Re: P&H Injector Interface - Wiring Approach

[garwillis@msn.com (Garfield Willis)]

On Mon, 29 May 2000 17:04:46 -0700, "Walter Sherwin" <wsherwin@home.com>
wrote:

>Whoa Gar! I think I see the problem here. We have vastly differing
>expectations for the wiring approach to this box thingy.

Bingo, Walt. Welcome to the party finally! :)

>Trying to drive "X" number of P&H injectors, through OE wiring which did not
>originally encompass one circuit per injector, is a bad idea IMO.

As I already said, driving each injector with it's individual driver is
ALWAYS better when possible. But the fact is, both cases of driving
multiple SATs and P&H injectors on one circuit are already an
accomplished fact in MOST of these OEM situations. What's being explored
is at most a doubling of the drive capacity on each circuit, a problem
with the connector pins, which we've already pointed out and addressed,
but NOT a problem with the existing wiring.

IF I was worried about the electrical engineering of such a combo, I
wouldn't have suggested it as possible to do that way. I don't see
anything you've mentioned in your response that would say it's a "bad
idea". I'd define a bad idea as something that wouldn't fall within the
bounds of good EE practice; I have no intention of producing anything
that's a "bad idea". If you mean by a "bad idea" that it doesn't seem to
you like the "best idea", well that's fine, everyone's gonna have an
opinion.

Frankly, I wish I *could* believe your suggestion that nobody's gonna
mind if they have to rewire their harness, in order to convert to P&Hs.
Certainly would make my life easier.

>Without
>going into detail, what about: 1)significant line losses

Not a problem, peak currents of 8A are easily handled by normal auto
wiring looms, especially the gauges that are used to wire ANY injectors.

>(2)current sharing issues within a cluster of injectors, arising from
>impedance shifts/variations that normally occur with injectors

I've never heard of solenoid aging causing an injector to significantly
change either it's resistance or inductance. Have you? If either has
happened, then some problem exists in the winding insulation or
connector contacts, a thing that can happen any time. We're not trying
to guard against injector failure, we're assuming you've got good ones.
The electrical characteristics don't change much; the injector may get
old and sticky or slower, but these are mechanical properties that don't
affect the impedance.

>3)unknown OE wiring quality and splice quality

Well, yeah, you could say that of any circuit in a loom. You can just as
easily have a problem with someone rewiring into an existing loom, and
not being savvy about how to make the very best connections, especially
with older wire that's been cooked underhood. Happens all the time.

>There exists more than just a few variants of differing "P&H" low impedance
>injectors, even within a given style family (ie: within the port family
>port, or within the TBI family). Therefore, it is impossible to sufficiently
>generalize clustered P&H injectors, and the equivalent circuits which
>result.

I don't agree with this at all.

Come on, how often do you see people mixing injector types within an
application, especially those that are driven together? Generally,
people are switching them ALL to one particular part no. I'm NOT trying
to cover every possible mongrel install. The same is true with present
setups that drive multiple injectors. You mix different kinds of
injectors in those 'clusters' and you've got the same thing.

Everyone knows that even if the electrical parameters were to exactly
match, open and close times can vary from one part no. to the next, so
nobody's crazy enough to do alot of mixing up of injector types.

While it's true that there are various impedances you run across as you
change brands and part nos., they almost always fall into one of two
categories, where you use either a 4A-peak/1A-hold or a
2A-peak/0.5A-hold driver. If this were not true, completely integrated
injector drivers wouldn't be possible. Another "existence proof" to
answer this "impossibility". WAY overstated, IMO.

>When you ponder the potential mixes, the above quickly become very
>real concerns. The OE's have in the past ganged 2 and sometimes 3 P&H port
>injectors in specific applications with a common current driver with good
>results, but, that is not to say that such applications were not
>problematic.

Ahh, well since I haven't HEARD of any such "problematic" situations,
I'll just have to consider the burden of proof on anyone suggesting they
ARE problematic. One thing I KNOW for sure, another "existence proof" of
sorts, is that NONE of the OEMs did any kind of injector matching per
se. All they did was use the same part no. and that appears to have been
just fine in paralleling them up. It's just that they didn't make the
PINS big enough to allow us to add more driving capacity. Always a
problem with connector schemes.

>For example, certain GM 4cyl ECM's had two 2.05 ohm 2/.5 amp
>Rochester port injectors ganged together in parallel, and driven via a
>common 4/1 amp driver in the ECM. However, as little as a .4ohm impedance
>shift or mismatch between the injectors in a pair would cause one or the
>other to not function properly.

Yep, they (GM) sure did/do. But if you want to worry about this, it
sounds to me like you're alleging they must have indeed hand matched the
injector pairs then? I don't think so. It's already been done/being
done, so you can't argue with the doability of it; at least I can't.

>Anyone who typically finds themselves requiring either more
>injectors, or bigger/expensive P&H injectors, during a DIY project has
>probably arrived there by previously performing a whole bunch of DIY
>swap/tweak/rework/reprogram/redesign stuff. So you are assuming that these
>same people for some reason either don't want to, or can't, add a wee bit of
>wiring?

Not exactly assuming, we "got there from here", by the following
progression. What's been happening is guys have been having their boxes
boosted with higher power drivers, replacing the originals. That's been
happening for a while. If you've been following the thread, you'd have
seen how we started there attempting to find an alternative that left
the harness just as untouched as with the drivers being replaced in
situ. But we found the pins were a limitation, and so had to settle for
at minimum, a hard break near the ECM. I stopped short of dictating a
complete rewire in order to solve the problem, simply because I don't
see it as necessary.

It's an absolute no-brainer to IMPOSE what you've suggested on the
product design. You just sweep all the difficulties and tradeoffs away
by saying, "if you want to convert to P&Hs, you'll have to rewire your
harness" to drive each injector separately. Fine, that makes it all
sooooo easy, except ask yourself this: if given two product choices, one
which requires re-wiring, and one which does not, which one is the more
likely one to be selected by the buyer?

>why the hang-up on rewiring? Make the box robust enough to live
>underhood, and run a few short lengths of wire from the injectors to the
>outputs of the box. The wires would not even have to pass through the
>firewall.

We've been over this in the thread as well. The requirements for living
underhood are alot more problematic, from both a thermal and general
environmental standpoint. Also, the wire runs for the individual
injector connectors aren't usually very long when they bust outta the
loom, so it's not going to be exactly just "a few short lengths" to take
them all to somewhere and then back. Wherever you put such a box
underhood, if it's centrally located, instead of broken up into chunks,
you're going to be increasing the wire run for an individual injector
QUITE a bit.

This scenario infact WOULDN'T require a rewiring; you'd simply put an
"extension cord" on your injector harness over to this centrally mounted
underhood device, and then another "extension cord" from the box back to
each individual injector. Like I said, we've been there and discussed
this.

>Let?s explore your approach to this same injector conversion???. If  I
>understand correctly, you would propose that one P&H current driver would
>simultaneously drive 4 low impedance port style injectors per batched bank
>through the existing OE harness.  Let?s also assume the use of Rochester low
>impedance port injectors having a typical impedance of around 2.05 ohms and
>3.2 mH and a design drive current requirement of 2/.5 amps.   Now model the
>equivalent circuit, taking into account all sources of voltage drop, and you
>will find that the net current will saturate at just over 10.5 amps at
>around 4ms at a system voltage of 13.2 volts

Never happens, if you've got 4 2A peak injectors, the driver is going to
be set to current limit at 8A and then foldback to 2A total. That's the
whole idea. There is most certainly some current variation between
injectors, but have you looked at the variation in driver current
limits? It shows that the nominal driver peak current limits are
purposely chosen to have a LOT of pull-in overkill margin in them
(somewhere around 30%). That's alot more variation than I'd expect from
differences in wiring conductance between ganged injectors. WAY larger.

>>Now, here's the other problem with your conjecture that "8 individual
>>4/1 amp current drivers would cover every probable conversion
>>configuration". You CANNOT safely drive a single PFI with a TBI driver.
>
>I completely disagree with you here, Gar.

Funny, the guys that make integrated drivers appear to disagree with you
Walt. They make BOTH current settings for their drivers, not just the
larger one. If it were perfectly OK to drive the smaller injectors with
the TBI-size drivers, they'd not likely offer BOTH in separate packages.
Remember, when you double the hold current, you're quad(4X)rupling the
heat dissipation in the injector winding. As an EE, I'm not at liberty
to say, "no problemo" if I overdrive a device at twice it's designed max
hold-in current.

>Not all low impedance injectors are created equally.  I agree with you that
>most of the TBI injectors will be around 1.2?ish ohms, and are typically
>driven via an individual 4/1 amp current limiting driver.  However, low
>impedance port injectors are more varied.  The Rochester?s will typically be
>around 2.05 ohms, but the offerings from Bosch (and others) as an example
>can be anywhere from 2.2 to 4.6 ohms.

I mentioned before the spread of impedances you get with various
injectors. They DON'T vary that much from one injector to the next, if
they're the same part no, AND they DO fall into the two PI or TBI
categories of current requirements.

>Yes, the Rochester?s had a design drive current rating of 2/.5 amps, but
>they will also operate quite happily (all day long) at 4/1 amps, using a
>conventional off-the-shelf 4/1 amp driver.  When driven by a 4/1 amp driver,
>at a system voltage of 13.2 volts, these injectors will typically attain 4
>amps at around 2.5ms and then transition to 1 amp holding current.
>
>How do some of the ?other? low impedance port injectors fare at 13.2 system
>volts?  For example, a Bosch ?912 @ 4.6 ohms will saturate at around 2.3
>amps and a Bosch ?170 @ 2.8 ohms will saturate at around 3.5 amps.  As long
>as your driver circuit contains a default current transition timer, then it
>is possible to drive these injectors (and others which are similar) via a
>conventional 4/1 driver.  Roughly 4 ms would be a typical timer value.

I'm not anywheres near as concerned with overdriving during pull-in
(altho I wouldn't design that in casually, either) since it's
transitory, as I am with overdriving during the longer hold-in time. The
timer doesn't help that at all.

>Not an issue.  Lots of examples of such configurations in use.

Name some, please. Examples of DESIGNED configurations in use. I'm
absolutely sure that we could "get away with" overdriving a PI P&H
injector with a TBI sized driver in alot of cases. But I can't DESIGN
something that way. People who design injector drivers, to this very
day, don't lump them all together as looking like TBIs, and have done
with the problem that way. That's a short-cut alright, but it's not
considered good design practice.

Once again, I would LOVE to hear from people saying "hey, no problem
rewiring my injector harness". Never heard that until Walt said it.

Gar


----------------------------------------------------------------------------
To unsubscribe from gmecm, send "unsubscribe gmecm" (without the quotes)
in the body of a message (not the subject) to majordomo@lists.diy-efi.org