Low z injectors, PWM vs Peak&Hold
Turbo Mini Forums

The first plot is from ages ago, just showing how the standard MS2 (V3.0PCB) deals with low z injectors



This is both channels of a two channel (normal siamese sequential) setup

You get one suitable length full voltage pulse (to open the injector) then a lot of short full voltage but not for long, pulses to keep it open. The short, but not very long, are the Pulse Width Modulation.


Now Peak and Hold (PandH)

But only showing one channel, plus its staged output, as I don't have a four channel scope......





Not the same pulse width because of the two different msqs, this one is showing the staged output of the first channel once the staging has kicked in (RPM in this setup).



and remember that a lot of the pulse width is the openning time set in the software so the plots aren't comparable, but the difference is very interesting....

So, ignoring the different pulse widths, what do we have......

In terms of limiting the current, PWM is VERY good as there are very little resistive losses in the drive transistors/FETs as they just switch on and off very fast and do not limit the voltage resistitavely.

For the Pand H the voltage drop is resistive in the drive transistor and equals heat to be dissapated onto the heatsink/case. The heat load would be a lot more in a normal PandH setup where the injector duty was at a normal 80% plus rather than our 20% for siamese.

My choise however, is Pand H....

No electronic "noise" to interfere with the CPU or other inputs.


A long time ago I had personal experience of the difference between PWM and linear (equivilant to PandH) drives on servo motors on a robot carrying very sensitive ultrasonic inspection equipment and we had to use linear drives to the servo motors to get rid of the PWM noise that the robot was originally provided with on its PWM drives.


But I'm running this test with a batch of wire wound resistors under my desk (plus a car battery.....) on an experimental PCB for six injectors (1 plus 2 staged per inlet) and the heat I'm getting from the driver chips and the resistors is really quite small....

Schrödinger's cat - so which one am I ???

While noise reduction is the main advantage of using linear P&H instead of PWM, there is another one when comparing the MS PWM and the LM1949 linear P&H.

The LM1949 goes from the peak phase to the hold phase using current detection so it will always have a 4A peak (as long as the battery voltage is high enough) and a 1A hold without having to tune or guess anything which is not the case for the MS PWM method. And you can use high Z injectors directly with the P&H driver without having to change anything.

Jean

http://www.jbperf.com/