V8 Ignition
- Thread starter stevehuckss396
- Start date
Help Support Home Model Engine Machinist Forum:
Here is the hall package I got from Farnellinone ref 139853. It has a integrated magnet and switching frequency over 200khz
View attachment hall.pdf
View attachment hall.pdf
- Joined
- Jul 31, 2010
- Messages
- 309
- Reaction score
- 70
stevehuckss396 said:I have been considering what to use for an ignition system for the small V8. Roy's module will only yield 12,000 sparks per minute.
(12,000 X 2) / 8 = 3000RPM
Unless my math is wrong, it will only support 3000 RPM on a V8 engine. Using the same math I think I need.
(8000 X eight) / 2 = 32,000 Sparks per minute.
Is there another module out there that will supply spark at that rate, or am I going to have to build my own. I have a drawing of a TIM type ignition that my pal Louis designed that might work. I won't need alot of energy (CDI) because the plugs will be so small.
This circuit will drive an automotive coil.
Hi All
Would a MJ10012, ECG98, or BU931R HV Darlington Power Amp Fast Switch transistor work for the NPN power transistor? Can these power transistor be driving by an 2N2907?
Has anyone used a MC3334 HIGH ENERGY IGNITION CIRCUIT?
http://www.jaycar.com.au/images_uploaded/MC3334R0.PDF
Or this one?
Hall effect pickup ignition controller
http://www.st.com/stonline/books/pdf/docs/1360.pdf
Thanks for any help.
Tim
Lakc
Well-Known Member
Interesting find there Tim.
Essentially, this is the "thinking" part of an aftermarket GM 5 pin ignition unit, as mentioned a few pages back. Note that it uses a reluctor input, so expect to have to fiddle with the circuit a bit for a hall sensor input. It also expects to be controlling a GM HEI coil, so your out of the optimization curve if you deceide to use something different. I am not sure if all this optimization provides us any benefit in the first place.
As this chip appears out of production, you limited to whats left out there, and who has it. Motorola spun off its semiconductor business into ONSEMI and FREESCALE. Freescale recognised the part number and said they transferred the part to ON back in 1997, ON didnt recognise the number at all.
As to your other question, the MJ10012 looks like it was made for this application, literally. The MC3334 datasheet calls out that as the driving transistor specifically. In real life, the answer, as usual, will be "it depends". Driver circuits depend mostly on what is being driven, so it may be overkill for smaller coils or not enough if your trying to drive a Tesla coil.
The 8mh coil it drives in that circuit is a lot of inductance. If you ever get hit by one of those, it ruins your day.
Essentially, this is the "thinking" part of an aftermarket GM 5 pin ignition unit, as mentioned a few pages back. Note that it uses a reluctor input, so expect to have to fiddle with the circuit a bit for a hall sensor input. It also expects to be controlling a GM HEI coil, so your out of the optimization curve if you deceide to use something different. I am not sure if all this optimization provides us any benefit in the first place.
As this chip appears out of production, you limited to whats left out there, and who has it. Motorola spun off its semiconductor business into ONSEMI and FREESCALE. Freescale recognised the part number and said they transferred the part to ON back in 1997, ON didnt recognise the number at all.
As to your other question, the MJ10012 looks like it was made for this application, literally. The MC3334 datasheet calls out that as the driving transistor specifically. In real life, the answer, as usual, will be "it depends". Driver circuits depend mostly on what is being driven, so it may be overkill for smaller coils or not enough if your trying to drive a Tesla coil.
The 8mh coil it drives in that circuit is a lot of inductance. If you ever get hit by one of those, it ruins your day.
- Joined
- Jul 31, 2010
- Messages
- 309
- Reaction score
- 70
Admiral_dk said:Well I do understand the concept - don't fix it, if it ain't broke ... ;D
But I'm also interested in improving things, so I went looking for an application note on the IRF homepage - none, googled it - again none .... So I looked at the datasheet and I'm almost certain that you're referring to figure 15 - fine, BUT it's NOT how to use it, but how to measure the device ....
I do not have a Hall effect sensor in my electronic simulation tool right now, so I can only show you how the modified circuit will look like .... that is if I succeed in attach the pdf .... ???
Your 20K resistor is replaced with R2 1K on my schematic, then follows the two transistors and another resistor and the IRFGH14C40L as you use, and the rest is just another way of drawing the same as your circuit. But it should be a much improved circuit .... but then again - it might not perform any better in your application -
Hi Admiral_dk
How does it work?
Points open, Hall effect without magnet.
R2 is holding the Base of Q1 and Q2 high.
In this state current is flowing through Q1, but not Q2.
Is this correct?
Why do you use an Emitter Follower Switch.
Can this be used to drive and BU941 NPN transistor?
Thanks
Tim
I recieved a mail from Roy over at S/S. He confirmed that he has had some good result with CDI modules at 30,000. I hope to talk to him more if he goes to Zanesville Next month.
"Yes, I've been working on several possible mods to the CDI to get the higher RPM's and not create problems in the distributor.
I have bench tested in the 30,000 area and with 4.8 up to 6 volts there are some real good looking results. I would be very happy to work with you to get you what you need. Feed back is very important along with real world testing."
Roy Sholl
"Yes, I've been working on several possible mods to the CDI to get the higher RPM's and not create problems in the distributor.
I have bench tested in the 30,000 area and with 4.8 up to 6 volts there are some real good looking results. I would be very happy to work with you to get you what you need. Feed back is very important along with real world testing."
Roy Sholl
Admiral_dk
Well-Known Member
- Joined
- Jul 15, 2010
- Messages
- 165
- Reaction score
- 0
My circuit is thought as an improvement of the circuit made by jpeter. Since he's using a IGBT you can't compare it with a circuit using a bipolar transistor - the Gate (input pin) on a IGBT isn't drawing any current at any point in time as such .... ???
Seen from the circuit in front of it, an IGBT is an (unwanted) capacitor. If you wish to change the voltage over a capacitor you need current to charge or discharge it - the faster you wish to change it the more current. This means that as long as you keep the Gate at a steady voltage it isn't drawing any current - the bipolar transistor needs a current running between it's Basis and Emitter pins in order to conduct any current between it's Collector and Emitter pins - the difference between those two currents are the Hfe (gain) of that device and most bipolar transistors suitable for use as the primary switch in a ignition system are low gain devices and therefore draws much more current that a typical Hall-Effect sensor can cope with .... For that reason you'll need at least one more bipolar transistor between those two devices.
The IRGS14C40LPbF IGBT is a device totally optimized for using as a primary ignition switch, including several internal protection circuits - one for protecting against the secondary not being connected (this rises the switched voltage well above the safe voltage for most transistors) and protection against to high voltages on it's input.
Ok back to my circuit. Since we want the switch (IGBT) to change as fast as possible, we need the Gate voltage to change fast and as explained this requires much current to charge and discharge the parasitic input capacitor. Apply power to the circuit and assuming that the magnet isnt near the Hall-Effect sensor (likely on a one cylinder engine, less and less likely the more cylinders), R2 will pull the Basis pin on both Q1 and Q2 high. Q2 will conduct a big amount of current, slightly limited by R1, for a very short time, in order to charge the parasitic Gate capacitor. This turns Z1 on and current starts to flow in the primary of the ignition coil. Turn the engine and the magnet enters the sensitive area of the Hall-Effect sensor. This makes the output of the sensor go low and this turns Q1 off and Q2 on. Q2 will very quickly discharge the Gate on Z1 and this makes Z1 stop conducting the primary current very fast => a very fast rise of secondary voltage until the voltage where the spark gap starts the spark.
Seen from the circuit in front of it, an IGBT is an (unwanted) capacitor. If you wish to change the voltage over a capacitor you need current to charge or discharge it - the faster you wish to change it the more current. This means that as long as you keep the Gate at a steady voltage it isn't drawing any current - the bipolar transistor needs a current running between it's Basis and Emitter pins in order to conduct any current between it's Collector and Emitter pins - the difference between those two currents are the Hfe (gain) of that device and most bipolar transistors suitable for use as the primary switch in a ignition system are low gain devices and therefore draws much more current that a typical Hall-Effect sensor can cope with .... For that reason you'll need at least one more bipolar transistor between those two devices.
The IRGS14C40LPbF IGBT is a device totally optimized for using as a primary ignition switch, including several internal protection circuits - one for protecting against the secondary not being connected (this rises the switched voltage well above the safe voltage for most transistors) and protection against to high voltages on it's input.
Ok back to my circuit. Since we want the switch (IGBT) to change as fast as possible, we need the Gate voltage to change fast and as explained this requires much current to charge and discharge the parasitic input capacitor. Apply power to the circuit and assuming that the magnet isnt near the Hall-Effect sensor (likely on a one cylinder engine, less and less likely the more cylinders), R2 will pull the Basis pin on both Q1 and Q2 high. Q2 will conduct a big amount of current, slightly limited by R1, for a very short time, in order to charge the parasitic Gate capacitor. This turns Z1 on and current starts to flow in the primary of the ignition coil. Turn the engine and the magnet enters the sensitive area of the Hall-Effect sensor. This makes the output of the sensor go low and this turns Q1 off and Q2 on. Q2 will very quickly discharge the Gate on Z1 and this makes Z1 stop conducting the primary current very fast => a very fast rise of secondary voltage until the voltage where the spark gap starts the spark.
This is done in order to increase the current during switching and still keep (almost) the same voltage at the output.
Yes . BUT the resistors should be off different values . And it still wont be a great match for a bipolar transistor as the switch so I wouldnt use it that way Its a perfect match for a MOSFET or a IGBT as the switches.
- Joined
- Jul 31, 2010
- Messages
- 309
- Reaction score
- 70
Thank you for explaining it. I was thinking that was the way it worked.
Tim
Tim
Hey Admiral_dk, I like your explanation but think about this: Keeping in mind the coil fires when the field calapses, what I tried to do was get the field to calapse as quickly as reasonable. Now, to get an igbt to conduct, build up the flux field in the coil, I wanted to excite the gate with at least 8volts positive which I got from the pull up resistor. This drove the coulombs out of the gate junction. At this point what I had was an igbt with a gate devoid of electrons, a large current in the CE junction and coil charged with flux; a large flux field, right? Here comes the magnet quickly switching the hall effect sensor into conduction mode, essentially grounding the igbt gate. Charges flood from B-, ground, through the hall effect switch, then through the diode (notice the direction of the arrow, the direction electrons flow) into the gate, thus clamping off the CE junction. This drop in current collapses the flux field and sparks the coil. The coil charging takes place between firings. Charging time is not too important as long as the flux field gets fully built before the next firing. If the engine rpm is too high to allow time enough between firings to charge the coil then more current is needed in the gate circuit, ie., smaller pull up resistor, to attempt to charge the coil more quickly. Right?
Admiral_dk
Well-Known Member
- Joined
- Jul 15, 2010
- Messages
- 165
- Reaction score
- 0
Hi jpeter
Your diode is connected so it "shorts" when you charge the Gate on the IGBT and it's off when the Hall-Effect conducts and discharges the Gate on the IGBT - that's the reason I suggested that you turned it 180 degrees.
You're right that the most important issue is how quickly the field collapses and this has everything to do with how quickly you discharges the Gate on the IGBT. The turn on of the IGBT is kind of unimportant to the spark, as long as the field is on full strength when you turn it off again .... The reason the turn on time matters, has to do with the heating of the IGBT.
P = U x I (Power = Voltage x Current)
So when the IGBT is off, you have 12v x 0 = 0, and when it's on you (ideally) have 0v x 3A (current is depending on coil) = 0 - fine so far .... But when the IGBT isn't turned all the way on (during turn on or off) you could have (values depending on coil data and time) 6v x 1.5A = 9 watt and this is more than enough to heat it up on a slow turn on.
You're right about the smaller resistor value at higher rpm.
Your diode is connected so it "shorts" when you charge the Gate on the IGBT and it's off when the Hall-Effect conducts and discharges the Gate on the IGBT - that's the reason I suggested that you turned it 180 degrees.
You're right that the most important issue is how quickly the field collapses and this has everything to do with how quickly you discharges the Gate on the IGBT. The turn on of the IGBT is kind of unimportant to the spark, as long as the field is on full strength when you turn it off again .... The reason the turn on time matters, has to do with the heating of the IGBT.
P = U x I (Power = Voltage x Current)
So when the IGBT is off, you have 12v x 0 = 0, and when it's on you (ideally) have 0v x 3A (current is depending on coil) = 0 - fine so far .... But when the IGBT isn't turned all the way on (during turn on or off) you could have (values depending on coil data and time) 6v x 1.5A = 9 watt and this is more than enough to heat it up on a slow turn on.
You're right about the smaller resistor value at higher rpm.
Hey Admiral_dk, thanks for your patience. I did have the diode backwards. I guess I had a brain fart. I was thinking the line was on the anode end. Dah. Before I did these tests I reversed it.
Anyway, here's some stats you might find interesting. I put the scope on the igbt gate to ground and here's what I found:
Total time per cycle = 3ms. That works out to about 2500 rpm on the v8.
Gate charges for a total of 2ms.
Gate rise time = .25ms.
The gate voltage levels off at 7 volts positive after .25ms and dwells there at least twice as long as as at 0 volts. The gate voltage drop time is almost too short to measure, no doubt because of the diode and dwells at zero for about 1ms.
I found these results very interesting. Obviously the long on time accounts for the warm igbt. Looks like I should reengineer the shutter in the distributor to shorten the on time. I was surprised the igbt gate charged so quickly through the 22k resistor. I added a small resistor in the switch lead to the sensor so I could measure the current through the hall effect sensor. Surprisingly, the switching current maxed out at about a mill, well within the hall effect switch specs.
Its been fun. I'm glad I did these tests. I learned a lot.
Anyway, here's some stats you might find interesting. I put the scope on the igbt gate to ground and here's what I found:
Total time per cycle = 3ms. That works out to about 2500 rpm on the v8.
Gate charges for a total of 2ms.
Gate rise time = .25ms.
The gate voltage levels off at 7 volts positive after .25ms and dwells there at least twice as long as as at 0 volts. The gate voltage drop time is almost too short to measure, no doubt because of the diode and dwells at zero for about 1ms.
I found these results very interesting. Obviously the long on time accounts for the warm igbt. Looks like I should reengineer the shutter in the distributor to shorten the on time. I was surprised the igbt gate charged so quickly through the 22k resistor. I added a small resistor in the switch lead to the sensor so I could measure the current through the hall effect sensor. Surprisingly, the switching current maxed out at about a mill, well within the hall effect switch specs.
Its been fun. I'm glad I did these tests. I learned a lot.
- Joined
- Jul 16, 2007
- Messages
- 3,028
- Reaction score
- 1,119
I have been following this thread from the beginning and when the electronic guys got to the point where their discussions were over the head of most of the people here I had to take a break.
The bottom line fellows is this, do we have a viable answer (ignition) for Steve's engine and can we get a finalized schematic for it?
gbritnell
The bottom line fellows is this, do we have a viable answer (ignition) for Steve's engine and can we get a finalized schematic for it?
gbritnell
gbritnell said:
I have been right there with you G! Sounds to me we are stuck at the 2500 RPM level.
What is the limiting factor? Is it the coil wont charge fast enough or IGBT cant switch fast enough? Sounds like you two are getting somewhere.
Pardon me for butting-into an electronics discussion with a mechanical suggestion, but it seems to me that if the electronic limit is insurmountable (which I don't believe it is, ultimately) then a way of effectively doubling the rpm would be to use two ignition drivers and take the pickups from the camshaft rather than the crank.
Can anyone see a reason why this would not work?
Can anyone see a reason why this would not work?
Peter. said:
For the record, butting in is allowed. That's how we learn.
I'm sure it would work but for simplicity a single ignition would be best. There are commercial units that can achive 7200 rpm. I will speak to the guy who sells them at the Zanesville show next month.
Lakc
Well-Known Member
Thats almost exactly what the auto manufacturers have done. There are plenty of electrical solutions to the switching speed problem, so they moved to multiple coils and waste spark.Peter. said:
The real problem, is that to store more energy in a coil, to have more spark energy, you need to increase its inductance. The more inductace a coil has, the slower the current inrush, and it takes longer to build up the magnetic field. There in lies the reason for multi coil systems.
Lakc said:
There has to be a way to do this with one coil. MSD has single ignition systems that deliver 60,000SPM (15,000RPM). I would never need to do that but I would think half of that would be attainable. If the IGBT cant switch fast enough, could something faster be substituted?
Lakc
Well-Known Member
Part of the trick is you cant Muntz the circuit.
http://en.wikipedia.org/wiki/Muntzing
Dont make your hall sensor trigger your drivers at all. Your hall sensor signal should be cleaned up with a schmidt trigger, and input to a microprocessor, or at least to a cascade/darlington type circuit. That then requires a lot of protection circuitry built into the system to shunt away the primary coil back emf after the field collapses.
Frankly, this may sound like heresy, but I dont think were best served by hall effect sensors to begin with. They work great in full size applications, but thats because they have room around them, and they dont move the magnet, they move the flux with mechanical windows. For our purposes, the ultimate trigger is probably the optical encoder found in any old ball mouse.
http://en.wikipedia.org/wiki/Muntzing
Dont make your hall sensor trigger your drivers at all. Your hall sensor signal should be cleaned up with a schmidt trigger, and input to a microprocessor, or at least to a cascade/darlington type circuit. That then requires a lot of protection circuitry built into the system to shunt away the primary coil back emf after the field collapses.
Frankly, this may sound like heresy, but I dont think were best served by hall effect sensors to begin with. They work great in full size applications, but thats because they have room around them, and they dont move the magnet, they move the flux with mechanical windows. For our purposes, the ultimate trigger is probably the optical encoder found in any old ball mouse.
Lakc said:
The problem with that is they are way to big. In the case of a V4 the trigger has to be in the distributor to fire correctly.
What is the smallest photo/sensor pair you know of?
Hey, don't misunderstand me. I never said mine was limited to 2500 rpm. What I said was I ran the time tests at 2500 rpm. With my system I can spark my v8 to 7000 rpm for sure. I don't do it much and for lots of reasons the engine won't free rev much faster. My test show that at 2500rpm the coil completely charges in .25ms and then dwells at full charge for close to 2 more ms. Taking these facts into consideration I have no trouble believeing I could increase the rpm by a factor of 7 and still be spark'n the plugs. That's 17000 rpm. I think most would be happy with that.
I have the entire hall effect unit; magnet, sensor, shutter for switching and zeners for protection, rotor and cap built into a 1 inch outside diameter distributor. The electronics board is in the base, under the engine. Regarding the schmidt trigger, the rise time of the hall switch is pretty fast and clean. My scope shows the addition of a schmidt trigger would be of little value. I switch the hall effect sensor with a steel shutter. I'm not sure how switching with multiple magnets would work cuz with the small size of the distributor the 8 magnet style cam didn't work; the magnet interfered with each other too much. To see how it works check the link.[ame]http://www.youtube.com/watch?v=B9zt3SF_Flc[/ame]
I have the entire hall effect unit; magnet, sensor, shutter for switching and zeners for protection, rotor and cap built into a 1 inch outside diameter distributor. The electronics board is in the base, under the engine. Regarding the schmidt trigger, the rise time of the hall switch is pretty fast and clean. My scope shows the addition of a schmidt trigger would be of little value. I switch the hall effect sensor with a steel shutter. I'm not sure how switching with multiple magnets would work cuz with the small size of the distributor the 8 magnet style cam didn't work; the magnet interfered with each other too much. To see how it works check the link.[ame]http://www.youtube.com/watch?v=B9zt3SF_Flc[/ame]
