Megasquirt users thread to discuss all things Squirty

[knightrous]

Hey guys,

Been doing some reading and going over eath points, anyone have a method for finding the location of noise?? Usualy when im using my headlights or indicators i get noise on my tach signal.. Since i changed alt i only seem to get noise whwn i use my headlights and indicators... I got the noise since i changed my front end back to stock.. Cnt figure it.. Any help bery very appreciated.. Thanks guys.. By the way cars on ms pnp

Extra info.

4AGE 16V Bigport + ITB's
DIYPNP v1.5
Stock 4AGE Bigport Loom
Stock Toyota VAST Igniter ignition setup (4-1 trigger, NE wire provides tach signal to MS, MS drives the IGt wire to fire coil)

Intermittent noise induced onto the NE signal (spikes like crazy, causes hell to the ECU as it suddenly sees 10000rpm then 1500rpm with only a pulse between) can last a few seconds or a few minutes.





Has become a rage inducing issue for Domma and myself! Random that the problem started after the front bumper was replaced and the alternator had packed it in, but I've no idea how either of those could cause an issue like this that is SO DAMN INTERMITTENT! If it was a dodgy alternator, it should be throwing noise all the time, but it's not.

[oldcorollas]

try grounding alt body direct to battery?
was there a noise filter cap on the old alt?

some of th eold wiring got kinked and cracked an earth wire somewhere?

[domma_aw11]

Will be checking the wiring soon, had to resolder a ground om the front left parker light.. Whats a noise cap filter on the alt?? One of my earth from the main loom is on the cam covers as its not long enough to go anywere else, but again, its only mad noise as i have changed to the stock front end.. Thanks arron for the additional info

[jabbatron]

..........

[knightrous]

Are you using shielded wire for your tach input, grounded at the ms end only?

It's using the factory shielded wire from the igniter to the MS. If this wasn't suitable, I believe we would see noise more regularly (when at specific rpm ranged or when certain auxilaries are used ie: lights or windows). I've some fancy shielded wire available to run into the loom, but the car ran for months without issue until the front end and alternator was replaced.

[oldcorollas]

so what was different with the alternator?
have you put a scope on the batt voltage? maybe one of the alts field diodes is whacky and voltage is doing funny things?

[domma_aw11]

It was just a remanufactoured alt, so a direct replacement.. I dnt know what a scope is, but when the cars is on, its 13.8v and is about 6v to crank... Otherwise my volt guage is allways stable... I should take the alt to get checked, but i might have an earth leak or somthing...

[domma_aw11]

Can anyone tell me if the 4age alt came with a noise filter cap?? What does it look like, were is it in the engine bay, were i can get a new one?? Anything like that??

Thanks heaps

Dom

[timbosaurus]

Not sure if this belongs in here, or the WPMO thread or the WMMH thread, but I'll put it here because at least you guys will be able to appreciate the frustration, lol.

So after a truckload of MS installs, I thought i'd do one on my own car for a change. Nothing unusual in the setup, straight 6, factory VR 24-1 through a JBperf conditioner, PWM boost control, RC 540cc Low-Z injectors on 2 banks of 3, 3 x wasted spark coils driven by offboard VB921 FET drivers, 6-wire stepper IAC, factory coolant/TPS and GM IAT sensor.

The first few MS2 installs I did, I would build a harness on the car checking everything along the way. But now the risk of f-ing up is low, so it's easier/neater to build & wrap it off the car on the bench before fitting and testing.

But this time it wouldn't barely start... way too much noise (the Achilles heel of MS) on the VR inputs and regular resets while cranking/stumbling. This was surprising, as I pay particular attention to grounding and cable sizing, and use aviation quality twisted & shielded cable on the VR signal, and had never had a problem with it.

So here begins the trials of getting through it, might help one of you guys one day

Day 1:

• First thing was to bypass the VR cable in the loom (in case it was coupling with the coils/injectors/starter cable) with a separate cable... no difference at all.
• Re-do the grounds at the block (all 9 cables!) in higher gauge wire. I suspected that the 6 low-Z injectors might be sinking more current than the high-Z i'm used to and lifting the ground plane. No donut there either.
• Start stressing that maybe the coils/leads are the source of noise, and throw all sorts of stupid ideas around my head to get around this. All too hard so go to bed mighty pissed off after losing a whole day with nothing to show.
• Start reading other peoples reasons for resets and sync losses.

Day 2:

• Next day try the interweb ideas on solutions to resets etc...
• Install cap in H1 boot location. Nothing.
• New ECU supply cable direct to the batt terminal. Nope
• It became obvious that most people having these problems are running low-Z injectors.... so I'm starting to get a feel that this is the elephant in the corner of room.
  
  At this stage my harness is staring to look like a roll of fishing line that's been rolling around in a car boot for 6 months, I'm getting frustrated and wondering if I'm better off simply replacing the bosch injector plugs with Nippondenso plugs and revert to the factory High Z injectors. But I'm stubborn so I persist.
• Fitted a giant 47uF cap across the injector flyback return, and hey preso.... huge improvement! Still some sync losses, but no more resets.
  
  After 2 days, this is call for a celebratory coffee while i think about further improvements to get rid of the occasional sync losses.
• Added a second cap across the second injector flyback. No better.
• Lifted the 12V flyback return pins and routed them direct to the battery. No better.

Day 3:

• Realising I need to get strategic, I start scoping various locations. I'm starting to think I've done everything I can on the flyback circuit, and that these sync losses are something else.
• Put the scope back to the inputs/outputs of the VR daughterboard and see that they just don't line up right. Every now and then I get 1 or 2 extra sync pulses out of the VR board (just leading up to the correct sync pulse) despite the input signal looking pretty clean.
• Start trying noise reduction on the VR board... cap across the power feed. nothing. 1K shunt resistor across signals, nothing. 10k shunt. No better.
• Scream, then go back to the webz and start reading about this MAX9926 conditioner (I'd previously only used the LM based Zeal boards for toyota VR sensor)

Turns out the MAX9926 is 'adaptive', triggering if it sees a signal of larger than 1/3 of the magnitude of the previous one. This sounds perfect, except the noise on the input is nothing near that sort of magnitude, yet it is still putting these extra pulses out.

Hmm... someone mentioned the adaptive timeout on this IC is 85ms, at which point it reverts back to default sensitivity of ~30mV (way more sensitive than the LM chip I'm used to). 600rpm is about 100ms per sync pulse, and I'm only seeing these triggers in this 85-100ms window. Bingo, the smoking gun!

But I've already gone from a 1k shunt up to a 10k shunt between the VR+ and VR- to reduce the sensitivity, so how do I go about fixing this? Well, in my haste, turns out I f-d up... to reduce the sensitivity on the VR you actually need to reduce the shunt resistor, not increase it! So after dropping it down to 470ohms, the signal is as clean as a whistle!



So, what can I tell you from this 3 day experiment...

• You are always learning with MS, despite the amount you have installed. You don't install MS to save money (unless you like to work for free), you do it because you enjoy pain.
• If you are using Low-Z injectors, put a 47uF electrolytic cap between pin 2 of the TIP42 and pin 2 of the TIP145 flyback IC's. Or if you're extreme, take the flyback circuits completely off board with something like JBperfs P&H injector board.
• If you are using a MAX based VR conditioner and a single sync input per revolution, be wary of the adaptive circuit resetting and picking up any stray noise on the input. Now that I know the issue, I assume it would have gone away above ~1400 rpm, but I still wouldn't set up a car like that. So either increase the number of sync pulses (24-2 maybe) to make sure the adaptive process never resets, or use a small shunt resistor (470R worked for me) to reduce the sensitivity.

3 x components to fix - 50c
3 x days to work out where to put them - priceless.

I'm off to re-tidy my loom...
Timbo

[NME308]

Hats off to you Timbo!
Thats some great persistence right there...
Glad to hear the right result was achieved however!

Cheers,
Jason

[oldcorollas]

great detective work!!

i'm going to have to check my setup.. it's never been that great at idling for some reason.. i'd always put it down to cams...

[timbosaurus]

Yeah lots of cursing had. I went with the MS2 rather than a 3, just because I thought I could. I don't think any of the additional hardware on the MS3 would have changed anything though.

OC, what's your config? Are u running a MAX based conditioner? A 24-1 VR wheel?

[Ken_T]

Thanks for that timbo. I'd half heard that ms units don't like noise. I got mine running on the CA and it just hates running. Afr's look fine, timing is confirmed with a light and it should be idling just fine. I've put a backward ramp on the timing and fuelling to try and stabilise idle but it's misfire city and idles like shit. Using just about the same settings on the vipec makes it run better and has a much stronger sounding note but still has plenty of misfires. It runs fine on stock ECU though.

I'm thinking the issue stems from having the ignitor in cabin near the ecu and all the noise it generates. I'm going to ditch the ignitor and switch to some denso coils to see if it goes away. Do you think this is an issue?

[oldcorollas]

OC, what's your config? Are u running a MAX based conditioner? A 24-1 VR wheel?

mate I honestly can't remember, it's been that long. will check it out next time i'm with the car..
I'll see if I can put the scope on and find any odd things

[timbosaurus]

OC, the only reason I ask is that the sync problem I had is only relevant to the MAX based conditioners and its time threshold before it jumps out of it's adaptive mode. I never came across it with the LM chips or the finicky discrete circuit on the V3 main board. Also, if you use something like a 60 tooth wheel with a missing tooth (or anything with more than a couple of pulses per revolution for that matter), you'll never come across the issue because you're not waiting more than the 85ms between signals.

You can see how the noise level is somewhat consistent, yet it's only in the last ~20ms where it has the phantom triggers. This was with the 1k shunt resistor in, once I dropped to the 470R shunt, the phantoms went away.



The reset problem (separate to the sync problem) is only relevant to the low-Z injectors and the PWM and flyback circuit it takes to drive them.