Changing injector pulse rate at idle?

[John T Stein]

Having lurked around this list for a while I realize that conventional wisdom
considers 1 to 1.5 milliseconds to be the minimum reasonable drive 
pulse width for injectors, and that this figure establishes the lower 
limit of the injector's dynamic range.

As part of  an effort to explain the ratty idle of my 2.5 liter TBI GM 
four, I looked at the injector drive waveform at idle and found it to be 
about 1.2 milliseconds in width, uncomfortably close to the minimum 
width for predictable operation but not surprising in light of the 
very light load at idle.

The disquieting aspect of the drive pulse was that while its duration 
was constant, the period was varying between about a value of  
about 40 milliseconds to one of about 45 milliseconds. Idle spped was 
esentially constant.

Is this a commonly-used technique to achieve an EFFECTIVE pulse 
width of less than 1 millisecond??  Is something more involved going on here??
Anyone have any similar findings (or opinions)?? 

John