- This topic has 11 replies, 4 voices, and was last updated 6 years, 8 months ago by
.
-
Topic
-
I have a 2001 CR-V (UK, 2l, 4wd). It died whilst driving along, and the recovery service diagnosed ‘no spark’. I have taken the distributor cap off, and seen that a grounded screwdriver held near the spring on the coil “tower” also has no spark.
I’m not particularly mechanical, but have some electrical/electronics knowledge and an oscilloscope. So I got a diagram of the ignition system (from Haynes) and set to with a meter. After a while I think I have narrowed it right down, but it leaves me thinking that one of the new parts I have is also faulty – at which point I suspect myself.
Here’s the relevant bit of diagram:I have many scope traces, but the most recent is looking at the PWR coming in, the IGN pulses, and the Primary Output Control.
In the following trace:
– Cyan is the PWR coming in, tapped from the wire between ICM and Coil (ultimately from fuse 9, HOT IGN ON OR START)
– Yellow is the IGN input signal (from ICM terminal of PowertrainControlModule)
– Pink is the Primary Output Control, tapped from the wire between ICM and Coil
– Blue is the TDC sensorThis looks like the primary of the coil is getting connected to ground for ~24ms.
I get exactly this trace whether I use a new or old ICM and whether I use a new or old coil, in all four combinations.
I would at this point be fairly sure the new coils is faulty, except I’m slightly surprised that the pink trace (ICM > coil connection) doesn’t stay at hard ground until letting go, like the IGN input from the PCM. I’m unsure whether to assume a new faulty part, or whether there is anywhere else I should look.
I guess my question is: does this look like the right coil input, with it’s non-hard rising edge? If so, and there’s no spark at the top of the “tower”, am I right to assume a faulty new coil?edit: I missed a bit of history. I bought an entire replacement distributor for speed and convenience. It did not fix the problem, so I put the old one back on and started diagnosing for myself. The coil and ICM in the new distributor have markings (manufacturer etc) which exactly match the originals, and are the new parts I have swapped in and out as part of diagnosing it.
-
Replies
-
On Hondas of that era, there are a few possibilities here. One; the driver for the ICM, in the ECM, has burned out. However, your engine would most likely run for a few seconds to a minute from dead cold and then turn off.
Secondly, Honda ignition switches are infamous for developing faults leading to certain key positions no longer working. Though, the fact your engine is turning over would likely negate this possibility.
Finally, and somewhat less likely by 2000, the Main Relay may no longer be providing power to the coil at times. (Something I’ve remedied by giving a good smack to the left side of the dashboard). They are known for developing cold solder joints.
In your case, I’d be curious to see what battery voltage looks like for the primary windings in IGN1 and START.Hi, thanks for taking time out to have a think about it..
I’ll answer your last point first because I think it answers 2 of the others by default. Here’s a different trace, which I’ve annotated. For this trace the scope was tapping directly into the connector on the side of the distributor.
The Cyan trace is at the wiring harness distributor PWR connector, the yellow trace is at the wiring harness distributor IGN INPUT SIGNAL connector.
It shows a similar pattern to the trace in my first post, but with the surrounding context of START > CRANK > START, and I had some measurements enabled (near the bottom).The Cyan and Yellow traces are for the same thing in both this trace and the trace in my original post, except that in the trace in the original post the PWR is tapped directly from the coil screw terminal itself (cap and rotor removed).
The original trace adds the PRIMARY OUTPUT CONTROL, in pink, which was also tapped directly from the coil screw terminal.Given they show voltage at the coil windings, I think it rules out the main relay and the ignition switch.
But you have made me think about “the driver for the ICM, in the ECM”. This controls the width of the pulse – the yellow trace in both captures – I am wondering if the pulse (see the trace in my first post) is too long. If it were shorter, then the pink trace shows the voltage at the primary would be lower when it was removed, and it seems to me that would make the field collapse stronger.I am also wondering – and this is now guesswork – whether adjusting the timing would adjust that pulse width, because it looks in my first post trace, to me, as though the TDC sensor is triggering the release of the GND condition. I think I can find that out with a bit more experimenting ..
For the curious, this is how I was tapping the voltages from the distributor, whilst everything was still together (though later some traces were with cap and rotor removed and taken directory from coil screw terminals)
And here is a nice trace (of purely academic interest) showing the relationship between PWR, IGN control, TDC and CYP (cylinder position – a new one to me).
And another of academic interest, this is the only one where I have the context of ignition Off > On > Start > On > Off in one graph, and it has a measurement on it if the peak voltage at the coil INPUT (not output), which presumably is the “flyback” you get when you remove power from an inductive load – peaking at around 340v – still clearly much smaller than the secondary side of course (or rather what the secondary side should be 🙂
edit: pictures in wrong order. fixed.
I am also wondering – and this is now guesswork – whether adjusting the timing would adjust that pulse width, because it looks in my first post trace, to me, as though the TDC sensor is triggering the release of the GND condition. I think I can find that out with a bit more experimenting ..
It doesn’t. Even if I turn the distributor as far as it will go, the pulse is always precisely 20ms.
Thanks for the graph, that is very informative.
All of the cases where I found a bad ECM driver, there was no spark at all. So I’d say you’re in the clear there. Also, by changing ignition timing you won’t change the dwell. Ignition timing is the amount of time before or after TDC, not how long the windings charge.Judging by the 342V output of the primary winding, it would appear the primary coil is in good nick. From your graph, I’d say the ICM is working too.
How about the secondary winding? I wouldn’t recommend shunting the secondary winding into your nice oscilloscope, but what about resistance between the negative post and high voltage out post?you won’t change the dwell
Ah, so that‘s the dwell angle – still playing catch-up here! Thanks.
I checked my
nicecheap oscilloscope ratings – nowhere near 🙂 I don’t have any fancy high voltage probes either.I pulled the connections off the coil and found the specs in the haynes manual (which I assume are correct), and measured based on that. The old coil is out of spec, and the new coil is in spec:
spec: primary 0.6 – 0.8 secondary 12.8k – 19.2k
old coil: primary 0.7 secondary 9.7k
replacement coil: primary 0.7 secondary 15.2kIn case I’m doing something fundamentally stupid, here’s a picture of how I decided there’s no spark at the output post:
I drew the red line because you can’t see the wire 🙂 It’s a decent quality test lead, and I held the screwdriver tip within less than 1mm of the spring tip while someone cranked, and got nothing at all.
I’m very tempted to get another coil, but given the new one is squarely in spec, it feels a bit like throwing parts at it gratuitously 🙁
I agree, the new coil is fine. As for your test method, I can’t say with 100% certainty that it is correct.
One thing that immediately comes to mind is that a plug wire typically has 8k Ohms per foot. Using a simple lead, with no resistance, isn’t a true test.
So I suggest this:
Use the same screwdriver and lead method you have now. With the key on and coil powered (may want to verify), shunt the coil ground for a second and then release it. If no spark is exhibited then your test needs some refinement.With the key on and coil powered (may want to verify), shunt the coil ground for a second and then release it. If no spark is exhibited then your test needs some refinement
I tried this, and still got no spark. With that, I took the coil out completely and tried to create a spark with just the coil and the battery, and still couldn’t get one, despite it looking very easy in a few YouTube videos.
At this point a quick search showed that there was a coil in stock locally for around £23 – and since I couldn’t get a spark with just coil wires and battery I decided to try another coil out of desperation. This was my last dice roll – after that I’d either succeed or give up and have it recovered off somewhere.
Much to my surprise, it worked – I have a runner! Despite the fact that the coil was in spec in terms of resistance, it obviously was in some way deficient. So all this trouble appears to be that the original fault was the coil, and the replacement distributor had a dodgy new coil in it too.
It’s hard to see how the new coil could be faulty but in spec; the only plausible thing I can come up with is that the insulation somewhere inside is breaking down at higher voltages, so that a low current DC resistance test is OK, but as soon as the voltage ramps up enough it’s jumping/leaking somewhere.
Thanks very much Tods213 for useful words of wisdom and helping me to focus in the right area.
Glad to help. This would be the first time I’ve seen a coil pass a resistance test and still be bad. We both learned something here.
A primary coil input of 340v is odd to me. The primary input should be battery voltage. You should see high voltage in the secondary side of the coil.
- You must be logged in to reply to this topic.