I'm using this circuit to provide 250V pulses to an ignition coil. https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 The IC is a Linear Technology LT1243. 12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 is ground. LTSpice says that this works, and indeed the real implementation does. Mostly. The difference is that after the circuit has been triggered once, it won't charge the capacitor again (above about 25 volts) for a varying period that can be up to half a minute. It's as if the wrong level is being used internally to compare with the feedback pin. In that regard, I'll mention that Q3 is used to pull FB up so that the circuit stops driving the FET, since otherwise the SCR never turns off. I've checked, and the problem is not that Q3 remains on, nor is it that the SCR fails to turn off. In any case, the 1.5uF capacitor C1 does get charged somewhat, as I mentioned above. When the issue resolves, it does so rapidly, within a fraction of a second. Connecting an oscilloscope probe to the RT/CT pin of the IC restores function, but only if the scope is set to DC coupling - not AC coupling - and functioning continues only as long as the probe remains connected - until the aforesaid varying period has elapsed. Touching the pin with a finger has a similar effect. That is, connect probe; works. Disconnect; stops working. Reconnect; works, etc. until eventually it starts working continuously. This cannot be capacitive loading - there's already a significant capacitance between RT/CT and ground. The fact that the coupling has to be DC made me think that some small current was flowing to ground through the scope, so I bridged C4 with a 100K resistor. It didn't make any difference. I tried changing the IC (a pain, given that it's surface mounted), but again it made no difference. Any other ideas about what might be happening here. Sylvia.
Ignition circuit issue.
Started by ●March 26, 2017
Reply by ●March 26, 20172017-03-26
On Sun, 26 Mar 2017 15:36:06 +1100, Sylvia Else <sylvia@not.at.this.address> wrote:>I'm using this circuit to provide 250V pulses to an ignition coil. > >https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 > >The IC is a Linear Technology LT1243. > >12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 >is ground. > >LTSpice says that this works, and indeed the real implementation does. >Mostly. > >The difference is that after the circuit has been triggered once, it >won't charge the capacitor again (above about 25 volts) for a varying >period that can be up to half a minute. It's as if the wrong level is >being used internally to compare with the feedback pin. In that regard, >I'll mention that Q3 is used to pull FB up so that the circuit stops >driving the FET, since otherwise the SCR never turns off. I've checked, >and the problem is not that Q3 remains on, nor is it that the SCR fails >to turn off. In any case, the 1.5uF capacitor C1 does get charged >somewhat, as I mentioned above. > >When the issue resolves, it does so rapidly, within a fraction of a second. > >Connecting an oscilloscope probe to the RT/CT pin of the IC restores >function, but only if the scope is set to DC coupling - not AC coupling >- and functioning continues only as long as the probe remains connected >- until the aforesaid varying period has elapsed. Touching the pin with >a finger has a similar effect. > >That is, connect probe; works. Disconnect; stops working. Reconnect; >works, etc. until eventually it starts working continuously. > >This cannot be capacitive loading - there's already a significant >capacitance between RT/CT and ground. > >The fact that the coupling has to be DC made me think that some small >current was flowing to ground through the scope, so I bridged C4 with a >100K resistor. It didn't make any difference. > >I tried changing the IC (a pain, given that it's surface mounted), but >again it made no difference. > >Any other ideas about what might be happening here. > >Sylvia.Just possibly the SCR is staying on, shorts the inverter, which thermal limits, and recovers later? It's awfully complex. Back when it wasn't so lethal to ride motorcycles on the street (and I wasn't too bright) I use to make my own CD ignitions. I liked this topology https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/CDI.jpg because it was easy to drive the SCR gate with a simple RC circuit from the points. A low value resistor from gate to ground ensures turnoff. You could elect to kill the boost converter during the gate pulse, too. To get a fat spark, you need a pretty wide SCR gate pulse, to let the whole mess ring a few cycles. I usually ran around 80 to 100 mJ per spark, which will fire a plug no matter how fouled. Barely needs gasoline. Actually, my HV supplies were open-loop 2-transistor DC/DC converters with full bridge rectifiers. It sort of relaxed for a while after being shorted, then started oscillating again, which was nice. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by ●March 26, 20172017-03-26
On 26/03/2017 4:33 PM, John Larkin wrote:> On Sun, 26 Mar 2017 15:36:06 +1100, Sylvia Else > <sylvia@not.at.this.address> wrote: > >> I'm using this circuit to provide 250V pulses to an ignition coil. >> >> https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 >> >> The IC is a Linear Technology LT1243. >> >> 12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 >> is ground. >> >> LTSpice says that this works, and indeed the real implementation does. >> Mostly. >> >> The difference is that after the circuit has been triggered once, it >> won't charge the capacitor again (above about 25 volts) for a varying >> period that can be up to half a minute. It's as if the wrong level is >> being used internally to compare with the feedback pin. In that regard, >> I'll mention that Q3 is used to pull FB up so that the circuit stops >> driving the FET, since otherwise the SCR never turns off. I've checked, >> and the problem is not that Q3 remains on, nor is it that the SCR fails >> to turn off. In any case, the 1.5uF capacitor C1 does get charged >> somewhat, as I mentioned above. >> >> When the issue resolves, it does so rapidly, within a fraction of a second. >> >> Connecting an oscilloscope probe to the RT/CT pin of the IC restores >> function, but only if the scope is set to DC coupling - not AC coupling >> - and functioning continues only as long as the probe remains connected >> - until the aforesaid varying period has elapsed. Touching the pin with >> a finger has a similar effect. >> >> That is, connect probe; works. Disconnect; stops working. Reconnect; >> works, etc. until eventually it starts working continuously. >> >> This cannot be capacitive loading - there's already a significant >> capacitance between RT/CT and ground. >> >> The fact that the coupling has to be DC made me think that some small >> current was flowing to ground through the scope, so I bridged C4 with a >> 100K resistor. It didn't make any difference. >> >> I tried changing the IC (a pain, given that it's surface mounted), but >> again it made no difference. >> >> Any other ideas about what might be happening here. >> >> Sylvia. > > Just possibly the SCR is staying on, shorts the inverter, which > thermal limits, and recovers later? It's awfully complex.I don't see any sign of that - the voltage across the coil is zero at this point, whereas the capacitor C1 has about 25V across it, dropping to zero each time the circuit is triggered, and then climbing back toe 25V. The scope shows a few steps, as it gets pumped to there, but it doesn't continue upwards.> > > Back when it wasn't so lethal to ride motorcycles on the street (and I > wasn't too bright) I use to make my own CD ignitions. > > I liked this topology > > https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/CDI.jpg > > because it was easy to drive the SCR gate with a simple RC circuit > from the points. A low value resistor from gate to ground ensures > turnoff. You could elect to kill the boost converter during the gate > pulse, too. > > To get a fat spark, you need a pretty wide SCR gate pulse, to let the > whole mess ring a few cycles. I usually ran around 80 to 100 mJ per > spark, which will fire a plug no matter how fouled. Barely needs > gasoline.Looks like I'm using a bit under 50mJ, which seems to be working, albeit with a plug in reasonable condition.> > Actually, my HV supplies were open-loop 2-transistor DC/DC converters > with full bridge rectifiers. It sort of relaxed for a while after > being shorted, then started oscillating again, which was nice. > >My application is the engine of a generator. It doesn't have points, but has some magnet/coil arrangement that produces a pulse at the correct point in the cycle. Certainly charging through the coil simplifies things, but I had/have no idea what the inductance of the coil might be. Indeed, it turned out that my coil had insulation failure in the HT side, and I had to source a new one, so even if I'd tried to measure the old one, the new one could have been different. Finally got the engine running today, using this ignition circuit, and LPG, after many hours spent cursing a PIC32MX that I was temporarily using to generate the trigger pulse - got caught out by yet another PIC errata - input capture doesn't work when the processor is in the idle state. Sylvia.
Reply by ●March 27, 20172017-03-27
On 3/26/2017 01:33, Sylvia Else wrote:> On 26/03/2017 4:33 PM, John Larkin wrote: >> On Sun, 26 Mar 2017 15:36:06 +1100, Sylvia Else >> <sylvia@not.at.this.address> wrote: >> >>> I'm using this circuit to provide 250V pulses to an ignition coil. >>> >>> https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 >>> >>> The IC is a Linear Technology LT1243. >>> >>> 12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 >>> is ground. >>> >>> LTSpice says that this works, and indeed the real implementation does. >>> Mostly. >>> >>> The difference is that after the circuit has been triggered once, it >>> won't charge the capacitor again (above about 25 volts) for a varying >>> period that can be up to half a minute. It's as if the wrong level is >>> being used internally to compare with the feedback pin. In that regard, >>> I'll mention that Q3 is used to pull FB up so that the circuit stops >>> driving the FET, since otherwise the SCR never turns off. I've checked, >>> and the problem is not that Q3 remains on, nor is it that the SCR fails >>> to turn off. In any case, the 1.5uF capacitor C1 does get charged >>> somewhat, as I mentioned above. >>> >>> When the issue resolves, it does so rapidly, within a fraction of a >>> second. >>> >>> Connecting an oscilloscope probe to the RT/CT pin of the IC restores >>> function, but only if the scope is set to DC coupling - not AC coupling >>> - and functioning continues only as long as the probe remains connected >>> - until the aforesaid varying period has elapsed. Touching the pin with >>> a finger has a similar effect. >>> >>> That is, connect probe; works. Disconnect; stops working. Reconnect; >>> works, etc. until eventually it starts working continuously. >>> >>> This cannot be capacitive loading - there's already a significant >>> capacitance between RT/CT and ground. >>> >>> The fact that the coupling has to be DC made me think that some small >>> current was flowing to ground through the scope, so I bridged C4 with a >>> 100K resistor. It didn't make any difference. >>> >>> I tried changing the IC (a pain, given that it's surface mounted), but >>> again it made no difference. >>> >>> Any other ideas about what might be happening here. >>> >>> Sylvia. >> >> Just possibly the SCR is staying on, shorts the inverter, which >> thermal limits, and recovers later? It's awfully complex. > > I don't see any sign of that - the voltage across the coil is zero at > this point, whereas the capacitor C1 has about 25V across it, dropping > to zero each time the circuit is triggered, and then climbing back toe > 25V. The scope shows a few steps, as it gets pumped to there, but it > doesn't continue upwards. > >> >> >> Back when it wasn't so lethal to ride motorcycles on the street (and I >> wasn't too bright) I use to make my own CD ignitions. >> >> I liked this topology >> >> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/CDI.jpg >> >> because it was easy to drive the SCR gate with a simple RC circuit >> from the points. A low value resistor from gate to ground ensures >> turnoff. You could elect to kill the boost converter during the gate >> pulse, too. >> >> To get a fat spark, you need a pretty wide SCR gate pulse, to let the >> whole mess ring a few cycles. I usually ran around 80 to 100 mJ per >> spark, which will fire a plug no matter how fouled. Barely needs >> gasoline. > > Looks like I'm using a bit under 50mJ, which seems to be working, albeit > with a plug in reasonable condition. > >> >> Actually, my HV supplies were open-loop 2-transistor DC/DC converters >> with full bridge rectifiers. It sort of relaxed for a while after >> being shorted, then started oscillating again, which was nice. >> >> > > My application is the engine of a generator. It doesn't have points, but > has some magnet/coil arrangement that produces a pulse at the correct > point in the cycle.I just sortakinda fell into this thread and I do not have the credentials to be jumping in, but I do have to say that that that "magnet/coil arrangement" is called a "magneto" on the backyard and shade-tree mechanic circuits I used to hand on, and it puts out a shocking pulse tall enough to fire a plug on a good day.> Certainly charging through the coil simplifies things, but I had/have no > idea what the inductance of the coil might be. Indeed, it turned out > that my coil had insulation failure in the HT side, and I had to source > a new one, so even if I'd tried to measure the old one, the new one > could have been different. > > Finally got the engine running today, using this ignition circuit, and > LPG, after many hours spent cursing a PIC32MX that I was temporarily > using to generate the trigger pulse - got caught out by yet another PIC > errata - input capture doesn't work when the processor is in the idle > state. > > Sylvia. > > > > > > > >-- quis custodiet ipsos custodes? -- Juvenal
Reply by ●March 27, 20172017-03-27
On 27/03/2017 4:54 PM, Larry Sheldon wrote:> On 3/26/2017 01:33, Sylvia Else wrote:>> My application is the engine of a generator. It doesn't have points, but >> has some magnet/coil arrangement that produces a pulse at the correct >> point in the cycle. > > I just sortakinda fell into this thread and I do not have the > credentials to be jumping in, but I do have to say that that that > "magnet/coil arrangement" is called a "magneto" on the backyard and > shade-tree mechanic circuits I used to hand on, and it puts out a > shocking pulse tall enough to fire a plug on a good day.There are such things, but this is not that. As originally supplied it was an ordinary capacitive discharge ignition. There's a separate winding that produces the power used to energise the capacitor, and then this magnet and coil device is used to produce a pulse that triggers the discharge into the coil. Sylvia.
Reply by ●March 27, 20172017-03-27
On Sunday, March 26, 2017 at 2:33:13 AM UTC-4, Sylvia Else wrote:> Finally got the engine running today, using this ignition circuit, and > LPG, after many hours spent cursing a PIC32MX that I was temporarily > using to generate the trigger pulse - got caught out by yet another PIC > errata - input capture doesn't work when the processor is in the idle state. > > Sylvia.Was that PIC quirk the problem? Otherwise, I'd ask o Whether C4 might be bad, open, or the wrong value? Or, that not being the case, maybe the LT1243's upset about having its FB pin pulled all the way to +5v (wind-up?)? Is the LT1243's Vref getting glitched from your reset circuit, upsetting the LT? Just some wild ideas. From your description it really sounds like it's RT/CT pin-related, and you've already put your finger on it (so to speak :). Cheers, James Arthur
Reply by ●March 27, 20172017-03-27
> > Is the LT1243's Vref getting glitched from your reset circuit, upsetting > the LT? >or are there invisible glitches on the feedback pin, try removing C5? m
Reply by ●March 27, 20172017-03-27
On Sun, 26 Mar 2017 15:36:06 +1100, Sylvia Else <sylvia@not.at.this.address> wrote:>I'm using this circuit to provide 250V pulses to an ignition coil. > >https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 > >The IC is a Linear Technology LT1243. > >12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 >is ground. > >LTSpice says that this works, and indeed the real implementation does. >Mostly. > >The difference is that after the circuit has been triggered once, it >won't charge the capacitor again (above about 25 volts) for a varying >period that can be up to half a minute. It's as if the wrong level is >being used internally to compare with the feedback pin. In that regard, >I'll mention that Q3 is used to pull FB up so that the circuit stops >driving the FET, since otherwise the SCR never turns off. I've checked, >and the problem is not that Q3 remains on, nor is it that the SCR fails >to turn off. In any case, the 1.5uF capacitor C1 does get charged >somewhat, as I mentioned above. > >When the issue resolves, it does so rapidly, within a fraction of a second. > >Connecting an oscilloscope probe to the RT/CT pin of the IC restores >function, but only if the scope is set to DC coupling - not AC coupling >- and functioning continues only as long as the probe remains connected >- until the aforesaid varying period has elapsed. Touching the pin with >a finger has a similar effect. > >That is, connect probe; works. Disconnect; stops working. Reconnect; >works, etc. until eventually it starts working continuously. > >This cannot be capacitive loading - there's already a significant >capacitance between RT/CT and ground. > >The fact that the coupling has to be DC made me think that some small >current was flowing to ground through the scope, so I bridged C4 with a >100K resistor. It didn't make any difference. > >I tried changing the IC (a pain, given that it's surface mounted), but >again it made no difference. > >Any other ideas about what might be happening here. > >Sylvia.Using a PIC when you should have gone analog >:-} <http://www.analog-innovations.com/SED/CD-Ignition-Basic.pdf> <http://www.analog-innovations.com/SED/CDI_Inductor.jpg> ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | Thinking outside the box... producing elegant solutions.
Reply by ●March 27, 20172017-03-27
Den tirsdag den 28. marts 2017 kl. 01.25.23 UTC+2 skrev Jim Thompson:> On Sun, 26 Mar 2017 15:36:06 +1100, Sylvia Else > <sylvia@not.at.this.address> wrote: > > >I'm using this circuit to provide 250V pulses to an ignition coil. > > > >https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 > > > >The IC is a Linear Technology LT1243. > > > >12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 > >is ground. > > > >LTSpice says that this works, and indeed the real implementation does. > >Mostly. > > > >The difference is that after the circuit has been triggered once, it > >won't charge the capacitor again (above about 25 volts) for a varying > >period that can be up to half a minute. It's as if the wrong level is > >being used internally to compare with the feedback pin. In that regard, > >I'll mention that Q3 is used to pull FB up so that the circuit stops > >driving the FET, since otherwise the SCR never turns off. I've checked, > >and the problem is not that Q3 remains on, nor is it that the SCR fails > >to turn off. In any case, the 1.5uF capacitor C1 does get charged > >somewhat, as I mentioned above. > > > >When the issue resolves, it does so rapidly, within a fraction of a second. > > > >Connecting an oscilloscope probe to the RT/CT pin of the IC restores > >function, but only if the scope is set to DC coupling - not AC coupling > >- and functioning continues only as long as the probe remains connected > >- until the aforesaid varying period has elapsed. Touching the pin with > >a finger has a similar effect. > > > >That is, connect probe; works. Disconnect; stops working. Reconnect; > >works, etc. until eventually it starts working continuously. > > > >This cannot be capacitive loading - there's already a significant > >capacitance between RT/CT and ground. > > > >The fact that the coupling has to be DC made me think that some small > >current was flowing to ground through the scope, so I bridged C4 with a > >100K resistor. It didn't make any difference. > > > >I tried changing the IC (a pain, given that it's surface mounted), but > >again it made no difference. > > > >Any other ideas about what might be happening here. > > > >Sylvia. > > Using a PIC when you should have gone analog >:-} > > <http://www.analog-innovations.com/SED/CD-Ignition-Basic.pdf> > > <http://www.analog-innovations.com/SED/CDI_Inductor.jpg>why even bother with CDI? most car do fine with regular coils, many coils even have the driver igbt buildin all you need is a 5V driver for a few ms of dwell time
Reply by ●March 27, 20172017-03-27
On Mon, 27 Mar 2017 16:31:19 -0700 (PDT), Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:>Den tirsdag den 28. marts 2017 kl. 01.25.23 UTC+2 skrev Jim Thompson: >> On Sun, 26 Mar 2017 15:36:06 +1100, Sylvia Else >> <sylvia@not.at.this.address> wrote: >> >> >I'm using this circuit to provide 250V pulses to an ignition coil. >> > >> >https://www.dropbox.com/s/1w33uss3eedem05/ignition2.pdf?dl=0 >> > >> >The IC is a Linear Technology LT1243. >> > >> >12V is provided on G$1 to the left. G$2 accepts a trigger pulse, and G$3 >> >is ground. >> > >> >LTSpice says that this works, and indeed the real implementation does. >> >Mostly. >> > >> >The difference is that after the circuit has been triggered once, it >> >won't charge the capacitor again (above about 25 volts) for a varying >> >period that can be up to half a minute. It's as if the wrong level is >> >being used internally to compare with the feedback pin. In that regard, >> >I'll mention that Q3 is used to pull FB up so that the circuit stops >> >driving the FET, since otherwise the SCR never turns off. I've checked, >> >and the problem is not that Q3 remains on, nor is it that the SCR fails >> >to turn off. In any case, the 1.5uF capacitor C1 does get charged >> >somewhat, as I mentioned above. >> > >> >When the issue resolves, it does so rapidly, within a fraction of a second. >> > >> >Connecting an oscilloscope probe to the RT/CT pin of the IC restores >> >function, but only if the scope is set to DC coupling - not AC coupling >> >- and functioning continues only as long as the probe remains connected >> >- until the aforesaid varying period has elapsed. Touching the pin with >> >a finger has a similar effect. >> > >> >That is, connect probe; works. Disconnect; stops working. Reconnect; >> >works, etc. until eventually it starts working continuously. >> > >> >This cannot be capacitive loading - there's already a significant >> >capacitance between RT/CT and ground. >> > >> >The fact that the coupling has to be DC made me think that some small >> >current was flowing to ground through the scope, so I bridged C4 with a >> >100K resistor. It didn't make any difference. >> > >> >I tried changing the IC (a pain, given that it's surface mounted), but >> >again it made no difference. >> > >> >Any other ideas about what might be happening here. >> > >> >Sylvia. >> >> Using a PIC when you should have gone analog >:-} >> >> <http://www.analog-innovations.com/SED/CD-Ignition-Basic.pdf> >> >> <http://www.analog-innovations.com/SED/CDI_Inductor.jpg> > >why even bother with CDI? most car do fine with regular coils, many coils even have the driver igbt buildin all you need is a 5V driver for a few ms of >dwell timeBut can it saw Plexiglas ?>:-} ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | Thinking outside the box... producing elegant solutions.
