Here's something I've been reading up on. I'd known that attaching a voltmeter to an O2 sensor can produce a nifty mixture gauge, useful even on completely non-computerized engines, but I hadn't given much thought to knock sensors. Many modern engines achieve impressive specific outputs on indifferent fuel by being designed close to the ragged edge and then using electronics to adjust away from impending knock when and as necessary. I got to wondering if a knock sensor might also have a use on a traditional-style engine, allowing similar extremes of tune. My first thought was of a circuit wired into the distributor LT signal lead, which shifts the phase of the signal in response to knock sensor inputs, but apparently it is not electronically straight-forward to match the delay to engine speed. If it were it would have been cool if such a unit were to have a manual override operable by a steering-column advance lever like on early cars, as a sort of bonus. From what I could learn, the J&S SafeGuard has cracked the delay/rpm thing, using digital choreography involving look-up tables and 555 timer chips and what not. The unit would certainly work on a points-and-condenser engine; indeed it is advertised specifically as such, and many on various forums have expressed considerable satisfaction with it. Especially nifty is its ability to retard timing on individual cylinders. Thus by all accounts there is nothing wrong with the SafeGuard, but I'd still prefer something which works on simple ****ogue servo logic, like a vacuum brake booster. For those who don't know, knock sensors are microphones which pick up vibrations in the block, head, or intake manifold. There are narrow-band sensors tuned to specific cylinder bore ranges, and wide-band units which need electronic filtering to isolate knock vibrations from all the rest. All of them need some basic filtering as well as signal amplification. There is some debate over whether any filtering is adequate at over 5000rpm. Some computerized engines pull timing and/or enrich the mixture at high engine speeds regardless of any knock sensor signal, but many don't. Perhaps it's just that nobody runs an engine at 6500rpm for half an hour at a time, so in real life there isn't all that much scope for high-rpm knock to do a lot of damage. Knock gauges, lights, and monitors exist. I especially like this inexpensive filter/amplifier unit, which outputs to an LED light, because it should be possible to use that LED output to trigger all kinds of things. The first instinct of the traditional-minded would surely be to retard the timing physically at the distributor. The obvious way to do this would be via the vacuum advance mechanism, for instance by inserting a floating solenoid and spring into the vacuum advance actuating rod, or having a fixed solenoid move the entire vacuum advance canister back and forth against a spring. It may or may not be necessary to modify the breaker base plate to allow for another 5-8° of rotation: the Ford CVH distributor I've got lying on the shelf here has definite rotation stops but as far as I could tell by a Google search many distributors might not. It would certainly be necessary to modify the vacuum advance rod and canister mounting to allow for more travel. Done this way, it would be possible to send current to the solenoid via a rheostat and thus rig an advance lever on or below the steering wheel, so as to enable static timing to be shifted ATDC just in case I want to start the engine with a hand crank. And in use it's pure servo logic: timing retards when there is a signal and reverts when the signal disappears. A bit of mechanical hysteresis or inertia should dampen that nicely. Another way would be to run a fuel spray bar or injector under the carb(s), fed via a solenoid valve as on a nitrous setup. I think we've all had occasion to calm a ping by giving it a shot with the accelerator pump. This would do the same automatically, using the same servo logic. A third way, on a forced-induction engine, would be to bleed boost or, in the case of a turbocharger installation, open the wastegate. And all these could be used together, with some way of prioritizing one over the others as may be necessary. At the very least it'd be nice to have the "knock" LED on the face of a big old Smiths or Jaeger tachometer. I'm sure someone has thought of all of this before. Perhaps someone has studied this and learned that none of it is possible. Any thoughts?
Interesting concept, but it just seems like trying to reinvent the wheel to me. On any engine covered by this forum the tried and true way would be to set it up on a distributor scope, set the advance curve and springs on the bench, then set timing and advance until it pings. I'm not sure anyone who is really interested in sticking with points distributors would be interested in the technology you are suggesting, and anyone interested in messing with knock sensors is probably more inclined to use a modern engine aswell.
From what I read a few years ago because of the way older blocks were constructed ,knock sensors aren't all that effective ?? They may have overcome that ??
Very interesting post. Innovation is the life blood of Hotrodding. I have a hard time tossing out innovation in order to be "period correct".
The second method- isn't that what GM did with the "Jetfire" Olds? Run methanol (I think it was basically windshield washer solvent) and inject it to avoid detonation under boost? https://www.carthrottle.com/post/oldsmobiles-jetfire-was-a-world-first-turbo-car-with-a-fatal-flaw/
Something I've been very interested in. Not so much for on the fly adjustments, just to be able to know it's knocking with my bad hearing and rattly heap. I'd put my foot in it a lot more if I was sure it was safe from detonation.
The Jalopy Journal
