I’ve seen countless threads on this but I haven’t found an answer to this issue. I bought a 1987 coupe back in July with a check engine light. Guy I bought it from ran the code for me at it came up 32 and 36. I disconnected the battery and code 32 went away but 36 still stays on. I’ve replaced both relays with the proper relays I bought from Eckler’s and I’ve cleaned the MAF countless times. It’s not the stock MAF sensor, it’s by some company named “Spectra,” or something along those lines. I’ve read around that these knock offs rarely work and have trouble communicating with the ECM in the long run. My car performs just fine, it ran rich before I replaced both relays but that problem has been solved. Code comes on at second start up after being cleared. I’m mostly asking if anyone has found a company that makes or a place to buy the MAF sensors that actually work well. Or if anyone has a “solve all” that’d be great. I just want this light gone once and for all, so any help would be fantastic.

Thanks!

 

One solution is to reprogram the ECM to eliminate the code check for the MAF burn-off circuit. This requires an adapter to allow an aftermarket chip to be inserted (Moates sells one) and a burned chip with the changes.

 

Thanks for your suggestion there, I wasn’t for sure if that was an option or not. So would I have to burn a chip myself or am I able to purchase one online? I’m not the best at programming things nor do I have the tech to do so easily. Also, if I alter the code in the long run will the ECM run off of pre-set settings? I’m assuming while the code is being thrown the car is using it’s stock settings since it can’t read the current conditions, and that’s working pretty well so far.

 

Correct on all counts. Removing the error code just removes the error code. But if the MAF is working properly, it's not critically important.

I can burn the chip for you and send it your way, and it'll be a hell of a lot cheaper than trying to find one online, that I can guarantee. I still have some of these chips laying around so I don't mind. However, you would still need to purchase the adapter to make that work. If that's a route you're interested in, feel free to shoot me a PM. Feel free to let others chime in with their experience though! I don't own an earlier car, but I have burned chips for other members with earlier cars and they all worked out. I don't do mail-order tuning, just making small changes to the code like changing error codes and fan activation temperatures and such.

 

The slightest vacuum leak will set a code 36. I have been down this road to despair with this thing many times but figured it out eventually. I had a vacuum leak on the manifold on one of my cars that was almost impossible to find. Broken hoses, relays , Cracked brake booster. a new MAF that did not have the burn off wire incorporated in it. Etc.. I bought a used MAF from Howell to fix that particular 36.

 

On an '87 (and other earlier years), code 36 only indicates the burn off circuit did not complete correctly the last time the ignition was turned off. This is a known and stupid code. Extra dumb thing, disconnecting the battery and starting up makes it appear code 36 is gone, but it's not, it won't show up again until the engine reaches closed loop and ignition turned off for 5 minutes or so. Replacing the MAF with common, workable MAFs found today does not solve this as they typically don't have burn off circuits. Even if they do, they are usually unable to make the ECM happy, the burn off circuit check is too fussy. Don't waste $600 on an original Bosch MAF, my opinion.

Best method is what Nomake Wan said, get the ECM MEMCAL modified, only change needed is disabling code 36. Good idea to update to flash as Wan also mentioned, as the old EEPROM can wear out eventually and few have the equipment to burn one anyway. I did the disable, though I had a pro change the MEMCAL for new fuel injectors at the same time. Works fine with a typical replacement MAF from Rockauto.

Make sure to ask for a digital copy (as in a computer file) of the original MEMCAL as well as the modified version, just in case.

Tons of other threads for this on the site, search around.

 

I think you are correct on 84 and 85 years. Car in the video is an 87. I have just been through this again on my current C4 which is an 87 like the OP and in the vid.
I had the same thing happen on mine and spent numerous hours trying to fix it. I had multiple small vacuum leaks on cracked vacuum hoses and also PVC hose like in the vid. I finally fixed it after fixing all the vacuum leaks.

A very small vacuum leak can cause an irritating redundant code 36 until 'Kingdom com'....

 

That is what I meant to imply. No code 36 issue on an 84 or 85. I owned an 85. You are in error if you believe a vacuum leak will not set a code 36 by itself. See vid I posted above.
I had this code 36 set on both my current 87 and 89 due to a VACUUM LEAK ONLY. Nothing else faulty.

More than once. You can replace MAF sensors and relays until one is blue in the face but if you have a slight vacuum leak you will most likely get a code 36. I just recently replaced my water pump and forgot to connect a vacuum hose under the throttle body.
Guess what ? Code 36....

 

Please stop propagating this nonsense about vacuum leaks pertaining to code 36. It is dead wrong.

Burn-off cycle explained:

Here's an example log to illustrate what is going on during the burn-off cycle from the ecm's perspective (use chrome to open link and wait for it to generate chart):

https://datazap.me/u/tequilaboy/burn...7-8&mark=64-67

After a 5 second delay after the ignition has been switched off, the burn-off cycle will be activated for one second and the resulting voltage will be monitored and compared against pre-set limits. This can be seen while scanning by observing the unlimited airflow signal. Converting from voltage to flow using the MAF tables below, the acceptable flow range is between roughly 5 gm/sec and 36 gm/sec (which corresponds to 400 mv - 1.9 v).

Look for the prominent IAC reset cycle to indicate when the ignition was switched off and note the drop in the unlimited airflow signal that occurs 5 seconds later (see the yellow markers).

In this log you will see the unlimited airflow signal bouncing between 27 and 29 gm/sec (or around 1.65 volts) during the burn-off cycle. This is within the acceptable range, so no failure was detected in this example. Unfortunately, I don't have any example logs showing what happens with a failed maf or with failed relays, but you are free to imagine how the signal will be affected.

Note: The normal maf control signal (mass airflow) is limited by the max maf vs rpm table, so you typically won't see the burn-off cycle by observing the mass airflow signal directly unless it should happen to fall below the limit value. In this particular case, the max maf limit was set to 26 gm/sec below 400 rpm. With standard tuning, this limit would be set to 23 gm/sec below 400 rpm.

Some mafs simply do not fall within the acceptable voltage range. If all else fails, you can re-tune the code 36 parameters, disable code 36, or use a modern maf that does not require burn-off and disable code 36 in parallel, or live with the code/ses lamp.

This data was obtained by scanning at 8192 baud with DataCat. Scanning with proper tools is essential for a thorough diagnosis. I can't emphasize this enough.

Acceptable voltage/flow range obtained from MAF tables 1 and 2 shown in green:

Volts Grams/sec

01.92 37.13
01.90 36 *
01.83 33.75
01.74 30.56
01.65 27.56
01.55 24.94
01.46 22.28
01.28 18.78
01.10 15.36
00.91 11.95
00.73 9.16
00.55 6.83
00.40 5 *
00.37 4.94

* You really need to interpolate between points to get the exact flow values at 400 mv and 1.9 v, but this is close enough for this discussion.

parameters involved:

SCALAR: MAF Burn off time 1.00 seconds
SCALAR: MAF Burn off delay 5.00 seconds

;--------------------------------------------------------
; ERR # 36 Params
; >> MAF Burn off Diag <<
;--------------------------------------------------------
LC221: FCB 3 ; 0.3 Second Dly prior to Burn off CAL = Arg * 10
LC222: FCB 6 ; 6 fails req for ERR #36 If HLM air meter
LC223: FCB 95 ; 1.90 VDC, Fail B/O test if HLM A/D GT 1.90 VDC
LC224: FCB 20 ; Fail B/O test if A/D < 400 mvdc

Note: this car has had some tuning of the MAF tables, so the values shown here do not necessarily reflect standard tuning. They are shown for reference purposes.

TABLE: MAF Table #1

Volts Grams/sec

01.46 22.28
01.28 18.78
01.10 15.36
00.91 11.95
00.73 9.16
00.55 6.83
00.37 4.94
00.18 3.59
00.00 9.16


TABLE: MAF Table #2

Volts Grams/sec

02.19 47.63
02.10 44.44
02.01 40.69
01.92 37.13
01.83 33.75
01.74 30.56
01.65 27.56
01.55 24.94
01.46 22.31


TABLE: MAF Table #3

Volts Grams/sec

02.93 82.03
02.83 77.16
02.74 72.30
02.65 67.76
02.56 63.22
02.47 59.01
02.38 55.12
02.29 51.23
02.19 47.66


TABLE: MAF Table #3

Volts Grams/sec

02.93 82.03
02.83 77.16
02.74 72.30
02.65 67.76
02.56 63.22
02.47 59.01
02.38 55.12
02.29 51.23
02.19 47.66


TABLE: MAF Table #4

Volts Grams/sec

03.66 134.47
03.57 124.98
03.47 117.07
03.38 110.74
03.29 104.41
03.20 98.61
03.11 92.81
03.02 87.54
02.93 82.27


TABLE: MAF Table #5

Volts Grams/sec

04.39 206.19
04.30 195.68
04.21 185.98
04.11 177.08
04.02 167.38
03.93 158.48
03.84 150.40
03.75 142.31
03.66 134.23


TABLE: MAF Table #6

Volts Grams/sec

05.12 254.00
05.07 254.00
05.03 254.00
04.98 254.00
04.94 254.00
04.89 254.00
04.85 254.00
04.80 254.00
04.75 254.00
04.71 254.00
04.66 247.03
04.62 239.06
04.57 232.09
04.52 225.12
04.48 218.14
04.43 212.17
04.39 206.19


TABLE: Maximum Air Flow vs. RPM

RPM Grams/sec

6400 255.00
6000 255.00
5600 255.00
5200 255.00
4800 255.00
4400 255.00
4000 255.00
3600 255.00
3200 255.00
2800 254.00
2400 181.00
2000 125.00
1600 85.00
1200 55.00
800 35.00
400 26.00
00 26.00

 

The vid I posted is not mine. Your words don't create reality.

 

The ecm is kept alive for 11.26 seconds after ign off. When scanning in 8192 baud mode, data will continue to flow during this time as can be seen in the log. This log simply demonstrates what happens when the ign is switched off from a running state while scanning at 8192 baud. Nothing special was done to record this log. It came from a customer's car and just happened to capture the ign off and burn-off cycle, so it is useful for this discussion.

As can be seen, the IAC cycles fully open, fully closed and then opens to the start position to prepare for the next start-up. 160 counts is the start position for 1989 cars (144 counts for earlier cars). This is all normal behavior. I've shown this signal because it serves as very obvious visual indicator of when the ign is switched off.

If scanning at 160 baud, you may not see anything (or only 1 or 2 samples) after ign off due to the slow frame rate and limited (25 byte) 160 baud data stream, so this may explain what you see with your scanner. I don't have much 160 baud data on hand to confirm, but I can do some digging. Sorry, I don't have a laptop available at the moment and I don't want to drag this pc out to the car to test (but have done so in the past). Its an all-in-one touch screen, so kinda like a big tablet, but also kinda ridiculous to use in the car.

Back to code 36: During the 1 second burn off, if the maf voltage is outside of the 400mv-1.9 volt range, a counter will be incremented. After 6 counts, the code 36 dtc will be set. 6* 6.25 ms = 37.5 ms, so it doesn't take long.

You should be able to see this behavior on the maf signal wire with a voltage meter or o-scope, but it is much easier to scan assuming you have the right tools available.

Note: The light green line (unlim airflow diag) is the important signal here since the mass airflow signal is being limited by the table and does not show the effect. The unlimited airflow signal directly follows the a/d conversion (including scaling) but without limitation. It is therefore the better indicator of the actual maf voltage.

Voltage out of range is likely due to bad maf sensor, relay or wiring issues. The behavior of the unlimited airflow signal (light green line in log) should offer a clue as to what has failed, but often it is the maf itself that is the culprit, especially with after-market or replacement mafs. No signal response would point to power issues, low voltage or high voltage out of range could be due to a short to ground or vbat, or a bad sensor.

Here's another log just for fun. Different car this time. It is equipped with a Blowerworks maf, so no burn-off and no code 36 (ever). This log includes maf voltage and exclusive maf air temp signals. Note the different behavior of the unlimited airflow (labelled as maf raw input) and mass air flow signals compared to the original log. The maf signals are not pulled up after ign off in this case due to input circuitry changes and there is no burn-off, so the signals decay naturally to zero with no flow.

https://datazap.me/u/tequilaboy/maf-...rk=28818-28865

 

FYI, my code 36 information was from my '87 and '87 FSM.

Only thought how "fixing" a vacuum leak might trigger a code 36 is if:
a) The burn circuit is failing check.
and
b) The car is not getting to closed loop due to the vacuum change. The burn circuit is only checked after the engine has reached closed loop and turned off. This means the vacuum change prevented closed loop, so no code 36 logged even though burn circuit didn't perform properly.

Seems unlikely this would happen, I'd expect the vacuum problem was properly fixed.

 

Despite people who 'gaslight' on this issue, it's happened to me several times in the past and just recently with a vacuum leak I caused from changing the water pump.

 

Ah, yes. Knowing exactly what the ECM does and what code routine runs for the MAF burn-off cycle and SES code set request is "gaslighting." Today I learned.

 

89 years ago. 87 now. Same thing. I don't know where you get Heinz 57 conglomeration.

 

Always lean something when I read your responses man.

 

Exactly! Which is why it's too bad he has this weird obsession with deleting his posts.

 

Yes I believe it’s because he doesn’t want to get in an argument with people who know less than him. This is a good forum but like all forums there are some that only believe their own ****. Anyway that’s a very quick generalization.

 

I need more information on a program chip for my 1987 Corvette with this code