SERT cable connection
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)
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.
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?
Trending Topics
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 whichare scheduled for deliverytoday. I won't be installing until Monday, so I can afford to be patient. Thanks again.
The Best of Harley-Davidson for Lifelong Riders
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?


