Gallons Per Hour Error?

[dcb]

Ok, it looks just like waveform Y in figure 2, here:
http://www.tomco-inc.com/Tech_Tips/ttt38.pdf

that lists it as 14 and 11 volts, not 5. Hmm.

If it were 14 volts then a 20k resistor in parallel with the zener should just about do it.

If it's really 4 volts and 5 volts, then maybe a 120k resistor in parallel with the zener would work.

(assuming standard zener and 100k resistor on injector lead)

[Coyote X]

I might have a bad 10x probe. I took it from the student lab since the guy who has the proper equipment was not there today. So the internal resistors could be about blown apart. I could also have a nasty ground loop throwing off my ground from the built in inverter

I need to get my own probes I just never think about it till I need them, then I don't want to wait for them to get shipped in.

[Coyote X]

Looks like the voltages are off on the probe. I hooked it directly to power and it was only reading about 5 volts. So I should just toss this probe in the junk pile when I take it back to school.

I will try a 20k tomorrow and see how it does.

[dcb]

Ok, I appreciate the effort and scoping very much too Removes much guessing and wood burning smells.

It looks like a balancing act on the actual resistor. A more refined guess would be 23500 ohms for 2.664 volts when injector is off, and 2.093 volts when in hold. And no promises if your alternator goes out But it really depends what is getting through to the CPU pin as you mentioned.

It might otherwise need some non-linear circuit that increases the spread between hold and closed.

[Coyote X]

I was thinking worst case would be it needs an op amp to drive the input. But really a pull down/voltage divider resistor should work fine. I drove it home with the scope running the whole time and the voltages never changed, so once I get a balance point above/below 2.5V it should be fine long term I figure.

Having to divide the voltage between 14 and 11V should not be that hard though really. The zener can protect the AVR from over voltages and the two resistors as a voltage divider should be all that is needed. I will try out this weekend first with a 100k pot pulldown and a lcd scope I have that will be good enough to see the voltages now that we know what we are looking for.

                                       R1
                                      ___
                to AVR ---+------+---|___|-- Inj Input
                          |      |
                         .-.     |
                         | |     z
                      R2 | |     A 5.1V Zener
                         '-'     |
                          |      |

                         GND    GND

One of these days I will quit making you add stuff to the schematic

[Coyote X]

Put a 22K resistor in for R2 in the picture and it seems to read properly now. Apparently I was only reading the very low end of the signal before because with the calibration I was using it is reporting 10gph at idle now

I will adjust the calibration now and hopefully it should be all good. I will keep the laptop showing the injector pulse width and keep track of the MPGuino for a while and make sure it is long term stable but it seems about right now.

[dcb]

Edit: This message is deprecated. I'm not happy with this solution. Original follows:


Ok, thx, that is encouraging.

Here is a picture of this very simple procedure (not a FL board) if anyone else thinks they have peak and hold and wants to try it. Please report your findings if you do. You can identify peak and hold by unplugging an injector and measuring the injector with an ohmmeter (2-7 ohms is peak and hold, 10-15 is saturation), they are in a fairly small percentage of cars these days, mostly from the late 80s and early 90s from what I can gather:





The zener in question on the spiffed board is the lower right one in this picture (I think):
kits:mpguino:04_place_220.jpg [SpiffieWiki]


Here is a source for the resistor used in this circumstance: 22K Ohm 1/4-Watt Carbon Film Resistor (5-Pack) - RadioShack.com

It could be added as an optional resistor to the schematic, not sure what to do with the pre-assembled units I'm planning on. It makes the signal on a saturated injection system less definite, and those seem to be in the majority so maybe I have to add some questions to the ordering process, and possibly add a jumper to the next revision.

[Coyote X]

If it is needed I can scope out a saturation injection system and show what voltages it is getting. I really don't think it will matter much if the resistor is on there for the saturation type systems.

Maybe trying to get the values higher would work better so the hold value is 2.3V and off is 2.8 or so would make it so that even the saturation injection systems can handle the voltage range easily and keep the system universal for all cars that way. I think with the off voltage of 2.6V it might not be accurate when the battery voltage drops if running without an alternator. So either putting a 30k mini pot on there to tune it with or trying to find the ideal single resistor would be better than having to make people figure out what they need. To keep it absolutely flexible for all circumstances I think it needs a potentiometer. Once it is adjusted and reliably reading then it will never need to be messed with again unless they want to swap it to another car.

100K-Ohm Horizontal-Style Trimmer - RadioShack.com Radio shack has a 100k trimmer. There is no need for one that high of a value but it is cheap and easy to get and will work just fine. Plus it could be turned to 100k and guarantee it won't affect a saturation injection reading since the zener would take over at that value.

[dcb]

Did you get a chance to measure peak and hold at the guino?

I don't know if there is a perfect resistor, I tried a 22k on a 3cyl metro, it worked fine. But I hate to make a revision right before shipping, that's how nightmares happen The current layout has been pretty well tested and how to explain the tuning of a trim pot escapes me at the moment (while doing a brake torque in your driveway, take this small screwdriver and... ).

I'll have to think about how to handle it. I have a stack of boards I can ill afford to make into coasters, and in most cases they won't be coasters. Implied in all of this has been user participation, so I think I want to look into addressing it on the ordering page (enter year and make, etc) and have to come up with a calibration calculation page based on year make model as well for folks to use at fillup that can maybe simplify the process and start gathering some vehicle stats. But I'm gonna take that off line and send you a PM about it.

[dcb]

I just tried watching digital 2 on my picscope and for some reason the 22k resistor seemed to allow the voltage to spike (or my cheesey picscope didn't like it for some other reason). Without it, the signal at the cpu pin was pretty clean.

Coyote (or anyone) If you get a chance with your fancy o-scope to measure the effects of the additional resistor at the CPU, I would greatly appreciate it. But at the moment I don't think we have a "peak and hold" solution.