Injector Control Options

[TonyS]

I am in the process of designing a Low-Z injector driver circuit.

Since I don't have a lot of knowledge about Low-Z injectors, I have done the usual "Google-ing" (also read the messages in this topic) and think I have a basic understanding of their operation, but I do have a few questions that I hope can be answered here.

1.
- The "Puma" board and the "Jbperf" products use the LM1949 in a standard linear mode (not PWM).
- The latest Mega products (through software tuning parameters) drive an IRFIZ34N in either saturated or pwm mode and have a two level "flyback" clamping mechanism to clamp to either BAT or ~36V (controlled through firmware).
Is my understanding of these products correct?

2. I think I understand the reasoning behind the Mega dual level clamp. Low Vclamp during PWM (longer injector current decay time during "off" time) and high Vclamp at the end of injection cycle to quickly close the injector.
Is my understanding correct?

3. Does the Mega firmware apply the high Vclamp during the ~4A to 1A transition, in other words, is it important to decay from the 4A level to the 1A level as quickly as possible?

Thanks

[jharvey]

- The "Puma" board and the "Jbperf" products use the LM1949 in a standard linear mode (not PWM).

Correct.

- The latest Mega products (through software tuning parameters) drive an IRFIZ34N in either saturated or pwm mode and have a two level "flyback" clamping mechanism to clamp to either BAT or ~36V (controlled through firmware).

I know the PWM approach has been talked about here before. There are pro's and con's. When done via software driven PWM, you rely heavily on not having hard to diagnose and find bugs. This can make the injector stability a bear to work with, and has a hight potentail for failure. When done with an external dedicated processor, it simplifys the process such that it's more likely to produce a reliable approach. I believe there are also some chips that nearly do the PWM thing, but they have some tunning concerns. Look for Delta, I seem to recall he posted a bit about it a while back.

[Fred]

1&2) Hey Huff, yeah, I think you have it spot on, and the circuit is good, BUT, it requires LOTS of processor resources, which is hugely counter productive overall. I imagine you could do something like the LM1949 with less specialised external components, and if you come up with one, and have permission to put it into the public domain, then please do share it here.

3) The only reason that it's required to have a hold current lower is to avoid heat generation in the injector while maintaining a rapid opening time. I can't see any reason why it would matter if the decay back to 1/4 current wasn't perfectly fast. The chances of burning out an injector this way have to be slim to none.

Fred.

[TonyS]

Thanks for verifying that my understanding of existing "designs" is correct.

jharvey - I have searched for a more modern alternative to the LM1949 (with / without PWM capability) but could not find any. I just don't think there is a meaningful market for IC manufacturers to develop new "peak-hold" injector drivers. From what I have read, most all new vehicles use high impedance injectors. Also, I did look at the design presented by Delta, and don't see it as a meaningful improvement over what you can / cannot do with the LM1949 configured for PWM mode.

Fred - I do like that the Mega provides a flexible way of dealing with most any injector type (even if it does take a lot of uC resources). One of the issues with the Delta / LM1949 PWM is that you would have to "tune" the hardware to work well with the specific characteristics of the injector being used.

I'm not sure I understand the "specialized external components" comment in your reference to the LM1949. Could you elaborate?

My question about the injector 4 -> 1A, current decay slope, was a bit of mis-thinking on my part. I was wondering about short injector "on" times and the effect on an injector that was in "slow" current decay mode on it's way to it's "hold" current state. Actually doesn't matter where it's at, when injection is done - clamp at Vhigh for quickest current decay.

I will ask Sean about design contributions to this site generally speaking, but I don't think that we are on the same path anyway as far as low-Z injector control goes as we are going the uC - PWM route.

-Huff

[jharvey]

I agree that I haven't come across a chip that does quite what's wanted. Could you be become interested in making a kind of sub MCU that can run an injector? Perhaps an ARM based PSoC? If so I'd be happy to brain storm some of the items I would see as beneficial with such a device.

One thing perhaps worth noting about fast and precise control of an injector, is that when you use the snubber diode, your decay is exponential, and the pin shuts off fuel in the low part of the your current decay. So with a 4 amp injector, lets say the pin shuts at .2 amps. The decay will have a vertical tolerance bracket, for talking purposes lets say .01 amps. The flat nature of the exponential decay at this point would cause the .2 amps limit to stretch over a reasonably large time range.

Typically that's not a real problem, you can simply tune it to run rich, and dump some out the pipe. However, if you're looking for precise control, I'd recommend an Over Voltage protected device. Those devices decay in a near linear fashion, not exponentially, so your horizontal bracket is shorted at this point of the decay curve, allowing for a more precise control.

Another note, Jbperf's board uses a 4 amp drive silicon. He does this to help limit the 4 amps a bit more naturally. It's a kind of interesting tuning approach. Some will try a 20 amp drive which causes some mild issue to keep in mind. So as an FYI, keep an eye out for the current range on those designs.

[Fred]

Huff, I just meant the LM1949 itself, really. IE, it's fairly expensive and uncommon. If we could do the same thing with opamps and/or other components, then great. If you're not doing such a thing, then don't bother asking Sean about it, there is no point.

I agree that I haven't come across a chip that does quite what's wanted. Could you be become interested in making a kind of sub MCU that can run an injector? Perhaps an ARM based PSoC? If so I'd be happy to brain storm some of the items I would see as beneficial with such a device.

An excellent idea. I wonder if we could find something that had the pin count and speed and resources to handle 8 channels of CPU driven P&H for a low price. The firmware to do such control would be trivial to write. It could even be programmable with on/hold times and PWM levels, or use some sensing setup or whatever. Nice!

Fred.

[EssEss]

Infineon has some Tricore app notes on doing this exact same thing here. There are also some direct injection appnotes (Evinrude/E-TEC ?) that outlined some of the same control methods in detail.

fpga's are also a very cheap/dedicated method for doing the same thing.

[Fred]

From what I have read, most all new vehicles use high impedance injectors.

Yeah, I was going to say, that's not true at all. Most low pressure injected vehicles use high Z injectors, low Z is rare and for hoons like me :-) However most new cars don't use low pressure injectors, they have DI and I didn't know how that worked until reading that document linked by EssEss, BUT, it doesn't surprise me at all. I knew they were high voltage, and it figures that they have a hold current seperate from the opening one.

Fred.

[TonyS]

jharvey - To summarize my understanding of your comments about injector current decay at shutoff - Yes, clamp inductor "flyback" at a high voltage when turning off the injector to dissipate the energy in the inductor as quickly as possible.

Another note, JPerf's board uses a 4 amp drive silicone. He does this to help limit the 4 amps a bit more naturally. It's a kind of interesting tuning approach. Some will try a 20 amp drive which causes some mild issue to keep in mind. So as an FYI, keep an eye out for the current range on those designs.

Not sure I understand. Can you elaborate?

EssEss - Thank you for the app note link. It was very informative.

Fred - Had no idea that the LM1949 was that pricey until I looked it up.
Also, Sean did say that I was free to contribute to this forum while using good judgement.

All - Since there appears to some interest in developing a Low-Z Injector board, should a new message thread be started? My initial thoughts hardware wise, would be to just use the cheapest Freescale S12 part that will do the job (S12 only because you are already working with the family). Use a protected MOSFET (to provide the high Vclamp at turn-off) and use something like the active clamp to BAT circuit that Mega uses to moderate the current decay in the inductor during PWM. I know, not much of a "spec" but maybe a starting point for discussion.

- Huff (by the way Fred, what's a "hoon"?

[jharvey]

Clamp good, snubber diode/device bad. Flyback terminoligy, common in MS land but shows a mechanic vs an engineer. Snubber is a better term.

I'd have to refresh myself about the 4 amp vs 20 amp drive silicon thing. I seem to recall that putting in a larger 20 amp drive, wold cause the driver to absorbe more energy or something like that. I'd have to look over the circuit again to refresh my memory. The big thing is to keep an eye on it when the time comes to choosing the power drive silicon. My feeling was that JPerf's 4 amp drive was a slightly better match than a 20 amp drive.

Several other chips can be programed with out specail hardware, so keeping to the S12 might not be a big deal. Perhaps that should be brought up for discussion. Perhaps the best thing is to ignore the MCU boundries and make a list of desired features, then let the technologies fill the voids.