Injector Control Options

[AbeFM]

Pretty awesome, thanks for asking. They only left a couple days back, and I've had all sorts of other weird stuff going on (including my first 2-3cm of actual beaded weld!), so I'm going to try to do this test now.

I also managed to find some power resistors which should be good for the job. I guess I'll go down and build it now - awkward, since I have ~2 weeks of work to do in 4 days here at work, and yet everyone's so busy I can't get the information I need to start.

[AbeFM]

Ah, this is a thread which I've let drop too much. I'd had some fuel system issues I'd wanted to get sorted in the past, but now seems a good time to pick this up once again. Certainly a circuit to try would be handy since there's a good chance I'll have some injectors to run in the near future.

Does anyone remember where we were at? It might have been as simple as my needing to heatsink the parts from the last circuit, I think I was waiting for feedback on the redesign verses heatsink issue.

[Fred]

Careful, if you look the other way for long enough, I may steal your FET ;-)

[AbeFM]

http://search.digikey.com/scripts/DkSea ... D16330P-ND

10 spares of these say you don't have to.


edit: Hmmm. 5A max? (15V / 16ohm) * 4 = 3.75 amps. Guess we should be ok.
edit 2: May have a spare of the other part, I'm not sure you need the protection. At least you'd have a spare.

[jharvey]

Just wanted to toss these chips in the ring. For low side drive,

http://octopart.com/parts/search?q=IPS1031SPbF&js=on

And for high side drive w/ current feed back.

http://octopart.com/parts/search?q=IPS6031&js=on

Some key extras include D2Pak and TO220 packages, so SMT it possible. Also this series can go up to 60 amps high side drive, and 85 amps low side drive, so direct drive of low frequency items like the fuel pump, ect is possible.

As a recap, I think the two key players at this point are VNP20N07 (or VNP5N07) and TIP122.

The VNP's have an Rds around .5 ohms for the 5 amp, and .05 ohms for the 20 amp. The IPS1031 has an Rds around .05 ohms, so low heat dissipation by the drive silicone. At 5 amps that's 1.25 watts, and at 1 amp it's .05 watts. In hold if used as peak and hold, it's around 2.25 watts max. Should allow for peak and hold with minimal heat sink requirements.

The IPS6031 has an Rds of .06, and includes current feed back, as a high side drive. This may be handy for measuring current in the hold state. Any thoughts about doing P&H on the high side?

[EssEss]

Any thoughts about doing P&H on the high side?

whats the application ?

[jharvey]

The application, I guess any fuel injection application. Either small engine, or large engine. Is that what you mean?

I know fuel is typically low side drive, but if done via high side drive, we can intergrate current sensing in the driving silicone, for use with a P&H function. This would decrease the number of parts. Less parts often means less breakage right, so perhaps a bit more along the KISS approach. I don't see much of a reason why it can't be high side drive, so perhaps this would work out better than a more traditional approach.

For the high current low frequency applictions, the current sense isn't really required, but could be handy for diag.

[EssEss]

typically (not all cases) :
1) high side switching is slower than low-side
2) current sensing is more expensive/involved

the 'pain' of all the above is relative to your application, which is why I asked first. (typical) high-side things driven from an ecu are mostly indicators/lamps/lighting which can tolerate low-freq (100's of hz) and crude current detect (simple open/short detection - NOT magnitude).

so against the application of driving injectors, I vote skip it since there are cheaper/simpler alternatives which achieve the result cleaner/better than high-side drive.

[Fred]

Such a detailed response! I was tending towards a response like this :

Any thoughts about doing P&H on the high side?

My thoughts are: NO.

:-)

[jharvey]

Thank you EssEss for clarifying. I see what you mean about slower switching times.

If we figure a max, RPM around 12,000RPM, it would correlate to 200 rotations per second. If we plan for one injector pulse per revolution, we have a min period of 5mS. If we plan for sequential injection, the max time we have to inject is around half of that. So a max pulse width of 2.5mS for 50% duty. If we want that pulse to be accurate to say 1%, we need the rise and fall to be less than 25uS. Other delays will of course have to be calculated out, but should be fairly constant. So I'll ignore them for this approximation.

[edit] If 4 stroke, the period is 10mS with max potential window of 2.5mS at up to 25% duty.[/edit]

I see the IPS1031 has a rise time of 20uS, however the tolerance goes up to 40uS, so it's out side of that ball park range. Also the IPS6031 is normally around 20uS, but under certain conditions can get up to 35uS. So also out side of that range again. Those chips really are pushing the slow side, and a faster chip would be better.

I see the VNP20N07 has a rise time of 165nS, while the VNP5N07 has a min rise of 50nS. These faster rise times will decrease the heat dissipated by the drive silicone during the transition periods. However, with a 20uS rise time, the transition time is around 0.8% of the cycle, so transition heating would be minimal, with any of the above noted silicon at full RPM.

It would appear we want a min on time of less than .2uS, but would likely be OK with just about anything below 25uS.

Does that range sound about right as a general guide line for speed?