[Jays_Truck]

francois, My design is a copy of ropecope's pcb design that you can find on page 287 of the AgOpenGPS thread. He has been a great help by supplying his cad files. I have attached a very cude schematic of the regulation side of my board. This is mostly in block diagram form so you will still need to dig into the datasheets to see how each component is terminated and the sizes and types of coupling caps to use.

Last year I only used prototype PCB boards and termination bars to connect the power/ground and it was a nightmare. There is a learning curve to building your own PCB, but overall it is so much easier and nicer looking too. Plus you don't have to worry about a random wire coming loose and causing issues. I am not ready to share my cad files just yet since I have not sent it out for manufacturing and with the Chinese new year coming up I probably wont have the PCB in my possession until late February to test out and make sure there are no unexpected oscillations. With the high gain of the TIP112 I was having bad high frequency ossications, 200mV P to P at 1 MHz, until I biased the transistor correctly.

 

[LarsVest]

francois, My design is a copy of ropecope's pcb design that you can find on page 287 of the AgOpenGPS thread. He has been a great help by supplying his cad files. I have attached a very cude schematic of the regulation side of my board. This is mostly in block diagram form so you will still need to dig into the datasheets to see how each component is terminated and the sizes and types of coupling caps to use.

Last year I only used prototype PCB boards and termination bars to connect the power/ground and it was a nightmare. There is a learning curve to building your own PCB, but overall it is so much easier and nicer looking too. Plus you don't have to worry about a random wire coming loose and causing issues. I am not ready to share my cad files just yet since I have not sent it out for manufacturing and with the Chinese new year coming up I probably wont have the PCB in my possession until late February to test out and make sure there are no unexpected oscillations. With the high gain of the TIP112 I was having bad high frequency ossications, 200mV P to P at 1 MHz, until I biased the transistor correctly.

I don´t understand the function of the TIP112 just before the 5v regulator? Can you explain.

 

[WTalen]

I am building a power supply PCB board with Power In, Power Out (Drive Motor), 8V, 5V, 3.3V and a filtered 5V output to sourced to the 16 bit ADC and source voltage for the WAS.

The attached screenshot of my oscilloscope shows three traces, the Yellow trace is the output of the DC-DC buck converter converting from 13V to 8V, the Blue trace is the output of my Low Drop Out (LDO) 5V regulator, and the Pink trace is the output after my low pass filter (capacitance multiplier) circuit (6.7V).

The vertical scale is set to 50.0mV per division for all three traces. As you can see the output of the DC-DC converter adds a lot of switching noise, almost 100 mV peak to peak. For any analog to digital converter, 10 bit on Arduino or a 16 bit ADC this would be troublesome. Putting the output through a 5V regulator does reduce the noise quite a bit but it is still in the 30 mV range, which may be completely acceptable for most applications. I wanted to reduce the noise even further so I put the output of the DC-DC converted through an "Active" low pass filter. This virtually eliminated all of the switching noise from the DC-DC converter. The input voltage to my filter is 8V while the output was at 6.7V for my given load. Since I will only be driving my 16 ADC and WAS with this source I will still have plenty of head room for my LDO 5V regulator to give a rock solid 5V.

My "active" filter consist of 1 x 56 Ohm resistor, 1 x 47uF cap, and 1 x tip112 darlington pair transistor. Your mileage may vary but this setup should work for my application.

Your knowledge is impressive! Pardon my lack of knowledge here, I'm just trying to learn what is going on... What is the 8V for for in this power supply? Also on a more basic level would the intention of this PCB power supply be to hook straight to the battery (with appropriate fuses of course)? Should we have a power output for charging the tablet/laptop as well? Has anyone traced or figured out how much current is is being drawn with all of these devices combined?

My guess is that we would need a fairly heavy gauge wire in to this power supply board if we are powering these high current items like electric motors/H-bridges etc. My knowledge of PCB manufacturing is basically nil, but is it common to have power supply PCBs with higher current devices?

 

[BrianTee]

francois, My design is a copy of ropecope's pcb design that you can find on page 287 of the AgOpenGPS thread. He has been a great help by supplying his cad files. I have attached a very cude schematic of the regulation side of my board. This is mostly in block diagram form so you will still need to dig into the datasheets to see how each component is terminated and the sizes and types of coupling caps to use.

Last year I only used prototype PCB boards and termination bars to connect the power/ground and it was a nightmare. There is a learning curve to building your own PCB, but overall it is so much easier and nicer looking too. Plus you don't have to worry about a random wire coming loose and causing issues. I am not ready to share my cad files just yet since I have not sent it out for manufacturing and with the Chinese new year coming up I probably wont have the PCB in my possession until late February to test out and make sure there are no unexpected oscillations. With the high gain of the TIP112 I was having bad high frequency ossications, 200mV P to P at 1 MHz, until I biased the transistor correctly.

Why not just use a 7805 5v linear reg right off the +12v? Input output caps, simple. The trouble with the 7805 and also the TIP circuit is temperature drift. If we are going the route of separate and filtered supply for A/D etc, we should then maybe just use a proper reference voltage source, but maybe that is overkill, the 7805 is pretty good all by itself.

 

[doppelgrau]

Looks great, but what about heatsink dimension for the motor drivers? The pcb cooling design (page 27 here https://www.st.com/content/ccc/reso.../DM00213876.pdf/files/DM00213876.pdf/jcr:content/translations/en.DM00213876.pdf ) doesn´t look of much compared to the big ribs on the IBT_2. IBT_2 goes to 43 amps, and the VNH7070ASTR goes to 15 amps each.

Thanks for the reminder to put a bit more copper down there.
Besides that I hope that there won't be a thermal issue.

Reasoning:
A typical motor like the Phidges draw between 2A (rated) and 11A (stall). Even if I assume 4A on average that means in a "bridged" configuration 2A for each VNH7070A, less than 15% of the rated current.
And if I understood the datasheet correct, the internal resistance is about 100 mOhm (total), so with 2A going through the chip, with P=RxI² => 0.1 X 4 => 0.4W amount of heat to be dissipated for each output, times two since both direction go through the chip.


Less than 1W of heat sounds possible to me, especially taking into account, that the steering motor won't run all the time.

 

[Jays_Truck]

LarsVest - The TIP112 is part of the "Active" Low Pass filter. Check out this youtube video

for the full description on how this works.

WTalen - I am using a DC-DC converter to decrease the VIN to the voltage regulators. Linear voltage regulators are basically variable resistors. There is an inherent voltage drop across a regulator which heats up the regulator due to the power consumption. A supply of 13V will need to be dropped to 5V. I plan on powering everything with my power supply board Arduino, roll, ADC, WAS, arduinosimple ublox zed-F9P, LoRa radio, some LEDs, and maybe some 5V relays. If all of this consumes 300mA or 0.3A the regulator will need to drop 8V @ 0.3 A. Power = Voltage X current so 8X0.3 = 2.4 W. The regulator will need to dissipate 2.4W of energy, this does not seem like a lot but even using a big heat sink of 16C/W results in a temp rise of 38.4 degrees C over ambient. If the tractor cab 21C the 7805 silicon temp would be almost 60C. 60C is an acceptable temperature but it requires a big heatsink, which takes up a lot of room on a 80x80 mm pcb (free version of Eagle PCB). I am stepping down 12 to 8V to reduce the power dissipation required for my regulators.

Brian - Correct the TIP112 will drift with temperature that is why I am filtering the output directly off the 8V rail. Even with 200 mA of load the output of my "active" filter was still at 6.7V which is plenty of head room for my LDO regulator. Probably don't need a reference voltage as long as the output is stable. My small transistor circuit is designed to filter out high frequency switching noise. I have only tested this circuit by powering it with my linear lab power supply, further testing in a vehicle will be next.

 

[ropecope]

Jays_Truck, I think
Power = V x I
But V is (Vin - Vout)
V = Vin - Vout
in your case:
V = 8 - 5 = 3
Power = 3 x 0.3 = 0.9W

 

[BrianTee]

The best way to filter switching noise is don't create it in the first place. Would an external linear 14.4v to 8 v supply make sense to power all digital and a/d circuits? Could also be fused separately.

 

[Jays_Truck]

The best way to filter switching noise is don't create it in the first place. Would an external linear 14.4v to 8 v supply make sense to power all digital and a/d circuits? Could also be fused separately.

Agree 100%! There are many ways to skin a cat. The best thing about open source is that you get a lot of ideas on how something should work with many different opinions on implementation. In the end differing opinions will only result in a better finished product. Just think how far this has come since last year. We are now talking about PCB prototypes with defined parts list to build a fully working setup.

In the end the KISS (Keep it simple, stupid) method usually wins out. A power supply with multiple linear regulators would be super simple cheep and very stable.

 

[BrianTee]

Just working on schematic for "default" board. This one has input protection, wheel angle sensor separate supply, switching 12v to 5v HW-411 module and can be used with either usb or ethernet, drive 1 Cytron for steer motor or 2 for hydraulic valve steering.Uses the common 3.3v MMA powered by nano 3.3v reg. Only one sided pcb for easy diy. Not all labeled yet, done in Eagle Cad Pro 7, so it can be loaded into the new trial version 9.1 that only has a size limit of 12 inches square if you want to do mods.


But this will give a good stable base that the required parts list can be made, a single INO, and make the wiki a whole lot easier. If people want to do their own thing, they can, but at least there is 1 that can be well documented for successful completion by anyone. I was tired of programming so I thought I would kick the dust off my pcb skills lol.



Thoughts?

 

[AlexS]

Yesterday I have the same idea, for an simple pcb for the Arduino & Co
It is an 2 Layer pcb for the parts: Arduino, MMA8452, BNO0, ADS, LM2596

 

[ropecope]

What do you think about this?


I did not want BNO and MMA to put it on the board because you never know how the board will stand in the box and how the box will stand in the tractor

 

[BrianTee]

I really like the separate power supply! So many ways to do this, it's great.

 

[ropecope]

Yes, regardless of whether we are changing something to an autosteer, the power supply will remain the same

 

[8582]

Very nice ropecope. Stackable boards would be the only thing to add to that. PITA to design, but nice and clean when done.

 

[ropecope]

Very nice ropecope. Stackable boards would be the only thing to add to that. PITA to design, but nice and clean when done.

Maybe something like this?

 

[8582]

Maybe

 

[BrianTee]

The switch inputs really need esd, short circuit, and over voltage protection. The WAS analog input also should have protection, and the PWM drivers should at least have inline resistors. If a 3.3v MMA is used the I2C lines need 300 ohms in series as well. 0.1 uf decoupling also should be on all Vcc lines adjacent to every device.

The tractor is a harsh environment and the arduino is designed to be used by children in a school lab environment. The addition of a few schottky diodes and resistors will save a lot of trouble and dramatically increase reliability. The arduino has power supply clamp diodes internally, but they certainly aren't big enough or fast enough.

 

[ropecope]

The switch inputs really need esd, short circuit, and over voltage protection.

power supply has BTS432E2

Description:

High side power switch with integrated vertical power FET, providing embedded protection and diagnostic functions.

Summary of Features:

Load dump and reverse battery protection1)
Clamp of negative voltage at output
Short-circuit protection
Current limitation
Thermal shutdown
Diagnostic feedback
Open load detection in ON-state
CMOS compatible input
Electrostatic discharge (ESD) protection
Loss of ground and loss of Vbb protection2)
Overvoltage protection
Undervoltage and overvoltage shutdown with autorestart and hysteresis
Benefits:

High load current capability
High inrush current capability
Very high energy capability
Capable for 24V applications
Integrated protection functionality
Integrated diagnosis functionality
Target Applications:

Automotive Body & Convenience
Power Distribution
Relay Replacement
Fuse Replacement
Heating Applications
Truck & Agriculture (24V)
Industrial Automation

The WAS analog input also should have protection, and the PWM drivers should at least have inline resistors. If a 3.3v MMA is used the I2C lines need 300 ohms in series as well. 0.1 uf decoupling also should be on all Vcc lines adjacent to every device.

The tractor is a harsh environment and the arduino is designed to be used by children in a school lab environment. The addition of a few schottky diodes and resistors will save a lot of trouble and dramatically increase reliability. The arduino has power supply clamp diodes internally, but they certainly aren't big enough or fast enough.

We'll do this

 

[ropecope]

The switch inputs really need esd, short circuit, and over voltage protection. The WAS analog input also should have protection, and the PWM drivers should at least have inline resistors. If a 3.3v MMA is used the I2C lines need 300 ohms in series as well. 0.1 uf decoupling also should be on all Vcc lines adjacent to every device.

The tractor is a harsh environment and the arduino is designed to be used by children in a school lab environment. The addition of a few schottky diodes and resistors will save a lot of trouble and dramatically increase reliability. The arduino has power supply clamp diodes internally, but they certainly aren't big enough or fast enough.

I did not understand immediately what you said, but I'm clear now.