DIY Fuel injection thread.

[yoeddynz]

I don't think so but I can check. I think I'd have remembered seeing them because I was sifting through my little box of speeduino bits a few nights ago.

[gibbon]

With enough inputs (air & fuel flows, map and EGT for a start) could an ecu do away with an o2 sensor but still run closed loop?

 

[h4nd]

10 hours ago, yoeddynz said:

3x bosch BIP373 ignitor drivers but its seems they've been out of production for a while. So the new alternative is this...

https://www.diyautotune.com/product/coil-driver-kit/

But postage is excessive. So I looked up OnSemi EcoSPARK IGBT and found them all over the place but there's various different models and my brain started to smoke when I tried to work out what to get.

Like this...

https://nz.rs-online.com/web/p/igbts/2410728?cm_mmc=NZ-PLA-DS3A-_-google-_-PLA_NZ_EN_Fallback_Whoop-_-(NZ:Whoop!)+All+Products-_-2410728&matchtype=&pla-293946777986&gclid=Cj0KCQjwnf-kBhCnARIsAFlg492mG4wDqPoHH3pRLzCB_sjehTvm24w4QkREUBR61FSGuxw4wSha4ooaAhzoEALw_wcB&gclsrc=aw.ds

Bugger, - I don't, maybe @Ned or @kempy?

If you've a little time, I can con a client into ordering some on their next buy (a month or so).

 

Anyways: (adjusts Barrie (tm) personalised cap):

Digikey has a really good parts selector thingy:

• you can see they have 50 kinds of those family IGBTs
  •  https://www.digikey.co.nz/en/products/filter/transistors/igbts/single-igbts/279?s=N4IgTCBcDaIPIDsDKBTAtgSwAQFEDGA9kgAoCCASgNJYCSA4gEIAqIAugL5A
• these guys reckon  https://www.msextra.com/forums/viewtopic.php?t=57426 
  • this goes one OK https://www.digikey.co.nz/en/products/detail/onsemi/ISL9V5036P3-F085/2509749
  • (though - no thermal protection - probably ok to just chuck a 15A fuse inline?)
• Narrowing for 400V ish to 100V, and 23A upwards in to-220 and the slightly bigger TO-262-3 gets you these 8 options:
  • https://www.digikey.co.nz/en/products/filter/transistors/igbts/single-igbts/279?s=N4IgjCBcpgbFoDGUBmBDANgZwKYBoQB7KAbRABYwBmMcgDhAMqvMpAF0CAHAFyhADKPAE4BLAHYBzEAF8CAJjAAGCNBDJI6bPiKlwSpQAIAgoxDyqJsxYB0AVisLyj8w9MLYL%2BbBsBOLwDsXnRe-u4gVEbhNC5U8rGW0c7RbmZUntFBmfaxATZZaSHRYWbkUaVgLuTx4eSJpcmlqUwZpQVMRaUlBHblPZXhdjVmdvU9jT3NIHatPe3TnT3dILB9KwNmsMMEsGMrEytTsLMr87CLK8sBawEbBAHbIAF7AQcBUwEnAfMBFwHLdDWdDuIDojzoezoBzoUzoJzoF18a18IN8j18e18B18U18J18iJKnBAvH4AFVxKIeAB5FAAWRwaCwAFdhDhZAplFQEOpUJhcARiJAyFQkYYAGppHES0oGGVMMAOSVMeRGZUUeRK0qReUUKhaphlXWseLq1gGii9Y2rY2A41i9W9NWbOXq2BW9XXZ33D1mJHekA4gPKV0cbh8SAgClU2kMpms9lyEAAWnIPI0WgFumF0yo%2Bt8HBkRaAA
• Anyways, if your MS is something like the one described here:
  • https://www.diyautotune.com/support/tech/other/troubleshooting-ms-injector-driver/
  • i.e. has Q14, R37 and Q15 and R38 (current limiting) section
  • and Q3, Q11 sections, then probably most of these will do (287 options)
    •  https://www.digikey.co.nz/en/products/filter/transistors/igbts/single-igbts/279?s=N4IgjCBcpgbFoDGUBmBDANgZwKYBoQB7KAbRABYwBmMcgDhAMqvMpAF0CAHAFyhADKPAE4BLAHYBzEAF8CAJjAAGCNBDJI6bPiKlwSpQAIAgoxDyqJsxYB0AVisLyj8w9MLYL%2BbBsBOLwDsXnRe-u4gVEbhNC5U8rGW0c7RbmZUntFBmfaxATZZaSHRYWbkUaVgLuTx4eSJpcmlqUwZpQVMRaUlBHblPZXhdjVmdvU9jT3NIHatPe3TnT3dILB9KwNmsMMEsGMrEytTsLMr87CLK8sBawEbBAHbIAF7AQcBUwEnAfMBFwHLdDWdDuIDojzoezoBzoUzoJzoF18a18IN8j18e18B18U18J18iJKnBAvH4AFVxKIeAB5FAAWRwaCwAFdhDhZAplFQEOpUJhcARiJAyFQkYYAGppHES0oGGVMMAOSVMeRGZUUeRK0qReUUKhaphlXWseLq1gGii9Y2rY2A41i9W9NWbOXq2BW9XXZ33D1mJHekA4gPKV0cbh8SAgClU2kMpms9kyOQrHmiAAm-AAtMoIOH%2BGYeABPLjsyNM5BJoA

Octopart is great for finding who has the thing, and at what MoQ and price:

•  https://octopart.com/search?q=ISL9V5036P3-F085&currency=USD&specs=1

 

Cheers

[h4nd]

9 hours ago, gibbon said:

With enough inputs (air & fuel flows, map and EGT for a start) could an ecu do away with an o2 sensor but still run closed loop?

@Roman would be a good man to tag here, but yes. You could closed loop control (means use feedback from some sensor to attain some goal) and have the system strive to give you the EGT you want, or the coolant temp you desire, or exactly 3000 RPM for your gen-set, etc. 

If you want good hp etc, you may find you need to use an O2 sensor to get a quality of results acceptable to you.

[Roman]

The problem with monitoring EGT is that it's a function of both lean/rich but also ignition timing. 

If you reduce ignition timing, the motor runs hotter - as less of the combustion force has been spent in the cylinder, more comes out the exhaust instead (as heat and more noise)

If you want to run as accurately as possible without closed loop monitoring with a wideband. Then using a MAF as the load source is 100% the way to go. 
The accuracy of a maf is crazy, through changing conditions without any compensation tables for altitude, temp, etc. (which is why OEM loves them)
That's as close as you'll get to closed loop for minimal headaches.

[yoeddynz]

@gibbon ask @kws just how much difference a wideband can make - if only to economy.

 

[yoeddynz]

10 hours ago, h4nd said:

these guys reckon  https://www.msextra.com/forums/viewtopic.php?t=57426 

• this goes one OK https://www.digikey.co.nz/en/products/detail/onsemi/ISL9V5036P3-F085/2509749
• (though - no thermal protection - probably ok to just chuck a 15A fuse inline?)

 

Ahhh yes- Cheers HandofWayne.

I had already found that thread a while back when looking. I'm not sure if no thermal protection is a big deal, especially if like you say I can use fuses instead.  So yeah - this one you've linked to could be the go..

Details here.. https://www.digikey.co.nz/en/products/detail/onsemi/ISL9V5036P3-F085/2509749

 

I've only just setup an account on RS. Not looked into digikey yet. I presume joe bloggs can create an account and order just a few bits ?

 

10 hours ago, h4nd said:

Anyways, if your MS is something like the one described here:

• https://www.diyautotune.com/support/tech/other/troubleshooting-ms-injector-driver/

 

That thread is for the 'lesser' models. I'm running a ms3x

I plan on building the same setup I had on my Viva V6 - three ignitors mounted to an external heatsink separate to the ecu. It worked well. The barely got warm.  I won't be asking much of them, same as on Viva it'll be a 6 post coil running wasted spark.

[gibbon]

5 hours ago, yoeddynz said:

@gibbon ask @kws just how much difference a wideband can make - if only to economy.

 

I was trying to see if it were feasible to have something akin to wideband feedback, without having an actual wideband sensor, for avgas reasons 

[tortron]

9 minutes ago, gibbon said:

I was trying to see if it were feasible to have something akin to wideband feedback, without having an actual wideband sensor, for avgas reasons 

My 3.8m8 ute has a nzdm spec buick V6, and no oxygen sensor as it has a nz leaded fuel tune still on it

It runs on intake air temp and a map sensor

[ajg193]

You'll likely be accurate enough with no sensor after a decent tune.

I disconnected the oxygen sensor on the Hilux and it runs better with no increase in fuel consumption.

Maybe I'll put a new one on since day, but it would never pay off and I can't picture it improving fuel economy below 9L/100km

[gibbon]

Ok so my takeaway here is that a wideband is more of a tuning aid than anything else? I guess cars trundle around in open loop from time to time anyway. Still I'm curious to know the answer from a hypothetical point of view - can an ecu extrapolate meaningful feedback with an EGT sensor rather than o2

[kpr]

Closed loop should only be there for a fine trim.  if isn't running right in open loop,  tune sucks or something has gone wrong.

As roman said.  too many variables to use egt to tune.  main use of egt probes is to see if you have a cylinder/s running different than the rest.   rather than having a wideband in each cylinder.

Just tune it right with wideband,  then run in "open loop"
 

[BlownCorona]

@gibbon one thing you need to bear in mind is, as you know avgas 100/130 and 100LL has lead in it and this will likely kill your sensors fast. i never worked with GA or aircraft that ran this kind of fuel so i dont really know, but the lambda/stoic numbers might also be different to auto fuel.

here must be av spec sensors but i can see the price tag now $$$$, would be interested to know what is done differently to handle the lead. 

edit - i re read your question and now i think this is exactly the problem youre trying to solve.

[tortron]

Yanks claim they last longer than you might expect, so no biggie for racecar

[Roman]

8 hours ago, gibbon said:

I was trying to see if it were feasible to have something akin to wideband feedback, without having an actual wideband sensor, for avgas reasons 

Out of curiosity, what are you needing to run avgas for? 

 

[gibbon]

@Roman it genuinely is a question of pure academic curiosity 

[gibbon]

Come to think of it we look after a couple of machines with fuel injected rotax engines that don't have manual mixture adjustment, can run avgas, and don't have o2 sensors... They must be able to compensate for altitude somehow and I can't imagine they do it without some kind of feedback. The ecu is definitely not connected to the altimeter 

[Roman]

Aircraft people do weird shit with running their engines.
My understanding is they adjust the mixture to peak EGT then back it off 10% richer (or something like that) 
Or some other witchcraft sorta stuff depending on if they are wanting best economy or power when running the motor in steady state slogging its guts out.

So in that case the reaction speed of an EGT sensor isnt an issue - they arent really in closed loop. Just using temps to manually trim the fuel. 
(At least that's my understanding) 

So for the purposes of what they're doing, ignition timing is essentially fixed. As the motor is at steady state.

By contrast the operation of a car engine is absolutely chaotic.
With rpm changes, ignition changes, fuel changes happening constantly. 
Which is probably why they're not commonly used for cars.

 

[gibbon]

yeah good point - the rotax would probably throw a fit if you started giving it the sort of throttle input that a car gets

also generally these EFI aircraft engines have a constant speed prop attached, so even if it did ran sub-optimally for a few seconds while it got it's EGT feedback, the prop would just fine out to keep the RPM constant, so you'd see a decrease in MAP instead and have a mathematical decrease in airspeed but you probably wouldn't even feel it. same setup in a car and there'd be bogging down, RPM decay etc

moral of the story is carbys rule

[Alfashark]

N/A leaning for altitude compensation is just leaning until you feel some very slight rough-running then yes, back it off on the rich side.

Turbocharged things will usually have you keeping an eye on the EGT gauge and the manual might specify 25° lean of peak, eg: lean it out, then keep going leaner once it's hit peak until you hit the number specified.