Transistor trick for 2GCDFIS.
04-04-05, 03:01 AM
#1
Transistor trick for 2GCDFIS.
I presented the idea a few days ago that you could trigger a 2nd gen leading ignitor with a J-109. It generated some interest so I've decided to start a thread about it where we can discuss our successes and failures (and hopefully there will be more successes than failures 
).
No one has actually done this before to my knowledge, but the theory seems doable. The basic idea is to keep the 2nd gen leading ignitor triggering the 2nd gen coil because it's more powerful and is capable of handling a lot more current than a J-109. For instance, it doesn't enter current limiting mode until later and keeps a long dwell period up to a much higher RPM than a J-109. We're talking about excellent sparks up past 9k RPM!
Testing a J-109 triggering a Diamond coil alongside a 2nd gen leading ignitor triggering its double ended coil showed major differences just sparking into the air. Granted there are some fundamental differences sparking at normal atmospheric pressure vs in a running engine, but the 2nd gen setup showed an observably stronger spark on both plugs at the same time next to the smaller, quieter spark output from the J-109 setup. We can assume the 2nd gen setup would outperform a J-109 sparking under similar conditions in an engine.
My MegaSquirt's dwell settings were a bit on the high side in order to trigger the J-109 and have it produce a dwell curve similar to when it is triggered by the variable reluctance sensor (pickup) and reluctor wheel in a stock 1st gen distributor. In other words, the J-109 was designed to trigger from a 0 to 100 (or more) volt alternating current and not necessarily a 5 volt square wave from an ECU. The resultant dwell curve was actually very similar to the curve I observed from each coil on the DLIDFIS setup in my REPU consisting of three J-109s and three Diamiond coils.
Now that I had a base to build from, I hooked the 2nd gen leading ignitor to the MegaSquirt without changing the dwell settings and was blown away by how much more powerful it was. I gotta say if I was going to MegaSquirt a 13B, I would have used the 2nd gen coil assembly on it right then and there (no need for EDIS). However since my MS is going onto a slightly different engine, and I hadn't even tried changing the dwell timings to some more conservative numbers, I figured I'd keep the 2nd gen assembly around for more tests. I tried other dwell settings and realised the 2nd gen leading ignitor had more or less the same end result no matter which numbers I entered; it continued to have a very long dwell period and very powerful sparks from below idle speeds to over 9k. It was actually still up to 50% at 9k! That's very nice for you bridgey and peripheral types. The 2nd gen ignitor looks much more capable of producing consistanly more powerful sparks well beyond the (observed) limited powerband of a J-109, and in my opinion, is worth figuring out how to hook it up in a 1st gen without an ECU.
Note: the J-109 peaked around 3-4k at 58% dwell duty cycle and dropped off above and below but the 2nd gen ignitor was way more broad staying above 50% duty cycle for a lot longer within the RPM range of a typical rotary (I tested up to 9k). Since changing the dwell settings didn't seem to affect the 2nd gen ignitor nearly as much as the J-109, I assume that as the dwell curve drops off above and below 3-4k, the same long healthy dwell period I observed on my dwell meter at high RPM will still be there. More testing is required.
The aluminum base of the 2nd gen coil assembly did get hot pretty quickly, even with lower dwell numbers, so since I'm not sure how much it normally heats up on a stock 2nd gen RX-7, I was unqualified to make any hard descisions at that time. I didn't run any lengthy tests for heat, but it still stayed cooler than the dizzy housing of a running 1st gen. Then again, Mazda mounted the 2nd gen coil assemblies on the firewall I assume for less direct heating from the engine, but how much heat is too much? Thoughts?
Ok, now on to the instructions. I first came across the idea about this time last year on the nopistons forum when 83turbo presented it in a DLIDFIS thread over there. I've learned quite a lot about ECUs and how they control ignitors since then, but as far as I know, no one has ever tried to trigger a 2nd gen leading ignitor with a J-109 yet. The theory is sound, but it just takes someone to do it. Is anyone interested in actually trying? I know there are a lot of 2GCDFIS users here. Wouldn't it be great to get more powerful sparks out of your setup with only a little extra work?
I know you're all getting anxious for me to spill it already, so here goes. These are the ultra-simple instructions lifted from the thread on nopistons as presented by 83turbo.
Get a PNP transistor (2N3906 or whatever). Emitter goes to +12.
Collector goes to red wire on the '86 ignitor. Base goes to a 1K
resistor. Other end of the resistor goes to both the J109 ignitor
output and to another 1K resistor. Other end of this resistor goes
to +12v. I haven't tried this but it looks like it should work.
Or, if you have an NPN PN2222A general purpose transistor like I do:
To use an NPN:
Emitter goes to ground.
Collector goes to '86 ignitor red wire and to a 1k resistor to +12v
Base goes to J109 ignitor output and to a 1k resistor to +12v
From what I've learned doing MegaSquirt stuff recently, those 1k resistors that connect 12V to the transistor and ignitor are pull-up or pull-down resistors. I actually added a few to my MegaSquirt for it to function the way I wanted it to for ignition input and output. The idea is simple enough.
Ok, so according to the instructions for an NPN transistor, I'd need to wire it up something like this.
Since the PN2222A can only handle up to 6 volts on its base, I added another 1k resistor to limit the current although I'm doubtful it'll be enough. What is a good resistor value to drop 12-14 volts down to 5-6 volts and not get too hot in the process? Of course it will be pulsed at no more than an observed 58% dwell duty cycle, but that was just on my setup. I'm sure others will vary to some degree. I'm thinking 60% is what the Mazda and Mitsubishi engineers aimed for. Either way, it will not heat nearly as much as if it were on continuously (if you're familiar with pulsed vs CW lasers, it's the same idea).
Well there you have it. I have no idea if the above circuit diagram will even work since I don't trust my ability to correctly wire up a transistor further than I could throw one. Any electrical or adventurous people out there care to try it? Standard disclaimers apply; use at your own risk.
).No one has actually done this before to my knowledge, but the theory seems doable. The basic idea is to keep the 2nd gen leading ignitor triggering the 2nd gen coil because it's more powerful and is capable of handling a lot more current than a J-109. For instance, it doesn't enter current limiting mode until later and keeps a long dwell period up to a much higher RPM than a J-109. We're talking about excellent sparks up past 9k RPM!
Testing a J-109 triggering a Diamond coil alongside a 2nd gen leading ignitor triggering its double ended coil showed major differences just sparking into the air. Granted there are some fundamental differences sparking at normal atmospheric pressure vs in a running engine, but the 2nd gen setup showed an observably stronger spark on both plugs at the same time next to the smaller, quieter spark output from the J-109 setup. We can assume the 2nd gen setup would outperform a J-109 sparking under similar conditions in an engine.
My MegaSquirt's dwell settings were a bit on the high side in order to trigger the J-109 and have it produce a dwell curve similar to when it is triggered by the variable reluctance sensor (pickup) and reluctor wheel in a stock 1st gen distributor. In other words, the J-109 was designed to trigger from a 0 to 100 (or more) volt alternating current and not necessarily a 5 volt square wave from an ECU. The resultant dwell curve was actually very similar to the curve I observed from each coil on the DLIDFIS setup in my REPU consisting of three J-109s and three Diamiond coils.
Now that I had a base to build from, I hooked the 2nd gen leading ignitor to the MegaSquirt without changing the dwell settings and was blown away by how much more powerful it was. I gotta say if I was going to MegaSquirt a 13B, I would have used the 2nd gen coil assembly on it right then and there (no need for EDIS). However since my MS is going onto a slightly different engine, and I hadn't even tried changing the dwell timings to some more conservative numbers, I figured I'd keep the 2nd gen assembly around for more tests. I tried other dwell settings and realised the 2nd gen leading ignitor had more or less the same end result no matter which numbers I entered; it continued to have a very long dwell period and very powerful sparks from below idle speeds to over 9k. It was actually still up to 50% at 9k! That's very nice for you bridgey and peripheral types. The 2nd gen ignitor looks much more capable of producing consistanly more powerful sparks well beyond the (observed) limited powerband of a J-109, and in my opinion, is worth figuring out how to hook it up in a 1st gen without an ECU.
Note: the J-109 peaked around 3-4k at 58% dwell duty cycle and dropped off above and below but the 2nd gen ignitor was way more broad staying above 50% duty cycle for a lot longer within the RPM range of a typical rotary (I tested up to 9k). Since changing the dwell settings didn't seem to affect the 2nd gen ignitor nearly as much as the J-109, I assume that as the dwell curve drops off above and below 3-4k, the same long healthy dwell period I observed on my dwell meter at high RPM will still be there. More testing is required.
The aluminum base of the 2nd gen coil assembly did get hot pretty quickly, even with lower dwell numbers, so since I'm not sure how much it normally heats up on a stock 2nd gen RX-7, I was unqualified to make any hard descisions at that time. I didn't run any lengthy tests for heat, but it still stayed cooler than the dizzy housing of a running 1st gen. Then again, Mazda mounted the 2nd gen coil assemblies on the firewall I assume for less direct heating from the engine, but how much heat is too much? Thoughts?
Ok, now on to the instructions. I first came across the idea about this time last year on the nopistons forum when 83turbo presented it in a DLIDFIS thread over there. I've learned quite a lot about ECUs and how they control ignitors since then, but as far as I know, no one has ever tried to trigger a 2nd gen leading ignitor with a J-109 yet. The theory is sound, but it just takes someone to do it. Is anyone interested in actually trying? I know there are a lot of 2GCDFIS users here. Wouldn't it be great to get more powerful sparks out of your setup with only a little extra work?
I know you're all getting anxious for me to spill it already, so here goes. These are the ultra-simple instructions lifted from the thread on nopistons as presented by 83turbo.
Get a PNP transistor (2N3906 or whatever). Emitter goes to +12.
Collector goes to red wire on the '86 ignitor. Base goes to a 1K
resistor. Other end of the resistor goes to both the J109 ignitor
output and to another 1K resistor. Other end of this resistor goes
to +12v. I haven't tried this but it looks like it should work.
Or, if you have an NPN PN2222A general purpose transistor like I do:
To use an NPN:
Emitter goes to ground.
Collector goes to '86 ignitor red wire and to a 1k resistor to +12v
Base goes to J109 ignitor output and to a 1k resistor to +12v
From what I've learned doing MegaSquirt stuff recently, those 1k resistors that connect 12V to the transistor and ignitor are pull-up or pull-down resistors. I actually added a few to my MegaSquirt for it to function the way I wanted it to for ignition input and output. The idea is simple enough.
Ok, so according to the instructions for an NPN transistor, I'd need to wire it up something like this.
Since the PN2222A can only handle up to 6 volts on its base, I added another 1k resistor to limit the current although I'm doubtful it'll be enough. What is a good resistor value to drop 12-14 volts down to 5-6 volts and not get too hot in the process? Of course it will be pulsed at no more than an observed 58% dwell duty cycle, but that was just on my setup. I'm sure others will vary to some degree. I'm thinking 60% is what the Mazda and Mitsubishi engineers aimed for. Either way, it will not heat nearly as much as if it were on continuously (if you're familiar with pulsed vs CW lasers, it's the same idea).
Well there you have it. I have no idea if the above circuit diagram will even work since I don't trust my ability to correctly wire up a transistor further than I could throw one. Any electrical or adventurous people out there care to try it? Standard disclaimers apply; use at your own risk.
Last edited by Jeff20B; 04-04-05 at 03:10 AM.
04-04-05, 07:46 AM
#2
Right near Malloy
iTrader: (28)
Join Date: Dec 1999
Location: Behind a workbench, repairing FC Electronics.
Posts: 7,825
Received 493 Likes
on
335 Posts
I keep trying to do the math to calculate how many watt resistor you need, but all I have to go in a 6 volts... I think I need another number, probably the current handling capability of the transistor base to emitter.
I can't find it on any of the datasheets I look at... All I can find are maximum collector to emitter currents (1 amp, Absolute Maximum.)
Gah. My electronics math skills have gone down the *******. I've been a machinist for far too long. I need to get back into discrete electronics. None of this RF bullshit that I do at work.
Sorry I have nothing really useful to add to this thread. I was just curious what you use for drawing schematics.
I can't find it on any of the datasheets I look at... All I can find are maximum collector to emitter currents (1 amp, Absolute Maximum.)
Gah. My electronics math skills have gone down the *******. I've been a machinist for far too long. I need to get back into discrete electronics. None of this RF bullshit that I do at work.
Sorry I have nothing really useful to add to this thread. I was just curious what you use for drawing schematics.
04-04-05, 01:08 PM
#3
Jeff,
It looks like that it could work. I could try to model that circuit a little later today or tonight. I will try a 14 volt square wave and vary the waveform form (700rpm/60sec/min)*2sparks per rotation = 23 Hz and then increase to 10,000 rpm = 333 Hz. This will tells the voltages, current, and power of all points in the circuit. I am pretty sure I have the data sheet for that transitor. I will try to modify the circuit some if it doesn't meet the requirements.
Kent
It looks like that it could work. I could try to model that circuit a little later today or tonight. I will try a 14 volt square wave and vary the waveform form (700rpm/60sec/min)*2sparks per rotation = 23 Hz and then increase to 10,000 rpm = 333 Hz. This will tells the voltages, current, and power of all points in the circuit. I am pretty sure I have the data sheet for that transitor. I will try to modify the circuit some if it doesn't meet the requirements.
Kent
04-04-05, 02:56 PM
#4
The pull-ups only need to be 1/4 watt I believe (it worked fine on my MS). The current limiting resistor should probably be a half watt at least.
I use Windows Paint program.
Here is a schematic of the transistor. http://www.fairchildsemi.com/ds/PN/PN2222A.pdf
Kent, go for it and let us know. If there's time today, I'll see if I can set it up too.
I use Windows Paint program.
Here is a schematic of the transistor. http://www.fairchildsemi.com/ds/PN/PN2222A.pdf
Kent, go for it and let us know. If there's time today, I'll see if I can set it up too.
04-05-05, 01:46 AM
#5
Jeff: I am having a bit of a problem with one of the pspice libraries. It only gives me problems when I have the transistor present. I will look at it more sometime tomorrow. We may what to add a zener diode at the output or maybe a 5V regulator to limit the voltage to 5V. I believe your circuit will pull the output up to 12V (don't we want a 5v wavefrom?). I will post up the results once I get it up and running. Do you any idea of the amount of current needed to trigger the 2nd gen igniter? I will see what I can come up with.
Kent
Kent
04-05-05, 05:36 AM
#7
Kent, I'd say 20mA or less... probably.
For testing purposes, 12V can be used, without harming the ignitor. I don't know if there will be any long-term problems using 12 volts though. Hmm, maybe a Radio Shack 7805 5 volt voltage regulator would be a good idea, like you suggested.
Wait a minute. Are you talking about turning the voltage reg on and off with the transistor? There might be some lag time. Would a PNP transistor be a better choice with a 5 volt reg?
Maybe we could add one more current limiting resistor in series between the transistor and the red wire of the 2nd gen ignitor. Thoughts?
Turbo SE, a MOSFET wouldn't be necessary since a very small amount of current needs to be switched here. You're welcome to try it though.
For testing purposes, 12V can be used, without harming the ignitor. I don't know if there will be any long-term problems using 12 volts though. Hmm, maybe a Radio Shack 7805 5 volt voltage regulator would be a good idea, like you suggested.
Wait a minute. Are you talking about turning the voltage reg on and off with the transistor? There might be some lag time. Would a PNP transistor be a better choice with a 5 volt reg?
Maybe we could add one more current limiting resistor in series between the transistor and the red wire of the 2nd gen ignitor. Thoughts?
Turbo SE, a MOSFET wouldn't be necessary since a very small amount of current needs to be switched here. You're welcome to try it though.
Trending Topics
04-05-05, 09:08 AM
#8
Rotary Enthusiast
Join Date: Jun 2002
Location: Yokosuka
Posts: 1,058
Likes: 0
Received 0 Likes
on
0 Posts
So, you're trying to make a circuit to mimic the 5v square wave controlling the coils that originally came from the 2g ECU, right? Do you really want to keep the j109 if you got an ignitor on the 2g coil? Can you just do something with the output of the dizzy itself?
I'm sure there are 5v timing circuits already in existence that will work excellent for this very purpose, too. What about a 555 timer IC?
Though, I only had 2 years of electronics way back in highschool, I want to help if I can. This seems very interesting.
I'm sure there are 5v timing circuits already in existence that will work excellent for this very purpose, too. What about a 555 timer IC?
Though, I only had 2 years of electronics way back in highschool, I want to help if I can. This seems very interesting.
Last edited by jayroc; 04-05-05 at 09:11 AM.
04-05-05, 11:41 AM
#11
Junior Member
Join Date: Jan 2005
Location: so cal
Posts: 25
Likes: 0
Received 0 Likes
on
0 Posts
What frequency (Hz) range are you expecting to operate the BJT? A mosfet *may* be a better choice for this type of switching circuit. Also, what bias are you running, or have measured? I might also advise you run the circuit in quiescent mode. Overall, good work 
04-05-05, 12:30 PM
#12
Jeff: No, I mean use a regulator or a zener diode on the output in a way to limit the range from 0 to 5v (that is what is done with the MS: the pull-up resistor goes to 5V). A PNP may work better, but we will see. I will try a few different things later and let you know what I come up with.
TurboSE: A MOSFET would work here as well, but I don't think it will be needed. I don't think there will be the current need or speed to require MOSFETs.
jayroc: You may be able to do something with a 555, but I am not sure if it will have the current capability needed.
J9R8: See my post above for the frequency range. I suspect the frequency will never be more than about 300 Hz or so. So we don't need anything real fast.
This will be a fun project. I will try to post some reults tonight.
Kent
TurboSE: A MOSFET would work here as well, but I don't think it will be needed. I don't think there will be the current need or speed to require MOSFETs.
jayroc: You may be able to do something with a 555, but I am not sure if it will have the current capability needed.
J9R8: See my post above for the frequency range. I suspect the frequency will never be more than about 300 Hz or so. So we don't need anything real fast.
This will be a fun project. I will try to post some reults tonight.
Kent
04-05-05, 02:28 PM
#13
This might seem a dumb question ... but i'm not all that much into car electronics yet so try to fill me in a bit ... First is this something that would only work with fuel injected rotary ... or will it also work for a 12a carbbed setup?
Also where is the J109 located in our cars. And where would the one from the second gen be. I would like to go to the junk yard tomorrow and pull them out so that I would be able to do some testing. Also i could use the teachers here in college to help me out with all this stuff
Also where is the J109 located in our cars. And where would the one from the second gen be. I would like to go to the junk yard tomorrow and pull them out so that I would be able to do some testing. Also i could use the teachers here in college to help me out with all this stuff
04-05-05, 02:34 PM
#14
It would work with either carbed or FI. The j-109s are on the distributor (two small black boxes (for '81-'85 models). They look like Jeff20B's avatar. The '80 uses j-105's mounted on the fender and the '79 uses a points system, so it doesn't use them.
The 2nd gens have the igniter attached to the coil. You will want to get the leading coil (one coil, with two output posts). The igniter should be on the back of it.
The 2nd gens have the igniter attached to the coil. You will want to get the leading coil (one coil, with two output posts). The igniter should be on the back of it.
04-05-05, 03:04 PM
#15
Well-put. I'd just like to add that to remove the coil assembly from the inner fender of a 2nd gen RX-7, there are three or four 10mm nuts and a couple disconnectable wires holding it in.
jayroc, the output of the variable reluctance sensor is sort of a lopsided AC or in other words, a dirty sine wave as you suspected. It can't trigger the 2nd gen ignitor correctly so you must either build a circuit that would condition the sine wave and then feed it to the 2nd gen ignitor (with hopefully the right dwell time and everything) or just let a J-109 do it.
Like I said above, I don't trust my ability to hook up a transistor further than I could throw it, and pull-up or pull-down resistors also give me trouble. Keep the ideas coming.
jayroc, the output of the variable reluctance sensor is sort of a lopsided AC or in other words, a dirty sine wave as you suspected. It can't trigger the 2nd gen ignitor correctly so you must either build a circuit that would condition the sine wave and then feed it to the 2nd gen ignitor (with hopefully the right dwell time and everything) or just let a J-109 do it.
Like I said above, I don't trust my ability to hook up a transistor further than I could throw it, and pull-up or pull-down resistors also give me trouble. Keep the ideas coming.
04-05-05, 03:42 PM
#16
For NPN, I guess we can think of it like this. The transistor will die with more than 6 volts on the base. Since it won't be on constantly, a simple resistor to drop the voltage would probably suffice. The 2nd gen ignitor trigger should be kept at 5 volts if possible, so if there is a way to stick a 5 volt regulator in the circuit with a pull-up or pull-down resistor connected somewhere in the circuit allowing the regulator to stay on, I think it might work.
04-05-05, 06:29 PM
#17
Rotary Enthusiast
Join Date: Jun 2002
Location: Yokosuka
Posts: 1,058
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Jeff20B
Well-put. I'd just like to add that to remove the coil assembly from the inner fender of a 2nd gen RX-7, there are three or four 10mm nuts and a couple disconnectable wires holding it in.
jayroc, the output of the variable reluctance sensor is sort of a lopsided AC or in other words, a dirty sine wave as you suspected. It can't trigger the 2nd gen ignitor correctly so you must either build a circuit that would condition the sine wave and then feed it to the 2nd gen ignitor (with hopefully the right dwell time and everything) or just let a J-109 do it.
Like I said above, I don't trust my ability to hook up a transistor further than I could throw it, and pull-up or pull-down resistors also give me trouble. Keep the ideas coming.
jayroc, the output of the variable reluctance sensor is sort of a lopsided AC or in other words, a dirty sine wave as you suspected. It can't trigger the 2nd gen ignitor correctly so you must either build a circuit that would condition the sine wave and then feed it to the 2nd gen ignitor (with hopefully the right dwell time and everything) or just let a J-109 do it.
Like I said above, I don't trust my ability to hook up a transistor further than I could throw it, and pull-up or pull-down resistors also give me trouble. Keep the ideas coming.
So it is easier, then, to use the J109, instead of doing something with the dizzy output and building a circuit.. I thought it would be kinda neat if you could just replace the j109 on the dizzy with a small circuit in a box, then ran the wires over to your 2g coil pack and be done.
BTW, Jeff. Where are you sourcing your info on the behavior of the j109 and 2g coil packs? I'd like to take a look too, if possible.
04-05-05, 07:53 PM
#18
Rotary Enthusiast
Join Date: Jun 2002
Location: Yokosuka
Posts: 1,058
Likes: 0
Received 0 Likes
on
0 Posts
Found some good info on the 2g stuff here:
http://www.teamfc3s.org/info/articles/demystifying.html
http://www.teamfc3s.org/info/articles/demystifying.html
04-05-05, 10:46 PM
#19
Okay guys,
Here is what I have come up with so far. The zener diode on the end limits the output voltage to 4.7v, we could also use a 5.1v zener in its place. I used R2 and R3 for a voltage divider. This puts a voltage of Vbatt*(R2/(R2+R3)), or in this case Vbatt*(1.5/2.5)=8.1 volts assuming Vbatt is about 13.5v. The pupose is to isolate the output from Vbatt and so the zener doesn't have to dissipate as much power.
Here is the circuit thus far.
Here is the input and output simulation voltages. This is at 333 Hz (about 10,000 rpm), about 50% duty, and Vbatt of 14v. You will see that the circuit is inverting. (J-109 low makes 2nd gen signal go high). We may want to use a pnp transistor or go through an inverter to correct this. (Ignore the little discontinuity about 1/3 and 2/3s across the plot on Vout. This is not real. It happened with the image capture. Jeff: Let me know if you want to see other duty cycles, frequencies, etc.
As far as power goes, it appears that R1 see the most current (about 12 to 14mA max). So if we say 14v*14mA = 0.196W. We may want to go with 1/2 watt resistor to play it safe. R3 and the diode see about 10 mA, but they never see full battery voltage so they are probably fine as 1/4 watt devices. The zener diode in the schematic is rated for 400 mW.
If anyone has any ideas that they want to contribute, post them up and I will add it to the design.
Kent
Here is what I have come up with so far. The zener diode on the end limits the output voltage to 4.7v, we could also use a 5.1v zener in its place. I used R2 and R3 for a voltage divider. This puts a voltage of Vbatt*(R2/(R2+R3)), or in this case Vbatt*(1.5/2.5)=8.1 volts assuming Vbatt is about 13.5v. The pupose is to isolate the output from Vbatt and so the zener doesn't have to dissipate as much power.
Here is the circuit thus far.
Here is the input and output simulation voltages. This is at 333 Hz (about 10,000 rpm), about 50% duty, and Vbatt of 14v. You will see that the circuit is inverting. (J-109 low makes 2nd gen signal go high). We may want to use a pnp transistor or go through an inverter to correct this. (Ignore the little discontinuity about 1/3 and 2/3s across the plot on Vout. This is not real. It happened with the image capture. Jeff: Let me know if you want to see other duty cycles, frequencies, etc.
As far as power goes, it appears that R1 see the most current (about 12 to 14mA max). So if we say 14v*14mA = 0.196W. We may want to go with 1/2 watt resistor to play it safe. R3 and the diode see about 10 mA, but they never see full battery voltage so they are probably fine as 1/4 watt devices. The zener diode in the schematic is rated for 400 mW.
If anyone has any ideas that they want to contribute, post them up and I will add it to the design.
Kent
04-06-05, 12:26 AM
#20
That actually looks pretty good and might work things out. I second using that 1/2W for the resistors, in an engine things are a bit different, how you're usgin the engine and how much electronics stuff can all affect your voltages and currents ... not to mention battery types. If we do get this thing working, I think that we should do a group buy on cuircut boards. so that things look neater that way. We could even get them made in a box as a group buy too and get some real connectors going. from my understanding, alot of places are building in their resistors right in the boards so there goes half our soldering job.
04-06-05, 01:51 AM
#21
Right near Malloy
iTrader: (28)
Join Date: Dec 1999
Location: Behind a workbench, repairing FC Electronics.
Posts: 7,825
Received 493 Likes
on
335 Posts
Originally Posted by gsl-se addict
Diagrams and screenshots!
How much does that program cost? I want it!
I wish I knew more about these mid frequency things. Transistor switching speeds and biasing and etc...
I know steady state DC and a little microwave RF stuff from work... Nothing in between.
04-06-05, 06:10 AM
#22
jayroc, a J-109 is certainly easier to use in this situation than an LM1815 variable reluctance conditioning IC and its associated supporting hardware. 
Kent, schematics always turn out so pretty when people use a program. Try 4000RPM and change the duty cycle to 60% since that's more or less what J-109 produces in a stock application.
Hmm, perhaps a PNP transistor would be best. If I understood pulled up and pulled down circuits better I would have caught this ealier. It was right in front of me on my MegaSquirt which has pull-up resistors on the LEDs for use as ignitor triggers (one 1k per LED connected from +5v to the cathode lead, which is also connected to the trigger input of the ignitor). Basically when the LEDs are on, the coils are not charging. Any time an LED is off, the ignitor is triggered to charge the coil and when the LED is turned back on, the coil fires.
Here's a stupid thought. Why couldn't we just hook the J-109's C terminal directly to the 2nd gen ignitor's red wire? It'd be 12 volts, for testing purposes, but wouldn't this keep both devices high at the same time? Then when the J-109 jumps low, it would do the same for the 2nd gen ignitor. Is it possible the transistor trick is nothing more than a trick and not really needed? Am I too tired and not thinking straight?
Kent, schematics always turn out so pretty when people use a program.
Try 4000RPM and change the duty cycle to 60% since that's more or less what J-109 produces in a stock application.Hmm, perhaps a PNP transistor would be best. If I understood pulled up and pulled down circuits better I would have caught this ealier. It was right in front of me on my MegaSquirt which has pull-up resistors on the LEDs for use as ignitor triggers (one 1k per LED connected from +5v to the cathode lead, which is also connected to the trigger input of the ignitor). Basically when the LEDs are on, the coils are not charging. Any time an LED is off, the ignitor is triggered to charge the coil and when the LED is turned back on, the coil fires.
Here's a stupid thought. Why couldn't we just hook the J-109's C terminal directly to the 2nd gen ignitor's red wire? It'd be 12 volts, for testing purposes, but wouldn't this keep both devices high at the same time? Then when the J-109 jumps low, it would do the same for the 2nd gen ignitor. Is it possible the transistor trick is nothing more than a trick and not really needed? Am I too tired and not thinking straight?
Last edited by Jeff20B; 04-06-05 at 06:12 AM.
04-06-05, 07:48 AM
#24
Rotary Enthusiast
I'd recommend spending some time figuring out how to control pulse width as well. Incorrect values will result in excessive coil and ignitor heating, which was evident from the temperatures reached on my test bench. I've posted some test results that might be of interest here:
http://www.msefi.com/viewtopic.php?t=7544
If you can't easily achieve the required pulse width, then an appropriately sized resistor added in series with the coil will help, or just carry a spare FC coil in the storage bin...
http://www.msefi.com/viewtopic.php?t=7544
If you can't easily achieve the required pulse width, then an appropriately sized resistor added in series with the coil will help, or just carry a spare FC coil in the storage bin...
04-06-05, 12:45 PM
#25
Pele: The program is called PSPICE. Basically you drag and drop components onto the schematic, give values (resistance, voltage, etc.), and then perform the anaylsis. It is a real nice program. PSPICE will do DC AND AC sweeps, calculate bias points, perform a parametric study (for instance how does the output change by varying supply voltage), calculates power, current and voltage. I am not sure about the cost, as I used it at my university. There is also a version called HSPICE for unix machines. It doesn't use the drop/drag schematics. You specify nodes and then what devices connect to those nodes.
One program you guys should look at (trial version) is called PROTEL. You create a layout and it will generate the circuit traces, the file for drilling the holes, and has SPICE built in so that you can perform an analysis on the circuit. You can get a 30 day trial, but I believe the license is like $1600 last time I saw. If we get a good circuit put together, we should use this program to create a circuit board.
Jeff: I will look into modifying the circuit some more today and run the case that you mentioned. I figured the circuit would probably be inverting, but my electronics are a little rusty, so I wasn't positive until I checked. You may be right about not needing a circuit. Correct me if I am wrong, but I believe the J-109 pulls down one side of the coil to fire it. The 12v or so you see when the coil is not firing it from the + side of the coil. So we may only need a couple of resistors and a zener diode to put 5v instead of 12v on that line (trigger on the 2nd gen igniter) and then let the J-109 pull it down to ground to fire the 2nd gen coil. What do you think?
Since we would have a 5v square wave, I wonder if there would be some features that we could add (digital tach to help set idle?).
Roger: Good point on the pulse width. I will see if we could control/limit it. The imporant thing is that the 2nd gen igniter uses the falling edge of the pulse to trigger (right?).
Kent
One program you guys should look at (trial version) is called PROTEL. You create a layout and it will generate the circuit traces, the file for drilling the holes, and has SPICE built in so that you can perform an analysis on the circuit. You can get a 30 day trial, but I believe the license is like $1600 last time I saw. If we get a good circuit put together, we should use this program to create a circuit board.
Jeff: I will look into modifying the circuit some more today and run the case that you mentioned. I figured the circuit would probably be inverting, but my electronics are a little rusty, so I wasn't positive until I checked. You may be right about not needing a circuit. Correct me if I am wrong, but I believe the J-109 pulls down one side of the coil to fire it. The 12v or so you see when the coil is not firing it from the + side of the coil. So we may only need a couple of resistors and a zener diode to put 5v instead of 12v on that line (trigger on the 2nd gen igniter) and then let the J-109 pull it down to ground to fire the 2nd gen coil. What do you think?
Since we would have a 5v square wave, I wonder if there would be some features that we could add (digital tach to help set idle?).
Roger: Good point on the pulse width. I will see if we could control/limit it. The imporant thing is that the 2nd gen igniter uses the falling edge of the pulse to trigger (right?).
Kent
