1989 TPI. 99% stock.
Code 34 & 36 Intermittent. It's driving me crazy.
New MAF, still happens.

The only mods are a Holley TB, and Accel injectors. But those were years ago.

Now it's difficult to start. Stalls out... Once I catch it with the gas pedal, she'll stay running. But something is not right...

What would cause an intermittent 34 & 36 ?

 

You might try rapping on the ecm with the engine running.
If it has a bad solder joint the flexing would cause bad running.
You might want to go through the flow chart as if the problem isn’t intermittent.
http://www.chevythunder.com/maf_code_36.htm

 

Similarly, you may want to inspect all connections in the MAF circuit, including the relays and their wiring.

 

I changed the MAF relay.... and the car is running much much better. Better than it has in years. Only drove 25 miles, but a lot, if not all of the intermittent problems seem better.

Time will tell, but I think I got it. Certainly a lot of weird **** going on drivability wise. So intermittent, and it got worse over the years. Hopefully the MAF relay got weak with age, and this is the solution. We'll see, and I'll update the post in a few weeks.

 

I hope you got it solved. A intermittent issue can be a real pain to diagnose.

 

More weirdness: since changing the MAF relay, the car runs amazing. All of the stalling, and hard starting is gone. Runs like new.

However, I am still getting a code 36. So I'm changing the ECM next......

 

Before you go that far, check out the MAF burnoff circuit and relay closely.

 

Although the burn off relay circuit tested good; I replaced the burn off relay, and the ECM.

Thinking about it.... about 5 years ago, the car stopped communicating with my auto x ray. I thought the auto x ray went bad. So I've been reading codes with the paper clip method.

Now... I can data log with the auto x ray. So I'm convinced the ECM was flaky. We'll see if codes come back, but the car runs nice...

 

Code 36 is haunting me. It comes back regularly.
New ECM.
New Relays.
New MAF.

Checked all vacuum lines.
Changed PCV grommet.

 

Check the auxiliary oil pressure switch.

 

The reason I mentioned the oil pressure switch is because that's the signal to the ECM to begin the burn off cycle. If the switch is intermittent it interrupts the cycle and sets the code. I encountered one of those once and will never forget it, although all the pulled-out hair has since returned.

One thing to note is that the burn off cycle does NOT occur at every engine shut down. Trip timer and number of starts factors into when the cycle is run. The ECM will look for a maximum MAF signal during the cycle to prove the burn-off was successful. If the connections at the MAF are poor, relays are unreliable, or circuit connections are faulty, the code will also set.

If the MAF itself were bad the ECM would tell you right away, since it only tolerates a MAF signal out of range for 600mS. Control is pretty tight that way.

The relays are the easiest thing to check, and in some years you can simply exchange the MAF power and burn off relays to test this. If the relay part numbers are the same they can be exchanged, but if they are both failing you may not learn anything. If you do this and clear the code you may have to wait several days to see if the code returns. Happiness is not instantaneous with this part of the system.

Another thing to note is that BOTH relays are necessary to complete the burn off since power to the MAF is supplied through the NC contacts of the power relay to perform the cycle.

 

I put the original OEM MAF back in the car last night. So far, so good. No codes.
I'm thinking.. possibly.. when I was throwing parts at it, to solve code 34. The MAF I brought failed in short order, and is causing this code 36.

I know I have to drive it another few days. But today was good, and the car runs amazing.

 

I'm reading... some replacement MAF's don't have the burn-off circuit. So it throws a code 36. Not sure how true that is.. but my thoughts are the aftermarket MAF's are not fully compatible.

 

Thick-film (TFT) MAF sensors do not have, nor do they need, a burn-off function. The best ones, however, have a circuit to "fool" the ECM into sensing a successful burnoff cycle. The old Wells SU-145 did that, and still does on my '86.

The unit now has a different part number:

https://www.rockauto.com/en/moreinfo...=1193496&jsn=3

 

Its not intermittent. Code 36 only throws on $6e after the ECU counts 6 burn-off failures in a row.

If its a MAF without the burn off function then burn a new PROM with code 36 disabled.

GD