Author Topic: Fuel injection systems  (Read 96553 times)

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #345 on: September 18, 2021, 07:50:54 AM »
I have been reviewing my work on fuel injection and have come up with a few problem areas to investigate:

Testing fuel:- I have been using cooking alcohol as a test fuel as it does not smell. Unfortunately it is not so compatible with the Viton seals and tends to absorb moisture causing some corrosion.

Delivery valves:- Although I have tried to improve the finish of the seatings I still don't think they are good enough.

Dead space:- There is too much deadspace between the top of the plunger at full stroke and the delivery valve making it difficult to get the air out of the system.

To resolve the first problem I purchased some lamp oil which is more like Diesel in consistency but doesn't smell. It also doesn't evaporate like the alcohol does and makes the 'ink blot' tests longer lasting. I tried a quick test with 100 strokes from the test pump which gave 1cc. The pump plunger is 2mm diameter and the stroke is around 4mm which gives a theoretical volume of 1.2cc so close enough.

<a href="https://www.youtube.com/watch?v=Ix2ve6jfvDk" target="_blank">http://www.youtube.com/watch?v=Ix2ve6jfvDk</a>

I then did another quick trial with the helix pump on the diesel. 100 strokes (500rpm for 12 seconds) gave 0.5cc so as the theoretical stroke is 2mm again not too far away from reality. I will do a more controlled range of tests on the helix pump to look at linearity and speed dependence.
Best regards

Roger

Online Vixen

  • Full Member
  • *****
  • Posts: 3075
  • Hampshire UK
Re: Fuel injection systems
« Reply #346 on: September 18, 2021, 01:08:09 PM »

To resolve the first problem I purchased some lamp oil which is more like Diesel in consistency but doesn't smell. It also doesn't evaporate like the alcohol does and makes the 'ink blot' tests longer lasting. I tried a quick test with 100 strokes from the test pump which gave 1cc. The pump plunger is 2mm diameter and the stroke is around 4mm which gives a theoretical volume of 1.2cc so close enough.

Hello Roger,

Good to have a calibration test. Is the lamp oil you used the same a Colemens lamp fluid? I believe Colemans is naptha based and you can run engines on it. It's quite expensive, but as you live in Switzerland so you will be used to expensive.

Mike
It is the journey that matters, not the destination

Sometimes, it can be a long and winding road

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #347 on: September 18, 2021, 05:10:38 PM »
Hi Mike,

As I understand Colemans is equivalent to the Alkylate Benzin (Petrol) that I use for my spark ignition engines (~CHF20 for 5 Litres at the local agricultural merchants). This lamp oil is a much heavier fraction somewhere between kerosene and diesel but virtually smell less  (~CHF3.60 for a Litre again at the local agricultural merchants).
« Last Edit: September 18, 2021, 08:14:03 PM by Roger B »
Best regards

Roger

Offline dieselpilot

  • Full Member
  • ****
  • Posts: 341
Re: Fuel injection systems
« Reply #348 on: September 18, 2021, 08:26:20 PM »
I'm fairly certain diesel calibration fluid is essentially lamp oil or slightly wider, heavier fractions. Also makes for excellent cutting fluid for aluminum.

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #349 on: September 20, 2021, 05:03:56 PM »
I have played a bit more with various pump strokes and speeds:

At 1200 and 2800 rpm the pump response is fairly consistent and linear. At 500 rpm the output has dropped off considerably.

There is another problem  ::) The metering helix is too narrow so at the nominal zero output and when turned beyond zero there is a significant delayed injection. I have tried to display this, the helix is 'unrolled' horizontally and the plunger moves vertically past the port. To the right of the diagram is the normal injection, to the left is this 'parasitic' injection. I will re-phase the rack to avoid going past 'stop' and try again.
Best regards

Roger

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #350 on: September 21, 2021, 07:24:50 PM »
After setting the rack to my theoretical zero I ran a full trial at 2800rpm and a couple of points at 500rpm. The 2800 rpm curve shows the mechanical stroke limit. The calculated fuel requirement is around 0.002cc for full output.
Best regards

Roger

Offline steamer

  • Global Moderator
  • Full Member
  • *****
  • Posts: 12697
  • Central Massachusetts, USA
Re: Fuel injection systems
« Reply #351 on: September 22, 2021, 12:57:27 AM »
Now I assume the volume should be linear with rpm....considering the volume pumped, that's AMAZING!

That's coming along nicely!   :cheers: When do I place my order for the 917?   8)
"Mister M'Andrew, don't you think steam spoils romance at sea?"
Damned ijjit!

Offline Admiral_dk

  • Full Member
  • *****
  • Posts: 3750
  • Søften - Denmark
Re: Fuel injection systems
« Reply #352 on: September 22, 2021, 11:24:41 AM »
I agree that it looks great  :ThumbsUp:

But to be a spoil sport - try to do a measurement with the "throttle lever" at the same position and do the plot for a big number of different RPM's - like 500, 1000, 1500, 2000, 2500, etc.  Or maybe better 500, 750, 1000, 1250, 1500, 1750, 2000, etc.

I expect that the liniarity increase with RPM until you get close to max pumping capacity for that position of the lever. Lower RPM will provide longer time for the fuel to leak past pumping parts ....

Why ...?... well if this is for a gasoline engine, you should have an injection that matches the throttle position, to get the right metering - if it's for your Diesel project it matters less.

Nice progress none the less  :cheers:

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #353 on: September 23, 2021, 05:13:32 PM »
Thank you both  :)

I made some initial trials with various speeds but my easy choice is limited. my battery drill/screwdriver gives 500 rpm, my big SDS gives 1200 rpm and the mains drill I use for starting gives 2800 rpm. The difference in output between 1200 and 2800 rpm is a few percent. With 500 rpm the output drops to around half (note the different scale on the 500 rpm graph).

I intend to make a new, more compact version, of the helix pump and am currently awaiting some MOD 0.4 rack and 20T gears. The current helix plunger was the first successful one I made and has seen some abuse. I will make a new one with a wider grove for the new pump.
Best regards

Roger

Offline Admiral_dk

  • Full Member
  • *****
  • Posts: 3750
  • Søften - Denmark
Re: Fuel injection systems
« Reply #354 on: September 23, 2021, 10:38:54 PM »
OK - this last graph look better than I feared - you are doing very well, from what I would expect to be Idle for most engines up to a rather big size off cc's  :praise2:

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #355 on: September 26, 2021, 12:01:35 PM »
The next step was too look at leaking delivery valves and injectors which I think is causing the loss of prime in the petrol systems. When I stop the engine heat soaks into the injector and pipework causing expansion of the fuel. If it can leak out air will be sucked back in as things cool.

I set up the test pump with the 20° injector I have been using on the twin. It gave a bad spray pattern and after just standing some 10s of seconds a drop started to appear from the nozzle. This is with a head of less than 0.5m  ::) The older style 30° injector I have been using on the horizontal engine did much the same. The 15° injector used in the last diesel trial did not leak at all  :)

Next I tried to test the delivery valve. I made an adaptor so I could pressurise the pump inlet with my foot pump to around 5bar and fitted a lower pressure gauge to the delivery valve union. With 5bar on the inlet there was around 0.5bar at the delivery valve as soon as I removed the pressure from the inlet the delivery valve pressure dropped to zero.

<a href="https://www.youtube.com/watch?v=weNjKGBM1DQ" target="_blank">http://www.youtube.com/watch?v=weNjKGBM1DQ</a>
 
The second test was just to pressurise the inlet. The pressure would slowly leak away, but if I moved the plunger forward to block the inlet port the leakage stopped.

Finally I removed the delivery valve to check that the spring compression was about as planned, which it was.

Lots of things to think about  :thinking:  :wine1:
Best regards

Roger

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #356 on: October 07, 2021, 05:01:26 PM »
While thinking about a solution to the delivery valves I decided to make a new injector and a new needle for the 20° injector.

The injector nozzle and body were made as before but I made a small change to the way I turned the needle. Even with a 0.5mm centre drill there is very little in a 1.4mm thread to take the centre so I made a small brass bush with a 1.4mm hole to support the needle when turning down the shank. As I wanted two identical 15° needles I used both ends of the piece of 2mm silver steel.
Best regards

Roger

Offline Roger B

  • Global Moderator
  • Full Member
  • *****
  • Posts: 6133
  • Switzerland
Re: Fuel injection systems
« Reply #357 on: October 07, 2021, 05:03:02 PM »
Next steps mill the feed slots, harden the cones, grind and lap.
Best regards

Roger

Offline RReid

  • Full Member
  • ****
  • Posts: 1668
  • Northern California
Re: Fuel injection systems
« Reply #358 on: October 07, 2021, 05:26:25 PM »
Hi Roger. That is some mighty fine work on those eensy-weensy parts! :ThumbsUp: :popcorn:
Regards,
Ron

Offline Dave Otto

  • Full Member
  • *****
  • Posts: 4693
  • Boise, Idaho USA
    • Photo Bucket
Re: Fuel injection systems
« Reply #359 on: October 07, 2021, 09:48:21 PM »
Yes, nice work on those fiddly parts.  :ThumbsUp:

Dave

 

SimplePortal 2.3.5 © 2008-2012, SimplePortal