93 Legacy throttle issues

[Bobs93Legacy]

My 93 Legacy just suddenly developed a weird issue.

It starts fine, but from a dead stop, when you try to accelerate it won't or if it does it does so VERY slowly, until it hits about 1500 RPM.

Above 1500 RPMs it runs and drives OK. But it is nearly impossible to get it to rev up from its idle speed of around 700 RPMs to around 1500 RPMs.

I've done a little research and figured out that it has a MAP sensor and a MAF sensor, but no throttle position sensor. Is that correct?
In my way of thinking it can't be a fuel issue (fuel pump, fuel filter, injectors) because it it were something like that throttle response would drop off at higher RPMs - not improve at higher RPMs.

Seems like I read somewhere that it might possibly be an O2 sensor failure?

So has anyone else experienced this issue? If so, what did it turn out to be? MAF, MAP, or O2 sensor issue? Something else?

Edited October 28 by Bobs93Legacy

[Bobs93Legacy]

No help for a newbie?

[moosens]

Tough call. And I’ve seen where a failing due pump does show some of what you’ve mentioned. But so would a cracked sensor or loose connection. 
 

I use my 94 Legacy as a daily beater. Can’t say I’ve really had this issue on those cars 90-94

 

Bit this also sounds similar to what I’m getting right now with my wife’s 2006. We have a broken cam follower sensor connector. So it’s going into default mode randomly but if I push my foot into the pedal I can get the throttle to act as it should and actually drive normally - until the next fit it throws. 
 

Just my thoughts so you’re not alone in the wilderness here. 

[Bobs93Legacy]

4 hours ago, moosens said:

Tough call. And I’ve seen where a failing due pump does show some of what you’ve mentioned. But so would a cracked sensor or loose connection. 
 

I use my 94 Legacy as a daily beater. Can’t say I’ve really had this issue on those cars 90-94

 

Bit this also sounds similar to what I’m getting right now with my wife’s 2006. We have a broken cam follower sensor connector. So it’s going into default mode randomly but if I push my foot into the pedal I can get the throttle to act as it should and actually drive normally - until the next fit it throws. 
 

Just my thoughts so you’re not alone in the wilderness here. 

FWIW, I grew up working on vehicles that had carburetors and mechanical diaphragm pumps for fuel delivery, and a distributor with points & condensers for spark.

This 93 Legacy was built when those types of mechanical systems were being replaced by computer controlled "engine management systems" - but before there was any kind of standardization (like OBDII).

That makes diagnosis of this kind of issue REALLY challenging for an old-school guy like me. 

This sort of driveability problem isn't simple, like rebuilding the carb or resetting the points.

BUT, at the same time it isn't a simple matter of connecting it to a diagnostic tool and letting the onboard computer tell you what is wrong either.

SO, any help or advice the experts on this board can provide would be VERY much appreciated.

[moosens]

I’m the same guy as you. Vintage 1963 youngest of three and middle brother was a mechanic before they called them technicians.

There’s online service manuals out there too. 
 

 

[lmdew]

Even it's old you can still pull codes from the computer.  Any Check engine light on?

Clean the grounds and connectors.

Be very systematic in your troubleshooting, documenting what you do and if the fault changes with any maintenance.

 

 

[Bobs93Legacy]

First, I found MAF sensors and MAP sensors on a couple of auto parts sites - but no throttle position sensor (TPS) - which made me think there isn't one. Does Subaru call it something else? Because there is definitely a sensor/rheostat on the shaft of the throttle body.

So, I did some troubleshooting yesterday.

Unplugged the MAF sensor (on airbox front corner of the passenger side of the engine compartment) and it died instantly

Unplugged the MAP sensor (on center of firewall?) and it died instantly

Unplugged the sensor on the throttle body (TPS?) and it kept running, but the throttle issue stayed the same. Plugged it back in and the throttle response issue seemed to go away - but only for a short time.

I replaced that sensor with one I pulled off of a parts car I have, but the throttle response issue remained. Though I don't know that the replacement part was any good either.

I used a pick to bend/tighten up the female connectors in the plug to the sensor on the throttle body, but that didn't seem to help either.

I'm really stumped on this one.

HEEEEEEELLLLLLP!

 

[el_freddo]

The MAF is required to run the engine. The “MAP” sensor as you call it is the ignitor, also needed to run the engine. 

The sensor in the throttle body is the TPS. They can get dirty contacts. It’s worth opening them up and cleaning the tracks that the fingers follow. Once replacing it make sure you’re following the procedure to have it properly installed. 

The MAF could need cleaning (VERY carefully!) or replacing. 

Also look at the knock sensor, back of engine near the bell housing on the LHS, got a 10mm bolt head holding it in place. It’s black and probably cracked by now. This won’t usually throw a code if it’s got an issue until it’s really proper dead. 

If you’ve pulled plugs off items when the engine is running you’ll have to reset the ECU if you want to accurately read codes. The 30min “battery dance” will clear the ECU memory of new and old codes. 

Start the engine again and let it idle up to temp. The idle speed will fluctuate - this is the ECU setting the idle etc. DO NOT touch the throttle once the engine is started until it’s up to temp. 

Drive it like you stole it and see if the issue persists. The issue will return, then check codes and start with the sensor the codes point to. 

I hope this helps. EFI systems are nothing to be scared of if you’re coming from the dark ages of carbs and dizzys. They will generally give you a starting point with a code or two  

Cheers 

Bennie 

Edited November 2 by el_freddo

[Bobs93Legacy]

Thanks Bennie.

Just curious why I can't find a TPS listing on Autozone, Oreillys, Napaonline, eBay, or even Rockauto.

Does Subaru use another name for it? Or is it a dealer-only item?

Edited November 2 by Bobs93Legacy

[moosens]

Random parts numbers form the 95/96 2.2 Legacy catalog if this helps at all.

The part that bolts onto the passenger side of the TB is called a Throtte Sensor - yup - we all call it the throttle position sensor, but who are we? Part number 22633AA110 and that’s for cars spit out from the factory from 03/94 through 05/95 and also 06/95 - going forward at time of print. 
 

Good luck with it. 

[Bobs93Legacy]

moosens - thanks for the info. Unfortunately, since mine was built before 03/94 (1993 model) it doesn't look like the one I come up with when I search on those part numbers. That part looks like it has a "tab" that engages a slot in the throttle body shaft. Mine has a round shaft that is flat on one side and the sensor has a hole shaped like the end of the shaft - it slips ONTO the shaft itself.

Imdew - sorry I missed answering your question. It isn't showing a check engine light - but it is possible that the bulb could be burned out.

FWIW, I found a video that shows there are two little black connector  under the dash to the right of the steering column that can be plugged together and make the check engine light "blink" the codes. I haven't tried that yet because I'm out of town until the 15th. That will be my next step. Hopefully the CEL bulb isn't burned out!

Edited November 5 by Bobs93Legacy

[forester2002s]

On most models, all the warning lamps are tested during start-up:

- turn the key to the ON position (just prior to turning the key to START the engine); the display should show all warning lamps illuminated, including the check-engine-lamp; this verifies which warning lamps actually work (or not).

[Bobs93Legacy]

On 11/4/2024 at 10:04 AM, forester2002s said:

On most models, all the warning lamps are tested during start-up:

- turn the key to the ON position (just prior to turning the key to START the engine); the display should show all warning lamps illuminated, including the check-engine-lamp; this verifies which warning lamps actually work (or not).

That's a helpful tip - thanks!

[el_freddo]

Pls delete

Edited November 6 by el_freddo
Pls delete

[el_freddo]

On 11/5/2024 at 4:04 AM, forester2002s said:

On most models, all the warning lamps are tested during start-up:

- turn the key to the ON position (just prior to turning the key to START the engine); the display should show all warning lamps illuminated, including the check-engine-lamp; this verifies which warning lamps actually work (or not).

^ this. Easiest way to ensure the check engine (CEL) light is working. 
And yes, those black plugs are the two you connect to read the codes. Such an awesome way to read codes for the DIYer! 

I found part number 22633AA060 for the TPS. It’s listed as “Sensor Assembly - Throttle”. This part is from a 1992 AUDM VIN search. This is also the same part for a 1992 USDM SS Legacy turbo (2.2L) and the 1993 AUDM RS turbo (2L). So I’d safely say that this TPS is used across the board on the Gen1s. I haven’t looked into the Gen2 Liberty/Legacy.

Cheers 

Bennie

[Bobs93Legacy]

el_freddo - thanks for the part number info. VERY helpful. A Google search for 22633AA060 pulls up several links - but unfortunately most of them say the part is "obsolete" or "unavailable". I was able to find a couple of places that sell that specific part. It seems that this particular TPS may have only been used a very limited number of 1993 Legacy models produced. As a result they are VERY expensive. Just my luck I guess 🙁

I'm still not 100% clear on what you previously referred to as the "30 minute battery dance" to clear the ECU codes. Can you tell me a little more about that?

Sorry for asking what may seem like such basic newbie questions, but this is the first early 90's Subaru I've ever had to troubleshoot. 

Thanks for all your help and advice guys!

Edited November 9 by Bobs93Legacy

[lmdew]

Disconnect the negative battery lead.  

30 minutes late hook it back up.

 

[Bobs93Legacy]

Thanks for clarifying that Imdew. I thought that disconnecting the battery for 30 minutes was what that meant, but it is very helpful to have that assumption confirmed.

FWIW, I've never seen or heard of it taking anywhere near that long for a computer to reset when the power is disconnected. I'm guessing Subaru must have put some kind of capacitor-based memory backup in their ECUs for them to require such a long power-down-reset cycle.

Most computers will reset in under a minute when disconnected from their power source.

[el_freddo]

Bob you’re probably right about that timing of the battery dance. I guess someone was playing it safe to ensure codes are cleared. 

30 mins is a good time to boil a kettle and have a cuppa tea & a bikky  

No worries on that part number. From what I can see it’s not just the 1993 model that uses it. I should find a VIN for a Gen2 and compare, I bet it’s the same though! 

Cheers 

Bennie

[nvu]

Looking at the TPS, I haven't seen those old square ones first hand, but it looks like you can take apart the cover on those?  They are probably just potentiometers. if you can take apart the cover, spray on some electronics contact cleaner.  It should work again at least until you can source a replacement.

[el_freddo]

Correct, they can be pulled apart and cleaned. 

I had one cleaned up by a mate and it worked well. I wasn’t there to see the process, I believe it wasn’t that difficult. 

Cheers 

Bennie

[Bobs93Legacy]

Well, I'm back in town and will be taking another crack at getting it running on Thursday afternoon (it's over at a friends house on the other side of town right now).

On your advice, I had him disconnect the battery so the codes will all be cleared by the time I get over there to try and get it running.

At this point I'm thinking it isn't the TPS since I've already tried swapping one off my spare parts car and it made no real difference. I'm thinking that if the MAF is bad then it wouldn't throttle up properly because it wouldn't be correctly measuring the increased airflow - and therefore the ECU wouldn't supply the right amount of fuel to maintain the correct air/fuel ratio. Logically, that would match the symptoms I've seen so far. But if I'm wrong about that or missing something, please feel free to correct my troubleshooting process. 

So, based on that theory, today, I pulled the whole top of the airbox assembly (including the MAF) off my parts car to make it quick and easy to swap out the MAF. That will be my next step in trying to troubleshoot this issue.

If swapping the top half of the airbox and MAF doesn't help, the next thing I'll try is to swap out the knock sensor (thanks for that suggestion el_freddo). I just watched a u-tube DIY video that showed me exactly where it is located and how to replace it. I'll pull the knock sensor off my parts car tomorrow and take it with me on Thursday. If swapping the MAF doesn't solve the problem then the knock sensor is next.

Thanks for all the suggestions and advice guys! You've given me some really good pointers on what to check. 

The absolute worst-case scenario is that if I can't figure it out & fix it, I'll use my AAA membership to have it towed home where I can work on it at my leisure. 

Stay tuned - I'll post back to this thread with updates on my progress and what I find. Thanks again for all the helpful suggestions...

Edited November 20 by Bobs93Legacy

[Bobs93Legacy]

Well, unfortunately swapping the top of the airbox and MAF didn't do the trick.

First thing I checked was to see if the CEL came on when I started it and it lit up as it should.

Next I disconnected the battery for 30 minutes as suggested to clear the codes from the ECM.

I then plugged the two black diagnostic connectors under the dash together and reconnected the battery and started it up.

No change in symptoms (of course).

I shut it off and then turned the key to RUN.

The CEL started flashing. It appeared to flash a code 31.

I say "appeared to" because I am unclear whether the CEL lights up and goes off before it starts flashing stored codes - or does it immediately start flashing codes without first lighting up to show it is working.

If it starts flashing the codes immediately then the code is 31 - TPS sensor/sensor wiring

If it lights up once (just to show it is working) and then starts flashing codes, then the code is 21 - Engine Coolant Temperature (ECT) sensor.

I'm kind of doubtful that the ECT would cause the lack of throttle response that I'm seeing. If I'm wrong about that hopefully someone will correct me.

So, I'm thinking that it is indeed Code 31 - indicating an issue with the TPS or wiring. However, since I have already swapped out the original TPS with one from my parts car, and nothing changed, I'm leaning towards it being an issue with the wiring/connector. While it is possible BOTH of the TPS units I tried are bad, it seems kind of unlikely.

Per el_freddo's suggestion I tried to remove the cover of the original TPS, and didn't have much luck. The end opposite the connector seems to have two tabs that can be pried open with a little screwdriver, but the cover still doesn't want to come off. I tried working a screwdriver tip into the joint between the cover and steel base plate on the sides, but it doesn't seem to have any kind of tabs or notches that can be pried apart to "unlatch" and release the cover. I'm afraid of breaking it, especially since they seem to be such a rare part. 

So, next will be using my VOM to trace and check the wires in the TPS connector plug and make sure they aren't broken somewhere between the ECM and the TPS.

I'm also open to any other comments or suggestions, as well as the question of whether the CEL should give one "preliminary" flash before it starts flashing the actual codes, and therefor whether the code it is flashing is actually a 31 - or is it a 21? Lastly, I'd appreciate any input anyone might have about whether a bad Engine Coolant Temperature sensor could cause the kind of issues I'm having.

Basically I still have more questions than answers at this point 😜  

 

P.S. I had it towed to the house using my AAA benefits. Since my membership renews at the end of the month and I hadn't had a tow yet this year, I had nothing to lose, and now it is home where I can work on it as I have time without making a trip across town.

Edited November 27 by Bobs93Legacy

[Bobs93Legacy]

After viewing a couple of online videos about how to test a TPS on the vehicle, I decided, just for grins to use my VOM to check the pin-to-pin resistance across every combination of the 4 pins in the original TPS, since it is off the car and handy.

For each pin combination I tested the resistance with the TPS turned full closed throttle to wide open throttle.

If pin 1 is on the bottom and pin 4 is on the top when the TPS is mounted to the side of the throttle body, here are the measurements I got

 

Pins                                       Closed - WOT

______________________________

1 – 4            measured           10.59k – infininty

1 – 3            measured           10.59k – 220 Ohms

1 – 2            measured            0 – infinity

2 – 4            measured            10.59k – infinity

2 – 3            measured            10.59k – 10.59k

3 – 4            measured            0 - infinity

For combinations 1-4, 1-2, 2-4, and 3-4 that produced an infinity (open circuit) measurement at one end of the scale, the reading changed instantly from its initial value to infinity as soon as the TPS left the closed throttle position.

For combination 1-3, the resistance measurement seemed to move across the range in a progressive and even fashion as the TPS moved from the closed to wide open throttle positions. 

For the pin 2-3 combination there was no change.at any point between the closed and fully open positions.

I still need to perform the testing procedure with the TPS on the car, but I suspect that the progressive linear increase in resistance between pins 1 and 3 - which should result in a progressive linear increase in voltage across those two terminals when it is installed in the car, means that the TPS is OK. I really hope so - new ones are unobtanium, used replacements are around $100, and a reman is around $250!

So I'm even more suspicious that the problem is in the plug or the wiring.

Does that seem valid? Any other thoughts?

Edited November 27 by Bobs93Legacy

[el_freddo]

Fast running out of lunch break atm!

From memory on the back of the TPS there are two black tabs that are kind of melted. Drill these out and the black cover should now come off to expose the fingers and their tracks they follow. 

Can’t reply about values you got. I do know there’s a switch or two that go to Open once it’s off idle. 

No check light with the self diagnosis mode with the CEL. It goes straight into throwing the codes then repeats. So if all you’re getting is code 31 over and over that’s the only code you have. 

Cheers 

Bennie