- The factory boost gauge will now work as a knock gauge. This is much more useful anyway, as everyone at this point should have a real boost gauge. When there's zero knock, the needle will stay at the bottom of the gauge. "0" on the gauge (in the middle) corresponds to roughly 15 knock sum, if I remember correctly. This will greatly help tuning if you don't have a datalogger or wideband O2 sensor.
- To make room for all the new tables and functionality, a lot of the self-diagnostic code has been removed. If something goes wrong, the ECU will still throw a CEL, but you won't be able to pull the codes from it. If you have a datalogger, and you certainly should, it should be relatively obvious what's wrong, though. Any errant sensors will usually make themselves clear just by taking a peek at the logger.
If this hasn't scared you off yet, here's a step by step guide to converting your 1g T/E/L's ECU to speed-density air metering. It's worth it.
Big thanks to "hakcenter" of www.ds-map.net for compiling all of the speed density information into one place and helping everybody out with their tuning.
Also to Nick Sanders (pudhed from dsm-ecu yahoo group) for the speed density implementation itself.
I didn't make this, I'm just trying help show y'all how to use it. :)
-
Step 1: Prerequisites.
Before we even begin to dig into any hardware installation bits, you're going to need a few things:
------------------------------------------------------------------------------------
1) First and foremost, you will NEED a socketed 1G EPROM ECU. No way around this.
Many 1g cars came with EPROM ECUs from the factory (mine did), and if you're lucky enough to have one, all you need to do is to remove it and get it socketed (and likely re-capped while you're at it.)
For information on how to check to see whether your ECU has an EPROM in it, have a look at TMO's page on the subject:
http://www.tmo.com/howto/ecu1g/checkeprom.htm
If you find you have an EPROM ECU, ask around on the board and you'll likely find someone able to socket and re-cap it for you. I did my own, and it's not that hard, but it's delicate work, and I happened to have most of the right tools lying around. It's not for the faint of heart.
---------------------------------------------------------------------------------------
2) You're going to want a datalogger. This isn't absolutely necessary, in the strictest sense, but trust me when I say you should have one.
Grab yourself a compatible palm and a serial cable, and you're off to the races. You don't need to spend the bucks on a TMO or Pocketlogger setup.
I've already made a complete guide to making/acquiring a 1g datalogging cable, so if you want, follow that, and just get a Palm.
http://www.labattz.com/phpBB2/viewtopic.php?t=14717
As of August 2006 when this was written, XS Cargo in Vancouver (www.xscargo.com) has Palm IIIX/IIIXe's that work fine for $29.99 to $39.99. The IIIXe has 8mb RAM instead of the 2mb of the IIIx, so is better suited to longer datalogs.
-----------------------------------------------------------------------------------
3) You're going to need some way to give the ECU the speed density code to run on. There are a few ways of doing this:
a) Cost: $10 or so and maybe a 6-pack.
If you're really really cheap, you can get someone to burn you a chip with a basemap that you will hope is close enough to your setup to work properly. Then, of course, you're stuck with it, however well it works. This is far from ideal, and I strongly discourage this. However, a blank EEPROM is only about $5 or so, and less in quantity, and burning it costs nothing if you have a friend or easily coaxed acquaintance with a burner.
b) Cost: $40 to $100.
You can buy your own EEPROM programmer, from eBay, moates.net, or any of a million different places, along with a spare EEPROM or two, and burn your own chips. This means, of course, that you can change the code if you decide you need to (and you will need to), but each time, you'll have to take the cover off your ECU, yank the chip, and reflash it. This is a royal pain in the ass, but if you're cheap, it works.
c) Cost: $175usd, plus shipping
You can get yourself an Ostrich, which is a realtime EEPROM emulator. It has a USB port on it, and hooks up to your ECU's EEPROM socket with a nifty little ribbon cable, and allows you to change any little bits of code on the fly. With this, you have almost all the functionality of a "real" aftermarket EFI computer, like a haltech E6K, AEM EMS, Electramotive TEC2, Megasquirt, and the like. This is what I have, and this is what you want. Accept no substitutes.
http://www.moates.net/ostrich-real-time-ecm-emulator-p-57.html
------------------------------------------------------------------------------------------
4) You're also going to need two sensors -- a GM 3BAR map sensor, and a GM IAT (intake air temperature) sensor.
Thankfully, these are relatively inexpensive.
a) The GM 3BAR map sensor is easily acquired from any MSD ignition parts dealer. Simply ask for MSD part number 2313. Mopac seems to be the cheapest place in the world for these, they're only $59.42cdn, which is cheaper than the cheapest place on eBay, the cheapest GM parts counter, or anywhere else I could find.
If you feel like getting ripped off, you can go to a GM parts counter and ask for part number 12223861, which should be a 3bar map sensor for an '89 turbo Trans Am, but you're better off at Mopac.
It'll look like this, although the little connector will likely be a different colour.
(http://www.msdignition.com/fuel/2311_big.jpg)
Unfortunately, you don't get the connector or wiring harness for it, and MSD doesn't seem to sell them.
In GM's infinite wisdom, they have also keyed the connectors for their MAP sensors, so you can't plug a 1Bar connector into a 3Bar or 2Bar sensor, and vice versa. Thankfully, a clever person can take an exacto knife and cut the right notches out of the 1bar connector and make it fit happily into the 3bar map sensor. There are millions and millions and millions of 1bar map sensors and connectors in any junkyard. Just find a 90's era GM car, with a sensor that looks like the MSD one above. Yank the connector along with a good bit of the wiring harness, and cut the slots in it to make it fit into your 3bar sensor. This will range from free to probably about $50, depending on which junkyard you happen to have the misfortune of visiting.
b) You will also need an industry standard GM IAT (intake air temperature) sensor.
These can also be ordered from mopac or any MSD dealer, by asking for MSD part number 2320. They even are nice enough to give you the connector this time.
(http://www.msdignition.com/fuel/2320_big.jpg)
However, these sensors were also used on a lot of GM vehicles, and while you're out at the junkyard grabbing your harness/connector for your map sensor, you might as well get one of these too. They tend not to fail with age, and mine's from the junkyard and works great. Then again, I'm cheap.
Most GM vehicles use a plastic-bodied sensor that looks like this, and it's just pressed into the rubber air intake tube. These have the same resistance characteristics as the metal sensor shown above, and will work in a pinch, or for temporary testing. I make no guarantees as to how well they'll hold up to high boost levels or heat though, because after all, they *are* plastic.
They'll look like this:
[img]http://www.ls1howto.com/content/partnumbers/pics/12160244.jpg[img]
A select few GM vehicles do use the metal-bodied, open-element air temperature sensors that look exactly like the MSD one pictured above. Try and find one of those if you can -- they're what you really want. I grabbed mine from an Oldsmobile that had the 2.4L Quad Four engine, and it seems most of the Quad Fours have the good sensor, and it's really easy to get to, right on top of the intake. Just make sure to bring a 19mm deep socket with you, or you'll have a hell of a time squeezing a 19mm wrench in there to get it out.
Oh, and just so you know, the connectors are the same between the two, and they aren't keyed. They're completely interchangable, electrically.
--------------------------------------------------------------------------------------
5) You need to ditch your EGR. If you're think you're hardcore enough to do this, you've probably already ditched all your emissions equipment, but if not, you will have to disable/remove your EGR system.
ECUs that use air volume or mass to determine how much fuel to inject don't need to care, for the most part, what the EGR system is doing. Regardless of whether the EGR valve's open or not, the ECU is monitoring how much air flows into the system by measuring it before it hits the intake, and if a little extra exhaust gas and a little less intake charge go into the cylinders, it still knows how much fuel to inject.
For a speed density setup like this, we aren't measuring the number of air molecules entering the engine directly. Instead, we're inferring how many should be going in, based on the pressure in the manifold, and rpm. Unfortunately, our inference is based on the assumption that everything in the manifold contributing to that pressure is air, with a specific oxygen content. When the EGR valve opens and starts letting inert gases into the intake, this is no longer the case, and throws all our fuel calculations out the window.
So, for the greater good, disable your EGR system.
This can be as simple as yanking the vacuum line going to the EGR solenoid from the intake manifold and capping the manifold nipple off, or you can take the entire EGR valve off and install a block-off plate if you're feeling ambitious. Up to you.
http://dsmloggers.com/DSM/taboo/emissions.htm
-
Step two: Sensor installation (electrical)
----------------------------------------------------------------------------------
So, if you've been following along, you now have shiny new (or used) sensors, and we have to splice them into our DSM's wiring so that our ECU can see them.
You need to decide first, exactly how you're going to do the wiring -- each way has its pros and cons, so think carefully before you go ahead and do it.
1) The first, and easiest way, is to purchase a pre-made harness online. This is simple, looks good, and is impossible to screw up. It also lets you simply plug your stock MAS back in, stick the stock code back into your ECU, and drive your car like it was stock, if you so wish.
(http://www.fullthrottlespeed.com/mmFTSS/Images/050sd1g.jpg)
http://www.fullthrottlespeed.com/itemdesc.asp?ic=050SD1G&eq=&Tp=
It's not cheap, though, at $39.99usd, plus shipping, taxes, and brokerage fees.
2) The second way is a bit harder, but will likely cost you less than buying the pre-made harness, looks almost as good, and still lets you retain your factory MAS wiring so if you screw something up, you can revert to stock in about five minutes. You can make your own harness just like the above if you find a donor 1g MAS and cut/hack the connector out of it. Just solder the wires to the right pins, and you can unplug it at will and plug your old MAS back in.
Wiring diagram to follow.
3) Third way is to strip back some of the factory split loom tubing and electrical tape near the MAS connector, and solder or quick-tap your wires for your MAP and IAT sensors into the factory wiring. This is the method I chose, but it's not the best for two reasons:
- This allows you to have both the factory Baro sensor AND the MAP sensor hooked up at the same time. If you do this, you will probably kill the Baro sensor, the MAP sensor, or both. Do NOT do this. This also allows you to have the IAT sensors both hooked up at the same time, but that won't hurt anything, although it will throw your IAT value completely out to lunch and the car probably won't run right.
- This also creates leaves extra wiring in the engine bay, that's unterminated, and could be susceptible to picking up noise (from the ignition, alternator, or whatever.) In practice, it likely won't make a difference, but it's something to think about.
4) Fourth way is just to hack off the MAS plug from your wiring harness and wire your new MAP and IAT sensors into the harness directly, and throw your stock MAS completely in the garbage. I advise people not to do this, especially if they're new to this sort of thing and the car is their only means of transportation. It's always good to have a plan B.
-----------------------------------------------------------------------------------------------
If you chose one of the latter 3 methods, you're going to be doing the wiring yourself. It's simple, and there are only five wires.
(http://www.vfaq.com/proj-pics/MAF/1G-2G/1G-plug.gif)
- MAP sensor:
The wires are labeled, both on the MAP sensor itself and the connector, at least on every one I've seen.
Port A on the MAP sensor is ground, and should be connected to pin 6 on the 1g MAS plug.
Port B on the MAP sensor is the actual 0-5v sensor output, and should be connected to pin 7 on the 1g MAS plug, which normally goes to the output of the barometric pressure sensor in the MAS itself.
Port C on the MAP sensor is the +5v power to the sensor, and should be connected to pin 4 on the 1g MAS plug.
- IAT sensor:
This is just a resistor (well, thermistor, technically) and polarity doesn't matter. Hook one of the wires up to pin 8 on the 1g MAS plug, which normally goes to the internal air temperature sensor in the MAS, and the other wire up to pin 6. (sensor ground)
-
Step Three: Sensor installation (physical)
--------------------------------------------------------------------------------
Okay, now that we have the wiring completed, we need to actually fit the sensors into the engine bay. This is pretty easy, thankfully.
1) MAP Sensor:
- Find a nice convenient spot to put the sensor itself, and secure it. The MAP sensor has nice little holes you can use to either screw it to something metal, or put zip-ties through.
- The MAP sensor needs a vacuum/boost signal. The best and easiest place to find this is to tee off the line going to your BOV. It's larger in diameter than all the other vacuum hoses in the engine bay, and fits perfectly onto the MAP sensor. Just get a 1/8" vacuum tee from any hardware store, and tee off the BOV vacuum line.
2) IAT Sensor:
- If you went really ghetto and got the plastic sensor against my recommendation, it should be pretty obvious what you have to do. Punch a hole into your (rubber or metal, stock or otherwise) intercooler piping, somewhere after the intercooler, and cram the sensor in, along with a good helping of JB weld. Then, hope the sensor doesn't crack, break, or get blown out.
- If you got the proper metal sensor from a Quad Four like I told you to, you'll notice that it's threaded. It has 3/8" NPT threads, and so you'll want to attach a 3/8" NPT female port to your intercooler piping somehow so you can screw it in.
If you have aftermarket pipes, you can get a 3/8" NPT steel bung welded in. This is strong, and won't leak, but for whatever reason steel NPT bungs are way more expensive than they should be, plus you have to pay someone to weld it in if you don't have a welder.
If you have an aversion to paying more than $10 for a little piece of steel (like me), there's a much easier and cheaper route. Go to any hardware store (I went to princess auto) and get a 1/2" NPT to 3/8" NPT bushing. I got mine for $1.79 at Princess Auto, and in 20 minutes, had it nicely installed in my stock BOV pipe.
I took my trusty RTX tool and cut the externally threaded part off, which left me with the perfect thickness to JB weld into my intercooler piping. I drilled a hole in my stock BOV pipe, and then JB welded the remnants of the bushing into it. After 20 minutes, it was good and strong, and the IAT sensor cage stuck into the pipe *just* the right amount. I couldn't have been happier with the way it turned out. It doesn't leak at all.
Total cost: $1.79, or add an extra $5 if you don't have any JB weld laying around.
(http://img1.putfile.com/thumb/9/24304551152.jpg) (http://www.putfile.com/pic.php?img=3297973)
(http://img1.putfile.com/thumb/9/24304545055.jpg) (http://www.putfile.com/pic.php?img=3297971)
(http://img1.putfile.com/thumb/9/24304542874.jpg) (http://www.putfile.com/pic.php?img=3297970)
-
Step four: The first start-up
--------------------------------------------------------------------------------
So, now that you have everything all wired up, installed, and ready to go, now's the time to go through our pre-flight checklist. So far you should have:
- An EPROM ECU, and a way to put new code on it (preferably an ostrich)
- Hopefully, you should have a palm and a working datalogger cable.
- Removed EGR
- Physically installed and wired IAT and MAP sensors
- Unplugged your stock airflow meter if you left it attached to the harness.
At this point, it's time to upload some shiny new speed density code to the ECU, and power up the ECU to make sure everything's okay in the datalogger.
1) Setting up the datalogger.
-----------------------------------------------------------------------------------
First things first, grab yourself a copy of the speed density project's datalogger, called PocketDyno, and install it onto your palm.
You will have to register for the ds-map.net forums before you can see the files and download them, and if you're considering doing this, you ought to do that anyway, since it's the best place to get help.
http://www.ds-map.net/phpBB2/viewtopic.php?t=18
Once you have PocketDyno installed, connect your logger cable to your ALDL connector and Palm, and start it up.
You should see this screen:
Click on the little icon that looks like, for lack of a better description, a piece of paper with writing on it.
(http://img1.putfile.com/thumb/9/24313563221.jpg) (http://www.putfile.com/pic.php?img=3299737)
After that, you'll want to click on the icon in the bottom right corner that looks like a big "P". You should be looking at a screen where you can pick which items you want to view/log.
Pick some good ones, particularly MAP and IAT, to make sure your sensors are working, and click OK.
(http://img1.putfile.com/thumb/9/24313565439.jpg) (http://www.putfile.com/pic.php?img=3299742)
You should now be looking at the viewing screen, and as soon as you turn the key in your car to the Run position (without starting it) the ECU will get power, and the datalogger will start to work. Don't mind the numbers in my screen below, I just made some random ones up. No, my car doesn't have 19 counts of knock at 468 rpm.
(http://img1.putfile.com/thumb/9/2431357279.jpg) (http://www.putfile.com/pic.php?img=3299747)
Keep the palm on the viewing screen, turn it off, and set it aside for now. When the time comes to test the car, just hit the power button and the logger will pop back up and still be ready to go.
--------------------------------------------------------------------------
2) Putting the speed density code into the ECU
If you've been following me so far, you should have an ostrich, or at least a chip burner or friend with one.
Grab yourself a copy of the speed density code, found here. I suggest grabbing one of the wideband ones. If you have an Innovate LC-1, grab the NTK version, it'll work great. I presume the non-wideband version works too, but I could never get it to work, so I just forced open-loop in the wideband version for tuning. More on that later.
Speed density ROM images (http://www.ds-map.net/phpBB2/viewtopic.php?t=5)
Upload the image to your ostrich (which should be connected to the ECU, of course), or burn it to a chip and stick it in your ECU.
Now comes the moment of truth. With the datalogger on, connected to the car, and set to view, and the MAP and IAT sensors hooked up and the airflow meter disconnected, turn the key to the run position, but don't crank the engine over.
You should see the datalogger come to life, and if everything's working properly, your MAP value should be pretty close to 1.0 Bar, and your IAT should be somewhere close to what the outside air temperature is if your car's been sitting for a while and is cold. The factory boost meter should go up to the middle like it normally does before you start the car, and if all these things are correct, then congratulations. You didn't screw anything up.
If you hit the key and crank the engine over, it might even actually start, depending on how close your car is to the default config.
It probably won't, and in which case, we need to fire up DSMap and start doing some editing and custom tailor the code to your car.
-
Step five: Fun with DSMap.
--------------------------------------------------------------------------------------
The next order of business is to download and install a copy of DSMap, and its associated .cfg file so we can edit the .bin file and change our VE tables, injector size, and all sorts of other goodies.
Download a copy of DSMap at the link below -- as of the writing of this article (Sept 1, 2006) the latest version was 0.03 stable.
DSMap Program Download (http://www.ds-map.net/phpBB2/viewtopic.php?t=16)
You will also need a copy of the latest .cfg file for DSMap. If you followed my advice and downloaded the wideband code, grab the wideband .cfg file, from the link below. As of Sept 1, 2006, the latest one was called "wideband beta"
DSMap .cfg file download (http://www.ds-map.net/phpBB2/viewtopic.php?t=17)
Put the copy of dsmap.exe and ecu.cfg you just downloaded in the same directory as each other, and fire up DSMap, and open the image you downloaded in step four. (the wb_ntk.bin file, if you're following along with the wideband code)
At this point, if the .cfg file is in the right spot, you should see a screen that looks something like this:
(http://img1.putfile.com/thumb/9/24314404881.jpg) (http://www.putfile.com/pic.php?img=3299884)
If so, great. Now we can start changing parameters to suit your vehicle.
In DSMap, use "home" and "end" to decrease or increase a value by one step, and "pagedown" and "pageup" to increase or decrease a value by sixteen steps at at time.
You cannot actually click on the box and type in the value you want with the number keys, and this makes sense if you understand a little about what's going on behind the scenes.
All DSMap is doing is changing numbers at a specific spot in the binary file that you open. In the ROM, numbers aren't stored as "29.4" or "450", but rather as an 8-bit binary value from 0-255 (because 8 bits gives you 256 possible discrete values) As such, each number has only 256 different values it can take on, and there's no way of representing numbers in between those. When you hit "home" or "end", DSMap is just incrementing or decrementing the raw binary value by 1, and then showing you what that value represents in something that you can understand, like "449cc" or "20 degrees." Because of the way things are stored in the ECU internally, it doesn't make sense for DSMap to allow you to enter values directly, because if you typed in, say, 451cc for your injector size, that would actually land between the only two values available, which work out to 449cc and 455cc. DSMap would just have to round it to the nearest step anyway, which would probably frustrate people even more, and cause more headaches and complaints.
So now you know why you need to increment and decrement the values instead of typing them in. Not that you really needed to, though.
Right, with that out of the way, the first order of business should be to get your injector size and deadtime values right, because without these, you don't have a hope in hell of being able to tune your vehicle.
Use the keys as described to set your injector size, and for the deadtimes, look up which injectors you have in the little database that ds-map.net so kindly provides, and add whichever deadtime number is shown there to each cell in the deadtime map.
For stock injectors, the deadtime map should look like this:
(http://www.ds-map.net/images/stock_dt.jpg)
For aftermarket injectors, simply add the deadtime number in the database to each cell.
Example for FIC 950cc injectors:
(http://www.ds-map.net/images/new_dt.jpg)
Now, if you downloaded the wideband code, but aren't using a wideband O2 sensor, or you don't know whether your wideband has the same output characteristics as the AEM or NTK sensors, then you need to disable closed loop mode, so that the ECU ignores the wideband sensor. You can still log the value, and watch it on your gauge if you have one, but the ECU won't take it into account for its fuel calculations.
To do this, go to the "CL TPS" map, and set all the values there to "0". This effectively disables closed loop mode whenever the throttle position goes above a given value, which in our case, is zero, disabling closed loop all the time.
From here on in, it's just a matter of tuning. Adjust your VE tables as necessary, and have fun and try not to blow anything up.
-
For completeness' sake, I'm going to do my best to explain what each item in the editor does, in case it's not obvious.
Fuel:
This map exists in the stock DSM ECU code as well, and represents what Air/Fuel ratio target the ECU wants, as a function of load and rpm. Since we've converted to speed density, our "load" is not grams/rev of airflow like with the stock 1g MAS, but instead a MAP value.
The ECU simply multiplies its calcluated fuel value by how many percent away from stoich (14.7:1) the value in the fuel table is.
For example, if you program an AFR target of 12.0 in a cell, then when the ECU sees that combination of map and RPM, it will try to give roughly 20% more fuel than it would normally need to stay at stoich.
--------------------------------------------------------------------------------------------
Timing:
This should be relatively self-explanatory. This works the same as it did in the stock ECU, with the only difference being that your "load" value is now manifold pressure instead of grams/rev of airflow.
--------------------------------------------------------------------------------------------
Soft VE:
This is an easy way to make big changes to your VE map. Ideally, you would have a proper Hard VE map, and a well setup fuel table, and you'd be able to leave your soft VE table at 100%.
However, if you want to quickly and easily add, say, 10% more fuel at a given boost level, at ALL rpm, just change the value from 100% to 110%, or 110% to 120%, and so on. This just acts as a multiplier for each MAP row in the hard VE table.
--------------------------------------------------------------------------------------------
Hard VE:
This is where we get into the meat and potatoes of tuning, and this is what will give you the most headaches. This is a representation of how much air your engine can aspirate at a given RPM and manifold pressure when compared to its theoretical limit.
You will spend a LOT of time fiddling with this, tuning out rich spots, lean spots, dead spots, and weird spots.
Ideally, your soft VE should be 100% across the board, and you should be able to tune the car such that the value in the "Fuel" table is exactly the AFR you see in the log on the wideband, and the Hard VE table should contain whatever numbers are necessary to make that happen.
Most people will fudge their numbers though, and this won't be the case. That should be what you're shooting for, though.
--------------------------------------------------------------------------------------------
Idle:
This works exactly the same as with the stock ECU -- it just raises and lowers the idle speed targets based on coolant temperature. If you have hairy cams that don't like to idle well, feel free to raise your idle speed here.
--------------------------------------------------------------------------------------------
Injector:
Self-explanatory. This is how big your injectors are, in cc/min.
--------------------------------------------------------------------------------------------
Deadtime:
Also pretty self-explanatory, this is how much time it takes for your fuel injector to open, as a function of battery voltage. See step five for how to get this right for your setup.
--------------------------------------------------------------------------------------------
CL TPS:
This map shows what throttle position is required to take the ECU out of closed loop mode and put it into open loop mode (ignoring the O2 sensor)
For tuning, set these all to 0% so you're in open loop all the time and your oxygen sensor feedback isn't messing with your Air/Fuel ratios.
--------------------------------------------------------------------------------------------
Dwell time:
Best to leave this alone. Even the guy who helped write this program admits he can't figure out if it does anything.
--------------------------------------------------------------------------------------------
Rev limiter:
This allows you to edit the rev limiter value in the code, although it's in hex, rather than decimal.
7500 –– F4
8000 –– D5
8500 –– B9
9000 –– A0
9500 –– 8A
10000 – 77
If you want a value other than these, just take a calculator, and divide 3,750,000 by the rev limit value you want. This will give you a number, which you will need to round to the nearest whole number, and convert to hexadecimal.
Use a scientific calculator (one comes built into windows) and it'll do it for you.
Example: 7650rpm rev limit:
3,750,000 / 7650 = 490.1
We want to round this to the nearest whole number, which is 490, and convert to hex, which gives us "01EA" - you take the last two hex digits, which are EA, and put that in as your rev limit value.
--------------------------------------------------------------------------------------------
Sbox
This works the same way as the rev limit, except the value here is used as the rev limit when the car is stopped, giving you a stutterbox launch.
-
ohh man, this is awesome.
Just a question though. It says we need a 3 bar map sensor but the max tunable boost in the rom editor program is 2 bar (14.8psi). Does the editor have "extended" maps for higher boost? I haven;t checked out the software yet, but i will be 2nite!
Props on a hawt writeup!
phil
-
Don't forget that, like the stock ECU, while the fuel enrichment map may end at a certain point, the ECU will keep injecting more fuel to compensate for higher boost levels.
So, for example, if you are at 2.50 bar, and the "FUEL" map ends at 2.0 bar, then from 2.0 bar all the way to 3.0 bar, your fuel enrichment from the FUEL map will stay constant.
If you're comfortable with running the same A/F, say 11:1 at both 2.0 and 2.5bar, then great - you don't have to do anything except have your VE table correct and the ECU will keep the same enrichment at 2.5 as it will 2.0 bar.
If you need more fuel at higher boost, which many of us do, you can add it either via the Soft VE table (increasing the fuel multiplier for that boost level, which is the best option) or fudging the numbers in your Hard VE Table such that your VE increases as your boost level rises. This will also inject the extra fuel you need.
In a sense, the "Fuel", "Soft VE" and "Hard VE" maps all do the exact same thing - modify the amount of fuel injected at a given load and rpm. Because of the way the stock ECU worked from the factory, we have an extra map that most speed density ECUs don't - the "Fuel" table.
With a megasquirt, or another aftermarket fuel injection computer, you likely wouldn't have the "Soft VE" or "Fuel" maps found in the DSM speed density project. With a megasquirt, you ONLY have a VE table, and if you need more fuel with higher boost, you just fudge the numbers in the table. You might have 100% VE at 1.0bar and 3500rpm, for example, but 150% VE in the table at 2.5bar, to reflect the fact that you need 50% more fuel to keep everything happy under boost.
It works out to be six one way, half a dozen the other way. Whatever you feel most comfortable with.
-
Oh, and for all you people who are stock, or pretty close to it...
Here's my roughly tuned base map for my car, which is completely stock except for a 2.5" downpipe, and 3" cat and catback exhaust.
Stock SMIC, stock piping, stock intake snorkel, stock bov, stock injectors, stock.... everything, minus the exhaust.
If you're still running the stock injectors, you should be able to stick this bin into your ECU and have it fire right up, first try.
http://www.sfu.ca/~rbc/1g_basemap_3inch_exhaust.bin
I've tuned for Chevron 94 octane, but if you have anything other than the stock SMIC cooling your intake charge, this should be pretty conservative.
I get only sporadic little counts of knock (under 5) at my boost level, which is 12psi for the moment, while tuning, and not all the time, either.
It's set by default to run open loop everywhere but idle, so it should idle pretty nicely, but be a little rich just putting around in vacuum.
It's a start, though.
-
In DSMap you refer to timing. Can it be adjusted or only observed using this Pocket Dyno program? Thanks
-
hey richard,
with respect to the timing maps, does the timing stay static after 14.8 psi or is it also "compensated" as boost rises past 14.8?
thanx
looking forward to playing with this!
phil
-
Richard (#2):
Timing can only be observed using the pocket Dyno program. It can be adjusted on the fly using a laptop and the Ostrich EPROM emulator though -- while driving, and without turning the car off.
Phil:
I haven't gotten a definitive answer to this yet, but it *looks* like for now, the timing stays static after 14.8 psi. This is the only drawback to this whole scheme I can find so far, although there should be a way to modify the code to extend that range a bit.
The stock ecu does interpolate between load points, so the best solution that I can see that doesn't sacrifice timing resolution in the lower rpms would be to change the last "load" point from 14.8 psi to 29psi.
Then, you can change that last load point to whatever you want your final timing to be at that insanely high boost level, and the ECU will linearly interpolate between the 12.2psi load point and that 29psi one as your boost level rises. It wouldn't be perfect, but it *would* give you at the very least, a basic "number of degrees of retard per psi boost", along the same lines as an MSD boost timing master box.
I just need to have a look to see whether that's doable, or whether the load points have to be evenly spaced or not.
For now though, looks like you're stuck with static timing past 1bar of boost.
-
Awesome writeup Richard. That must have taken some time :)
I've been up and running with it for about a month and a half now and I have nothing but good things to say about the conversion.
Unfortunately the analog outs died on my wideband 3 days after getting everything running, I then had to sent it to Innovate for replacement. Even petty much guessing with regard to the tuning I've now got it to the point where I can say it's running pretty well. I still have a bit of knock here and there at certain throttle points, however it's very consistant. There are no surprises like when I was running a maft.
Just got the wbo2 installed again tonight so hopefully it will be running in all it's glory soon :wink:
One thing I can say for sure it I'm just loving the throttle response. It rolls on very smooth and reacts immediately to any changes in throttle. Huge Huge improvement. 2Gmaf is unhooked but still in place so the improvements are definately not a result less intake restriction either.
I definately recommend this to anyone as even with a crappy tune my car is running better than it ever has. However in my opinion an ostrich is pretty much a necessity. Burning chip after chip to get a good tune would be way too tedious.
EDIT : Just wanted to add for those without a keyboard, the mouse scroll wheel works great for raising / lowering values in dsmap. I use a trackball for the purpose :)
-
Hey guys, I just wanted to add a little note:
I was a little annoyed at the whole fuel/timing tables only going to 14.8 psi thing as well. I did a little research, and a little digging in the disassembly of the speed density code, and I figured out why, and it makes perfect sense.
The speed density binary actually uses the map sensor and IAT to calculate a load value just like the stock code did before, in grams/rev.
If anyone here remembers playing around with the stock code, the stock maps stop at 2.06 grams/rev (or some number very close to that.) This hasn't been altered in the speed density code, and the value of 14.8 psi in DSMap is an approximation for that load value.
If we want our timing/fuel maps to go further, one needs to only apply the "extended timing/fuel maps" mod to the .bin, just like one would to the stock code. The maps are still in the same place, and used the same way.
If you're ambitious, I'm sure a few of you (Miles, Artem, et al) can figure out how to do it yourself. For those who aren't, I will do the conversion myself and post a turnkey .bin file for y'all to use in a little bit. I might also make a few modifications to DSMap so that the fuel/timing tables reflect the actual load value along with the "estimated" psi value so people don't get confused.
-
refreshing information, grassroots racing at its finest
nothing quite like reinventing the standalone, although you dont get the same mad props
-
Hey there guys, this topic has been dead so I figured I'd bring it back to life. I've been running the speed density code for a while now. I've been through 88 iterations of the code and am still modifying.
I've now figured out how to implement the extended maps to the speed density bin. Now I have much more adjustability.
Features of the custom Busta SD chip are:
-water injection activated via the purge solinoid at greater than 3000 RPM, greater than 75% tps and greater than 10 psi.
-studderbox and NLTS with knock elim on studderbox.
-timing retard to 10 degrees while on the studderbox to help build boost.
-CEL shift light.
-knock type fuel cut (if more than 20 knock counts then ecu cuts fuel).
-phantom knock elim code using map instead of E1 (g/rev).
For those interested in converting to speed density I recommend you get an Ostrich emulator and the Tunerpro or dsmap software.
For everyone's references I'm running about 30 psi no knock with the extended maps Busta custom SD chip. Mods on the car are:
3" full exhaust
680 cc injectors, 255 wally re-wired, spoolin up FPR
kinetic FMIC
water/meth injection
GT-14 dual bb turbo at ~30 psi
Did I say that SD is the sh*t! 8)
Later, Vlad.
-
What kind of timing and fuel type with 30 psi? What is the rate of water/meth injection? Are you not worried at all about cutting fuel due to knock leading to more knock?
-
Hey there Martin, the timing maps being used are the ones included in the extended maps guide. As for fuel, I'm running 94 and currently the fuel map is set to 14.7 below the 4th row and 11 anything above. The car is currently locked in open loop.
I think I have a McMaster Carr #5 nozzle with a Sureflow pump set to 100 psi. I was just running straight meth but I might go to a 50/50 mix to be safe/cheap.
That's a good point about cutting fuel when knock is detected. I was thinking if the car hits fuel cut most of us would take our foot off of the gas :wink:
Later, Vlad.
-
well, if it cuts fuel, there is no danger anyways, because to have damaging detonation, you have to have fuel. If its cut fully, even if you kept your foot in it, you will just be letting air in, but you will have no fuel or boost pressure.
What you might want to do to make it less sudden and more freindly is to have it cut spark to 1 cylinder when it hits say 10 counts of knock, 2 cylinders at 15 counts of knock, 3 cylilnders at 20 and all 4 at 25. This way the guy won't have to fully lift, just let it breathe and go back to driving. It will make the car more usable, not just sudded nothing.
not sure if its technically possible, but if so, it might be quite a nice feature.
If not individual cylinders, maybe a pulsed ignition and fuel cut thing?
-
OEM waste spark would not allow a single cylinder shutdown.
-
ahh, for sure :)
Ok, then howabout, at 6 counts, pulse spark to 2 cylinders, then the next 2, so its fairly soft, not a harsh thing at all, but gets your attention. Then at 10-15 or so counts, it alternates between 2 sets of cylinders being killed, more harsh and noticable, but still not pulling your eyeballs out of your eyesockets. And if you still are being a dumbass and keep your foot into it, at 25 counts, well, fuck ya, full spark cut :)
-
Is the meth injection providing enough fuel to make up for those 680cc injectors? I found at much lower power numbers I was at 100% duty cycle, so I assume the shortfall is from the 50% injected methanol?
Do you know how what percentage of fueling you are getting from the methanol at 30 PSI?
-
Kevin, you are maxing out 680's? Wow, thats some serious power.
With water injection, typically you can go leaner even if you are just injecting pure water as the detonation threshold is much higher. You may be adding fuel just to get rid of knock?
-
Haha, I was at 85% duty cycle on my Big 28 on C16 at roughly 11:1 -> 11.5:1 A/F and that was "only" at 25 PSI. That should be significantly lower airflow than a GT-14 at 30PSI. (I was running stock base fuel pressure with a rewired "Supra" Nippon Denso pump.) With those duty cycles I was trapping 107-108MPH with a completely full weight 2G pig.
I am fairly confident that I would be close to overrunning 650/680 cc injectors fairly easily now with water injection and 94 octane pump at much lower boost levels. (Mind you I have extra displacement now.)
-
FWIW a 2.3L, 8.5:1, GT35R, 720cc, 255hp, 20psi, 92 octane will hit 85% IDC easy.
-
The exhaust side will be a big factor in that equation. The -14 hot side is brutal and wont flow nearly as much as a full T3 or T4 GT35R...I had one.
-
I tried turning up the boost to 30 psi but I would get a studder at RPMs above 5000. Problem is that I'm blowing out my spark I think cause when turning it back down to 25 psi it's all good.
Looking at my logs I'm running the injectors at about 18 ms at WOT top of 2nd gear. AFRs are at 11.
Later, Vlad.
-
Vlad, gap your plugs down to .25 or .26 Give it a shot, I have had a spark blow out problem too and have read about lots of others.
Its easy, free and will probably take care of your problem.
-
Hey there John, I've tried gapping my plugs to .25 but no dice. Kinda hard to run 30 psi lately anyways so I'll wait for drier conditions 8)
Still trying to tune in my cruising VE but WOT is pretty much bang on.
If you're scared about having to adjust for mods added and such maybe an AFC can be used off of the map sensor input but instead of using hz set the afc to use bar. Just a thought.
Later, Vlad.