Harley Davidson Forum

SERT cable connection

  Printable Version
Harley Davidson Forums >> General Harley Davidson Tech Forums >> Electrical/Ignition/Tuner/ECM/Fuel Injection >> SERT cable connection Page: [1]
Login
Message << Older Topic   Newer Topic >>
SERT cable connection - 5/9/2008 6:34:36 AM   
madcitycustom


Posts: 867
Joined: 6/24/2006
From: Wisconsin
Status: offline
When I bought my SERT, they threw in a couple of cables.  The black one connects the SERT to the bike, and a gray one that supposedly connects the SERT to the computer.  My problem is twofold.  First, they gray cable is a nine pin serial cable, but my computer takes a 15 pin.  Second, the open end of the gray cable would be female, while the computer is also female.  Best Buy says I could use a serial to USB converter, which I may try after seeing what you all use.  Any suggestions?
Post #: 1
RE: SERT cable connection - 5/9/2008 6:54:44 AM   
Doc 1



Posts: 1844
Joined: 5/9/2007
From: Florida
Status: offline
Here is a good explaination of why the SERT uses the old serial port by one of my customers that does extensive computer programing. You probably have a moniter 9 pin serial port on the back of your computer which is the same connection as the cable....a searial port card would need to be installed in your computer, if this is imposible then the Biliken usb adapter will work. Forgive the text and spacing...I don't have a clue why it came out like this. (it was a copy/paste)

Interesting trip through serial port communications. I had some fun communication adventures loading the map into the ECM and found out some things about the SERT software.  As I worked to get the map loaded it became obvious that the software was designed specifically for a good old fashioned serial port, as I had issues when trying to use a USB-to-serial adapter.  I believe this is due to the adapter using software rather than hardware for some functions. The RS232 serial port found on older laptops and desktops primarily uses IC's and hardware to operate.  This minimizes the load on the PC processor and assures very close timing for the data signals.  The USB-to-serial adapter uses only one IC to translate data from the computer into RS232 signals.  The function of several hardware IC's are "emulated"(simulated) by software that requires the PC processor to handle.  These functions are in the drivers that must be loaded for the adapter to be recognized by the PC.  The big problem with using software to replace the serial port hardware is that the timing of the data signals is not as precise. RS232 serial ports have some signals to act as "traffic control" to keep the computer from (1) sending data when the receiver is not ready and(2) make the computer send data immediately upon demand by the receiver.When the old hardware is used this happens very quickly and smoothly, but a lot of USB-to-serial adapters show delayed reactions to this "handshaking". I tested three different types of USB-to-serial adapters and found significant delays in the handshaking.  Referring back to some old texts, the delays were exceeding the specification for RS232 communications. I expect that for most equipment these delays are no problem, but when programming the ECM it becomes very important. I expect that the guys who wrote the upload program for the SERT are using the serial port handshaking to ensure that the data reaches the ECM in a precisely timed order.  When a USB-to-serial adapter is used that exhibits poor handshaking timing the program kicks out the bootloader error. This prevents the SERT from trying to load a map when it knows something is wrong with the interface.  Incorrectly loading a map to the ECM could be disastrous, so I'm happy that the programmers went the extra mile to besure the upload doesn't occur unless everything is working right. For myself, I had to use a desktop computer with a hardware serial port to load the map successfully. 
(quote by J.Pope)

_____________________________

HDMD Tuning Center
EFI Training
352.242.9616
hdmd88@yahoo.com

(in reply to madcitycustom)
Post #: 2
RE: SERT cable connection - 5/9/2008 7:12:03 AM   
madcitycustom


Posts: 867
Joined: 6/24/2006
From: Wisconsin
Status: offline
Thanks for the response.  I guess I can try some things out and if they don't work, hope the stores have a favorable return policy. 

(in reply to madcitycustom)
Post #: 3
RE: SERT cable connection - 5/9/2008 7:50:32 AM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
A 15 pin female connector would be a VGA monitor output connector. If there is no nine pin male connector then you don't have a com port. A common situation on most laptops sold in the past few years. 

(in reply to madcitycustom)
Post #: 4
RE: SERT cable connection - 5/9/2008 8:05:47 AM   
Doc 1



Posts: 1844
Joined: 5/9/2007
From: Florida
Status: offline
Frisco42... I agree.

Mad City, becareful what you use...the best is the Biliken usb adapter, there are some cheap adapters that have been known to lock the ECM beyond recovery and that means it will have to be recovered by the Dealer with a Digital Tech....this will cost you more than the Biliken adapter would have.

_____________________________

HDMD Tuning Center
EFI Training
352.242.9616
hdmd88@yahoo.com

(in reply to madcitycustom)
Post #: 5
RE: SERT cable connection - 5/9/2008 8:13:37 AM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
Doc, I've got one of those noname USB to serial adapters and it seems to work fine for data logging from the SERT. Should I be wary of using it to program.
I've got 2 laptops. One is an older 400Mhz with a com port, the other is a 1.2GHz Dell with no com port. The USB adapter works with either one for data logging. The com port on the old laptop also works for logging. When I decide to load a map I will probably go with the real com port machine. Any thoughts?

(in reply to Doc 1)
Post #: 6
RE: SERT cable connection - 5/9/2008 8:26:07 AM   
Doc 1



Posts: 1844
Joined: 5/9/2007
From: Florida
Status: offline
I would use the one with the serial port my self just to be safe. I know a lot of people tell me they have no problem with the cheap usb adapter....BUT..... let your ECM lock up just once will make your butt pucker...(sorry)..., I have seen ECM's lock up before and it is a MAJOR pain and it will ruin your day. It just isn't worth it to me. When I go to dealers to train the first thing I do is max out their memory in their computer and install a second Serial port....(one for the dyno and the other for SERT).  The new EFI Super Tuner (Fisher Price) along with the new V-Tuner have been programed to work with usb connections but the older SERT....well your taking a BIG chance. If you do lock up your ECM there is a ''ECM Recovery'' section in the ''Tuning Help Manuals''....just click on the Help in the menu bar.

_____________________________

HDMD Tuning Center
EFI Training
352.242.9616
hdmd88@yahoo.com

(in reply to frisco42)
Post #: 7
RE: SERT cable connection - 5/9/2008 8:45:03 AM   
madcitycustom


Posts: 867
Joined: 6/24/2006
From: Wisconsin
Status: offline
Well, I definitely do not want to lock up the ECM. 

I went over to a computer store and they had a usb to serial port converter made by Sabrent.  Is this one of the el cheapos, or a decent brand name?  It says that it "Supports Windows Xp, Connects an additional serial peripheral in seconds," and "Supports baud rates from 1,200 to 115.2k."  I don't recall the baud rate off hand for the tuner, but I'll check before making a go with it. 

UPS lost the air cleaner anyhow, so it's a few days behind the pipes which are scheduled for delivery today.  I won't be installing until Monday, so I can afford to be patient.  Thanks again.

(in reply to Doc 1)
Post #: 8
RE: SERT cable connection - 5/9/2008 8:47:54 AM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
Doc, to change the subject slightly.
I've been doing a LOT of data logging in an effort to understand what's going on. The bike is an 07 FXSTC.
Here is a graph that I put together showing VE front and VE New Front plotted against RPM.

http://people.delphiforums.com/frisco42/VE.pdf

The hair on the curves is because at any given RPM there are data points at different MAP values, but the overall trend lines are obvious. VE New is 4 or 5% lower than the programmed VE values. VE rear is similar. Given this, when the ECM is in open loop as it is most of the time above about 4500 RPM can I assume that the system is running rich given that VE New is the same as VE under those conditions?

(in reply to Doc 1)
Post #: 9
RE: SERT cable connection - 5/9/2008 11:42:39 AM   
whittlebeast

 

Posts: 281
Joined: 3/2/2007
Status: offline
Drive and data log for about 20 min and post a sert data log here or send it to me.  I think you will be amazed at the things you can see with good graphing software.

AW

_____________________________

When the racers get done doing the impossible the engineers get a chance to correct the errors in the model.

***My other addiction, running Megasquirt EFI ***
http://www.youtube.com/watch?v=ZXH4WWhqZ_s

(in reply to frisco42)
Post #: 10
RE: SERT cable connection - 5/9/2008 12:36:08 PM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
The data file from which that graph was made is at:

http://people.delphiforums.com/frisco42/Run3.DM2

Show me something. I would be interested in any good tools.

(in reply to whittlebeast)
Post #: 11
RE: SERT cable connection - 5/9/2008 12:58:20 PM   
Doc 1



Posts: 1844
Joined: 5/9/2007
From: Florida
Status: offline
As the data is collected by the New VE (Adaptive Fuel Values) it actually interpolates the information (data) and helps in the Open Loop areas too.
Is this what you where asking?

_____________________________

HDMD Tuning Center
EFI Training
352.242.9616
hdmd88@yahoo.com

(in reply to frisco42)
Post #: 12
RE: SERT cable connection - 5/9/2008 1:12:48 PM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
Something like that. It looks to me that in the open loop areas Ve and VE New are pretty much the same unlike the 4-5% difference in the closed loop region. This leads me to wonder if there is any adaptive learning in the open loop mode.

(in reply to Doc 1)
Post #: 13
RE: SERT cable connection - 5/9/2008 1:37:05 PM   
whittlebeast

 

Posts: 281
Joined: 3/2/2007
Status: offline
PM me with your email and I will send you a few files and links.  I can tell you that the bike is tuned fairly well in the closed loop section of the tune.  I will post a screen shot of the data later this weekend.  The injector duty cycles are maxing out at close to 70% or about perfect. 

Note: it is better tuned than the Harley factory got my stock bike :(

AW

_____________________________

When the racers get done doing the impossible the engineers get a chance to correct the errors in the model.

***My other addiction, running Megasquirt EFI ***
http://www.youtube.com/watch?v=ZXH4WWhqZ_s

(in reply to Doc 1)
Post #: 14
RE: SERT cable connection - 5/9/2008 2:49:09 PM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
The ECM got the 'torque smoothing download' at it's 1K service. It's had another 6500 miles to self tune.

What I would REALLY like is a way to upload the map from the ECM to the computer. If I do any changes I would like to start with it. I've been reverse engineering the map from the datalogs, but there are some holes yet to fill.
Comparing what I've gathered so far to Sert map #67, the AFR table is identical. The VE tables are about 5% lower than 67. The spark tables are similar from 2500RPM up but very different below that. Example: map 67 has 20 deg. advance at idle, my bike idles at 12 degrees.

(in reply to whittlebeast)
Post #: 15
RE: SERT cable connection - 5/10/2008 9:10:25 PM   
Steve Cole

 

Posts: 110
Joined: 1/26/2008
Status: offline
Trying to reverse engineer the map from a data log is a lot of work and will not be correct. Basically it's a waste of time. Several reasons why but the first is that when the ECM learns, it is correcting the mixture to something you do not know. If the Bias tables have been changed 14.6 AFR is no longer 14.6. HD themselves sets the bias tables to keep the bike richer than 14.6 so you really have nothing to go by. Some of the calibrations event have the bias changing at the same RPM but at different loads so unless you know whats going on in the bias tables your fighting a loosing battle. The ECM applies what it learned in closed loop to the open loop areas as well so whether the correction is right or wrong the open loop areas get it. The best thing you can do is install the base calibration that is close to the parts you have and go from there. This way you can see what it's trying to do and make the proper adjustments.

(in reply to frisco42)
Post #: 16
RE: SERT cable connection - 5/11/2008 8:19:06 AM   
frisco42


Posts: 82
Joined: 10/23/2006
Status: offline
Steve,
My current level of understanding is that the VE (not VE New), AFR, and Spark Advance tables are static tables and do not change. Am I wrong? Those are the basic tables that I'm trying to recreate from several hours of datalog data.
This would all be unnecessary if you and yours would just supply a utility that would read the maps in the ECM and create an MTx file. How about it. Are you prevented doing it by one of those contracts you refer to?

(in reply to Steve Cole)
Post #: 17
RE: SERT cable connection - 5/11/2008 2:38:16 PM   
Steve Cole

 

Posts: 110
Joined: 1/26/2008
Status: offline
Your wrong and right, sorry but that the only way I know how to answer your question, and yes there are things I know that I cannot say. The bottom line is how do you know what value to use when the ECM doesn't go point to point. It issues an average value from a calculation done within the ECM from the values that are in the table your looking at. Take a look at a VE table in one of the maps and just image that the ECM grabs the 4 nearest cells for the input and averages them, I hope that gives ou the idea as I cannot go much further. That's why I tried to tell you it's a waste of time as you will not end up with the correct values. Spark is event worse to do as it has many other things that come into play.

(in reply to frisco42)
Post #: 18
Login OR Register now to post a reply to this forum topic.
Page:   [1]

 
Harley Davidson >> General Harley Davidson Tech Forums >> Electrical/Ignition/Tuner/ECM/Fuel Injection
Jump to:



Featured Sponsors
Advertising Info

New Vendors
Warehouse Direc
Baker Drive Train
Haulen Ass
Pickard USA
Harley Goodies
Pickard USA
AMSOIL - Performance Oil Technology
Brain Bucket Magazine
HDForums.com Sponsors
AMSOIL - Performance Oil Technology

HDForums.com is not affiliated with or endorsed by Harley Davidson, Inc.