Injector Control Options

[Delta]

So to be CLEAR, please fill this out:

v1= injector signal voltage - TTL levels so 0-5v: 0v = injector off, 5v = injector on

v2= Regulated 5V

v3= Battery Voltage

The opamp uses +/- what voltage? Your using an auto fet to drive the injector - so either will be fine - I'd like to see it on regulated 5V just for testing but whatever.

The dual FET chip doesn't take power? No

....All grounds are common, and all listed voltages in reference to that same ground? Yes

....and the drive signal comes in where? V1 is the driving signal.

I guess I made more assumptions in building that from the drawing than I realized. I powered the opamp off +12V (for more swing), too.
These things happen - I should have written better instructions. Even after a few years teaching I still forget that other people won't necessarily read my diagrams and get out of them what I put in hahahha. Best of luck for round two.

[AbeFM]

Awesome, thanks. It's what I got in words from the earlier posts, but now there's a handy reference. Certainly the V1-V2-V3 thing was open to some interpretation, and the op-amp being 0-5V obviously was since you've already changed your mind on it. Not saying anything bad - I was having a similar discussion with a friend who's company had stuff made by a subcontractor. They worked for 6 months from nothing but drawings, and SURPRISE there were misinterpretations.

Ok, there's a chance I have stuff going on tonight I can't get out of, but otherwise I'll rewire up this circuit and give it another go tonight or tomorrow.

[AbeFM]

http://abefm.smugmug.com/gallery/592505 ... 6623_wNaas

No time for explanation yet. It seems to work, but the FET gets hot very quick! I was running something like 3 seconds on, and 4-5 seconds rate... Anyway, the resistor was ~60*C on the side, and the FET - about the time it started shutting off at a 50% duty cycle, but say a 2 second cycles... Anyway, the back of it was at ~130*C

Pics above, show how it was opperating. The first batch are with oscope probes reversed relative to the later. Oscope probes on channels 3, 4, with aligator clips on channels 1, 2. Um, when you see 1 pic with four traces, then 4 pics in a row with 1 channel each, that's just me showing each without moving any of the scales.

[jharvey]

Looks like some of those pictures are spot on with the spice. I didn't see a picture that indicated a switching frequency, which also makes sense that the FET gets quite hot.

Keep it up. That proto is a good chunk of knowledge.

[Delta]

Not surprising it gets very very hot without any heatsink hahaha. The thermal resistance of case to ambient is 62.5 degC/W according to the datasheet - so even 2W continuous will put the temp up to 125deg above ambient.

Did you try it with high Z as well to compare and see how well it will run normal injectors?

[AbeFM]

Oooo - no, I didn't try it with high Z. I did notice (in retrospect) that I got a little different on times with the fuel pump on.

At least the high Z shouldn't be hard to try at all.Just hang one injector off the end.

Do the currents look right? I didn't even bother checking them, I just wanted to get some clean traces for you guys and get on with some other stuff!

What I found odd, shape wise, was the linear increase in current, I'm not sure if it's enough. I got fuel flow, so maybe it is. I'm trying to remember with my MS P&H testing, I think the car really didn't run well unless I left it on for a while. I don't think those are the fastest injectors (some volvo 300CC greentops). I have my set of lancer evo injectors I haven't tried yet on this test. Same resistance, but I think they open faster.

What's a reasonable heatsink to put on there to do some steady state testing? I don't want something that's going to over represent what will be acheivable in a realistic freeEMS case.

Er... Is there some componant swap which would make it switch more/faster? Guess I better go view the traces myself at this point. Anyway, any comments on testing, or suggestions, will be appreciated

[AbeFM]

Also - how come the FET gate never really turns off? And the current through the injector never really goes to zero? Maybe it's that 100 ohm resistor. Maybe... I should read the thread.

Anyway, it looks like 9 amps peak, cuts down to 6. Isn't six a bit much for a holding current?! I think you could go to 4 without issue, and it would mean less heat dumped? I can't find my pictures for the currents through my lowZ injectors.. Too bad. Perhaps I could remeasure them, though... Put a small cap across a 10 mohm inline resistor and see what I get?

[Fred]

Earlier in the thread it was stated that it is roughly 4:1 peak to hold current ratio. I confirmed it with RC engineering who said they have seen as high as 8/2.

Use a piece of 1"x1/4"x6" flat bar only don't use it that long, make it 2" instead and see how you go. (ally)

[baldur]

Let me rephrase my initial thoughts posted earlier in the thread, being a little more detailed this time.
The thing here is that you don't really have any control over the PWM frequency, not without extending the time each switchon/switchoff takes.
If the PWM frequency is too high the FET will spend more time in transients where it heats up rather than fully on/off where it stays cold.
Still more efficient than a linear current limiter but less efficient than a clocked PWM with a high speed FET driver.
Possibly the circuit could be improved by adding a stronger full-bridge driver (perhaps a 33152 between comparator and FET or some comparator with better load driving characteristics) because the comparator is a very weak driver (33152 will source/sink 1.5A while LT1215 comparator will only do 50mA).
The reason the circuit oscillates is because of the characteristics of the comparator when driving a capacitive load, it'll overshoot the target because the capacitance causes a current/voltage phase lead in the output.

[jharvey]

Hmmm, perhaps I missed something, I didn't see a clear PWM at least not below the injector on/off time. If the injector is running at 1KHz, I'd expect to see a switching frequency at the FET gate of at least a decade higher than that. I didn't see that in the pictures I looked at. Perhaps I missed some? What I saw looked like it was fairly linear and not oscillating.