CDI Ignition Easy and Cheap

[Jeff JorgensenParticipant @jeffjorgensen88000]

Been developing this for quite a while and have it pretty much as I originally imagined.

System uses largely off the shelf items and some modification/additions as required.

Basic parts are the 4 wire 12V DC "black box" CDI units, GY6 Ignition coil and the KY_003 hall effect module. For the HT coil wire, I cut the 1/4" thick HT lead close to the coil, strip back and add whatever length of Belden high voltage cable. This is around 4 mm thick and scale to some degree. In a couple of units for my mate George Punter whom some of you may know, I added a 1A diode in the power supply (Just in case……💥

Additional parts are a 2.5mm spacing 0.1uF capacitor, LED and 1k0 resistor, HK-19F DC12V SHG 8 pin relay (only uses the coil in the circuit) some small bridging cable, a miniature switch on off, and the add on board. Naturally as in the beginning, some veroboard could be used.

Seen all over in places like Ebay etc. and when I started out the above bits were around AUD$15 in total, now perhaps double that but still worthwhile.

The original CDI black box ran a pickup coil which was large and cumbersome for model work hence the need to convert to hall effect pick up. I have just completed a changeover on my Atkinson engine and it uses points. Although the unit is originally designed for 12v DC, this latter application works well on a 2s LiPo @7.4v DC.

To get the unit to function properly, it needed addition of some type of "coil kick" in the sensor. I played with this for a quite a while and during some discussion a fellow model maker, PaulC from homemodelenginemachinist literally fell on the idea of an inductive "reed coil" in the hall circuit and hence the CDI trigger which made all the difference. He added a transistor stage to the hall output to drive the reed coil but I have since found this unnecessary.

I simply added a miniature oftheshelf complete mini relay and use just the coil in circuit leaving the rest of the unit intact. The hall effect module has an onboard LED to use for timing settings.

The add-on board I designed, carries the above relay, connections for a power LED and resistor, connections for the hall effect module and connections to the CDI black box unit.

In the latter I now tend to keep power to the CDI black box and on off switch separate as those pcb traces seem mighty small when one gets to actually see it in real life, different to on the screen design. (EasyEDA)

If using in this manner then just battery power pos and neg to the add on board for the power LED, wiring to the KY-003 hall module and the trigger input wire from the add on pcb to terminal P1 on the CDI black box.

Important to emphasise the need to run a good earth from the battery negative to the engine frame for the spark return for obvious reasons.

Anyhow, I'll drop the files for the add on board and the circuit diagrams for both hall effetc and good old points in below. Points btw need just the hall effect signal wire and terminate the unused pos and neg.

Cheers Jorgo

Tried to upload jpg images of the schematics and a zip file with Gerber files for the add on board……no luck

ok then try this….go to this link….. go to post #184 

https://www.homemodelenginemachinist.com/threads/model-engine-cdi-easy-and-cheap.31128/page-10

Edited By Jeff Jorgensen on 10/08/2022 00:48:26

[Jeff JorgensenParticipant @jeffjorgensen88000]

[Jeff JorgensenParticipant @jeffjorgensen88000]

Video here……. **LINK**

[mike TParticipant @miket56243]

Jorgo

I built two of these 4 wire 12V DC "black box" CDI ignitions for my big Bristol Jupiter radial engine. Each CDI seems to work well on it's own, but when I try run both CDI units together (the Jupiter has two spark plugs per cylinder) there is a lot of mutual EMC interference between the two CDI units.

Full credit should be given to PaulC for his relay coil modification. Without that relay coil the 4 wire 12V DC "black box" CDI units just will not work with a Hall sensor input.

Mike

Edited By mike T on 10/08/2022 11:11:06

[Jeff JorgensenParticipant @jeffjorgensen88000]

Mine did work, and I have the video evidence to prove it, however, I came to notice the spark was much less intense than when I did the mod.

In case you didn't notice, I have given PaulC full recognition for his modification in the drawings and elsewhere also.

No idea what you have there so can't comment really.

It was never initially meant for multi cylinder or multi ended spark coils however in the future I will try it on the V8 which will be via distributor, perhaps two and my standard design.

Don't know about the UK but here in Aus, the access to small ignition coils as were used in the old style motorcycles with kettering system, just don't exist and even if one could track one down, cost is expensive.

I have had success using a Ford BA Falcon on the plug coil stripped down as old Doc did in the states on one of his V-twin Harley engines but he used 6V and a large wattage resistor pack all driven by a 555 multispark unit. I used an off the shelf 18650 LiPo cell instead with basic points so perhaps that route might suit you better. I see one can buy a six pack of those coils online at Ebay for around AUD$40.00 so might be worth your trying and let us know how it goes.

Cheers Jorgo.

Edited By Jeff Jorgensen on 10/08/2022 23:46:57

[John ATTLEEParticipant @johnattlee20632]

Dear All,

For many years I have been working on the REME Museum's Rolls Royce / Rover Meteor M120 petrol engine fitted to the Museum's Conqueror ARV. The Meteor is an automotive version of a Merlin engine.

I have developed what I have named a "Discombobulator" and is applicable to a magneto or coil and distributor system.

The rotor arm and points etc are replaced by a rotor / chopper disc that has a transparent sector. As many photo transistors as there are cylinders are accurately placed radially around the circumference so that the chopper disc switches each photo transistor on and off. The angle of each sector is arranged so that the coil will be charged for 3 milli seconds at max desired RPM.

The mechanical (and vacuum advance if fitted) mechanism is retained and works in the usual way.

Current through the coil is controlled by International Rectifier, Ignition, Insulated gate, Bi-polar Transistors, (Ignition IGBT), type IRGS14C40L. Again, there are as many IGBTs as cylinders.

When the photo transistor turns the coil "ON" via the IGBT, the coil is charged up. When the photo trani turns "OFF", the IGBT stops the current going through the coil and a rather large spark is generated.

I am using six Renault Clio, four cylinder coil packs for the V12. The IGBT takes care of all the nasty high voltage spikes. Obviously, there is no HT distributor. There are two plugs per cylinder which are fired at the same time by the same coil.

I could have used Hall Effect switches but with 12 of them at about £15 each, they were too expensive.

I am pleased to say that the system works very well indeed.

Neither the Meteor nor the Meteorite (the V8 version) have any form of vacuum advance, so timing is independent of load and is only controlled by a simple mechanical advance system. I think that it is inconceivable that an improvement in torque and fuel efficiency cannot be achieved by carefully mapping the advance curve. My next development will be to mount the base plate with the photo transistors on a baseplate running in bearings. The position of the baseplate can be controlled by a servo motor or stepper motor controlled by an Arduino and suitable driver.

I would describe a Discombobulator as an ignition system comprising a rotor, position sensors, IGBTs and coils for each cylinder. Sensing can be by photo transistor or Hall Effect. The system can operate at crankshaft speed using the 'wasted spark' system.

I should point out that the Meteor M120 already has electronic fuel injection controlled by an Arduino using the Speeduino system.

John

[Jeff JorgensenParticipant @jeffjorgensen88000]

Posted by John ATTLEE on 11/08/2022 09:22:20:

I could have used Hall Effect switches but with 12 of them at about £15 each, they were too expensive.

Few dollars for 50 on Ebay…work just fine….

**LINK**

[SillyOldDufferModerator @sillyoldduffer]

Posted by Jeff Jorgensen on 10/08/2022 23:44:53:

Mine did work, and I have the video evidence to prove it, however, I came to notice the spark was much less intense than when I did the mod.

…

Apologies if I missed an explanation in the link, but what does the relay coil do?

I hope this is the right circuit:

My simplified version of the above suggests the relay coil is being used as an inductance working with the 0.1uF capacitor to shape the pulse applied to the CDI's trigger, but might add energy too.

When the points are closed, the inductance is grounded and current flowing in the coil creates a magnetic field. Opening the points causes the magnetic field to collapse, causing a slowly changing voltage to appear on the CDI Trigger. It rises until the magnetic field is completely consumed then falls again. I guess the effect is to remove contact bounce and stretch the operating pulse but it makes my brain hurt.

I'd expect a CDI to trigger as soon as the input pulse reaches about 2/3rds of the supply voltage, 8V or so. If correct, any pulse lengthening is irrelevant unless it causes a malfunction by merging pulses together at high rpm, and the extra energy shown in the video is due to the CDI working correctly; no energy wasted by false triggers.

Mike T's comments are interesting – he says his CDI won't trigger from a hall sensor unless a coil is present. That hints that maybe pulse lengthening is important. His other observation about mutual interference is typical. The amount of EMC generated by a spark is considerable, especially close up as is unavoidable on a small engine. It's very likely that sparks are triggering one or both CDI's as well as the contact breakers, and at the wrong time. Chaos.

The cure is often difficult requiring several techniques. Keep wires short, and the low-voltage wiring well away from the high voltage outputs. Apply shields, twisted pair cabling, filter capacitors, toroids and ferrite beads. Use an AM radio, tuned off station, to listen to the racket whilst applying stuff. Often possible to hear counter-measures having a useful effect on the radio before they add up enough to fixing the CDI.

Has anyone used an oscilloscope to see what waveform appears on the CDI trigger with this circuit? Although I don't understand how the relay coil works, it's probably genuine – not just electronic moonshine.

Fascinating what other people get up to!

Dave

[mike TParticipant @miket56243]

Dave,

The 4 wire 12V DC "black box" CDI units are fully encapsulated, however when destructively dismantled, they are found to contain a high voltage capacitor which is discharged into the ignition coil by a silicon controlled rectifier (Thyristor) similar to a type BT 151 SCR.

The SCR is switched on by a positive going pulse but needs a negative pulse to switch it off again. The pulse from a Hall sensor or from contact points is only a positive excursion; there is no negative element to switch off the SCR.

The addition of a small coil (a relay coil) rings(oscillates) when the Hall pulse or the contact points pulse disappear; The negative going oscillation switches off the SCR.

BTW, the modified 4 wire 12V DC "black box" CDI units work well individually, It fires one set of spark plugs on my Bristol Jupiter radial engine through one of the two distributors. It's only when I have two CDI units, running side by side in the same enclosure, firing the second set of spark plugs, that the EMC problems show up.

Mike

 

Edited By mike T on 11/08/2022 17:36:12

Edited By mike T on 11/08/2022 17:37:51

[noel shelleyParticipant @noelshelley55608]

In 1970 I worked for the firm that made Luminition, for a V12 IF the dwell was long enough it would be just a case of making a 12 segment chopper. I can't remember if they did one for the Jag V12 if so then you would be home and dry(ish) ! I ran them on MGBs with no side electode on the plugs. They claimed cylinder to cylinder accuracy of .25 of 1degree. The coil was an ordinary Lucas sport coil. The good old days !!! Noel.

PS, Hall effect transistors about £2.50 each.

Edited By noel shelley on 11/08/2022 18:09:50

[Jeff JorgensenParticipant @jeffjorgensen88000]

Posted by noel shelley on 11/08/2022 17:57:45:

 

PS, Hall effect transistors about £2.50 each.

Edited By noel shelley on 11/08/2022 18:09:50

As shown above, even a link supplied, one can get 50 for a few dollars on Ebay (3144)

Also as shown above, the original trigger sensor was a large coil so self explanatory really.

Edited By Jeff Jorgensen on 11/08/2022 23:42:40

[SillyOldDufferModerator @sillyoldduffer]

Posted by mike T on 11/08/2022 17:35:17:

Dave,

The 4 wire 12V DC "black box" CDI units are fully encapsulated, however when destructively dismantled, they are found to contain a high voltage capacitor which is discharged into the ignition coil by a silicon controlled rectifier (Thyristor) similar to a type BT 151 SCR.

The SCR is switched on by a positive going pulse but needs a negative pulse to switch it off again. The pulse from a Hall sensor or from contact points is only a positive excursion; there is no negative element to switch off the SCR.

…

Mike

Thanks Mike!

Made a false assumption. Years ago I fitted a kit built CDI to my car which used an oscillator to charge a capacitor up to about 150V and then discharged the cap through the coil whenever the contact breaker opened. Worked well. The kit's rather complicated circuit must have had some way of resetting it's SCR.

Luckily I read Mike's post before thoughtlessly wiring up the circuit I was going to use to test the LC combination with a scope! I planned to simulate a contact breaker with a Nano and the back EMF from the relay coil might well have fried it. Only got as far as measuring the inductance of a similar 12V relay coil: 528mH.

Incidentally, if anyone uses an Arduino relay for this circuit, make sure the board's snubber diode is disconnected. Relays switched by any kind of transistor usually have a diode wired across the coil to kill the back EMF, but the back EMF is needed in this circuit!

Dave

[mike TParticipant @miket56243]

Hello Dave,

Pleased to have been able to help. Actually it was the work of Paul C in Australia, who discovered how add the relay coil to make these 4 wire 12V DC "black box" CDI units work off a Hall sensor or c/b points.

The 4 wire 12V DC "black box" CDI units are available from many vendors on E-bay for little money as are the GY6 ignition coils they are intended to work with. Make sure to order a 4 wire 12v DC CD unitI; the 6 wire versions are for use with AC systems only.

I did not use a PCB for the extra components, there are not many, so I simply glued then to the back of the CDI unit with silicon rubber. A single CDI unit on it's own, seems to work very well with a powerful spark. It's worth trying.

MIke

Edited By mike T on 12/08/2022 17:49:42

[Jeff JorgensenParticipant @jeffjorgensen88000]

Posted by mike T on 12/08/2022 17:31:50:

. Actually it was the work of Paul C in Australia, who discovered how add the relay coil to make these 4 wire 12V DC "black box" CDI units work off a Hall sensor or c/b points.

 

And how many times I have to acknowledge PaulC for making the improvment which arrived at todays design I will never know.

It might be noted that PaulC ran with my original idea and added and extra transistor stage and a reed coil but omitted to use a pulse capacitor as shown in my final circuit, but no mention of that anywhere.

Also, (for the Nth time for Mike T benefit)  the system DID work but PaulC addition/mod in the form of the coil certainly made an improvement.

No requirement for the extra transistor stage.

Use the circuit as you will but it was designed originally for those who build single cylinder engines and require spark ignition at both an affordable cost and with parts readily available off the shelf.

There is no requirement to use other than the parts I mentioned especially the coil type, no need for any "Arduino type relay". The HK19F-DC12v-SHG works fine, no changes needed OR if you REALLY want you can use a Jaycar 12v Reed coil (SY4038) @ around AUD $4.00 a piece if that floats your boat.

Made conversion on my Atkinson engine a couple of days back so tried and tested as they say.

With all that in mind 100% success.

Response from Mike T has ensured I will keep my ideas to myself in future. Seems to have a rock in his boot because the system for one reson or another ( only known to him ) would not work with some multicylinder engine. We never did receive any input from that to try to determine why , other than "it won't work" but there ya go.

Edited By Jeff Jorgensen on 13/08/2022 01:05:56

Edited By Jeff Jorgensen on 13/08/2022 01:08:10

Edited By Jeff Jorgensen on 13/08/2022 01:16:55

Edited By Jeff Jorgensen on 13/08/2022 01:20:29

[Jeff JorgensenParticipant @jeffjorgensen88000]

For the benefit of MikeT who seems to doubt the system ( pre PaulC mod) did actually work, see the video from over 2 years ago.

**LINK**

[John HaineParticipant @johnhaine32865]

Posted by mike T on 11/08/2022 17:35:17:

………..

The SCR is switched on by a positive going pulse but needs a negative pulse to switch it off again. The pulse from a Hall sensor or from contact points is only a positive excursion; there is no negative element to switch off the SCR.

…….

Hmm. This is only true of a "Gate Turn Off" (GTO) thyristor – see **LINK**

A conventional type or SCR will not turn off by the application of a negative pulse to its gate, it continues to conduct until the forward current drops below a threshold value. For this reason the SCR was great for phase-controlled AC circuits but not so good for DC, hence the development of the GTO type. For a CDI ignition system the obvious choice would be a conventional SCR as it will switch off when the capacitor has discharged and only needs a unipolar trigger pulse. These CDI modules must use a GTO type which complicates triggering (as people have discovered) and is really a daft design decision. Even then it ought to be possible to avoid the use of a coil to get the negative trigger and use a capacitor instead.

I built two CDI systems for my old Triumph Herald. One was a design from Practical Electronics which was a total dog and very complex, using a 2-transistor inverter to charge the cap. Several of us where I worked built it with little success. The other was a design from an RCA power MOSFET application note that used a single MOSFET blocking oscillator with a simple transformer, single rectifier diode, and a handful of other passives. Triggering from the points used a couple of resistors and a capacitor. The circuit worked beautifully and gave me good service for several years until something (I think the thyristor) went wrong. The friend who pointed me to the circuit recommended to someone he met who had a Ferrari Dino to contact me to get one built for him but he never did – so I never got the chance to work on a Ferrari…

[SillyOldDufferModerator @sillyoldduffer]

Posted by John Haine on 13/08/2022 10:13:39:

Posted by mike T on 11/08/2022 17:35:17:

………..

The SCR is switched on by a positive going pulse but needs a negative pulse to switch it off again. The pulse from a Hall sensor or from contact points is only a positive excursion; there is no negative element to switch off the SCR.

…….

Hmm. This is only true of a "Gate Turn Off" (GTO) thyristor – see **LINK**

… For a CDI ignition system the obvious choice would be a conventional SCR as it will switch off when the capacitor has discharged and only needs a unipolar trigger pulse. These CDI modules must use a GTO type which complicates triggering (as people have discovered) and is really a daft design decision. …

 

Be interesting to open one of these black boxes up and see exactly what's inside. The SCR type number would be telling.

What I know about electronics is both amateur and out-of-date! Plus I have a bad memory. So I had to look this up. None of my textbooks cover the basic operation of SCRs or Thyristors – they assume the reader already knows! But Wikipedia says this of the GTO, which might explain why it's been used rather than an ordinary SCR, my bold:

GTO thyristors suffer from long switch-off times, whereby after the forward current falls, there is a long tail time where residual current continues to flow until all remaining charge from the device is taken away. This restricts the maximum switching frequency to about 1 kHz. It may be noted, however, that the turn-off time of a GTO is approximately ten times faster than that of a comparable SCR.

A maximum switching frequency of 1kHz is plenty for a single cylinder 4 stroke engine, where I think 12000 rpm would spark every other revolution, so 6000/60 = 100 sparks per second, or 100Hz. But is that fast enough for all engines, 2-stroke and multi-cylinder?  What's the general relationship between spark rate and engine configuration?

For example, say I decide to model a Bristol Centaurus (I wish), which has 18 cylinders in 2 rows:

How many sparks per second are needed to run an engine like that? What's the firing sequence? Don't ask me – it's already too hot for thinking this morning…

Dave

Edited By SillyOldDuffer on 13/08/2022 11:24:37

[John ATTLEEParticipant @johnattlee20632]

Dear All,

Going back to the cost of the Hall switch, we are not comparing like with like! I should have said that I was considering the Honeywell 4AV series VANE switch. This would enable me to position the 12 switches accurately so that the 12 cylinders would all fire at the correct moment. Instead I used the Sharp GP1A51HRJ00F photo transistor which has a similar mounting arrangement but is much cheaper

The IGBT is about £1. You need a 1K resistor between the Hall switch and gate of the IGBT. Provided there are 3 ms of time available to charge the coil up, I don't see what the advantage of a sizeable CDI package is. Clearly the Hall switch is a huge improvement over the points system which is a failure waiting to happen. Does the CDI give a higher voltage spark or a fatter one?

John

[mike TParticipant @miket56243]

Advantages of CDI …….according to Google

The advantages of CDI include the following.

• The major advantage of CDI is that the capacitor can be fully charged in a very short time (typically 1ms). So the CDI is suited to an application where insufficient dwell time is available.
• The capacitor discharge ignition system has a short transient response, a fast voltage rise (between 3 to 10 kV/ µs) compared to inductive systems (300 to 500 V/ µs), and a shorter spark duration (about 50-80 µs).
• The fast voltage rising makes CDI systems unaffected to shunt resistance.

Disadvantages of CDI

The disadvantages of CDI include the following.

• The capacitor discharge ignition system generates huge electromagnetic noise and this is the main reason why CDIs are rarely used by automobile manufacturers.
• The short spark duration is not good for lighting relatively lean mixtures as used at low power levels. To solve this problem many CDI ignitions release multiple sparks at low engine speeds.

[Author]

Posts