Injector Testing Fred/FreeEMS Style

FreeEMS topics that aren't specific to hardware development or firmware development.
Locked
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Injector Testing Fred/FreeEMS Style

Post by Fred »

Pretty soon, as soon as I get some orings for the injectors that I have, I'll be doing some testing. In typical Fred style, I won't be doing it like almost anyone else on the planet. So I thought I'd do a write up on exactly what I'll be doing with a pictorial guide. This first post can be for "best laid plans of mice and men" and later posts with pics and progress and so forth.

Comments: viewtopic.php?f=41&t=2404

The Plan

The plan is to gather a lot of data for a single pressure (whatever the regulator is setup for) at various voltages and drive methods. To achieve this I'll be using some old equipment that I built for making noise. My list of gear:
  • 500w variac/autotransformer
  • 500w 12V stepdown transformer and rectifier setup
  • 1 Farad capacitor network with brass bus bars
  • One truck battery with idling hotel attached to it
  • Bosch 130lph pump as detailed in hotel thread
  • Starion turbo TBI throttle and fuel rail/regulator
  • Two unknown 1 ohm injectors that appear identical - let's prove that
  • Four identical small GSXR600 injectors that it'd be nice to know the flow and dead time data for
  • Four 6A FETs on a Jaguar 0.6-alpha board
  • Various resistors to limit current through the FETs
  • One can of asparagus spears :-p
  • 10 litres of 91RON pump gas
  • 30 litres of pure toluene
Goals
  • Minimum current that will open the injector at nominal pressure
  • Minimum current that will keep the injector open at nominal pressure
  • 3D Table of dead time vs. battery voltage vs. drive resistance
  • 3D table of opening time vs. battery voltage vs. drive resistance
  • Injector flow at nominal pressure
  • Non-linearities in flow at low pulsewidths
From the above data we can use 6V as our lowest working voltage and choose a resistor that provides the minimum current, and some margin, for holding it open. We can also know what the opening and dead times will be like at the chosen drive resistance that limits peak current to 6A at 15V. V = IR, 15/6 = 2.5 = R, R - 1 = 2.5 - 1 = 1.5 = R-ballast. Running through that again for a 5A limit we get 3 ohms, and 2 ohm ballast. This seems like a good place to be for opening the injector. Holding it open will need to be determined through empirical tests.

Method

In the past I gave Sim some bad advice and he got some average
data from this. Sorry Sim! This section is wrapped to your taste to
make up for my failing in that area. I plan to look at what I told him,
what he got, and figure out what the right approach is, and post
that here.

Aside from that, the fuel pump will be powered from the big battery
in the idling hotel under FreeEMS control, this will ensure consistent
pressure for all runs, regardless of voltage being fed to the injectors
under test.

The asparagus tin has smooth sides, and a nearly flat bottom. Due
to this I can calculate the precise volume from the internal diameter
and depth of fuel in the can and not worry about trying to read
graduations and find/buy fancy measuring cylinders, etc.

The variac will power the transformer which will fill the capacitors
which will power the injectors under test. Using this setup I can
generate a smooth and reliable voltage anywhere from about 18
down to 2 or 3 or so.

My voltage break points will be 6, 7, 8, 9, 10, 11, 12, 13, 14, and
15 volts. I'll use resistances of 2, 4, 8, 16, 32 ohms to start with,
and some intermediate values once I know the first value that fails
to get and/or hold the injector open.

I'll use a variety of pulse widths and pulse counts such that the
results of measuring fuel volumes can be plotted in neat curves
and used to calculate the dead time and flow rate for the given
circumstances in the particular test. More detail on this once I've
fully analysed what is required and where I failed last time.

This thread shall be locked for the time being, start a comments thread if you wish, otherwise I'll probably open it up after I've completed testing and posting.

EDIT: Someone did: viewtopic.php?f=41&t=2404

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Fuel Pump Testing Interlude

Post by Fred »

I thought I'd start things off by testing the pump out pretty thoroughly, so I hereby present full voltage versus flow curves for the Bosch 0 580 464 070 Pump!

Toluene was used as the test fluid because I have some and because it's a standard liquid, unlike petrol which varies a lot. The viscosity is similar, but different, so these flow numbers are not what you'll get running pump gas, but they ARE what you'll get running straight toluene, which I actually did once, albeit briefly :-)

The pump:

Image


Click these thumbs for 1280 resolution images:

Image

Image


Ultrasonic cleaning the injectors:

Image


Stop watch and beer-tin based test rig. Capacitor bank in box under amp. Volt meter ah la China bought in BCN. Homemade 8ga jumpers used as switch. Synchronised switch/start of stopwatch/nokia.

Image


Forgot to put the hose in the bottle for the current test, lost about half a litre onto the drive. Took a break while it evaporated:

Image


Transformer+rectifier box, variac power source and some wiring visible in this shot:

Image


Aerial view of the good bits:

Image


Variac front panel:

Image

Touch of nostalgia for the Apple fanboys out there (at least one, you know who you are):

Image

Data and graphs:

Image


BIG version of the important graph:

Image



Overall I'm very happy with this data and analysis. The errors are realistic and insignificant. The pump is strong and does not seem to overwhelm the regulator. A good result with no leaks from the fuel rail or injectors used in the pressurised testing in the second run.

Sadly I don't know what the rail pressure was. Using an air compressor, I could raise the rail pressure by various increments to see how the pump responds to increasing pressure, however it's irrelevant to my cause, so I'll live with this data this time. In future I'll build a serious rig with various instrumentation and voltage regulation to a finer degree.

We conclude this post with compelling data for the 0580464070 Bosch EFI fuel pump by stating that it's good for about 400 horsepower naturally aspirated.

Next up, injector testing! :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Injector Testing Fred/FreeEMS Style

Post by Fred »

Today's goal was to determine the drive characteristics required for these injectors. Goal achieved. If I were to P&H them resistively, I'd need about 6 ohms @ 30 watts to open them and about 12 ohms at 15 watts to hold them open. This is for fixed NA use, not boosted use in case you have a starion turbo and are reading this. Because I CBF doing any proper P&H mods and because the injectors open pretty convincingly even at very low voltages/currents, I'm just going to rock 6 ohm 30 watt drive resistors and treat them like high-z injectors. This shouldn't overheat them and is perfectly fine for the FETs involved (~2amps). I may drive the secondary set from 4 ohms instead so as to provide a different dead time requirement to the firmware. Tomorrow, if I get time, I will do flow testing of these units with various drive resistors and various voltages and various pulsewidths and determine the flow rate and the dead time for each drive configuration. Today was cut short by a requirement to drive an hour north to pick up father and fish and fishermen. Lost 5 hours but had a nice afternoon. Pretty tired now. Data below, screenies are the same bar the min voltage requirement field which affects various other calculated fields.

Image

Image
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Injector Testing Fred/FreeEMS Style

Post by Fred »

Initial static one litre flow tests performed:

1:47.31 to a litre for one (This is primary injector, scribed with '2') and
1:45.94 + ~0.50 stopwatch fail error =
1:46.44 to a litre for the other (This is secondary injector, scribed with '1')

These equate to injector flow figures of 559.1279606174 cc/min and 563.6980595063 cc/min respectively. A far cry from the 850cc/min the internet tells me :-p

Tolerance on those figures is about 550.3511496124 - 568.0095528235 for the former and 554.8410736575 - 572.6609864733 for the latter.

EDIT: Also tested one of my 1998(?) GSXR600 ITB injectors and got 212.68cc/min with min of 209.59 and max of 215.81, note, this has a higher rail pressure than the TBI setup. I don't know what pressure either is, only the relative order of them.

Next up pulsed testing for dead time and small-pulse calibrations. EDIT: Next turned out to be retesting the pump at the higher pressure.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Injector Testing Fred/FreeEMS Style

Post by Fred »

Data and graph with three pump tests together:

Image

Image
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Injector Testing Fred/FreeEMS Style

Post by Fred »

Injector driver resistors manufactured!

http://stuff.fredcooke.com/injector.driver.resistors/

One is theoretically 5.4667 Ohms (30 Watts) and the other is theoretically 4.1 Ohms (40 Watts). 5.45 Ohms is the correct value to ensure consistent opening down to a 6V supply during cranking. 4.1 is simply a different value to ensure I have a realistic situation to deal with in the firmware. I guess NZI stood for "New Zealand Insulators" in days long gone by. This album uploaded partly for a certain person who has a fetish for my chunky soldering exploits. You know who you are.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
Locked