digmorepaka wrote: ↑
Mon Apr 12, 2021 11:17 am
… do I understand it correctly that the inaccuracy comes from not incorporating the closing time?
No, dead-time (injector latency) measured by flow-testing contains both opening and closing times. A live setup should assume using both values, as that is what is seen during real-world operation. During the pulse-width, fuel flow is reduced while waiting for the opening, but fuel is still flowing (added) while waiting for the injector to close after the end-of-signal. While the separate values are useful for other purposes; the combination of both values is how we arrive at useful "dead-time" when operating.
As a side-note, many other factors can play into latency besides voltage, such as fuel pressure or current capacity*
, so be sure to rate, graph or compare with all the same parameters. Many online references fail to include factors such as rail pressure and inadvertently ends with inaccurate results. Likewise, the injector usually flows differently at low pulse-widths (non-linear flow range), where latency values are not correct. So, be sure to stay out of that range unless intentionally testing for the variation.
Finally, and no matter how awesome the bench tests the final results are what count, and so testing latency on the running engine is always a good idea. Because I always test running latency, I only use latency references or defaults as a starting point, to be tuned in the real-world. If any running parameters are different than tested parameters, then tuning tests are a requirement. For basic tuning tasks with typical injectors, dead-time tuning makes injector bench testing unnecessary for most users.
If using a small battery or power supply, the supply voltage may read correctly when unloaded, but actual voltage and current drops when the circuit is loaded. Verify your test conditions relative to installed/running with a 'scope before relying on the results.