• Protip: Profile posts are public! Use Conversations to message other members privately. Everyone can see the content of a profile post.

Need help sourcing this chip for NA1 engine ECU 1995-1996

Joined
29 July 2023
Messages
53
Need the overall collective knowledge of the forum to source this IC in the 1995-1996 engine ECU. It is the flat inline IC with the labeled X4HT BX8371 9447 right in the middle of the pic.

This IC does not appear to be in the NSX 91-94 ECU but could be in the post 96 NSX ECU, or in other Honda/Acura ECUs from the same era. Background on why I need this chip below.

IMG_20231021_115850.jpg

If you search the forum, you will see I have been trying to get a 95-96 MT NSX ECU for a while. I purchased a 95 MT NSX and it was throwing P0141 and P0161 codes. The car came from Colorado and had a clever cats delete (it had the cats but the cats were hollow and the CEL light bulb was removed). I need a functioning emissions system to pass CA SMOG so went about gathering up a complete emissions system from ebay, offerup and the forum. After putting the emissions back to stock, the P0141 and P0161 codes remained. Went through the factory service manual checks and the only thing left to do was to replace with a known working ECU. Sourcing a 95/96 ECU is difficult given there were slightly over 1,000 95/96 delivered and I am not sure how many are left. I was eventually able to get a 96 auto ECU from ebay and a 95/96 MT ECU from a member on this forum.

The 96 auto ECU is identical to the MT ECU in all ways except for additional components to communicate with the auto transmission controller and different codes to do that. When I plugged in the auto ECU, the engine fires right up and the only code was P1607 (communication fail to the auto transmission controller). This gave me hope and I did some surgery to remove one of the auto ECU CPU (the one with a different version label than the MT CPU) and replaced it with the MT CPU (this involved desoldering a 80+ pin surface mount chip and putting a socket on the board, I have pics if anyone is interested). Again I was able to fire up the engine with the modified ECU and the P1607 code went away, replaced by a different code for ECU side A failure. This is probably due to the CPU pulling code from an eeprom somewhere else on the board. At this point, I gave up on the surgery idea and put the MT chip back into the MT ECU.

The other MT ECU I got from a forum member also threw two O2 sensor codes (P0141 and P0155). P0141, P0155, P0166 are all related to the O2 heater circuit which run on batter voltage. I suspect the many issues people had with 95/96 O2 sensors were probably actually caused by the ECU - therefore the extended warranty. In desperation I systematically moved good components from the AT ECU to the MT ECU. I had hints on where to start by tracing the pins that connect to the Port A connector that drove the O2 heater circuits. By replacing the IC at the top of this posting, I was able to fix the P0141 code. Replacing additional chips around the same area fixed the other code. I can source the other chips but cannot find the big one anywhere.

I would like to source the BX8371 IC to help other 95/96 NSX owners avoid the difficulty of getting a working ECU. I think the best bet would be finding the same IC in other engine ECUs that are much cheaper. BX8371 appears to be discrete components soldered to a board and then encapsulated - a long way to say it is not a logic chip that is more widely used.

Drop me a line if anyone can help me help other members keep their CEL off.
 
It looks like a resistor array (search for "sip resistor array" or "sip resistor network" to see more).

You could desolder the device and measure the resistance between pairs of pins -- that would tell you if it really is a resistor array (could be an array of other components) and how the resistors are wired and what their values are.

With that information, it should be pretty straightforward to find a replacement -- either another SIP array or a handful of individual resistors.
 
It is for sure not a resistor array since it is bumpy in odd places. The board is littered with resistor arrays but that one is not.
 
You are doing great! I probably would have continued swapping parts from the MT to AT ECU.

That is likely a Hybrid IC (HIC) in a ceramic package. Probably made by NAYUTA. The instrument cluster has a similar HIC.

Is your car ODBI or OBII? If OBD1, you might consider using a previous year...though that might cause the cluster connectors to have to be repinned (the cluster pinout is different, not sure on what side of the ECU it is changed though)

I also recommend a Legend ECU to rob parts, but have no specific knowledge of that HIC.
 
95/96 are obd2 with 4 O2 sensors. I did get a 95 Legend ecu for parts but it did not have that IC.
Unlike the 91-94 NSX ecu, most of the components on the 95/96 NSX ecu are surface mount. The 95 Legend ecu still had mostly discrete components. Let me look up nayuta. Thanks.
 
Also, T3TEC keeps a supply of similar IC units for their gauge cluster repairs. Might be worth reaching out and seeing if they will sell you one or point you to a distributor.
 
I find jotrin.com just scrapes everything and presents it on a search. When the RFQ is sent IS when they start looking for it and want to charge you a fortune if they find it. I also suspect they start off with a big number to see if it is worth their time to even look and then the number gets bigger due to all the work.

I'm a cynic after spending days looking for the cluster HIC. All these false leads.

If anybody sees a T3 HIC in the field, I would like to see how they solved it. I am collecting cluster HICS from JDM Accords for a future reverse engineering effort. I suspect that nearly the entire cluster can be reproduced.

**

Re: You might also consider a 97+ ECU prior to DBW, it can't be all that different. I've seen 3.2's dropped into 3.0s without issue, so I suspect you can drop a 3.2 ECU into a 3.0 enviro; but I may be over simplifying it. I would put a good chance on the ECU's being effectively the same.
 
Dont the 95/96 already have drive by wire? Pretty sure my MT 95 is dbw and the auto ecu is nearly identical. If anyone has a Pic of 97 ecu, I can quickly tell if it has the same part.

Given I was able to start the car with an ecu failure code from my cpu swap, I am not even sure if the cpu is needed to fire the car up (but I won't try) if the spark plugs are handled by some brain stem function.
 
You most definitely do need the ECU for the engine to start up. Starting fuel pulse widths and the ignition timing calculations are all done in the ECU.

As you note you appear constrained in terms of donor ECUs to 1995 and 1996 model years because 1995 is the start of OBDII and Drive by Wire and 1997 is the roll out of the C32. The 1997+ models with auto continued to use the C30 engine; but, there were also emission control system changes in 1997 which probably makes a 1997+ auto ECU impossible / exceedingly difficult to back date to 1995 if you need to comply with emission tests. I also expect pin compatibility issues for the later ECUs because of the additional emission functions; but, that may not be an insurmountable problem.

I am a little unclear as to your current status. The following

By replacing the IC at the top of this posting, I was able to fix the P0141 code. Replacing additional chips around the same area fixed the other code. I can source the other chips but cannot find the big one anywhere.

kind of implies that the ECU is now functional and not generating any error codes? If so, that is great. Before taking the car for an actual compliance test I would use an OBDII scanner to check that the actual emission monitors in the ECU have all completed and achieved ready status. Absence of error codes does not always mean that everything is fine.

If you are now just asking about a source for the BX8371 device, I have the following observations. Depending on what your investigation into the circuit board traces showed, that might be the current sensing element for the 4 heater circuits. It looks to have 16 pins so 8 pins for the in and out on the 4 O2 heaters and 8 pins for the 4 sense lines to the CPU, so that is a possibility. Depending on how many A/Ds the CPU has there may be a front end multiplexer chip between the sense elements and the CPU which allows an A/D on the CPU to handle more than one function. Unlike the mission critical stuff on the ECU (ignition and fuel PW control) the software monitors that check things like the heater status and the evap and other things do not run continuously so sharing A/Ds is sensible. If that is the sensing element, it is probably custom configured for the heater circuit currents and the firmware coded error limits specific to the 95/96 ECUs. If, as you suggest, there were issues associated with the heater current monitoring on the 95-96 ECUs, that chip might have been the problem or part of the problem in which case it is probably an orphan. So, I expect that you will not find a matching number replacement for that chip. However, if the architecture of the later ECUs (which might include non NSX ECUs) is similar, they might continue to use the same style sensing element. If that is the case, those chips might be backward replaceable for the BX8371 device.

That is a lot of ifs and maybes!

My son has a 2005 RSX. A couple of years ago while we were on a mission in the local pick and pull getting arts for another project car he spotted a wrecked 2005 RSX. The ECU was not damaged so he pulled the ECU 'just because'. He is also an Electrical Engineer so that is all the excuse he really needed, and it was only about $35. I will ask him to bring it over some time and we will pop the cover off to trace out the heater circuits and see if the ECU has a similar device in the heater circuit. If so, that might suggest that although there are no exact part number matches there may be acceptable part substitutions.
 
Wow, the forum shines again. Always impressed with the depth of knowledge and info sharing. As you said, I now have a fully working 95/96 MT Ecu but 1 code throwing MT Ecu and 1 was good auto ecu but now code throwing. Seems to be a waste due to one unobtainable IC. I will keep looking but zero luck so far.

Moved on to the EPS controller and the two relays (code 22). That is a story for another time but sharing a pics here.IMG_20231027_064734.jpg
I removed the relays and cracked them open to see the contacts. One has carbon deposit while the other is clean (top is clean). Got the high capacity relays coming from digikey at $4 each. If this fixes the issue, I will be pretty happy. The eps controller is still available new but at over $2,000.
 
Last edited:
Wow, the forum shines again. Always impressed with the depth of knowledge and info sharing. As you said, I now have a fully working 95/96 MT Ecu but 1 code throwing MT Ecu and 1 was good auto ecu but now code throwing. Seems to be a waste due to one unobtainable IC. I will keep looking but zero luck so far.

Moved on to the EPS controller and the two relays (code 22). That is a story for another time but sharing a pics here.View attachment 185219
I removed the relays and cracked them open to see the contacts. One has carbon deposit while the other is clean (top is clean). Got the high capacity relays coming from digikey at $4 each. If this fixes the issue, I will be pretty happy. The eps controller is still available new but at over $2,000.
The EPS cuts out around 35mph anyway so if youve got normal arm strength a 95 drives just fine wo EPS, reverts to a full time manual rack. I would opine it may even be better that way.
 

Anyone thinking about replacing the eps relays should read the above. The high capacity relay is not a drop in. The original relays were very difficult to remove. I still damaged the trace during the removal process but was able to repair.
 
Last edited:
Turns out my Son now has a collection of RSX ECUs. I had a look at a pre 2005 and a 2005+ ECU and the physical architecture has clearly evolved since the mid nineties NSX ECU. In both series of ECUs, there is now only one single in line module and it is STA464C which happens to be the 4 injector drives. The NSX ECU has a similar module STA460C which you can see in the first photo; However, STA460C only contains two transistors so there must be 3 of them on the ECU board. The rest of the RSX board is SMD components.

The 2005+ ECU has evolved to a 256 pin 64f7058 family RISC chip (I am not floating that one off the board with my rework station!) plus a Toshiba TB9293 and a proprietary Keihin 100 pin SMD of some ilk. I have the case covers off; but, have not pulled the boards out of the cases to inspect the boards; but, they appear to be multilayer boards. If that is the case, physical tracing of the O2 sensor heater circuits is likely a no-go. In fact, as soon as I saw the 64f7058 chip I figured this is way beyond my pay-grade.

Based upon my initial look the architecture of the early 2000 ECUs has evolved far beyond the NSX ECU. If you are looking for a donor for parts it might be best to stick to other Honda products in the 1995 - 1996 vintage which may have used similar architectures as the NSX to implement the OBDII requirements for O2 heater monitoring.
 
Thanks again. Had a bit of trouble today. Was driving 50+ miles to setup for the CA smog test. About 17 miles from home the dash lite up and the car rpm surged up and down. I was lucky to get off the freeway near the LA Convention center. If you don't know the area, it is not a good place to have a bright yellow NSX disable. Called AAA and waited over an hour for a tow truck and then another 30 mins for a flatbed. Got it home, swapped out the ecu with my backup (the one with P0141 and P0161) and the car cranked right up. I moved the BX8371 to the backup ecu and the two O2 sensor codes went away. It is highly likely the situation you stated above, that all the O2 sensors go through that chip.

I think putting the cpu in a socket was a bad idea. Only have to miss a single clock to doa the car. Now I need another backup ecu.
 
Last edited:
Back
Top