Spark mod changes on DIY's Site
 

"Option 1: Spark Output to factory Miata ignitors: (Recommended)
Parts Needed for PCBv3 Output Mod using factory ignitors: (2) 270 to 330 ohm 1/4w Resistors"

I have been going through my megasquirt stuff, specifically miata mods and it seems that the required resistors on the spark outputs has change from 1K Ohm, to 270/330 ohm.

Anyone know why, and what difference this produces?

I think it produces a smoother signal. IIRC Joe Perez mentioned the change in one of his threads.

Yup. The igniter turns out to be a surprisingly capacitive load. When you first start driving it high, it consumes a lot of current and drags the signal down. So with too big of a resistor in the pullup, you get an ugly rising edge on the trigger signal. IOW, the voltage on the line starts very low, and works its way up as the igniter's current requirement tapers off. It doesn't seem to cause any actual harm, but it isn't "right", it probably affects dwell just a tiny bit, and it's likely on the hairy edge of working vs. not working.

So the change to the resistor value was to address this concern. By using a smaller value resistor, more current is available to power the igniter, thus there is less voltage drop, and so the rising edge of the trigger signal is much sharper.

The only real concern with resistor size here is the amount of current being sunk through the drive transistors (Q6 and Q8) when the coils are not on (the CPU turns on these transistors to pull the circuit low when it wants the igniter off.) The devices in question are the 2N3904 (on the 3.0) and the MMBT3904 (3.57) both of which are rated for 200ma continuous collector current. The LED itself is going to pass about 10ma, and so even if you went with a 100 ohm resistor in this circuit (which is what I'm using) you'd still only be passing 60ma total, which is well within the transistor's comfort zone. If you do go with 100 ohms, make sure to use 1/2 watt resistors as the power dissipation in the resistor will be exactly 0.25 watt and you don't want to cut it too close.

Effects dwell huh. This should be quite an interesting change.

I don't have anything as low as 330 on my desk, I have 470 and 100 ohm 1/4 watt. Ill start with the 470 and see what happens.


Thanks Joe, ill report back.

It's a tiny effect. A while back I scoped the trigger voltage vs. coil current on the stock ignition system, and found that even though the trigger rising edge was a bit grungy, it didn't seem to have a huge negative effect on the timing of the coil current. Tenths of a millisecond maybe. Like I said, it works as-is, this is mostly to satisfy purists like me who don't enjoy seeing ugly signals.

Here's a later scope capture of the same thing. This was with the EMU driving the stock igniter and coils. Yellow line is trigger voltage, blue line is coil primary current:

http://img34.picoodle.com/img/img34/...lm_3f63c21.gif

As you can see, the trigger is being significantly drawn down, and never gets anywhere near 5v. The stock ECU was slightly better in this regard, but still showed the same effect. Despite this, the current profile of the coil is close to what I'd expect to see in a healthy system. Note that in this image, it's only dwelling for about 3.7ms and thus the coil is not reaching full saturation, although it is tapering slightly. Never could figure out why the EMU did that...

Interesting stuff.

I wish I had one of those USB interfaced oscilloscopes.

I picked up 330ohm resistors because im just ---- like that, and will put it all back in the car this afternoon.

It would be interesting if this "tiny" effect, produces a big difference, at least for me.

Just an update

"Tiny effect" = 3.5 ms of running dwell.

After putting the computer back in the car I was FINALLY after about 8 months of shit was able to start and run my car with 3.5/2.5 dwell, vs the 6ms I was using since winter started and 5/4 I was using in the summer.

Also, now my fuel readings are totally off, very rich during low map.
I even had to reduce my warm up enrichments by almost 30%.

Thanks for the explanation Joe, it really helps explain why the new resistors have helped to reduce the required dwell on the megasquirt.

P.S. I sent the same question to Joe Cramer, and he said that they made the change because on a select few cars they were getting weak spark with 1k ohm resistors. It just so happens that with my Jewish luck, I am one of those.

I'll be doing this when I get back to cookeville. Maybe the '99 coils don't suck so bad after all.

Wow! I guess the 1.8s are a lot more finicky about their trigger voltage than the 1.6 cars. I'm really surprised that it had such a huge effect.

Still, glad it's working for you. :bigtu:

But now lets figure out why exactly.

Other then the external ignitor on the 1.6 what else is different with the 1.8 cars? Unless that is the only difference and now that I think about it, probably explains why it works on the 1.6 cars and not on the 1.8.

Just as Chad says, I am now also curious about the worthiness of the stock coils. If the megasquirt was producing a weak spark, and now it isnt, the stock coils should be just as good as the COP's if the prior effect of the larger resistors was effecting them as well.

Chad I am assuming you built my MS with the higher 1K Ohm resistors since it is identical to yours. If so, while my car is offline, I want to pull my MS out and do this as well. I've got access to a good soldering station so it should be cake.

So, to be sure I have this right, the overall potential effect is stronger spark with less/same stress on the coils?

Yes

To be more correct, its a stronger trigger voltage. Which I assume means it gives a longer fuller signal to the coils, if the signal is weaker the coil isnt sparking or maintaining the spark for the correct length of time due to voltage drop.

I have an OG (early) MSpnp and Savington is finishing my COPs...does this affect me?

Call DIY and ask them. I honestly don't know if the MSPNPs have always shipped with the lower value resistors or if that change was made some time into the production run. Only they can tell you.

I called DIY for clarification. Matt pointed out that this hasn't been tested on any NB coils. But he also didn't indicate any major worry, i.e. our cars won't blow up. I hope. :)

He said that "for dummies" people like me just need to put one 330 ohm resistor from pad 5 to pad 3, and another from pad 5 to pad 1.

If doing so would really make my car's coils work better and produce a stronger spark, negating the need for COPS, then I'd be very happy.

John, your MS was built with the 1k ohm resistors. I didn't know about this mod until recently.

Thanks. I'll be doing this as well then, though I suspect given my current engine state that you will be testing it before I. :)

One other note: evank has a V3.57 that was using the built in 1K pull-ups, so removing the normal pull up resistors would be something of a pain.

So do 1.6 users need to modify this as well? Or just 1.8?

Huh- I never even noticed that the 3.57 had built-in pullups, but looking at the schematic, there they are! What a thoughtful addition. And as it happens, I have pdexta's MSPNP sitting on my desk right now, and sure enough, R58 and R60 are 1k on his as well. I'll be changing those. :D

Looks like they used 1206-size parts, which are pretty easy to work without special tools. I'm used to dealing with 0402s, but these things are massive by comparison. If somebody with a 3.57 board wanted to do the mod, there's no need to remove them however. Installing a 330ohm resistor in parallel with the existing surface-mount device will yield an effective 250 ohms, which is right in the ballpark of where we want to be. With a small 1/8 watt resistor, it should be pretty easy to tack it onto the sides of the existing device.



This mod will be of theoretical benefit to everyone. It's quite possible that the variance in igniter performance is not just from year to year, but from car to car. When I scoped mine, the actual dwell wasn't being severely affected, but there was some measurable effect.

Come scope mine :D

Did I mention I ripped the caps off my harness after performing this mod. They are useless as well.

We're currently using this sort of mod across the board in the MSPNP lineup.

I have one of the really early mspnp's. I'm about to drop $600 in dyno time for my car. Do I need to go get it and send the computer back to DIY to get this shit fixed, or not? I just paid for COPs, so I don't really know what to do here. I'd like to dyno the car once because I'm hardly rich.

You're planning to dyno tune the car just prior to making a significant change to the ignition system? Fail. :D

I didn't think the cops were going to get made this year.


I have the car at BEGi, Stephanie is going to strangle me.

Last night my car started in 10F without a single hiccup. CRAAAAAZY.

-----------

Deleted to avoid confusion and further questions or concerns.

Because too much fuel is being injected into the engine to achieve stiochometric combustion with the amount of air present?

Deleted to avoid confusion and further questions or concerns.

I'm going to be so pissed if I can run 18 psi of BOOOOOOSTTTTTTT with my stock coils after doing this mod.

I'm going to have to destroy some nawzz spraying saturns in the name of Joe Perez just to make things right.

LOL'ed at "nawzz spraying saturns".

Nothing is wasted, my friend. COPs are cool in their own regard.

EDIT

Everything I had here before was wrong. I assumed that DIY was populating lower value surface-mount resistors into the R58 and R60 position on the main 3.57 board, however as it turns out they are leaving those at 1k and instead installing a second set of pullup resistors on the daughterboard. For the 9093, these are R1 and R2. You can't see the labels with them installed, but here they are:

http://img34.picoodle.com/img/img34/...lm_adfb9a1.gif

The brown-black-brown code indicates that these are 100 ohm resistors. The fourth (gold) band is a tolerance indicator, it merely means that these resistors are guaranteed to be within 5% of rating.

These are taking +5 from the Vref pin coming across the DB37 and using it to pull up the spark output lines directly at the daughterboard. This is fine- it's basically the same as doing it on the mainboard. With the 100 and 1k resistors effectively in parallel, the resulting equivalent pullup is 91 ohms, which is just fine and dandy.

If your daughterboard has different values in this location, you might benefit from replacing them with 100s if you are having trigger problems or unstable dwell.

Joe, once again, you are the man. Thanks to Sam for bringing this up.

Yo G wach out fo da naws SaturnZ. They$ stoopidz fa$T. Youz need 30 turboz of boostz fo sho.

Hey guys, disregard all my concerns with the richness after the mod(i also deleted those posts).

Im stupid and changed my injector opening time, didnt realize it right away and have been running around with the wrong value.

All is well and the car runs like a swiss clock now, better then it ever has. <knock on wood>

Joe, we actually put a second set of resistors in parallel on the mainboard - the MSPNP does not have those resistors swapped out. The changeover happened before we even brought out the MM9495, so all '94 and later ones are using a lower resistance value, and so do most '90-'93 models. The highest total resistance we've ever used on an MSPNP is 500 ohms.

Ah ha. You guys are installing additional pullups on the daughterboard. Silly me, I figured they'd have just put lower values on the main board.

The unit in question is a 9093. Through a little reverse-engineering, it appears that R1 and R2 are the pullups, as the DB-37 side of each show continuity to 1G or 1H which are the ignition triggers. The other side of R1 and R2 (the side facing the ECU connector) both show continuity to pin 26 of the DB37, which is the Vref +5 supply. You guys are using Vref to do the ignition pullup?

Anyway, these resistors are both 100 ohm, so I guess I didn't need the pullups I installed across R58 and R60 on the mainboard. Damn, I drove 30 miles to buy those things...

Oh well. With your circuit (1000 and 100) the effective resistance is 91 ohms for If = 55ma on the drivers (plus the LED). So there you are, folks. It's safe to use astoundingly low value pullup resistors on your ignition trigger.

Just easier for us to get all the V3.57s set up the same way.

Makes perfect sense from a manufacturing standpoint. Minimize the number of hand operations required on the mainboard and let the pick-n-place robot put the pullups on the daughterboard.

The above writeup has been amended.

So I did the resistor mod today. Seemed like a good idea even though DIY hasn't tested it on NB coils (nor do I know if MSM coils differ from ordinary NB '01+ coils).

Then I went for a long drive and datalogged. When I got home and did the VE analysis in MLV, I decreased the minimum RPM from 1800 to 500, to see if the resistor mod would have any impact on my car's idle (where I often get misfire CELs). Analyzed on the "hard" setting.

Attached are before-and-after MSQ files.

The car feels much better with this mod and the new tune! It started quieter and more easily. Also, before with the frequent misfires, if I sat for a few minutes at idle, I could distinctly feel serious vibrations through the shift knob and even through the damn steering wheel sometimes. AFR at idle was too rich, around 10.5, because the car was way too rough at 11-12. Now, after doing the spark mod, the car idles in 11-12 almost as smooth as when it was bone stock! So I'm very happy about that. I also did a 50 to 100 pull (don't tell the police in Millburn, N.J. :) ) at full 15 psi and the car felt as smooth as it ever has. Nothing but constant torque. Now I'm wondering how much I can increase the spark plug gap and still not have resets (currently it's at .020) .... I guess I'll keep things as-is for now, then try .025, watch that for a while longer, then .030, etc.

I'll also have to keep an eye on cold weather starts to see if it now has too much fuel.

Bottom line, here is one DIY customer who'll be going to sleep happy tonight.

Thanks also to Sam for bringing this up.

PS - My car and I have a long history of declaring problems "fixed" based on short-term results, and then having the problem come back worse than before. So we'll have to see what happens ..... I hope this is as good as it seems though!!

Awesome thread, thanks for the heads up joe et al.

yes joe, thanks again. You are the wind beneath my wings.

Update .... yesterday afternoon I drove 2 hours on the highway to visit a friend for the weekend .... got there and then drove another hour to show off my car's latest mods to the friend .... then drove 2 hours back home tonight ..... and all's still great with this spark mod.

I only had one CEL yesterday - P0507, "RPM higher than expected". Indeed, sometimes my car's idle hangs around 1,500 for a few seconds when I bring the car to a stop, before settling down to normal idle. But the good news is no misfires. So there really is a stronger spark. Both car and owner are giddy. :yippee:

Threadjacking: I've heard of the "hanging idle" problem in Miatas before but I never saw any authoritative fix. How do I make the idle settle down more quickly without changing the idle RPM itself?

Im having some serious issues with resets/spark blow out with my toyota cops and it seems to me this mod would help. Can some please be a pal and outline what needs to be done to a v3.0 board to get this mod one? I have a DIY MS built by Brain...

Thank you Very Much you may bring me off this ledge by helping me. LOL

it'll be easy. as long as you have a soldering iron.

remove the two posts on the little db9 connector, then unscrew the four screws on the opposite side and pull the board out.

If you look on the bottom of the board there are two resistors below the two LEDs. Those right now are 1K resistors, swap them out for 270ohm resistors. You can take one from R32, R30, R31, R36, R34, or R35....you don't need them on the board as long as 1. You have Hi-Res code or 2. You're never going to use the PWM circuit.

I've been building all my latest MSes using 270ohm resistors on the spark pull-up (I'm left with 6 leftover the way I build the board) and been running it myself for some time now.

Mike are you running the Hi-Res code? Do you have the .1uF cap mod performed as well?

Mike, just go buy a pack of resistors at radio shack, 1.07 after tax.

I dont want to go over Scott's head, but scavenging parts from other circuits is just asking for more problems. Not worth it.

I'm willing to bet 0% of us here use the circuit. I don't even build it anymore, with the hi-res code it's rendered useless anyways.

http://boostedmiata.com/MS/built/lue...005%20copy.jpg

I just built this yesterday, notice the missing R32, R30, R31, R36, R34, R35, Q12, Q9, Q13 and Q10?

I did the spark mod without the .1uf cap a few days ago and still get the misfires at idle a bit, so it is definitely needed with the high res code.

yes, the .1uF cap is absolutely necessary when running hi-res code.

Yes and yes... Got a more specific picture of the resisters i need to swap?

Bottom right corner:
http://diyautotune.com/images/cartec...t_mods_med.jpg

http://boostedmiata.com/MS/built/bryce/CIMG1272.JPG


aw im too slow.

Got it! Thanks guys. I will stop by the shack on my way home tonite. I'm not really optimistic that this is going to solve my issue but hey u never know. Ive had the same ms,cops and chassis with an NA 1.6, a turbo 1.6, and an NA 1.8. now all the sudden turbo 1.8L wants to give me resets like a motherfucker? I don't get it.

Switched the resistors for 270 ohm ones and swapped back to stock coils and the car run beautifully. I dont think im going to mess with Cops any more. I just going to sell the two sets i have and run stock ignition...

cripes, I've been getting resets as well. Maybe I'll give this a shot if it becomes more frequent. I only seem to get resets off throttle which is wierd. It has never happened with my foot on the gas...

AAAARRRRGGGHHHH .... after several days of running better than ever from doing this mod, my car is now running TERRIBLE again. :(

All was well until I sat in bumper-to-bumper traffic a few days ago, for about an hour .... every time I came to a stop the idle went a little bit higher ..... got way up to 1600 before I was finally able to exit the jammed-up highway. Then the idle went back to normal. I did feel some misfires though .....

.... The next morning I drove to work and the misfires were very very very bad. Been that way for the past couple of days.

I can't even explain how frustrated I am right now.

Someone several months ago suggested my car's problems could indicate a bad head gasket. So I searched last night on M.net and found a semi-related post which could be the problem:
MX-5 Miata Forum - View Single Post - No. 2 Misfire Mystery, part 2

Here's the post from someone whose car had similar problems to mine, i.e. if you don't want to click the link:

"It turned out to be a wire in the Injection Harness, which is also the harness that goes to the coil pack. After replacing the Oxygen Sensor and that not fixing the issue we were poking around in the engine bay while the car was running and moved the injector harness to look at something and that cause the car to run fine. Move the wires back and it ran rough. Move the wire... fine... you get the idea. We ended up replacing the injector harness by taking off the top half of the intake manifold (don't forget to plug the bottom half of the manifild with a rag, or you will lose one of the injector grommets down the hole like we did). This DID fix the problem."

I guess that's worth my time to check out. Dumb question, when I open the hood, where is this injector harness? I'm not familiar with that. I'm also not clear what the poster meant by "replace" the injector harness .... is he saying get a whole new harness? What's involved for that kind of thing?

Someone educate me please.

(Sorry for the thread drift.)

the only way your idle is going to go up to 1600 is if the stock ecu decided the idle valve needed to be opened more to compensate for something. were you perhaps heatsoaking and you don't have a good corrections table setup?

I have no idea if it was heatsoaking. Without a real temp gauge and without having my laptop on, how would I know?

I don't know anything about that topic.

Did the log indicate that MAT was rising to implausible levels during this period?

What makes you think I was logging? I was only driving home from work.

FWIW I log driving home from work on a regular basis. It is a good way to get some real-world data.

I checked to make sure my soldering from the spark mod is still intact (it is) and I checked to make sure the MS MAP gauge is working correctly (it is - sitting at 100 with key on and engine off).

Paul says I'm not using a corrections table. He sent a long and detailed email about why such tables are bad. :)

So right now the idle is a long-shot secondary concern; the misfires are my primary concern.

To reiterate: I previously changed and checked everything there is to change and check, but still had random misfires ... then I did the spark mod and life was heaven ... then a week later the random misfires suddenly returned. I tried the M.net guy's suggestion (linked in a recent post in this thread) of double-wrapping the coil plug wires in electrical tape for a few inches where they're near the spark plug wires .... that didn't accomplish anything.

I suppose it's plausible that the problem is with the injector connections, but if so, then doesn't explain why the problem magically fixed and then unfixed itself with the spark mod!!!

Also on an NB you have to remove the stock top half of the intake mani to even GET to the injector connections. So obviously the engine can't be on while I do that. How, then, will I know if wiggling the injector wires find a bad connection or not?

Thinking aloud: what's the best way to test that each injector is getting the proper voltage supplied to it? And what is the proper voltage anyway?

Sorry to resurrect. I have done the mod with 270 ohm resistors, but am now also converting to COPs. Can I just take care of the lower value resistors with dwell settings, and if so what dwell should I use? Or do I need to reinstall the 1Kohm ones?

Thanks