work in progress Common Rail MC33816/PT2001 dev board - 4 injectors GDI

Hardware inside and outside of the ECU
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Common Rail MC33816/PT2001 dev board - 4 injectors GDI

Post by kb1gtt »

Update: now https://github.com/rusefi/rusefi/wiki/GDI4


This first post will detail what this board is and what it does. The goal is to make a board which can be added to Frankenso which will allow control over high pressure common rail injectors.

This board can control most common rail direct injected, injectors which are commonly found in automotive applications. This accepts the 12V battery, it generates the higher voltage required to open the injectors rapidly, and it controls the low side of the injector for precision open and closing times.

This board is based on the KIT33816AEEVM with SPI connector, such that we can use the USB and software for that board. However we'll have a significant start at our own board, which would be integrated into it's own ECU or what have you at a later date. Here is a snippet of the schematic.

https://rusefi.com/docs/ibom/Common_Rail_MC33816_latest.html
https://rusefi.com/docs/ibom/Common_Rail_MC33816_0.3.html

https://github.com/rusefi/rusefi/blob/master/hardware/Common_Rail_MC33816/Common_Rail_MC33816_schematic.pdf
Common_Rail_MC33816_top_iso.png
Common_Rail_MC33816_top_iso.png (96.63 KiB) Viewed 58676 times
Common_Rail_MC33816_bottom_iso.png
Common_Rail_MC33816_bottom_iso.png (72.38 KiB) Viewed 58676 times
PDF copy of schematic and source files found here
https://github.com/rusefi/rusefi/tree/master/hardware/Common_Rail_MC33816

As of 02/16/2018, there is a BOM and footprints are selected. Next step is to print 100% scale and verify the footprints would accept the chips.
Last edited by kb1gtt on Mon Feb 19, 2018 2:21 pm, edited 6 times in total.
Welcome to the friendlier side of internet crazy :)
User avatar
AndreyB
Site Admin
Posts: 14280
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: Common Rail MC33816 dev board

Post by AndreyB »

Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

Does anyone know what the Y1 clock chip is on the KIT33816AEEVM board? The Freescale datasheet has neglected to put that in the BOM. I can not read the markings on the pictures I was able to find.
Welcome to the friendlier side of internet crazy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

I have footprints and BOM information entered for most components. The ones that remain are mostly the resistors, caps and inductors for the boost circuit. Those are noted at not recommended for new designs, so I plan to change them to something more current. I also need to choose a component for that clock that's missing from the eval sheet's datasheet. However most components are selected and ready for layout.

Does anyone have a MC33816? I would like to do a 100% scale paper test before getting to far into the layout.
Welcome to the friendlier side of internet crazy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

BOM created and footprint modules are selected. Next step is to order key parts and verify the footprints match the real world parts.
Welcome to the friendlier side of internet crazy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

Some more progress. Picture added to first post.

Does anyone know why the demo board uses 3.3V for the external XTAL? I'm tempted to remove the 3.3V and make it 5V. As well I still don't have a crystal for this. I do not know the MHz which the crystal should operate.

Should I expect the SPI to work with the discovery? I think we might have a 3.3V vs 5V issue.

Any how I made progress, and I might make some more tomorrow morning.

On a side note, I rumble grown moan about how FreeScale decided to do the pin out. It seems every pin is jumbled and requires many many signals to cross over each other. I'm tempted to encourage that another chip be used. This one just seems messy. It's a real pain to layout it out properly. So I rumble groan moan about this layout.
Welcome to the friendlier side of internet crazy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

Layout done. First post updated. I'm trying to collect parts for a test bench now. If someone wants to test this let me know. It will be a while before I've collected the parts for a flow bench to be able to test it myself.
Welcome to the friendlier side of internet crazy :)
User avatar
AndreyB
Site Admin
Posts: 14280
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: Common Rail MC33816 dev board

Post by AndreyB »

What is the next step? Is anyone planning to build the flow bench with actual fuel and actual injector?

Any ideas how and who would get the board assembled? I believe one can mail components to https://www.smart-prototyping.com/PCB-Assembling.html they might be available to assemble manually. One might also ask @ to assemble but that's a risky route :)
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

I know how much @likes the small circuit boards, so I made sure to make this extra small :lol:

It is small, no doubt about it. The MC33816 is a fairly fine pitch and it has a huge solder slug in the middle. So you need a 300W iron, which can delivery very fine heat. I'm honestly a bit less than impressed by this chip.

I'm starting to shop ebay and ask around local garages about scrap injectors and scrap injector pumps. I'm not sure how long it will take me to assemble a flow bench. I also need to do the 100% scale paper test.
Welcome to the friendlier side of internet crazy :)
dist01
Posts: 24
Joined: Thu Jan 25, 2018 6:20 am
Location: Czech rebupblic

Re: Common Rail MC33816 dev board

Post by dist01 »

Realy nice I am very happy :)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

I'm still collecting parts and I haven't purchased the board yet. I also have another project that getting prioritized. So progress is slow. Let me know if you have interest or capability of getting this to a flow bench. I would priortize this more if others expressed interest in getting it done.
Welcome to the friendlier side of internet crazy :)
dist01
Posts: 24
Joined: Thu Jan 25, 2018 6:20 am
Location: Czech rebupblic

Re: Common Rail MC33816 dev board

Post by dist01 »

kb1gtt wrote:I'm still collecting parts and I haven't purchased the board yet. I also have another project that getting prioritized. So progress is slow. Let me know if you have interest or capability of getting this to a flow bench. I would priortize this more if others expressed interest in getting it done.
Hello any news?


Sent from my iPhone using Tapatalk
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

No :cry: Lots of other projects and life getting in they way.
Welcome to the friendlier side of internet crazy :)
Jghidulu
Posts: 3
Joined: Fri Feb 08, 2019 12:23 pm

Re: Common Rail MC33816 dev board

Post by Jghidulu »

Hello everyone Can somebody shed light on the purpose of Spi on mc33816?
User avatar
AndreyB
Site Admin
Posts: 14280
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: Common Rail MC33816 dev board

Post by AndreyB »

Jghidulu wrote:
Sat Feb 09, 2019 3:41 pm
Hello everyone Can somebody shed light on the purpose of Spi on mc33816?
Please confirm that you have opened MC33816 datasheet. Which pages have you tried to find the answer?

We are happy to help but a the moment we are a bit puzzled and confused if you are a troll or not :)
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

I recall in one of these threads was a link to another forum where a fellow claimed you could pull up two config pins, or pull down two config pins, which would put the chip into an INIT state. In that INIT state you the power supply works and injectors can be pulsed. However you cannot get extra data, and you do not have the watch dog features, etc.
Welcome to the friendlier side of internet crazy :)
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

Should we test this?

How about the firmware part to control it, is it much work?
User avatar
AndreyB
Site Admin
Posts: 14280
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: Common Rail MC33816 dev board

Post by AndreyB »

960 wrote:
Fri Feb 15, 2019 6:11 pm
Should we test this?
But how? We need high pressure in the rail as well.
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

russian wrote:
Fri Feb 15, 2019 6:36 pm
960 wrote:
Fri Feb 15, 2019 6:11 pm
Should we test this?
But how? We need high pressure in the rail as well.
Oh, it doesn't have the pump/rail pressure circuit?

Was looking at the PT2000/2001, and saw that was integrated.

I have all the needed data for the TFSI now, so need to get something up and running :-)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

The bottom left of this schematic has pump control. I believe that is the same circuit used on the PT2000. The key problem with knowing if this will control the pump, is that are do not have a pump to control. A control bench or an engine would resolve those kinds of details.
Welcome to the friendlier side of internet crazy :)
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

kb1gtt wrote:
Sat Feb 16, 2019 1:07 am
The bottom left of this schematic has pump control. I believe that is the same circuit used on the PT2000. The key problem with knowing if this will control the pump, is that are do not have a pump to control. A control bench or an engine would resolve those kinds of details.
Yes, I looked at the schematic now.

I dont see why it shouldn't control it.

But do you know how to control the IC?

This one should work at VAG TFSI with no problem.

Looks like this one:

https://www.nxp.com/docs/en/user-guide/KT33816UG.pdf
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

As noted in the first post of this thread, this board is based on the KIT33816AEEVM.

I do not expect this to work with the VAG style of injectors. The VAG (and other injectors similar to VW's injectors) use a higher voltage. The R0.1 version of this circuit generates a 72V rail. If this works, I expect the next spin will modify it such that we can change this voltage to a higher voltage.

I expect the R0.1 version of this board will work with Fords Eco boost, and similar series of injectors.

In the below article, it notes a voltage range of +/- 70V, whcih is a total range of about 140V.
https://www.tiepie-automotive.com/en/articles/direct-petrol-injection-voltage-current

While the injector in the below notes a voltage range up to 80V.
https://www.tiepie-automotive.com/en/articles/direct-injection-servohydraulic-voltage-current

I understand some older injectors or performance injectors can use up to 300V, but 95%+ of the injectors out there are 72V or 120V.

Also note that if you have low fuel pressure, around 50 PSI, these injectors can be opened with 12V. This is important for things like cleaning. If you have a cleaning bench which is low pressure, you can simply connect it to the injector and it will allow fluid to flow. While you could technically try to inject into a cyl during the intake stroke with a low pressure version of the drive circuit, this is a bad idea for many reasons. Keep the low pressure 12V trick to only cleaning and diagnostics. When these injectors have around the 2k PSI normal pressure, they require the higher voltage to open properly.

About SPI, that's still in the works. I believe we can make it operate without the SPI, by putting a jumper on RESETB and DRVEN, holding them in a fixed position. When done like this we loose many features like watch dog, DC-DC booster, and some diagnostics features. I saw that in a post forum post once, but I am having trouble finding it again. I believe it was pin2 DRVEN pulled up to 5V, and pin3 RESETB is pulled down. If you do that I believe you'll see the 72V for a prolonged amount of time. If I have the wrong, it's only 2 pins, all combinations can be checked fairly quickly.

In the below link, on page 16, figure 5. If RESETB stays low, you can see VBOOST is active. However notice that SPI is not ready in this configuration. If you do not have SPI coms, you do not get the (10) which activates the DC-DC converter. I believe this means you either do not make it down to the 5.5V or you cannot do fast pulses. However it would be functional with out the (10) step happening. The big thing is that you should be able to get basic operation of this chip by leaving the RESETB pin low. Then if DRVEN is high, it should allow the injectors to be pulsed.
https://cache.freescale.com/files/analog/doc/data_sheet/MC33816.pdf
Welcome to the friendlier side of internet crazy :)
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

So why does Bosch Motorsports own driver for them say 65 Volts and max 72V?

http://www.motorsport-systems.co.uk/wp-content/uploads/2016/05/Manual_HPI5_2012_02pdf.pdf


And Link ECU has made a GDI version with max boost of 65V?

This one works fine with Vw and Audi injectors.

If you open the base map for Golf GTI 2006 2.0 TFSI, the boost are set to 60 Volts.

https://www.linkecu.com/products/wirein-ecus/g4-force-gdi/
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

Also the oem driver IC in a bosch ecu I have here says 65V.

http://www.bosch-semiconductors.com/automotive-system-ics/engine-management-systems/cy335/
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

I'm ignorant about where the CY335 is used. I could be wrong about all VW's using the 120V injectors, perhaps it's only some. Perhaps only the diesels which use the high voltage injectors, and not GDI, I do not know. The MC33816 allows the upper rail voltage to be programmed, with 72V being the max.

My understanding is that the higher voltages are needed to open and close the injectors quickly. The higher the voltage, the fasted you can move the pintle. I also under stand some performance injectors pulse up to 12 times during combustion. However most injectors are 5 pulses or less. I think it is worth while to have the ability to generate higher voltages, and it is good news that it appears that many of the VW injectors use a lower voltage.

I understand the voltage is largely driven by the internal technology, and to a lesser extent, the desired performance. I understand the Piezoelectric Injector needs a higher voltage, while the electro-hydraulic actuator requires lower voltage. Then if you need fast open, you can increase the voltage to some extend, but the key difference in voltage is the general technology inside the injector.

I also understand that electro-hydraulic actuator goes by many names, some times it's Servo Hydraulic injectors, and probably other names. The naming convention doesn't seem to be very standard.

As a vague and incorrect rule of thumb, diesel typically uses the Piezoelectric and has a rail pressure as high as 2700 bar (39k PSI) more typically around 1400 bar (20k PSI). Then GDI typically uses the electro-hydraulic and typically has a rail pressure closer to 140 bar (2k PSI). I understand that there are commonly exceptions to pretty much all of that. So read this paragraph with questioning eyes.

Here's a generic picture of the two most common types of direct injectors.
dfi3.png
dfi3.png (199.7 KiB) Viewed 57368 times
This picture originally found here
https://www.autoblog.com/2008/09/08/on-the-verge-of-liquidation-delphi-intros-new-direct-piezo-dies/

Of course videos like 2011 video could be very wrong. It notes 100V to 400V rail voltage.


Any how, I think it is a good idea to plan for higher voltage, such that this circuit can drive what ever injectors come along.
Welcome to the friendlier side of internet crazy :)
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

Just to clarify, I am very sure about it.


https://www.ecstuning.com/b-genuine-volkswagen-audi-parts/fuel-injector-priced-each/06f906036a/

http://www.bosch.co.jp/tms2015/en/products/pdf/Bosch_di_folder_HDEV5.pdf


There are almost no market for diesel, so only driver for GDI are needed.
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

Seems like the PT2000 are a better option.

Finally found the differences:

https://www.nxp.com/docs/en/application-note/AN5203.pdf
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Common Rail MC33816 dev board

Post by kb1gtt »

Thanks for sharing the comparison.

I see
-- 2015 MC33816 was release to production. This could account for why the MC33816 is more readily available.
-- 2018 MC33PT2000 was released to production.
-- 201? PT2001 has not yet been released. The eval board is listed but no stock at the time of this post.
https://octopart.com/search?q=FRDMPT2001EVM

I wonder about a comparison between the PT2000 and PT2001. I see the PT2000 used a 12MHz clock, and the PT2001 has gone back to the 6 MHz clock.

It looks like the PT200# can be easily made from this existing MC33816 schematic. The key difference is that it includes another bank, so 2 more injectors. Same circuit, just a 3rd bank. Some day perhaps we'll spin a PT200# board. They are very similar, with the PT200# costing more per cyl, and adding a couple new feature like 12V battery sense. This feature on the PT200# is handy as it could free up an analog channel on our STM32.

It looks like the SPI microcode is pretty much the same, so when and if we jump to the PT200#, that effort should be the same. It also seems like a new GDI chip comes out each year. This is an emerging technology, and a changing market.
Welcome to the friendlier side of internet crazy :)
User avatar
AndreyB
Site Admin
Posts: 14280
Joined: Wed Aug 28, 2013 1:28 am
Location: Jersey City
Github Username: rusefillc
Slack: Andrey B

Re: Common Rail MC33816 dev board

Post by AndreyB »

Do we have a thread suitable to rail pressure control logic discussion? I assume this thread is about the board. Am I missing something or are we supposed to start by figuring out what logic is needed to do what exacly with high pressure pump control?
Very limited telepathic abilities - please post logs & tunes where appropriate - http://rusefi.com/s/questions

Always looking for C/C++/Java/PHP developers! Please help us see https://rusefi.com/s/howtocontribute
960
contributor
contributor
Posts: 336
Joined: Mon Dec 10, 2018 1:22 am
Location: Norway

Re: Common Rail MC33816 dev board

Post by 960 »

kb1gtt wrote:
Sat Feb 16, 2019 9:52 pm
Thanks for sharing the comparison.

I see
-- 2015 MC33816 was release to production. This could account for why the MC33816 is more readily available.
-- 2018 MC33PT2000 was released to production.
-- 201? PT2001 has not yet been released. The eval board is listed but no stock at the time of this post.
https://octopart.com/search?q=FRDMPT2001EVM

I wonder about a comparison between the PT2000 and PT2001. I see the PT2000 used a 12MHz clock, and the PT2001 has gone back to the 6 MHz clock.

It looks like the PT200# can be easily made from this existing MC33816 schematic. The key difference is that it includes another bank, so 2 more injectors. Same circuit, just a 3rd bank. Some day perhaps we'll spin a PT200# board. They are very similar, with the PT200# costing more per cyl, and adding a couple new feature like 12V battery sense. This feature on the PT200# is handy as it could free up an analog channel on our STM32.

It looks like the SPI microcode is pretty much the same, so when and if we jump to the PT200#, that effort should be the same. It also seems like a new GDI chip comes out each year. This is an emerging technology, and a changing market.
The Gerber files and schematics are actually on the site.
I tried just uploading the gerbers at JLCPCB, and It came out 47$ for 10 boards :-)

I am currently making a symbol for the PT2000.

I havent looked into what code are needed to drive it yet. I know the SPI are only needed at every startup.

I think I found a comparisation, and the main or only difference are the third bank.
Post Reply