Power board for the Lattepanda based ISOBlue.

[fabi...@gmail.com]

Hi team,

I am working on the design and of a power supply board for the Lattepanda x86-based minicomputer (https://www.lattepanda.com/products/lattepanda-delta-432.html).

The main goal is to implement a system like ISOBlue on a x86-based board with an expansion board that will provide the Lattepanda with CAN and miniPCIe ports.

The features of the power board are the following:

• Over-voltage protection.
• Reverse voltage protection.
• Filtering.
• Three power rails (12V to power the main board, 5V & 3.3V for peripherals running on the expansion board).

However, I have some questions regarding the enviroment where the board is going to be deployed (Ag machinery).

1. What are the most critical electrical hazards present on the Ag machinery electrical system for a small board?
2. This board is designed to supply power to another board that has peripherals for the Lattepanda board. Is there
3. How well does the electrical system on the machines regulate power? Have you had power issues with deployed ISOBlue units?

Thank you for your attention.

Fabio A. Castiblanco.

[wweis...@gmail.com]

Hello Fabio,

Andrew asked me to comment.  I still work part-time with InDesign, a contract engineering company, and some of my customers were automotive, including Ag.  I currently volunteer in the senior design lab, when ever it picks again.

There are some general requirements for the various transients, etc., and most companies will have their own internal test requirements, usually more severe.  Usually there are different levels of passing for each test; must survive, may reset, must remain operational, etc.  Different tests will require different requirements to pass.

The following articles discuss the transients on the battery voltage.  This was a quick search to get something out:

     https://www.industrologic.com/autotransients.pdf

     http://www.ti.com/lit/ug/tiduc41a/tiduc41a.pdf

     http://www.ti.com/lit/an/slyt707/slyt707.pdf

The nominal 12V battery voltage is between 8-16V, and can go up to 18V.  Cold-crank start can sag the voltage to 4V.  Jump-start can push it up to 24V, and reverse battery down to -24V.

The worst transient tends to be load dump - the load on the battery suddenly drops, resulting in a low impedance high voltage spike, up to 125V, of fairly long duration (40 - 400ms).

Some of the other transients can be fast and/or higher impedance and capacitance can be suppressed with capacitors.

Some systems will provide a "protected voltage" where a module will block reverse polarity, and reduce some of the transients, including load dump.

I will continue to think and search on this subject.  Please let me know if you have any questions, etc.

Thanks,

Wayne

[wweis...@gmail.com]

Hello Fabio,

The ISO 16750-2 standard can be found here:

     http://www.compel.ru/wordpress/wp-content/uploads/2017/05/ISO-16750-22010E-.pdf

This gives a better, formal set of requirements.  I am use to seeing (customer requirements) some additional tests, including mutual inductance due to the cable bundles.

I find the coupling requirements here:

    https://www.smd.ru/upload/medialibrary/a3d/iso_7637_2.pdf

Most of these are emulating coupling in the cable bundle, so often the pulses are brief and handled with a capacitor.

Sorry about the ramblings, but I no longer have my old information and I am reconstructing things from memory and searching.

Thanks,

Wayne

[wweis...@gmail.com]

Hello Fabio,

One last item that came to mind.  To prevent discharging the battery when the system is off, which can be for extended periods for Ag equipment, there is often a maximum current drain for the system.  Usually a portion of this system leakage current was allocated to the section that I worked on, often in the mA range.  You will need to determine if your application will get switched battery (easier to do) or not.  Some of this is determined if your system needs time to "tidy up" at shut down, will power up when the engine is off, etc.

Thanks,

Wayne

[wweis...@gmail.com]

Hello Fabio,

My earlier post seems to be gone, likely my lack of experience with posting.

The biggest issue tends to be load dump, where the load on the system drops and a large spike up to 125V can result for up to 400ms.  Sometimes this is clamped to a lower voltage (>80V) in the system.  The voltage range can vary from 18V for a overvoltage conditions (regulator fails)  to below 5V during cold cranking start conditions.  Jump-start can go up to 24V and reverse battery to -24V.

Many of the coupling transients are at a finite impedance and short enough in duration to be handled with sufficient capacitance.

Thanks,

Wayne

[wweis...@gmail.com]

Hello Fabio,

Some of the articles that got lost are:

     https://m.littelfuse.com/~/media/electronics_technical/application_notes/varistors/littelfuse_suppression_of_transients_in_an_automotive_environment_application_note.pdf

     http://www.ti.com/jp/lit/pdf/slyt707

     http://www.ti.com/tool/TIDA-01167

     http://www.ti.com/lit/an/snva717/snva717.pdf

Thanks,

Wayne

[fabi...@gmail.com]

Hi Wayne,

Thank you so much for this information. This is going to be helpful. I am currently reading the documents and will let you know any question that I might have.

Regards,

Fabio A. Castiblanco

[wweis...@gmail.com]

Hello Fabio,

Load dump tends to be the most severe test.  ISO 16750-2 section 4.6.2 has the load dump without central suppression peak at 101V, 4ohm source impedance (or 79V and 0.5ohm) and pulse duration up to 400ms.  The several classic approaches: 1) clamp the voltage to a lower value that the system can tolerate, 2) high voltage/current transistor to make a type of linear regulator (possibly with shut-down capabilities) to regulate down to the system tolerant voltage, 3) high voltage /current transistor to open-circuit on transient (system allowed to reset during transient).  There are transient devices to clamp the voltage and the standard assumed 35V clamped peak.  I am not sure if you can assume a central clamp or not.

There are various TVS clamp devices available (some below):

     https://www.littelfuse.com/~/media/electronics/application_notes/littelfuse_tvs_diode_automotive_circuit_protection_using_automotive_tvs_diodes_application_note.pdf.pdf

     https://www.taiwansemi.com/upload/files/products/Application_Note/AN-1003_A1901.pdf

I found an example design that uses a classic high voltage P-channel device to act as a linear regulator at higher voltages and keep the load below 22.7V:

     file:///C:/Users/waw46/Desktop/PURDUE-AG%20ELECTRICAL/M139_EN-Surge_Stopper.pdf

     

This may not be what you ultimately use, but is a classic approach explains this approach well.  It may serve as a start as you work towards a solution.

Thanks,

Wayne

[wweis...@gmail.com]

Hello Fabio,

     I see that I had the wrong pointer for the article.  3 that may make the most sense:

          https://www.analog.com/media/en/technical-documentation/technical-articles/M139_EN-Surge_Stopper.pdf     Discusses load dump and has AD's approach

          https://www.mouser.com/pdfdocs/Littelfuse_TVS_Diode_Automotive_Circuit_Protection_Using_Automotive_TVS_Diodes_Application_Notepdf.pdf     Discusses Littelfuse TVS devices

          http://www.ti.com/lit/an/snva717/snva717.pdf     Discrete protection by TI

     The transient occurs at the battery, and the resistance/inductance of the cable allows the TVS to clip the peak.  The diode is needed for reverse polarity and may be replaced with a high voltage transistor ideal diode.  If your max power supply is lower than the clamped voltage, then an additional protection circuit will be required (or different supply).

Thanks,

Wayne