I've worked around an engine builder that was a firm believer that the stud threads/block threads needed to be under their normal strain load while the loctite cures.
That is one hell of a good idea :thumbsup:
ARP makes this point very clear in their literature also. I'm not sure how long the various grades of loctite take to set up, but this approach sounds problematic to me.
Seems like a soft or semi soft slow-setting sealant on the threads would be best.
I wonder how to test exactly what the leak mode is... iow, is the coolant forcing it's way past the thread sealant? And if so, is that any different for bolts? Or is the leakage occurring under the shoulder of the bolt/nut, where the TTY bolts have the ring of orange silicone pre-applied? If this last bit is the culprit, seems like installing the studs, then carefully applying a ring of RTV at the stud:washer:nut joint might do the trick.
How often do TTY bolts leak coolant into the valve covers? Don't think I've read any reports of this, but maybe it would be difficult to notice? Just as reported with the studs... not sure I'd have caught a slightly overfull dipstick reading if not forewarned. (THANK YOU chevydiesel !!) I'd be watching for milky oil and thinking everything was fine.
So for thread sealing:
TFE sounds like it would ideal, but the results aren't consistent given the different reports I've read about here.
RTV sounds like the consensus best compromise solution, but I wonder... what's it like trying to chase the threads next time the heads come off? Sounds near-impossible to get it all out.... maybe using a chasing tap, then very stiff round wire brushes similar to those used for engine oil galley cleaning? Seems like you'd push a fair amount of RTV scraps into the block. Maybe flushing with the rad hoses disconnected would prevent sending all that debris into the core tubes.
For a simple problem, this sure gets complicated.
We need an accelerated test routine. I'm willing to conduct some testing, as I'm going to use studs in this next motor. Anyone have suggestions about how to structure the testing to really nail this down?