First drive with VP 44 project

[Great work!]

First drive today no psg or ECM. Using my own.  See it on YouTube 

 

[Tractorman]

That is one impressive video!!!

 

- John

[wil440]

WOW

[Mopar1973Man]

Very impressive to build both a PSG and ECM. 

[W-T]

GW I've always wondered what the clock speed is on the factory ECM and seeing your board, I do see the crystal at the 1:09 time, it is marked 8.000; is this 8Mhz as a reference and is this fundamental or are you deriving or extracting a multiple clocking speed?

 

At 14.3 volts @ 7 amps I see your heat sink and fan...I'm assuming the driver(s) are being heat dissipated here....I must ask at 107 watts how warm is that heat sink after an extended period of run time. 

 

This is impressive and I have many questions. This is very exciting!

 

W-T

[Mopar1973Man]

Biggest challenge is creating CCD bud and sending the gauge data back on the bus.

 

Then creating error codes for out of range stuff 

[Great work!]

@Mopar1973Man CCD will be coming along I have to decide if it will use the current ST micro or a Mega. The ST while vastly more powerful is limited to 3.3volt inputs which makes 5volt sensor interface a pain especially the high impedance temperature sensors.

 

@W-T the factory ecm has multiple Crystals an EEPROM scratch pad and a large flash memory chip I suspect one of those has failed. The stm32f303 module has 8mhz crystal pll'd up to 72 mhz clock speed. The power dissipation on the bosch unit is insane it's linear regulation plus high wattage 2ohm resistor cooled by fuel. The inj. Solenoid is less than 1 ohm, quite a challenge to control. Mine uses a high speed switching regulator, much more efficient but still makes some heat. I may not be able to mount it under the hood because of high temps.

Edited July 5, 2022 by Great work!
Clairty

[W-T]

@Great work!yeah, I was thinking the CPU would be clocking faster than 8MHz to lengthen the sample rate bits. Linear regulation is wasteful and kudos on transferring to switching but, that spawns another curiosity, is it quiet? I work with a Hewlett Packard 8921A and I've eliminated several switching supplies from my equipment nest due to noise in the spectrum. This test instrument is good to 1 Gig and sensitivity is extreme.

You mentioned the cost of the multi-pin connectors and I had not done any looking around Mouser or DigiKey , so I do not know who manufactured them (Molex?) and the reason was, I wanted to transfer the location of the ECU. I find the decision to bolt that damn thing to the engine block absolutely INSANE! Truly, it's a poor location for so many reasons... so, I thought I'd hunt down the connectors and create a high quality "relocation" harness. Your endeavor is beyond my thoughts however, I'd prefer to have a controller of this nature in a much friendlier environment. Just my two cents...

I hadn't even contemplated the PSG so, again I'm following your work closely because of the nobility in nature.

Yeah, I had to show off my HP....you know these sold new for 25K ....it's one of the reasons the last wife left.

 

W-T 

[Great work!]

@W-Tthe micro runs at 72 mhz internally and sampling rate is programmable. I set conversion time to 25 microS. The switching frequency is around 150 khz. Most of the noisy stuff is the huge fly back spikes when the sol. Is de energized (you can see it in the sensors) and with its low resistance and high current even the clamp circuit gets hot.

The current has small blips that occur when the plunger moves, they must be detected in order to calculate the dead time from energized to movement and vice versa on release. To much noise hides the blips. The driver circuit has to be fast and the fly back at the end must extinguish rapidly,  mine is clamped at 51 V. I think bosch took the other approach for speed and less noise, they even used the main driver in linear mode and active clamping.

That's living on the edge.

Don't bother on relocating. The ECM is pretty tough and that connector is only made for through hole pcb mounting, no wire crimping. So that would probably create more points of failure. 

[W-T]

13 hours ago, Great work! said:

The driver circuit has to be fast and the fly back at the end must extinguish rapidly,  mine is clamped at 51 V

Again @Great work!you're accomplished in this endeavor. Observing the limited scope reference of active duration for the "event" of opening and closing the plunger does depict the extremes of overshoot. The characteristics of any lumped inductance (solenoid coil) with substantial "windings count" would account for many micro-henries storing the DC charge in the surrounding electromagnetic field of force and collapsing into itself once the DC source is removed. A back-EMF, is what I recall from studies, ( I'm not trying to school you here) your clamping at 51 volts is my question? Not seeing an actual schematic have you incorporated an additional snubber circuit to soften or curtail this 51 volt spike? Again, please forgive my curiosity, I have the most outright respect for what you're creating here.

 

As for what BOSCH has in it's original design layout, I've not even looked at it. You've opened my eyes to a heck of a lot and I'm barely worth the dirt under your finger nails at this point however; my sophomoric observations are just that. In the past, I've been quite assertive in simple aspects of basic DC and limited AC theory to assist fellow members to avoid or eliminate electrical noise that was introduced via factory design. These little faux pas cause errors elsewhere in these trucks. 

 

Onward...the original question pertaining to the switching supply regarding noise generation was my first thought, now you've also brought my attention to the internal solenoids of the VP-44...that 51 volt spike collapsing causes me concern from my humble view point. Not so much for noise but, for the aspect of longevity for proximity associated circuitry.

 

I confess to having never done a direct panoramic spectrum analysis of the VP-44 in it's operation just to see what it's generating from DC to daylight. I'll share my observations with you once done, perhaps it maybe of value.

 

P.S. I'd like to volunteer as a test subject in beta for what you might be producing. 

 

Thank you,

W-T   

   

[Great work!]

The back EMF or free wheeling I sometimes call fly back dates back to the old TV lingo. It can be delt with recirculating, snubber or clamp or a mix. Recirculating diode was about 4X slower and a snubber became really large to be effective and about 2X slower. So I used a zener - resistor clamp with a small amount of capacitance. The back EMF can be low voltage over a long time or a high voltage over a short time. The solenoid valve must respond in both directions as fast as possible to have a good clean and accurate injection. Straight away high frequency PWM was tried but resulted in high noise, skin effect and large eddy currents in the plunger. 

So I used a buck converter with minimal capacitors for fast response while balancing low ripple. The timing control valve is more conventional 10 ohms and high winding count so a snubber was used there.

 

To be less technical think a cordless drill  suddenly stopping when the trigger is released. That freezing of the armature is an example of extreme back EMF clamping ( a dead short across the motor terminals) we do not want that. We want the solenoid loosely clamped so it will "coast" easily. 

[W-T]

20 hours ago, Great work! said:

I used a zener - resistor clamp with a small amount of capacitance.

GW, is it possible to view a schematic of this?

 

I'm guessing this portion of the component count is being applied directly at the solenoid terminals or very close to it. I'm trying to picture the configuration and I'm guessing this is occurring at the top of the VP-44 as a modification directly to the PSG unit.

 

Using a Zener and selecting it's value to accommodate the desired voltage at the knee of avalanche is a common practice however; applying Zener dynamics to eliminate the reverse EMF of a collapsing field is unorthodox. I would appreciate learning how that looks in schematic form and how this actually works?

 

   

[Great work!]

W-T I'm sorry i think i misunderstood your question. I have a simplified hand drawn drawing that may clear things up. I not really doing anything un -conventional, except the low side switch buck converter. BTW the solenoid is only about 30 turns (7 feet) of 25 ga. magnet wire so the inductance is comparatively low. 

Inductance delays current, but current is what moves the plunger. So it was designed with low inductance to get things moving fast. Unfortunately that makes it very hungry for current. 

 

[W-T]

GW, I've asked enough of you and I do not wish to drag you through the mud. This is a remarkable project you've undertaken and I can only assume you're developing this for self satisfaction?

I appreciate the effort in hope that we CTD enthusiasts may benefit with an added development and avoid obsolescence or discontinued critical control devices.

There is no telling how long our platforms can be supported when companies discontinue supplying or servicing critical modules, @Auto Computer Specialistis certainly an answer to prolong the desire to operate these vehicles and I'm grateful that such facilities exist. Yeah, old school P7100 systems do work but, the advanced drive-ability of the modern electronic controlled VP-44 is superior in many aspects. 

I look forward to seeing further development, seeing a 'working beta platform' actually working(!) is exciting, you needn't have a nerd nit-picking circuit aspects that can be addressed when necessary. 

I truly appreciate what you're doing!

 

W-T

   

[wil440]

As @W-T says this is a major achievement and all credit to you @Great work! I hope that you can figure out how to put all your work into a plug and play box. Put me down for one.  

I can fix rust, and anything mechanical and I'm good at it too but I cannot do what you have done.  

Get it into a P&P box and market  it

[Mopar1973Man]

Biggest thing that qould capture my attention to remove all ECM limitation for fuel and boost mapping. If it could be expanded to say 100 PSI of boost and create a fuel map that would cover that wide boost table.

[Hutch24v]

Holy s@*! I’m a degreed engineer but I seriously need to go back to school to follow along with this project. Love the innovation.  

[Great work!]

Thank you all. The plug and play solution is not going to be easy. The Duetch ECM connector is not available in a wire friendly format so a pcb would be required to convert it to another form. The mating connector is but that's already on the harness. Second the connection to the vp speed sensor is delicate. It would probably take a new VP top cover to do it right which would require machine work. We would still be at the mercy of Bosch for the sensor although one could be made with a 3D printer, still thinking about it. The solenoid wires are not very long so putting plugs on them is not easy. As for massive fueling the EPA has been cracking down hard on the diesel tuner makers "tunes" is a bad word now, granted are trucks are not subject to the stricter standards as the newer ones but the EPA is highly politicized. 

[JAG1]

Just for the sake of discussion, questions coming from my misunderstanding... lets say the Gov't mandates that our second gens ECM/ PCM can no longer be manufactured nor remanufactured. Is there a simpler method perhaps by allowing manual control on timing and other automated electronic controls needed to keep it running down the road, or am I in the dark here? Thank you a lot for your work and W-T's input.

[Mopar1973Man]

1 hour ago, JAG1 said:

Just for the sake of discussion, questions coming from my misunderstanding... lets say the Gov't mandates that our second gens ECM/ PCM can no longer be manufactured nor remanufactured. Is there a simpler method perhaps by allowing manual control on timing and other automated electronic controls needed to keep it running down the road, or am I in the dark here? Thank you a lot for your work and W-T's input.

Basically comes down to @Great work! ability to make the software rich enough to allow easy reprogramming of the timing and fuel tables.