Mazda MX-6 Forum banner
1 - 7 of 7 Posts

2,697 Posts
Discussion Starter · #1 · (Edited)
I am trying to figure out how they wire and how to use them but I am having a hard time understanding Vin + (Voltage in) and Vref - (Voltage reference).

In this image the Vi is - and the Vref is +
How can two voltages be compared when one is a ground or does +- represent sides of the comparator and not polarity?
Every wire diagram I find has one of them grounded Vi or Vref.

I know that comparators come in different voltage ranges, come in two types inverter and non-inverter and can be fixed, variable and programmable, I can figure out what voltage range I need and how to get the output signal I need. I can't figure out how to wire it.

I want to monitor and use an extra 88-92 GT knock control unit and knock sensor.
The knock control unit has 4 wires, it receives 12+ and Ground. I generates a voltage between 3.3 and 5V with the ignition on/idle it sends that voltage to the knock sensor and the knock sensor sends it back to the control unit which sends it on to the ECU. The ECU receives that 3.3-5V signal constantly. During knock the sensor vibrates creating resistance (I am guessing) and voltage drops between 1.3 and 2.6V (not sure if it varies with intensity or if it varies based on sensor input voltage 3.3-5V).

The scenario:
There is a second Knock sensor installed on the engine block and a knock control unit wired independently. It's output signal the would go to the ECU is what I want to monitor and work with.
-I use a 1.7 - 5.5V comparator.
-I want to run this 3.3V to 5V + signal into the comparator. In Voltage
-I want to use a pot resistor, variable rate resistor or potentiometer to set a reference voltage of 2.6 or 2.7V
-I want the out signal of the comparator to be OFF when the In Voltage is above 2.6 or 2.7V and ON when the signal voltage is below 2.6 or 2.7V to trigger a timed relay. (However an ON voltage above 2.6 - 2.7V can be used to turn off a normally closed relay and the relay will turn on below 2.6-2.7V).
-The timed relay will power a warning light and cut boost at one of the boost control solenoids.

Would the wiring be?
- Vcc+ = Voltage supply (5.5V) (regulated down from 12V+)
-Vee- or Vcc- = Ground
-Vin = Knock control unit output signal (3.3-5V)
-Vref = Pot resistor adjustable voltage piggybacked of the 5.5V for the Vcc+ and set between 2.6-2.7V
-Vo or Vout = relay + coil voltage signal.

The engine turns up to 6250 rpm, that's 3125 cam rotations per minute, the cylinders each fire once during 1 cam rotation that is 12500 spark events per minute. There are 60 seconds in a minute 60 divided by 12500 = 1 spark event every 0.0048 seconds at 6250rpm, engine knock probably happens for less time than that.
Can a Comparator pick up such a short signal change? The knock control unit possibly lengthens the signal duration.
*Comparator reaction time is measured in ns (nanoseconds) 1ns = 0.000000001 sec or μs (microseconds) 1μs = 0.000001 sec. They have different reaction times but those I have researched react fast enough to pick up 0.48 ms (millisecond) signal change.
Can a relay timed or not react to such a short input signal?

The car still runs the factory knock control and the ECU can pull up to 6degrees timing during knock.
Because the solenoid that the ECU uses to lower boost during knock is also used by the ECU to vary boost throughout the rpm range and does so through 1 wire leaving the ECU it is impossible to separate RPM and Knock low boost so the solenoid can not be used on performance builds.

The idea of the piggy back system is that it can alert me of knock conditions so I can add more fuel or water meth or less boost or pull more timing and it can lower boost to protect the engine in case I miss the warning light or ignore it.
I think it would make a great tuning tool allowing me to fine tune the air fuel mix without guessing. I have air fuel ratio readings but nothing to warn me of their limits, well pre ignition/detonation but thats not a warning.

I have looked in to aftermarket performance knock control, audible knock listening devices, stand alone engine management knock control set - ups... They are all really expensive and most require dialing in the hertz frequency of knock for each specific vehicle (The 88-92 knock frequency is nowhere to be found). Many of them require filtering out background noise and other noise to isolate engine knock, this would mean the only way to set them up is by causing engine knock. So you pay a fortune then have to figure out how to generate engine knock at low rpm and low engine load to isolate the sound and hope that engine knock under heavy load and high rpm gets picked up by the knock control system because you can't generate knock at high rpm and load to set the device properly without a stack of engines to swap.

The Factory F2T knock system was built to pick up F2T engine knock frequencies, I have many spare knock sensors and knock control units. Free and ready to go, can be dialed in and tested on a bench sounds better than expensive and needs setting up, custom tuning, induced engine knock or driving around with head phone listening to amplified rotating assembly sounds.
I am sure this is possible to do using the ECU knock signal but I would rather run an independent system that can't jeopardize the knock signal to the ECU if anything fails.
This could be of benefit to anyone running a factory knock set-up and has cheap access to another.

Reasons and rant:

The car runs 94 octane, this is inconvenient because it can't be found everywhere, either I have to drag around 94 octane or 110 racing fuel to mix with 91 octane if I'm going where no 94 is available.

I plan to either tune around 91 octane and downgrade octane rating and take advantage of 91 octane over this engines designated 87 octane by leaning out the air fuel ratios from idle to 3000 rpms when the car is in closed loop mode, figure I can run A/F's of high 15's to low 16's without issues and get less bad gas millage.

Or continue running 94 octane and tune properly around that attempting to achieve better than the amazing 8.9 miles to the gallon average the previous version of this car got.

One change I am making to the build is that I will upgrade the intercooler, the car was running a factory intercooler with 2.5" pipes, although this did the job and functioned up to 30 psi boost (with water meth injection) cooler charge air temps will require less fuel to cool and yield better gas millage, I have 4 larger intercoolers to choose from, I didn't want to custom install an intercooler on this build but the effort could be worth it.

I am also contemplating reinstalling the EGR pipe between the exhaust and intake manifolds, although I hate the thought of hot dirty exhaust air in the manifold and every F2 intake I have disassembled had carbon and soot blockeing EGR passages, on this set-up and with the amount of traffic and cops on the roads running EGR could save in wasted gas when wasting gas:
If you don't have a shit eating grin from ear to ear, feel like your unstoppable, invincible and you are burning gas you are wasting it.
If you can achieve this sensation in a smart car you are easily amused and will most likely always be happy.
If you achieve this sensation simply because the idea of something that can move is so mind baffling it's like magic you are simple and will most likely always be happy.
If you achieve this sensation from sniffing gas you are an idiot, can't remember what happiness is and will probably be dead soon.

Water injection could be used to run and compensate for 87 octane but the Idea of running alcohol injection all the time doesn't appeal to me ideally running 2 water/meth injection kits would be perfect, the first kit runs only distilled water and runs from X-21 psi, the second kits runs water and methanol, comes on above 21 psi and the first kit shuts off. First kit would run just enough to lower combustion temps and reduce chances of detonation, the second kit would run more cc's to make up for the lower octane fuel and advanced timing. This set up would be far more economical than running 91 or 94 octane but running and setting up two water injection kits and fail safes to default to the highest safe boost on 87 octane are also impracticable at the moment.

Yes I have joined the ranks of those weird creepy guys who join a car forum and their only goal is saving money by burning less gas with some crazy hair brained skeem. My cheap ass unrealistic dream is to achieve 13 to 15mpg average and 7 mpg on a good day.
A- "why did he say that? 13 to 15mpg is horrible!"
B- "I think there is something wrong with his head or his foot"
A- "I don't understand why 7 mpg would be a good day"
B- "Because it means he was going really really fast most of the time"
A- "What is wrong with him?"
B- "I don't know, just eat your second desert and we will go back to watching cable TV"
C- "Each to his own, live and let die"

A = Asshole (too dumb to understand whats going on)
B = Bitch (smart enough to understand too weak to act)
C = Carnage (don't give a shit about A or B)

Right so does anyone know comparators?

*Edited to add answers and info.

204 Posts
I don't know much about comparators but I'm curious how the boards are set up on these old school ecus. Did triacs exist back then? I forget how new that technology is but modern day control boards utilize triacs instead of relays which can respond to that sudden change in signal. You're very smart to be able to come up with the theories and when you ask questions I know it's good. This will be the first time I've ever heard the word comparator!!

2,697 Posts
Discussion Starter · #3 ·
I don't know much about comparators but I'm curious how the boards are set up on these old school ecus. Did triacs exist back then? I forget how new that technology is but modern day control boards utilize triacs instead of relays which can respond to that sudden change in signal. You're very smart to be able to come up with the theories and when you ask questions I know it's good. This will be the first time I've ever heard the word comparator!!

I worked with triac in building controls, they are fast acting solid state devices that can switch power in both directions.
The old ECU's are digital and analog direct signals, the ecu either carries the full load of that signal (injectors, idle valve) or triggers an external device to amplify or generate a signal (relay, igniter).

There is a good chance the the knock control unit conditions the signal to the ECU.
A comparator will pick up the shortest electrical change and react but I am not sure the relay trigger time will be long enough.
I found this relay:
BU509TD - Beüler BU-509TD Time Delay Relay, Delay to Off - Miller Technologies

It is triggered (stays on a set time after getting a pulse signal).
Adjustable (2 seconds to 3 minutes).
Activated by low voltage signals(3V low amps).
Shielded from negative feedback (if the trigger source goes negative/ground it wont reset the timer ).
Reacts as a normal relay to constant signal (the timer will reset with every trigger signal and restart the count down, so if it's set to 15seconds it stays on 15seconds after the last signal).

I plan to run an Led off the comparator to flash knock events and the relay with a second LED to identify boost cut relay on.

There is also Hysteresis that stabilizes the control voltage of a comparator (Vref) and a schmitt-trigger with IC that is a comparator with built in Hysteresis and more options.

Until last week when I looked in to after market knock detection and control again after 5 years hopping for lower prices and advancements did it dawn on me the factory Knock system detects F2 knock, I gathered the limited workshop manual info on knock sensors. I searched and guest and imagined solutions using the knock control output and finally posted a somewhat intelligent question on an electrical engineering forum and heard about comparators and everything else I've mentioned for the first time, electronis is so complex product descriptions explain nothing without a good understanding of electronics. Lets just say figuring out what sized turbo is best for your build/power output goals and designing a build around it (intercooler, exhaust, engine management...) with no knowledge of cars is easier than figuring out how to monitor a voltage drop and use it as a trigger without programming a controller without electronics knowledge.

Here is how I understand a comparator so far. It has a power + - in this is fixed and different with all comparators. (5v-12V- 120v...)
The power +- is what powers the comparator and the signal the comparator puts out +V or 0Vground.
It has a comparative signal range rating also + - but both require positive signals.
These are reversible so if your Power in is 12V and Vin is the 3.3-5V knock signal and Vref is a 2.7V dialed in voltage the comparator will put Vout 12V above 2.7V and 0V ground below 2.7V . If your Vin is the dialed in to 2.7V and 3.3-5V is Vref the Vout will put out 0V ground, when Vref falls below 2.7V the comparator Vout will be 12V.
On some the power +- can be reversed to change the output from from + to -.
Most have 2 Vouts that reverse signal between states, one is ground one is positive and they reverse during signal voltage change.

So it acts like a relay but instead of positive and ground turning it on voltage change turns it on. Instead of putting out + or - it puts out both. It can be N.O. and N.C. at the same time. It can turn on 2 devices at once or turn them of or turn one ON and one OFF simultaneously.

Very smart? For the amount of time I have spent researching and learning about something as simple as a comparator and all the surrounding components someone with a mind for electronics could learn and build their own ECU. My mind has a hard time with electronic circuit diagrams because they use representations and not physical wiring connections.
Just like the + - wiring on the comparator and the fact diagrams show positive and negative signals connecting to each other make it hard to physically visualize the complete wiring and function of a device I have never used. I felt pretty dumb during my first hour looking into comparators and op amplifiers (same symbol, similar function, often used as a comparator but op amp is much slower to react to input signals).

If I was smart I would have learned to program an arduino and have it operating water/meth, boost control, knock control, lean boost condition failsafes ... instead of building and modifying a bunch of piggy back stand alone devices in the car, I will have more than 20 extra relays in this car to run all this crap and I keep adding more. I don't trust electronic controllers I like mechanical devices, relays, pressure switches... the old ones where build to last for ever.

The split second engineering injector controller for the MX6 is all programmable fuel maps, they had to use different values than other engine management systems in their maps and the controller has no wideband connection, so you have to program fuel maps in millisecond injector pulses. The number of hours spent figuring out how much extra fuel is needed at X boost and Y rpm above 15psi and how to translate that to the programming have been numerous. I have several maps written and the controller is still in my basement. I am fairly confident I got 1 of the maps perfect, I will start with the richest one and know from the results if the others will work or not. Even if I datalog air/fuel ratio's on the 626 and try to match them in the controllers map I would have do do it at each psi (16-30) to properly program the maps and I still have to figure out extra injector duty cycle based on air fuel ratios after the factory injectors max out. So bench programming the extra fuel controller has proven a pain in the ass. Making things better through technology my ass, or is this so much better than the simple digital systems EIC controller with two dials, turn the start dial to the boost you want it to come on at and the gain at a high setting, floor the car and turn the gain down until you get desired A/F ratio, it's set, don't need to adjust it again if you lower or raise boost, dialed in in 10 seconds. Buy the time someone figures out how to program the Split Second EIC, the guy with the SDS EIC has already got 3 speeding tickets.

But I think this coming winter it's time to learn to program a controller or engine management. It really does seem overwhelming to me in a way. I have downloaded and played with tuner studios, but learning and understanding enough to bench tune the entire engine is a lot.
Can a speeduino or megasquirt reverse datalog? Could I weld 2 distributors together wire one to standalone and have the other run the car on factory ecu, connect the standalone to the injectors pulse, add a map and maf sensor, and monitor ignition pulse signals to basically copy the factory FT2 engine management as a base map? This would chart out air pressure and air flow signals to rpm, injector duty cycles and spark timing to 6250rpm.

There was one guy, Nally who tuned factory subaru ECU's, he knew this shit properly, they would tell him the turbo size, injector size and boost and he would write the maps at home. He would connect his laptop to the car and install the program, they would then sit in the car on the side of the road for 30 to 45 minutes with the car idling and talk to each other while the customer watches because otherwise the costumer didn't want to pay 400$, when the costumer wasn't there he would flash the ECU with the car in the shop and leave. He never had to road test and they never had to call him back because of issues.

The gap between understanding the tuner studio settings enough to fill them out and then dial them in VS knowing them and being able to dial them in before running them are worlds apart. I don't want to install standalone and spend the next two years playing with the programming, I want to install stand alone and spend the next two years playing with the car (driving it).
I guess I am old school, software is intangible and to me can't hold a monetary value but the world has changed, people can be conditioned, look a music, movies and games, they use to be purchased on a tangible media storage device, 8track, record, tape, cassette, disc... you bought and owned a copy, you could touch it, copy it, lend it and sell it... now people pay as much for the intangible data and they can only listen, watch or play the music, movie or game.
I would feel a sense of achievement building an injector pulse emulator, running pressure switches and some extra injectors (mechanical extra fuel management) butt will only feel a sense of completion when I get the split second EIC dialed in.

Once I have completed all the learning required to set up this piggyback knock controller, I will order what is needed and bench assemble and test it and post a functional set-up diagram. with options for 5V or 12V output from the comparator and wiring for reversed output settings Knock = 5V /12V and Knock = 0V

I looked at installing and setting up knock sensors with MS, they use GM sensors and control units for most of the kits, knock control seems pretty bad with MS2, MS3 pro allows you to dial in engine knock bandwidth frequency, but who knows what that is. There are online calculators and they say bore dictates frequency, but there are a hundred other variables that dictate frequency output of knock. It would suck to have engine management pull timing and diagnose knock to engine noise (trying to tune out valve train noise would be chasing ghosts) or even worse filter out and ignore knock all together.
Getting the sensor to vibrate by tapping the block does not duplicate the frequency of knock and is probably at the frequency of noise the knock control unit filters out which is why the knock sensor is tested at the sensor and not the control units output.

I feel using the factory knock sensor and controller is the best way of isolating/monitoring knock events, and although MS can't use the 3.3-5V to 1.3-2.6V signal from the factory knock controller it can read 0V knock, 5V no knock signal that a simple comparator can generate from the factory signal. This makes the lack of frequency dial in with MS2 issue disappear, and having to dial in the frequency with MS3 obsolete.

So this is useful to all F2 modifiers and there is a possibility that the comparator draws so little current from the signal it's comparing that it can be run off the factory knock control output without affecting the signal going to the ECU, so F2T modifiers can rig a timer relay that powers a warning LED and the factory boost solenoid to lower boost during knock without it lowering boost in low and high RPM through the ECU. Knock warning and added knock protection for the price of some electronic components.

I am pretty excited about running this and my makeshift high boost lean a/f boost cut set up using the 0 - 5V programmable aux output from the innovate motorsports LC-1 wideband controller programmed 0V below X A/F, 5V above X A/F and a pressure switch off bellow X boost to turn off a boost solenoid when lean. The LC-1 can only deliver usable readings to a gauge through the stereo jack if using the 5v aux out for this and only a few gauges can connect this way.
But the LC-1's 0-1 volt aux out can be easily programmed to give accurate wideband readings on the autometer narrow band o2 gauge, the instructions are on innovates site. (Have two or 3 autometer narrow band gauges for sale if anyone is interested).
  • Like
Reactions: juanpac_tupac

204 Posts
I'm excited to read more. I'd love to use a setup like that with my ms3x. Tbh I still don't have a knock detection system set up and run the best gas I can get. I went the standalone route off the bat. Used an ign map I found for a low boost f2 and scaled it for my dynamic compression ratio with the cam, basically added a degree or two in the right places to aid with compression before the turbo kicks in. The VE table was easy to dial in sitting in the passenger seat. I've studied the system a lot but overall I'm confident with my tune up to 18 psi. I'll full blown race the thing and let's see how much I can beat em by. I'd love to get a real knock system set up but as I've done test pass after test pass after test pass I'm confident with my ability to tune on my ms3x. When I make changes to the system I just add it to the program. When I build my black mazda it'll have 8 fuel injectors and multiple stages of water meth injection all controlled off the ms3x using one pickup in the gt distributor. Gm sensors for IAT and coolant temps. 4 injector plugs and a couple splices into the main harness and you're good. I'm sure if you can figure out a way to read a 5v signal for knock you can send that signal to the ms3x and I'll be doing exactly that! Def looking forward to this. I was just going to run an ms knock module and figure it out. At this point I have hours of tuning in and my ve table is pretty good. I need to hit a dyno to trim up some random spots I get into during high Rev deceleration. Seriously how smart you are you'd prob find out that a standalone ecu is the way to go. Cheers! 🍻

P.s You can program outputs for lights for warnings and other fun safety features (overboost, ign timing trim table based on IAT, read your o2 wideband and you can set up a trim + safety table, oil pressure safety, etc) and I use my phone for data logging and gauges via Bluetooth adapter that's really quite fast it'll keep up with the Rev counter.

2 Posts
The Best Headphone Or Speakers For The cars. The first thing you need to consider is what type of headphones you want. There are three main types: earbuds, on-ear, and over-ear. Earbuds are the smallest type, and they fit inside your ear canal. On-ear headphones sit on top of your ears, and over-ear headphones cover your entire ear.

2,697 Posts
Discussion Starter · #7 ·
The Best Headphone Or Speakers For The cars. The first thing you need to consider is what type of headphones you want. There are three main types: earbuds, on-ear, and over-ear. Earbuds are the smallest type, and they fit inside your ear canal. On-ear headphones sit on top of your ears, and over-ear headphones cover your entire ear.
I am pretty sure this is a spam but driving with headphones on is illegal here in Canada.

To update the Knock control, here is the diagram I came up with:
Product Rectangle Slope Font Parallel

The problem with electronics is that sometimes to learn to do one thing you've got to learn about a whole bunch of shit you don't want to have to know.

Question 1 : what is the power consumption of Vin? Would it affect the R/Y wires signal going to the ECU? Can the knock system in the car be monitored using a comparator or would a second independent knock controller set up be required?

Question 2: can the comparator's power (+Vcc) run off the cars 12V power that can reach 13.7V or does it have to be regulated to 12V only?

Question 3: is the comparators Vout signal strong enough to operate an LED bulb and trigger a relay?

Question 4: what is the best way to supply the +VDC Variable rate resistor with 5Vdc that isn't effected by the fluctuating voltage from 12-13.7V DC.

Question 5: does the Comparator require Hysteresis ?

Rectangle Slope Font Parallel Pattern

Rectangle Slope Parallel Font Triangle

Funny how a handful of tiny components that cost almost nothing can require so much information to work with.
1 - 7 of 7 Posts