Code 52 help

I’ve scoured the forums for days and I’ve tried every solution I could find. I’ve been dealing with code 52 for weeks now. I’ve installed a new knock sensor, the ats one. Torqued to 33ft lbs. I’ve ran new shielded wire into the ecu grounding on pins e1 and e2 with no luck on either. Tried grounding to the screws on the ecu case and nothing. I haven’t been able to find a straight answer on whether code 52 can be an actual knock warning or only a bad circuit but I’ve replaced everything so I don’t see how the circuit can still be bad. My timing is right on at 10btdc with the terminals bridged. Compression is 180 around the board. Car is mostly stock so I don’t know what could be causing detonation if that’s what it is. I run 93 octane gas and have for 2 years with no hiccups. Boost settings are at wastegate pressure only. Anybody know what’s going on with this thing?

Given code 52 is for an open or short circuit in the knock sensor signal (not a knock control fault in the ECU), it is possible your ECU has an internal fault and can't process the signal it gets from the circuit. Capacitors in the ECU on these have been known to go bad (and they aren't always visible by burnt legs or leakage on the board.

If you can get your hands on another ST185 ECU and see if the problem still exists, that would probably been your best test to see what the issue is (at least if your present ECU has an internal problem).

Know of any way to test the ecu? Getting another one to test out won’t be easy, I will need a JDM one. I am not very adept at wiring or what to look for.

I don't have a wiring diagram for a JDM ST185. I do for a USDM ST185.

BTW, my manual says the torque to tighten the knock sensor to the cylinder block is 27 ft-lbs (37 N-m). I don't this 33 ft-lbs would have caused any damage though, but it may have (see this thread: http://www.mr2.com/forums/threads/57950 ... n-think-of). This thread also talks about the shielding possibly wearing through and contacting the sensor wire itself. Your type of testing would not pick this up. You need to check your sensor wire to ground to make sure it is an open circuit. If it is not, then your wiring is the problem.

On the USDM ST185 wiring diagram, The knock sensor input wire (white) to the ECU is on pin E5 (the 16-pin connector to the ECU). The shielded part of the sensor casing is grounded to E1 on the ECU, and also through a brown wire associated with the ground on the intake manifold. Is your intake manifold grounded to the chassis? Also, per the wiring diagram, the knock sensor ground is not run back to the ECU - it just shows a physical ground in the manual - but doesn't describe the ground location, but it could be the threaded sensor itself is the ground. Interesting that the service manuals don't describe how to test the knock sensor or even what it does with the ignition system through the ECU.

Hmm. 33 is what I’ve read on most of the forums. I doubt a couple extra would’ve killed it.

The ground on the back of the intake manifold is good. As I understand this is the end point for all the ecu ground wires so you should be able to tap into any ground that goes to this point. I know for certain, based off a diagram I’ve seen that the e1 pin goes to this. I have a colt meter but I’ve never used it and frankly don’t know where to begin. I guess I can check for continuity from the sensor plug up to the ecu and check that I don’t have a circuit between the conductor and the ground wire at the other end. I can also try grounding to get another different source for the shielded wire and see what that does. This was brand new wiring and the sensor is brand new as well so it has to be how it’s hooked up, but I don’t know any other way to hook it up really. I found a pin out Diagram for the JDM models, which most pins are the same across several layouts anyways so I know I’m getting the right wires.

I know this issue plagues a lot of owners but it seems like every solution I’ve found that worked for them isn’t working for me.

Assuming your new knock sensor was a Toyota OEM part number, and your shielded wiring is properly grounding the shield wiring, and that the connection technique isn't somehow grounding the shield wire with the signal wire, then one should assume your circuit is good all the way through. That only leaves your ECU. And I'm sure the internal circuitry in the ECU has capacitors - so it may be an ECU capacitor issue causing an open circuit within the ECU.

If you are not comfortable removing your existing ECU and looking at the circuit board, your best bet is to look for a replacement JDM ECU. Even if you did look at your circuit boards in the ECU, visual inspection itself isn't going to pinpoint failed capacitors. Some capacitors fail internally without leaking and without leg stems corroded off. These are tough things.

Between my son's two USDM ST185's, we have 3 ECU's (one spare). We haven't checked that spare yet to make sure it is OK, just hoping we don't need to. But I'm pretty sure a USDM ST185 ECU isn't compatible with a JDM ST185. And there were differences between 1990-1991 ECU's vs. 1992-1993 ECU's.

Here is a 1990-1991 JDM ST185 ECU for sale on ebay: https://www.ebay.com/itm/1990-1991-Genu ... SwO9RbHvH3. Only thing on this one is there is no warranty on it to safeguard the buyer against it being defective. So the purchase carries a risk.

If you want, you can give me the VIN on yours, I can look it up to make sure what the Toyota OEM P/N is for it. I'm not saying with certainty this is the problem, but if you are certain everything else is fine, there isn't much more to be at fault than the ECU itself. I would be checking the circuitry of what you originally had and what you have now to see if the problem is there first - and the techniques are hard to explain in words in a forum. I would normally put an ECU fault low on the order of things to check - but you've replaced everything else. I just don't know what you've replaced the originals with. If the original knock sensor failed, and you bought a non-OEM knock sensor, you may have purchased a knock sensor that is not compatible with the frequency range necessary for a 3S-GTE engine. So your replacement knock sensor may not work either, causing the same issue as your failed original. And I've put in shielded cable and accidentally got the shielding interconnected with the signal wire, causing a short to ground, which would cause a 52 code. So I know I've messed repairs up so I can't assume my repair attempts are perfect. These are tough. Not to many Toyota techs out there that can handle these 32 year old vehicles anymore either. So I understand.

The new knock sensor is the ats units. Gm sensor with an adaptor for our engines. There’s controversy over how good it is compared to oem but the general consensus is that it does work. I am going to reach out to them again as well. I tested it this morning and I’m getting continuity between the pigtail and the knk pin on the connector. I’m not getting anything from the pigtail to the shield wire so I know my wiring is good now. It’s either the ecu or a brand new ks at this point. The only reason I’m having this nightmare is because I broke the connector on the stock ks while doing a head gasket job. I didn’t have ks issues before that.

As far as the ecu, I think the chances are low but I might have to take it out anyways. But, if it’s not visible there’s not much I can do. Can you do a continuity test on the capacitors? I’d be a afraid to fry the whole board or something. I’d be weary of buying an ecu on eBay. Especially if it’s beat up looking. Heck, I’d almost just get an aftermarket ecu over paying 400+ for a stock one.

Vin is st1850023705. Pretty sure my pin out is correct. I found a picture of every 3sgte configuration and the st185(jp) matched mine exactly as far as missing wires and connection of knock sensors wire, which stands out.

I see from your VIN that you have a 1993 MY Celica GT-4 (JDM) wide body. Your original knock sensor P/N is 89615-50010. This does look like it is still available from Toyota, as many USDM Toyotas used this same part number through 2004, including:
- 1990-1993 MR2 (w/3S-GTE)
- 1990-1993 Celica (w/3S-GTE)
- 1992-1996 Lexus SC300/400 (w/1UZFE)
- 1993-1997 Lexus LS400 (w/1USFE)
- 1989-2000 4Runner (w/3VZE, 3RZFE)
- 2002-2004 Hilux (w/3RZFE)
- 1995-2004 Tacoma (w/3RZFE, 2RZFE)
- 1993-1998 T100 (w/3RZFE, 3VZE)

For our 1990 USDM ST185's, our knock sensor part number is 89615-20020. But the 1993 USDM ST185's show 89615-50010 for the knock sensor. I don't know if these two part numbers are identical (just a different manufacturer) or if the knock sensors themselves are somewhat different.

I would say if the knock sensor code surfaced after you broke your original, then most likely the issue is with your replacement knock sensor.

You might want to check if you have continuity of your shield wire and ground. You should. If you don't, that may be your problem (your shield wire isn't grounding properly).

@transneptunian
all capacitors go bad eventually. A lot of numbers get thrown around, but personal experience is 25 years in automotive for good caps. You can get new ones soldered onto your ecu by a professional (I guess you could attempt it yourself, but I don't have that steady of a hand so I've never considered doing it myself )
That could be affecting it.

Good call on checking the wiring and making sure the shield isn't contacting the signal pin.

I think it's the right move to see what ATS says - maybe they had a bad batch recently.
Barring any activity from Aaron, I suggest getting an oscilloscope. You can find USB ones for ~100 usd, maybe cheaper if you look. You can use it to see the signal coming in from the knock sensor... if there is any.

A bunch of possibilities here. Does the code pop up immediately when you rev the engine above the threshhold RPM, or only sometimes? If the latter does it get more frequent as the engine heats up?

If you broke the original plug lots of other Toyotas use the same one. I took one from a 93 5SFE from a Celica to make my replacement wiring. I had both the wiring fail internally and a defective knock sensor (it was brand new). Troubleshooting that after a big overhaul was tricky. Don't assume anything works until it has been proven to work.

ECU caps will fail with age. That's just a given, so best to replace them around this age before they leak and cause damage. It's about $10 in parts and took me an hour to do so if you're not able to do it yourself it should be cheap to have an electronic shop do it for you. You can always grab an ECU from the junkyard to practice on before working on the one in your Celica too, that way you can get the hang of it with no risk.

Just to update anyone on the outcome of these issues, looks like the problem was possibly the ATS sensor the whole time. When I went to take it off and inspect it, the gm section of the sensor snapped off. I took it off and there was something jingling around inside. I did notice that with the ATS unit installed, it rubbed right up against the tvis controller switch and I’m wondering if the sensor was pushing down on the sensor and causes it to crack under the stress.

Anyways, I sent that unit back to ats for a refund and bought an oem sensor. I had already cut the factory wiring so I cut off the factory pigtail and spliced it into my new wiring. My final routing was to the knk pin for the signal wire and I scorch locked the shield wire into the e1 pin. Long story short, everything works now. No code 52 and no more limp mode. I know ATS is highly respected in the community but after this, I wouldn’t recommend using their gm based knock sensor. Just spend the extra 50-70 bucks and get an oem one.

I appreciate everyone’s input and am glad I didn’t have to open up the ecu. Hopefully this thread becomes helpful to anyone suffering similar issues on their car in the future.

transneptunian, thank-you for providing your follow-up on how you got to a successful ending. It's important to the forum to hear how you got there, and what works as well as what doesn't. Thanks for taking the time to complete the story behind getting your situation resolved successfully.

For sure. Part of the issue I had searching the forum is too many unresolved threads. Hopefully this saves someone else several weeks worth of troubleshooting and ordering various parts.