Distributorless 5K Engine

[coln72]

Tried a Holley Red pump once in the engine bay. Didn't like sucking much. Would stop randomly but was soooooo much quieter.

[Banjo]

Thanks !    Is this the one ?

They are under $ 100 on ebay, so if I need one permanently, & it is quieeeet, it might be the go,.

Cheers Banjo

Edited September 29, 2018 by Banjo

[coln72]

8 hours ago, irokin said:

Noisy bastard things though if you're sitting in traffic not flowing any fuel.

Thats what the radio is for ;)

 

3 hours ago, Banjo said:

Thanks !    Is this the one ?

They are under $ 100 on ebay, so if I need one permanently, & it is quieeeet, it might be the go,.

Cheers Banjo

Something like that. Took it off after having to suck fuel through the line on the side of the road to re prime the pump. Got old real soon. don't know if it would work better mounted under the tank.

Edited September 29, 2018 by coln72

[altezzaclub]

The main injector pump in The Girls KE70 hangs from a cut-down VL Commodore mount. Its a metal cross-piece with two detachable rubber mounts that bolt to another cross-piece, so it literally is suspended on rubber. You don't hear the pump in the car at all.

The only photo is this one, not much of an illustration I'm afraid.  I put the zip-ties on because I didn't trust the rubber, but if the system didn't work there would've be a lot of dead VLs on the side of the road 30years ago..

Ah-  here..

http://holdenpaedia.oldholden.com/File:Fuel_pump_pressure_reg_.JPG

[Banjo]

Love you work !   Always attention to detail.  Never apologise about the use of zip ties.  I never go anywhere without a resealable plastic bag, with a selection of different lengths therein. In the glove-box, in my briefcase, in my tool box.  One of the most useful everyday inventions ever made.  I recently was given a packet of special wide heavy duty ones for water hoses on automotive engines.  I've used them on one engine.  One size fits all size hoses !  Just zip it up & cut off the excess.   I believe they are using them on the latest model Volvo cars as standard.

Cheers Banjo 

 

[Banjo]

All input trigger, teeth, & synch are critical in an electronic ignition system, but none more so than the synch pulse, as it is the starting point, for setting the sequence & start of the other pulses being counted.

Although the synch pulse sensor I created on the timing case cover, with a rare earth magnet, attached to the camshaft sprocket, worked perfectly, I was not happy at all with the amount of work required to install this sensor.  Simply put, the timing chain cover has  to come off, which can be a frustrating job, if the engine is still in the car, and even if the engine is out, usually means removal of the sump.

The whole idea of this project, if it was successful, was that it could be replicated, by anyone, wanting to rid their K Series Rolla of its distributor & points, without any great heartache.

My first thought was to pick up this pulse from the camshaft lobe for the fuel pump.  Sounded simple !   Replace the fuel pump with an electric one, which worked perfectly.

$ 12.00 on ebay, with free postage.

Fashion a blanking plate to cover the opening in the block for the fuel pump.

Check to see whether a Hall Effect proximity switch mounted through the blanking plate, could pick up the fuel pump lobe on the camshaft.  Well it didn’t  -   reliably. The reason was that the lobe is pretty much a circular, offset from the centre of the camshaft, so didn’t have a sudden on & off, cycle, as you would obtain with a “pointier” lobe of say, a valve camshaft lobe.

So I sleep on it, & came up with the idea of reinstating the fuel pump body but “gutting” its internals, and fitting a Hall Effect proximity switch though the lid, sensing a small “button” type rare earth magnet, attached to the centre top of the diaphragm.

The casing for the Proximity Switch is a threaded tube, so it would easy to move it up & down to obtain a point, where its ability to produce a pulse was ultra reliable.

So here is the finished product.

Some of the mechanical fuel pumps used on K Series Rollas, were a sealed unit, which don’t lend themselves to being modified.

However, on my shelf, I found a olde dicast fuel pump, which was perfect.  You can still purchase these dicast type on ebay.

It had a lid, valve section, & an actuator body, in three (3) separate parts.  I removed the lid, & middle section, leaving the base & diaphragm intact.  I punched out the valves in the mid-section, & then “gutted” the dicast framework for the two valves with a rotary carbide cutter, which took about 30 minutes.

I then drilled a 12mm dia hole in the middle of the lid, to take the sensor.

I tried several 10mm dia. circular rare earth magnets, but only needed a very thin one, 2mm in depth. This attached itself magnetically to the centre top of the diaphragm.  However, I noted the metal centre of the diaphragm was slightly domed.  I then removed the diaphragm, & filed the centre piece completely flat. I then reattached the magnet, simply using the magnetic attraction, although intending to Araldite it in place if it all worked.

All reassembled, it worked perfectly. I screwed the sensor in,  & stopped when the indicator LED on the rear of the sensor was “On” permanently. I then marked the adjusting nut, & rotated it slowly outwards, so that the proximity switch cam away from the magnet (greater air gap). 1.5 turns later, the pulsing stopped altogether (LED Off).  I then turned the adjust ¾ of a turn, until the proximity switch was in the midway position, & it is running perfectly, over the full rev range.

I was so carried away with how well it worked, I had forgotten to Araldite the magnet in place.

I removed the lid to glue said magnet, & it had not moved at all, probably because being so thin; it didn’t have too much mass, & those rare earth magnets are so strong.

I will now convert my daily drive KE30 to electric fuel pump, & substitute this Synch Pulser into my car, & give it a good thrashing for a week, to make sure the design is robust.

The synch pulser I fitted to the camshaft timing chain sprocket was adjustable, as to where it produced the pulse in the 720 degrees of a full engine 4 stroke cycle.

However, using the fuel pump lobe, I didn’t have that luxury.

I checked it with the degree wheel on the flywheel, & the synch pulse is produced 71 degrees BTDC No: 1 cylinder, which would make it perfect for other ECUs, that “require” the synch pulse to appear prior to cylinder no: 1 TDC.

So after I run it for a week, on my daily, I let you know how it went.  I’ll just hook it up to a tacho, & it should read  the actual engine revs, divided by 4.

Cheers Banjo

 

Edited November 28, 2018 by Banjo

[altezzaclub]

Now THAT is clever! 

I would've only thought in terms of rotation, the crank, the flywheel, the cam..

Quote

it should read  the actual engine revs, divided by 4

Four??

[Banjo]

Quote

Four??

I also had to think about that one !

Most commercial auto tachometers are impulse type, & pickup the pulsing signal, derived from the collapsing field of the ignition coil.  This pickup point is usually the  negative side of the coil, although in modern cars with an ECU, this signal might come directly from the ECU.

Depending on whether your engine is 4, 6, or 8 cylinders, the number of pulses created in 2 revolutions of the crankshaft, is therefore 4, 6, or 8 pulses.

All multipurpose commercial tachos, usually have a "cylinder selection switch" on the rear, so the meter can be scaled appropriately.

In our 4 cylinder Rollas, the points open & close 4 times, for each rotation of the distributor, which is 1:1 ratio with the camshaft.

As the fuel pump lobe only "opens & closes" once per camshaft revolution, then 4 "distributor points" operations would have occurred in the same period of time.

As the tacho I hook my synch pulser up to, would only receive 1 pulse instead of 4, in the same time period, I would expect the rpm, to be 1/4 of what the engine rpm actually is.

If I've got that wrong somehow, I'll soon find out, as soon as I hook it  up.

Cheers Banjo 

 

Edited November 28, 2018 by Banjo

[altezzaclub]

I was thinking the cam runs at half the engine revs and it gives one fuel pump pulse per cam rev.

So you should be counting half the engine rpm.

Edited November 28, 2018 by altezzaclub

[Banjo]

You are correct, but a commercial tachometer is internally "scaled to count" 2, 3, or 4 pulses per revolution of the crankshaft, if they are 4, 6, or 8 cylinder engines respectively.

Think of it this way. The turns ratio between camshaft & distributor is 1:1.

It takes 1 turn of the dizzy/camshaft, to produce 4 pulses at the coil -ve terminal.

If the pulse is derived from the fuel pump lobe, it takes 4 turns of the camshaft, to produce 4 pulses, from my synch pulser.

Therefore, the tacho would display a rpm 1/4 of what it actually is, if using a tacho that would normally be fed a signal from the coil.

[altezzaclub]

true, true..   I was thinking of the sync pulse. 

So the pump pulse can tell the computer which cylinder is firing and when?  It says "Cyl #1 is firing NOW" and then the other three are fired by inference from that. Two rpm later it can say "Cyl #1 is firing NOW" and repeat the sequence.

[Banjo]

Nearly there !

The logic is that once the synch pulse is noted by the ECU, it indicates to the ECU, that the very next "trigger signal", (derived from the crankshaft trigger sensor) will be to fire at BTDC No: 1 cylinder.  The next three (3) following trigger pulses must be 3 - 4 - 2, as that is the firing order. 

With the COP system, where we have a separate ignition ignitor/coil per cylinder, it doesn't really matter if the synch pulse comes along before another cylinder, as each ECU cylinder ignitor output lead can be allocated/wired permanently to any particular COP.

If the synch pulse happens to be located before cylinder No: 2, then it will fire cylinder number 2, and the next three (3) pulses must be for 1 - 3 - 4 in that order. 

If the synch pulse happens to be located before cylinder No: 3, then it will fire cylinder number 3, and the next three (3) pulses must be for 4 - 2 - 1 in that order.   

If the synch pulse happens to be located before cylinder No: 4, then it will fire cylinder number 4, and the next three (3) pulses must be for 2 - 1 - 3 in that order. 

I've also used a similar technique when replacing a dissy on a K series, without going through the whole process of determining when the crankshaft pulley mark lines up with the zero mark on the timing chain cover;  is either TDC no: 1 or TDC no: 4 ?

I simply line the marks up without determining above.  Put the leads on the dissy cap in the standard rotation position.  You have a 50:50 chance of getting it right the first time.  If it back fires when you try to start it, then just "swap" the leads directly opposite each other, on the dissy cap, & it will start.  Great when you are working in the dark.

Cheers Banjo

    

[Banjo]

Distributor Rotor Arc

Just doing some calculations on where to allow the ECU output to switch on each COP, with their own inbuilt ignitor.

My 5K flywheel has 106 teeth on the flywheel, so there are effectively 212 teeth being counted over 2 revolutions, which is a full 4 stroke cycle. As 212 teeth span a rotational angle of 720 degrees (2 revolutions) the distance between teeth of each count is 720 / 212 = 3.396 degrees.

 

So the synch pulse is generated, & that starts an 8 bit counter (can only count up to 255), to start counting from zero, and at count number, say 22, Cylinder no: 1 is at TDC. 

The second cylinder to be at TDC will be no: 3, which  180 degrees further on in rotation.  The 8 bit count at that time will read 75, ([180/3.396] + 22).

Cylinder no: 4, will be the next one at TDC, which again is 180 degrees further on in rotation.  The 8 bit count at that time will read 128, ([180/3.396] + 75).

The last cylinder to be at TDC will be no: 2, which again is 180 degrees further on in rotation.  The 8 bit count at that time will read 181, ([180/3.396] + 128).

At a count of   212, a full cycle is completed, & the counter can be reset, & it will then await the sync pulse again, to repeat the full cycle count gain.

So I want an "arc" or period of rotation, where a cylinder can be fired, anywhere between TDC & back to a point lets say 40 degrees BTDC.

Most K Series distributors, limit the advance to about 35 degrees.  So lets be generous and allow a cylinder to fire anywhere between say 50 degrees BTDC & say 10 degrees ATDC. That is an arc or period, of 60 degrees of rotation, during which the spark plug can be fired.

This got me thinking, as the arc on the end of the rotor in the distributor is made that way, for the same reason, so the tiny gap between the end of the rotor contact & the post in the distributor cap are always opposite each other, throughout the entire advance period.

However, the arc on the rotors I have here don’t appear visually, to be anywhere near 50-60 degrees ?

So I collected about 3 or 4 rotors in my box, from 3K, 4K, & 5K distributors, and all arcs were exactly the same, when I held them against each other, back to back.

So I decided to measure up a rotor, & determine what angle or arc, the rotor actually was.

The picture below was a 1:1 drawing I made.

The angle was then measured with a protractor at 35 degrees.

Then the penny dropped !   The distributor rotates at half the speed of the crankshaft, so every degree of distributor rotation is actually 2 degrees of crankshaft rotation. All degrees for BTDC & ATDC are in crankshaft degrees.

Therefore the 35 degree arc of the rotor end contact, is actually 70 degrees of crankshaft rotation.

So my 60 degrees allowance, which I thought was generous, is not even as great as what Toyota designed all those years ago.  Maybe there was a bit extra in there for “tolerances”.

So now with that 60 degrees nominated, that translates into a count of (60/3.396) = 17.67 say 18 counts.

From the above calculations at an assumed count of 22 before TDC of no: 1 cylinder, we can set the limits of firing for each cylinder, in counts.

From the above calculations, we have already the counts for TDC of each cylinder.

TDC no: 1 = 22

TDC no: 3 =  75

TDC no: 4 = 128

TDC no: 2 = 181

So the count for the beginning of the allowed firing period, is each of those TDC count numbers, less 50 degrees which equates to (50/3.396) = 14.72 (say 15)

The count for the end of the allowed firing period is each of those TDC numbers, plus 10 degrees, which equates to (10/3.396) = 2.945 (say 3)

So now we have a set of counts based on flywheel position, for a particular synch pulse position, prior to & after the TDC positions of each cylinder.

They are . . .

Cylinder No: 1

Start of firing period 50 degrees BTDC = 22 – 15 = 7 counts

TDC = 22 counts

End of firing period 10 degrees ATDC = 22 + 3 = 25 counts

 

Cylinder No: 3

Start of firing period 50 degrees BTDC = 75 – 15 = 60 counts

TDC = 75 counts

End of firing period 10 degrees ATDC = 75 + 3 = 78 counts

 

Cylinder No: 4

Start of firing period 50 degrees BTDC = 128 – 15 = 113 counts

TDC = 128 counts

End of firing period 10 degrees ATDC = 128 + 3 = 131 counts

 

Cylinder No: 2

Start of firing period 50 degrees BTDC = 181 – 15 = 166 counts

TDC =181 counts

End of firing period 10 degrees ATDC = 181 + 3 = 184 counts

So the count of 184, is well before the end of the 212 tooth of the flywheel.  212 -184,  means there are 28 counts, or (28 x 3.396) = 95 degrees of crankshaft rotation left.

This will help us determine what the design limits are. as to how far before the beginning of the first firing period, the synch pulse can physically be.

We can just take all 12 points we have calculated in counts, and move them backwards or forwards in rotation around the flywheel by adding or subtracting a particular count (angle) to every count number listed.

So lets work backwards. We know the last count on the flywheel is 212, because that is how many teeth there are.  Lets say we reset the counter on count 210.  Lets say the last count for the "switch off" the cylinder no: 2 is 208 counts. As our calculation above was 184, that is effectively rotating everything (208–184 = 24) counts, or (24 x 3.396) = 82 degrees.

So our new set of switching points are as follows by simply adding a count of 24 to each previously calculated number.

Cylinder No: 1

Start of firing period 50 degrees BTDC = 46 – 15 = 31 counts

TDC = 46 counts

End of firing period 10 degrees ATDC = 46 + 3 = 49 counts

 

Cylinder No: 3

Start of firing period 50 degrees BTDC = 99 – 15 = 84 counts

TDC = 99 counts

End of firing period 10 degrees ATDC = 99 + 3 = 102 counts

 

Cylinder No: 4

Start of firing period 50 degrees BTDC = 152 – 15 = 137ounts

TDC = 152 counts

End of firing period 10 degrees ATDC = 152 + 3 = 155 counts

 

Cylinder No: 2

Start of firing period 50 degrees BTDC = 205 – 15 = 190 counts

TDC =205 counts

End of firing period 10 degrees ATDC = 205 + 3 = 208 counts

 

So the earliest that the synch pulse could physically be located is 30 counts, (lets say 28), before the earliest possible firing of cylinder no: 1 (50 degrees BTDC).

A count of 28 correlates to ([46 - 28] x 3.396) = 61 degrees BTDC no: 1.

This is ample room to achieve a result from any physical position of the synch pulse relative to the first cylinder to be fired thereafter.

So off to do some programming now, & then attach the CRO oscilloscope & see if the theory all works out in practise.

As I have a timing light, I can trigger it from any of these individual 8 points, I have noted above ( 31, 49, 84, 102, 137, 155, 190, 208 counts), & see what the degree wheel on the flywheel actually reads.

Cheers Banjo 

 

 

[altezzaclub]

woo!  Every second it has to count over 14000 teeth as you hit 8000rpm..

I thought you could run it on a timer. You measure the time taken for the last revolution via the fuel pump sensor, adjust the time for the next spark based on that plus whatever throttle inputs there are and fire the spark so many milliseconds before TDC. Then you're only counting 133 times a second.

Ah- 14KHz is pretty slow in electronic terms though, isn't it.

[Banjo]

Quote

Ah- 14KHz is pretty slow in electronic terms though, isn't it.

You are so right.  kHz is slow, relatively.  The microprocessor handling all the counting & switching is running at 32MHz.

The only limit is the sensors themselves.  That's why I use Hall Effect sensors, as they have the ability to switch very quickly, & produce nice clean square wave pulses.

Here are the ones I am using for trigger & synch pulses, off ebay.

They are built into a threaded tube, so are very easy to install & adjust the distance between the "blue" tip & the rare earth magnet.

Believe it or not, these are about $ 4.00 ea. delivered, stocked in Australia.

Better still are the specs.

It can switch at 320 kHz,  which means it could handle the 106 flywheel teeth counting application up to 18,000 rpm !

All good !

Cheers Banjo

 

Edited November 30, 2018 by Banjo