can ms3x run a common rail diesel ?

[frank_ster]

feaseability of running a diesel engine with the ms3x. injectors running sequential injection. i don't see why it would not be possible, only concern is currant draw and speed of diesel injectors. since diesels have a very short window for firing injectors, precision would have to be very high.

reason being looing to stand alone vw diesel engines, ford and duramax.

[racingmini_mtl]

What I see as being one of the main problems is the hardware to drive the injectors. There doesn't seem to be one universal and widely available method of driving the injectors. I have seen a few generic sketches but not seen any actual circuit and confirmed testing of DIY injector drivers.

By comparison, driving gasoline (petrol) injectors is trivial.

Jean

[frank_ster]

well as everyone knows the price of gas is going up and diesels are very efficient and can be powerfull.

well is there some thing i can do or say ? i have pretty much unlimited resources. what can i do to make this happen ?

[jsmcortina]

i have pretty much unlimited resources. what can i do to make this happen ?

Send me your unlimited resources by Paypal and I'll see what I can do?? (joke)

The real trick as Jean says is to get the injectors working correctly.
I believe that some of the Ford engines use an external driver box, this was discussed before but was claimed to be unreliable.

James

[frank_ster]

link please ? what was figured out already ?

[racingmini_mtl]

A quick search turned this up: viewtopic.php?f=94&t=42338&p=297839&hilit=common+rail+injectors#p298393

Jean

[frank_ster]

well it seams for this there could be a seperate module that would be programmable it would fire 1,2,3 or 5injections by its self from one command of megasquirt.
it would aslo control the peak hold and such stuff. i found several 2.0tdi engines in montreal for about 1000$ cdn i'll pick one up and see what i can make work.
at work we often use pcb 123 for circuit boards and have a guy that has wave soldering machine. so i could slip a coupple in. i allready have my boss interested in putting a v 10 tdi into his porche. so he will be a big help.

[racingmini_mtl]

If the 1,2,3 or 5 injections need to be related to engine position then it would be better to leave these to be controlled by the MS3 CPU. Once the code has the angle clock code, it should be doable to have something like that. However, the peak&hold control should be done in external module.

By the way do common-rail diesel engines need more than one pulse per cycle or is that for emission and NVH requirements? If initial tests can be performed using a single pulse, that would simplify the process.

Jean

[frank_ster]

well i'm sure they would run fine on 1 injection per cycle,

i would think at higher rpm's the common rail motors would aproach a single long injection.



so think 100 volt 60 amp mosfets would work? i'm looking through digikey tring to find out what there cycle time is.

[frank_ster]

so most of the mosfets are 40ns switching time

VI-B0B-CX-F2-ND digikey
that is a 10-20 volt dc-dc 95 v converter 75 w. charge up a few capisitors should be able to run a few injectors expecially since there never open longer than 10 ms.

[frank_ster]

So i came up with this,

the top right is a timer that controled the resitor bypass for high and low currant. so when the pusle from mega squirt is recieved it will give a predetermined amount of time at high currant then drop down.

the left is per injector coding since this will be piezo electrick driver so its reversing currant but can be disabled.

center is reversing , after the signal from ms is droped the reverse will activate with a timer for predetermined amount of time to close injector.

[SymTech Laboratories]

You could probably do testing with a single pulse per cycle, but multiple injections are definitely something to keep in mind while designing. It would also be a good idea to think about variable voltages.

From our understanding, some systems drive piezo injectors with an H-bridge, most likely made with IGBT's. Other systems use IGBT's/MOSFET's arranged as common emitters with twice the voltage. Most designs have chokes in series with the injectors. For both systems, ou will need high-voltage gate drivers to interface the IGBT's with the controller.

Instead of using a step-up transformer, you might consider individual boost converters for each injector. The DC-DC converter you found at Digi-Key can't deliver enough current. A piezo injector is like a capacitive load, so you need brief spurts of high current (~10A) at ~90V to ~160V or more.

[frank_ster]

from reading the injectors are about 60 volts and 1.5 amps. so about 100 watt. but also 100 watt for what half milisecond ? then lower to hold position. and then realease. the overall duty cycle would be about 2% but even 10 amps at 100 volts would be a 1000 watts at 2% duty cycle so 1000 watts divided by 50 would b e about 20 watt average. so i;m sure an array of capisitors would making things happy.

thanks for the input, i will also draw up a sketch as you say, see what looks nicer!

[SymTech Laboratories]

from reading the injectors are about 60 volts and 1.5 amps. so about 100 watt. but also 100 watt for what half milisecond ? then lower to hold position. and then realease. the overall duty cycle would be about 2% but even 10 amps at 100 volts would be a 1000 watts at 2% duty cycle so 1000 watts divided by 50 would b e about 20 watt average. so i;m sure an array of capisitors would making things happy.

thanks for the input, i will also draw up a sketch as you say, see what looks nicer!

It sounds like we're talking about different things. Are you talking about standard solenoid injectors or piezo injectors? The two have very different requirements.

[frank_ster]

piezo,

from my understanding they open with a peak then hold, and then reverse to "push" closed,

but the total injection length would be about 5 ms at high loads. and probably 500 ns or less at idle.

so they are high current but for a very short amount of time. since the start and finish of the fire needs to be with in a few degrees of tdc.

[SymTech Laboratories]

piezo,

from my understanding they open with a peak then hold, and then reverse to "push" closed,

but the total injection length would be about 5 ms at high loads. and probably 500 ns or less at idle.

so they are high current but for a very short amount of time. since the start and finish of the fire needs to be with in a few degrees of tdc.

Our understanding is different (though it could be wrong): piezo injector drivers don't use anything like solenoid injector peak and hold drivers. Since the piezo elements behave like a capacitor, the current draw is naturally high before exponentially tapering off. You do, however, need to limit the rate at which the load charges and discharges (we've seen either 10A maximum draw or 180uS minimum charge/discharge time). Once the injectors are charged, they hardly require any current.

Piezo injectors also need to be driven with much higher voltages than solenoid injectors. 60V definitely isn't enough; 120V-140V seems like a more common voltage. You're right in that the pulse is usually split between two polarities; the H-bridge handles this task nicely.

[frank_ster]

well i plan on making peizo driver, that the reversing circuit can be disabled for regular type injectors.
the first engine i would test on is a volks tdi, and those are suposedly piezo.

but you do agree that the pusle is very short time wise. so a 1000 watt or 1500 watt for 50 ns or even 500 ns would net a low ultimate watt hour useage.

do you have an example h bridge ? part number i could look at ?

[subwoofer]

Half bridge drivers are plentiful, but SiliconLabs Si8234 is easy to work with and and automotive temperature range. Just make sure to add a voltage regulator zener to the high-side switch gate supply capacitor if you use a bootstrap circuit with an inductive load (not shown in any example circuits). I think I killed five or six of them before I figured out the reason why they went poof (in minutes).

[SymTech Laboratories]

This simplified circuit should get you started. You will need one of these circuits for each injector. A lot of the major semiconductor manufacturers, like Fairchild and Vishay, make suitable components for this application.

Like subwoofer noted, in reality you will probably choose an integrated half or full H-bridge. You will also probably pick a dedicated boost or flyback converter IC.

piezo_injector_driver.png (35.62 KiB) Viewed 1373 times

[subwoofer]

The dedicated ICs also provide dead-time control, so you don't have to think about the sequencing of the high side and low side switches in order to avoid shoot-through.