[Supras] Factory shielding of knock sensors

[berniek at technicaldevelop.com]

Dear Arlene:

Please see in-line responses, hopefully in blue:

Lanman, Arlene wrote:
>
> Bernie,
>
> Perhaps, I may be wrong – I have been known to be wrong several times 
> in my life. So have I, more than I care to remember!
>
> Knock Sensor Shielding: What is the primary source of the stray 
> signals that will be taken to ground by the Knock Sensor wiring shield 
> ground? I suspect, they will originate from the engine compartment, 
> most notably by the spark plug wires, and nearby coil and other 
> electrical devices (alternator, etc). So where is the ground for all 
> of these nearby electrical signal generators – Ground B on the engine 
> block. So, does it not make sense to ground the shield of the knock 
> sensor wire to Ground B – as Toyota has done? I would believe your 
> statement to be correct, since the engine is the electrical datum for 
> all EFI sensors as well as the ignition system. I'm not sure I fully 
> understand why the ECU shield ground works. I'm going to bond the two 
> together with copper braid.
>
> ECU: As others have said, Reg has the most knowledge, due to his 
> development of his piggyback module. Numerous ECUs arec listed within 
> the electronic parts catalog. Some of the major variations are:
>
> 5F (manual 5-speed transmission, floor shifter) – also, referred to as MTM
>
> 4FC (4-speed automatic transmission, floor shifter) – also, referred 
> to as ATM
>
> TWT – Twin cam turbo (turbo vs. NA)
>
> CAL – California 2^nd O2 sensor
>
> Location (assume fuel maps are different due to emission regulations 
> and starting maps may be different due to climate differences):
>
> G = Europe (1983)
>
> E = Europe
>
> CND = Canada
>
> USA – United States
>
> Difference in connectors between 87, 88, early 89 and late 89 on.
>
> Altitude sensor is external to ECU for 87, 88, early 89 and internal 
> late 89 on. Yes, I've read the same information, and 8908 to 8910 date 
> codes carry different part numbers from 8808 to 8908. But comparing my 
> mid-'89 and '91 ECUs internally, there is no difference except for the 
> formerly stated input diodes and resistor
>
> An additional balancer was added to the crank on late 89 on and, 
> hence, redline is lower on late 89 on than previous; also, change in 
> differential to compensate, some what – given this, I would suspect 
> fuel maps between early 89 and earlier is different than late 89 on. 
> Could be. Perhaps emission limits changed in that period as well.
>

Hope you have been well and not working so hard. Did you ever send your 
car off to Reg for updating? Take care.

Bernie
>
> I hope this helps
>
> Arlene Lanman
>
> 88T automatic with a few mods
>
> Stillwater, OK
>
> Tele: 405-744-7131
>
> arlene.lanman at okstate.edu <mailto:arlene.lanman at okstate.edu>
>
> Message: 1
>
> Date: Mon, 23 Jul 2007 17:53:30 -0400
>
> From: "berniek at technicaldevelop.com" <berniek at technicaldevelop.com>
>
> Subject: Re: [Supras] Factory shielding of knock sensors
>
> To: Hacker J <jonbhacker at yahoo.com>
>
> Cc: supras at supras.com
>
> Message-ID: <46A5235A.6060709 at technicaldevelop.com>
>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> Jon:
>
> You are absolutely correct about ground loops. Grounding in this 
> instance is a conundrum of sorts, because of ground noise elevation of 
> the engine block with respect to the ECU. A major factor in the 
> equation is internal capacitance of the sensor, which if low would 
> indicate grounding to be better at the ECU. (wait) I just grabbed a C 
> meter and found the capacitance to be about 8000 pF, far from low, but 
> characteristic of piezo devices generally. Also, I lightly clamped the 
> sensor in a machinist's vise by the threaded end and tapped both the 
> vise and the body of the sensor with the plastic handle of a #1 
> Phillips screwdriver. This is just a qualitative test, obviously. The 
> sensor developed about 1V-5V peak to peak on the oscilloscope, with a 
> ringing frequency of about 7 KHz. That frequency may be mechanical 
> resonance arising from the sensor mass and cantilevered mounting 
> arrangement from the threaded end.
>
> Maybe use of substantial ferrite beads on the coax from the ECU would 
> serve for assuredness of high frequency common mode decoupling with 
> the shield grounded at the ECU. I've ready of many instances where 
> grounding at the ECU is better, but if so, why did the factory get it 
> wrong for so many years?
>
> BernieK
>
> Hacker J wrote:
>
> > Bernie,
>
> >
>
> > The knock sensors shielding should be grounded from the ECU end only 
> to avoid ground loops. The actual ground used by the factory is the 
> main ECU ground (so-called B ground in the wiring diagrams) which 
> ultimately connects to the bunch of ground wires that get bolted to 
> the intake runners. I would not deviate from this arrangement.
>