Methanol injec. help

[85maro]

Not sure where I should've put this thread, anyway I'm cobbling together a methanol/water injection system for detonation suppression and I want to use a knock sensor connected to a relay as the switch, problem is knock sensors only put out a ~1volt signal, obviously not enough to actuate a relay. I want to keep it as simple as possible (no computer), anyone know of a relay that can operate using a voltage that small? Or am I looking at this completely wrong? Thanks.

[Dave96z34]

If its already knocking its too late for it to realy help out. You could put it on a tps switch or some thing that ties in to a map sensor.

[Joseph Upson]

If it's for a boosted application you may want to look at a pressure switch since as previously stated the objective is to prevent detonation.

Radio Shack sells relays that are often found in electronics equipment that will likely operate in that range. I used one about the diameter of a pencil eraser and maybe a half inch long connected to a basic stamp micro computer to fire an injector that used a different power supply.

[smilinguy99]

An electromechanical relay would have a lot of latency (delay between the signal and the starting of the pump or whatever it is you want to switch). In this situation it could prove fatal for your motor at higher speeds.

Better to use something electonic like a transistor circuit. There must be someone on this board who could design a nice simple reliable little thing you could use.

[85maro]

Dave96z34- It's not knocking (infact it's bone-stock) but I plan on running about 12.5/1 compression on race gas when I build it up so I need something to make it more forgiving. Your setup could probably benefit from a methanol system, some guys say the right setup will give you another 10-20% more power.

joseph- I don't plan on boosting it so a hobbs switch wouldn't help. I'll check out radio shack in the mean time, but like I said I want to keep it basic.

smilinguy99- How much delay do you think it will have? Couldn't be more than a few milliseconds, right? Although it may take a little extra time for the pump to build pressure at the nozzle even with a check valve. It's also tied to a switch on my dash so if I feel the need I can flip it on manually. Low rpm and high load situations are really where an engine's prone to detonation.

[Dave96z34]

Then a tps switch would prob suite you more.

[85maro]

A tps switch would'nt do me any good, I need it to activate during detonation, not at a set throttle position. I checked with radioshack and the smallest relay they had was a 3.5volt one, maybe an electronics manufacturer will have what I need. Thanks for the help everyone.

[jwvess00]

Hi there!

You can't simply connect a knock sensor to a relay for a lot of reasons. The sensor's signal needs to be filtered and it does not have enough drive to fire a relay.

On newer (!) GM cars, the knock sensor is directly connected to the ECM, and the ECM contains the filtering circuitry. Older GM cars' knock sensors were connected to an external knock filter that connected to the ECM. They're simple to use -- they get 12V, ground, and knock sensor in, and they output a 5V signal when knock is detected. For a V6/60, you could get the sensor and filter from, say, a 2.8V6 TBI S-10.

I don't know if the older ESC knock filter has enough drive to fire a relay but that's the system you'll want to use to provide your methanol injection system the knock signal.

[smilinguy99]

smilinguy99- How much delay do you think it will have? Couldn't be more than a few milliseconds, right? Although it may take a little extra time for the pump to build pressure at the nozzle even with a check valve. It's also tied to a switch on my dash so if I feel the need I can flip it on manually. Low rpm and high load situations are really where an engine's prone to detonation.

The system you are describing with all mechanical and electromechanical components is going to be very slow. The response chain is serial, meaning one thing will have to respond after another has responded, so latency stacks up. K/S response + relay response + pump delivery response = total response time.

It is also going to be hard to find a relay to pull in using only 1V. That knock sensor is going to have to have its signal amplified so that some part, whether mechanical relay or solid state relays can do the switching. There are many SSRs out there, but none that use less than about 3.5V input. Many SSRs have 15-20µs of latency in the relay itself and an SSR will be faster than any e/m relay.

I would be most concerned with detonation around the torque peak where 1) cylinder pressures are highest and 2) where more detonation events will occur before any system be it human, electromechanical, or electronic can respond to it. At 5000 rpm for example, the time between combustion events on a single cylinder is 24µs. Not much time.

EDIT---That ESC knock filter described above might be able to drive a SSR.---/EDIT

What do the commercial kits use to switch this thing on? Are there any knock-sensor driven kits? Seems I've heard of the TPS switched stuff that Dave96z34 mentioned.

[85maro]

I thought about using an esc, but the problem is the esc sends a signal to the ecm until knock is detected, then it drops the signal. Not sure why GM programmed it that way, seems kinda backwards.

I'm sure my system will be slower than just about any computer controlled system, but even if it were computerized it would still have lag from whatever sensors it used for input, processing time and activating the pump. Infact most current methanol/water injection systems don't function anything like the system I have in mind, rather they operate off 'threshold points', ex. above a certain boost level or throttle position it activates the pump and increase from there. A noncomputerized system is just easier in my mind, although finding the right relay and a way to activate it properly is starting to seem difficult. I can rely on the ecm's timing retard function for immediate detonation prevention with the peace of mind that if things continue to get out of controll the methanol/water system will pick up the slack. I can also switch it on manually if I want a quick boost in performance.

[sharkey]

the thing that we are trying to get at here is that if you knock happens, it doesnt matter how quick the system turn on time is, the damage from detonation is already happening. you dont need a way to supress it, you need to prevent it. that is why all the systems out there opperate on a boost threshold, or you could make it run off tps threshold.

[85maro]

I realize that, however a throttle position switched system would also compromise overall protection. A throttle switched system can prevent detonation off idle, but at idle and low throttle (stop & go traffic, highway cruise) the system is most likely ineffective. The safest system would be one that constantly supplied methanol/water, but it would run the tank dry quick. The best compromise would be a system that supplies some methanol/water off idle and kicks up the flow if knock is detected regardless of throttle position. I'll put a throttle switch in the system too, keep the ideas coming.

[jwvess00]

Hi there!

I thought about using an esc, but the problem is the esc sends a signal to the ecm until knock is detected, then it drops the signal. Not sure why GM programmed it that way, seems kinda backwards.

Yup -- a quickie check says that the ESC module drives the sense line high, and pulls it low if knock is detected. This is a common electronics design point -- the signal is probably pulled low by an output transistor in the ESC module, and has a pullup resistor to keep the line high when the output transistor is off.

If you were using a relay for your injection system, you would power one side of the relay coil, and connect the ESC output to the other relay coil terminal. When the ESC system grounds the output, the relay would close.

I don't know if the ESC module could sink enough current to fire a traditional mechanical relay, but there are probably solid-state relays that would work. Or, you could build / have made a very simple transistor-based circuit to run the pump. If you're not into electronics design, hunt down your local ham radio club -- there is very likely someone in the local ham club that would be happy to help.

[85maro]

I was thinking of using the esc's ecm signal to activate the pump directly, but obviously that wouldn't work. Just using the esc as a knock sensor filter like you're suggesting would solve that problem, but do you think it would work without powering the ecm signal side? Could a transistor based switch be run off a knock sensor? I'm not big on electronic design so thanks for the help.

[85maro]

Oops, double post.