fabrication question

General engine tech -- Drag Racing to Circle Track

Moderator: Team

Post Reply
Belgian1979
Guru
Guru
Posts: 3728
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: fabrication question

Post by Belgian1979 » Thu Mar 22, 2018 3:28 pm

No it is a general i/o board that lets you do AD conversion. What you have there seems to be a board that adds voltages then averages them and sends out a voltage signal.
already have 8 digitized map signals to work with. I assume that I can work from there. The problem is however that MS only allows me to use a certain number of channels for map. So I need to figure how I can make the code do the averaging and then send the raw adc to a certain channel that could be used by ms.

mk e
Guru
Guru
Posts: 5374
Joined: Mon Sep 17, 2012 3:19 pm
Location: Elverson, PA

Re: fabrication question

Post by mk e » Thu Mar 22, 2018 3:43 pm

Belgian1979 wrote:
Thu Mar 22, 2018 3:25 pm


I might need that. I have an add on board that lets me hook up 8 maps already. I just need to figure out how to average all of them and then output the signal for MS to be used as map input.
You can't just average them. The signal in each runner goes up toward attmophetic like 2/3 of the time and then down toward runner MAP the other 1/3 sort of, so an average gets you anout 1/4 to 1/3 actual MAP, same problem you have now with the plunum doing the averaging mechanically.

The circuit I posted will pull off the peaks, well valley actually, for you so the strongest vacuum of the 8 is what the ecu will see, then you just need to do a little smoothing on the signal and you're good. My engine idles at 35kPa with this setup.
Mark
Mechanical Engineer

Belgian1979
Guru
Guru
Posts: 3728
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: fabrication question

Post by Belgian1979 » Thu Mar 22, 2018 3:45 pm

Thanks if you can mail it. I'll see what I can do with it.

user-23911

Re: fabrication question

Post by user-23911 » Thu Mar 22, 2018 4:22 pm

Software never works on broken hardware .


Like the diagram Mk e posted.
Didn't allow for a 0.6 volt drop for the diodes leaving a max of 4.4 volt.

But WHY are the diodes there?

Using MAP sensors like that, it's the wrong operating conditions for them, ideally they'll be seeing a fairly constant pressure, not a pressure which fluctuates wildly with crank angle.
Expect them to fail prematurely.

Belgian1979
Guru
Guru
Posts: 3728
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: fabrication question

Post by Belgian1979 » Thu Mar 22, 2018 5:31 pm

Marc,

how did you deal with the fuel pressure regulator ? This would also require a vacuum source to regulate fuel pressure ?

mk e
Guru
Guru
Posts: 5374
Joined: Mon Sep 17, 2012 3:19 pm
Location: Elverson, PA

Re: fabrication question

Post by mk e » Thu Mar 22, 2018 6:32 pm

joe 90 wrote:
Thu Mar 22, 2018 4:22 pm
Software never works on broken hardware .


Like the diagram Mk e posted.
Didn't allow for a 0.6 volt drop for the diodes leaving a max of 4.4 volt.

But WHY are the diodes there?

Using MAP sensors like that, it's the wrong operating conditions for them, ideally they'll be seeing a fairly constant pressure, not a pressure which fluctuates wildly with crank angle.
Expect them to fail prematurely.
Blah. Blah, blah....you are a very cranky man Joe.

The way the circuit works is the main buss pulls down to the lowest voltage sensor (active cylinder) + diode offset
, then there is ANOTHER diode on the output so it is buss-diode offset and the output matches the sendor with a few mV.

It works very well. The pcb has a switching circuit to double the the 5v because the buss needs to be about 6V for it to work right. Or you can use a 12v input and skip the dounler parts but most harnesses have signal, sensor gnd, sensor 5v so the double is nice.
Last edited by mk e on Thu Mar 22, 2018 6:47 pm, edited 1 time in total.
Mark
Mechanical Engineer

mk e
Guru
Guru
Posts: 5374
Joined: Mon Sep 17, 2012 3:19 pm
Location: Elverson, PA

Re: fabrication question

Post by mk e » Thu Mar 22, 2018 6:47 pm

Belgian1979 wrote:
Thu Mar 22, 2018 5:31 pm
Marc,

how did you deal with the fuel pressure regulator ? This would also require a vacuum source to regulate fuel pressure ?
I use a pressure sensor and correct flow rather than try to hold a constant delta pressure.....you're spraying into runners where the pressure is always changing so i figured it wasn't worth bothering with plus it's naturally aspirated so delta p will always be more not less, just read the pressure and correct if your ecu does that?

I had a failing fuel pump last time I ran it and didn't realize until it stalled with pressure under 10psi :shock: so it seems to work, and now I light the check engine light for low fuel pressure so I know :)
Mark
Mechanical Engineer

user-23911

Re: fabrication question

Post by user-23911 » Fri Mar 23, 2018 1:53 am

From someone who likes to tell the world you're a mechanical engineer.......you're not an electrical engineer.

Nuff said.

Belgian1979
Guru
Guru
Posts: 3728
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: fabrication question

Post by Belgian1979 » Fri Mar 23, 2018 3:03 am

Averaging apparently doesn't work in the firmware of either the MS or IO device, so I'm stuck there.

user-23911

Re: fabrication question

Post by user-23911 » Fri Mar 23, 2018 3:19 am

joe 90 wrote:
Fri Mar 23, 2018 1:53 am
From someone who likes to tell the world you're a mechanical engineer.......you're not an electrical engineer.

Nuff said.
Trust me....I know what I'm doing............hahahahahahahahahah

mk e
Guru
Guru
Posts: 5374
Joined: Mon Sep 17, 2012 3:19 pm
Location: Elverson, PA

Re: fabrication question

Post by mk e » Fri Mar 23, 2018 5:28 am

Belgian1979 wrote:
Fri Mar 23, 2018 3:03 am
Averaging apparently doesn't work in the firmware of either the MS or IO device, so I'm stuck there.
As I said, averaging wouldn't work anyway. You need a way to pick off the valleys then average those.

The smoothing or noise reduction function or whatever they call it is basically a running average, it works like 0.9 x oldvalue +0.1xnew reading....it adds lag but it works. I have mine setup to use different factors based on rpm so its always about the same number of cycles worth of lag but that's probably overkill. That part is easy.

Grabbing the valleys without ending up with an electronic version of a check valve is not so easy and involves timing the readings or looking in lowest/cycle or similar which brings other problems......which is what got me to looking for a HW solution. The multiMAP works very well, I'll send you the design stuff.
Mark
Mechanical Engineer

Belgian1979
Guru
Guru
Posts: 3728
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: fabrication question

Post by Belgian1979 » Fri Mar 23, 2018 5:57 am

OK thanks. For me I don't have that much an issue with the actual map values used for the fueling as the current setup runs quite well fueling wise. I rather see it as a way to minimize cylinder cross talk.

I tried to set the fuel pressure based on closed loop, but that wasn't very succesful. The reason was at the time I had both the computer doing closed loop and the FPR still in there and they were fighting eachother. I choose the easiest solution which was open loop control with an FPR.

n2xlr8n
Expert
Expert
Posts: 557
Joined: Tue Jun 28, 2005 9:01 pm
Location: Bama

Re: fabrication question

Post by n2xlr8n » Fri Mar 23, 2018 8:15 am

aka a comparator.

The MS forum (and Mke) is likely where you'll solve this issue.

There are quite a few very sharp EFI folks there.

I use the MS "PimpX" in my 2.3 turbo Ranger- but it's not IR.
He who is in me is greater than he who is in the world.

mk e
Guru
Guru
Posts: 5374
Joined: Mon Sep 17, 2012 3:19 pm
Location: Elverson, PA

Re: fabrication question

Post by mk e » Fri Mar 23, 2018 9:38 am

Belgian1979 wrote:
Fri Mar 23, 2018 5:57 am
I tried to set the fuel pressure based on closed loop, but that wasn't very succesful. The reason was at the time I had both the computer doing closed loop and the FPR still in there and they were fighting eachother. I choose the easiest solution which was open loop control with an FPR.
I don't closed loop my fuel pump. I set it up with a table of battery V/MAF with the cells being own I send to the pump ...I've not tuned it yet but the purpose is just to minimize fuel heating so pwm is just set high enough that pressure is not affected and rhe regukaor remains is full control.

On my injector tab in the middle is a table of injector flow vs diffential pressure. I read MAP and fuel pressure thrn subtract to get the difference and lookup the flow rate at that pressure and used that in the calculation of the injector pulse. This way there is no real need to have the fuel pres regulator hooked to the intake trying to keep the differential pres constant to keep the flow rate constant....i just let the flow rate change and don't care.

I change the deadtime or offset based on voltage and pressure too then correct it with a "short pulse adder" to fix any nonlinearity, so that is also correct regardless of fuel pressure.

Last I have the MAP error of each cylinder as a % difference from cylinder 12. 12 it the one my actuator is attached to, so it's the one I sync everything to match.....this makes doing the sync pretty fast and I also use the cylinder map data to apply a cylinder trim so the mixture stays pretty close even when the tbs are overdue to be sync'd....this seems to work but I've not fully tested it yet.

There are many ways to attack the problems, these are just what made sense to me for the problems I was having and I have the flexibility to change the ecu code to match what I want to try so I'm a bit soiled.
You do not have the required permissions to view the files attached to this post.
Mark
Mechanical Engineer

mk e
Guru
Guru
Posts: 5374
Joined: Mon Sep 17, 2012 3:19 pm
Location: Elverson, PA

Re: fabrication question

Post by mk e » Fri Mar 23, 2018 10:26 am

Here is the layout and BOM but it seems I don't have the gerber that will let you order the pcb :(

I'll pm you the email i have for the guy who drew it, or last i knew he can be found at the rusEFI forum user name kb1gtt.
You do not have the required permissions to view the files attached to this post.
Mark
Mechanical Engineer

Post Reply