Troubleshooting a Honda P28 Chip Installation
This situation is more common than you think, as many people choose to install a chip them selves, even if they have no experience with electronic parts soldering. These ECMs are small and somewhat simple, but still, are computers with fragile sophisticated and sensitive circuits. Once in a while, even professionals face some problems doing these installations, so here are some basic methods to troubleshoot your installation.
NOTE: This article covers not only the Honda P28, but also the Honda P05, P06, USDM P30, P72, P74, P75 and OBD-I PR4 computers family
The installation consists of adding some parts to the circuit of the ECM, where the space where those parts belong is empty in the circuit board as an unused option built in when the ECM was manufactured. There are even footprints in the circuit board, indicating where parts belong.
The parts added are; two capacitors (for filtering electronic noise), a resistor, the new EPROM chip (a socket will be installed), a small 20-pin chip (74HC373 latch) and a jumper (J1). The jumper “J1”, enables the new chip when installed and disables it when cut, just like a switch. When J1 is installed, it will make the ECM microcontroller (MCU) ignore its internal program (engine calibrations) and will look at the external EPROM chip installed. So if the installation is good and all parts are good and installed correctly, you may switch between “Power” (external chip) and “Stock” or “Normal” (MCU internal program). The only thing is that it will work just like that, but the engine must be turned off and then on for the changes to take place every time you switch programs.
That being said, one good test to do if you have problems with your installation is to cut the jumper. The ECM should go to “stock” mode and all should run fine. If it does, it means that the installation is presumably right, but the chip is either damaged or the program on it is not right or not compatible with your car. On the other hand, if problems stay after cutting the jumper, then something went wrong (physically) when installing the kit.
Possible common causes are:
- A short circuit was created with soldering.
- A broken trace (copper trace) created by excess soldering heat.
- The 74HC373 latch chip is either bad or installed backwards
- The 27C256 EPROM is either bad or installed backwards
- The program in the 27C256 EPROM is no good or not compatible.
1. Short Circuits by Accidentally jumping two or more connections with soldering.
The first thing to check if having problems with an installed chip is the soldering. You should check for accidental bridges, jumping contiguous pins or connections of the installed parts. Do this first by visually inspecting the installation, using a good magnifier. If nothing wrong is found, then you should use an ohmmeter to check for continuity between every pair of contiguous pins you suspect may have been jumped. This is important as the circuit board is a multi-layered board, meaning that it has connections above, below and in between its layers, making some connections not visible. When too much solder is used, it tends to absorb it through the integrated tubular connections that join all the layers and possibly accumulating an excess, jumping two or more points. So, if any two contiguous pins of any of the installed parts shows a resistance reading of less than 1 or 2 ohms, then somehow those two connections got accidentally jumped.
2. Broken Connections.
The next thing to check is possible broken connections. This often happen on these jobs due to overheating when soldering. This computer circuit boards are VERY sensitive to excess heat and this accounts for the majority of the problems that show after an installation.
If a connection was broken during installation, it must be between the chip, the latch (small 74HC373 chip) and the MCU (64-Pin microcontroller unit), because that is the area involved on the job.
Below is a chart that will help you corroborate all involved connections. All the connection should measure continuity (less than 1 ohm) with an ohmmeter or continuity checker. Any reading over 1 ohm is indicative of a faulty connection. If any of the connections displayed in the table fails the continuity test, it means a broken connection and must be corrected before going further with the test. Do so by checking the circuit to see if the connecting trace is visibly broken so it can be fixed by soldering. Otherwise, just run a thin insulated wire, soldering between both points that failed. The wire should be kept as short and close to the board as possible. This Will prevent the wire from picking or generating electronic noise.
Before the chart, there are two pictures. One identifying the involved components and the other, showing a guide for identifying the pin numbering on every respective chip.Identifying involved components:
Pin Identification of involved chips:
Connection test chart:
|EPROM Check Points|
|EPROM Pin#||Connects to:|
|1||#20 of Latch|
|2||#13 of MCU|
|3||#15 of Latch|
|4||#12 of Latch|
|5||#16 of Latch|
|6||#19 of Latch|
|7||#9 of Latch|
|8||#6 of Latch|
|9||#5 of Latch|
|10||#2 of Latch|
|11||#3 of Latch|
|12||#4 of Latch|
|13||#7 of Latch|
|14||#1 & 10 of Latch|
|15||#8 of Latch|
|16||#18 of Latch|
|17||#17 of Latch|
|18||#13 of Latch|
|19||#14 of Latch|
|20||R54 (at left side)|
|21||#11 of MCU|
|22||#23 of MCU|
|23||#12 of MCU|
|24||#10 of MCU|
|25||#9 of MCU|
|26||#14 of MCU|
|27||#15 of MCU|
|28||#20 of Latch|
The only connection that is left out that table is the following, so please also test that one:
Pin#11 of the HC373, connects to Pin#22 of the MCUWhen “J1” Test Fails and all connections are good:
If cutting “J1” test fails and all connections on the table test OK, we have seen that sometimes the latch chip (74HC373) goes bad by overheating when soldering it or by static electricity. Also, as mentioned before, backwards insertion of any of the chips can damage either one or both, but being this one the more susceptible to such problem.
The less possible but has happened before, is the MCU getting damaged for the same reasons.
The procedure then, is to replace the 74HC373 first as it is quite inexpensive (about $2.00) and do a test. If problems persist, then it is best to replace the ecu, as a spare MCU is hardly found, adding to this that it is a discontinued product. Also, replacing it will take a greater effort than the the whole installation job and might really ruin the ecu, loosing the money spent in that part anyway.FAQ
Q – Will a chip with an incompatible program damage my ecu?
A – No. The ecu will activate the check engine light steadily and go to protection (limp home) mode to avoid engine damage. To achieve the purpose of protection, it won’t allow the engine to rev more than 4000 RPMs and fuel mixture will go very rich to avoid destructive lean conditions.
Q – I have had many problems buying chips like these on eBay and other places. Do you know where to buy a real good chip?
A – Yes. We know about problems with eBay. Most chip sellers don’t know about mechanics or racing, neither programming. The ones that sell a chip that doesn’t fail, are just selling the same copies of the same chip, nothing new, and those chips are only fairly good. We only recommend our chips. They are tested and will make your car pull real hard: