My derivative of the VESC 6.4 HW from the posted schematic

I have @stewii ESC hw6. 4, he calls it ESCape

I was just about to order one of these now Im sceptic

im still waiting for someone to do a proper bench test. I need to see how far the Erpm goes.

Where is the schematic posted.

Just a little update. Changes are the addition of a TVS diodes to the signal line of the servo port, CAN transceiver, and one near the MOSFETs. Added pulldown resistors to the low side MOSFETs. The low-side MOSFETs and the TVS diode can be found in the VESCX design. Reduced PCB size from 70x70 mm to 68.3x64 mm after rearrangement of components. I also figured out the MPU9150 is mostly used in applications regarding drones and airplanes but not much for eskates, so this stays off. There is space for it on this board should it be used.

The TVS diode near the solder pad for power and the pull-down resistors are not required for operation. The TVS diodes is for protection against transient spikes and low side resistors for negating the shoot-through condition MOSFETs can encounter.

Iā€™m planning to run my design at redline such as 58 volts. Itā€™s pretty much done and ready to be sent to Oshpark.They are just I just have to fix a few things with the silkscreen layer and slap my own makerā€™s mark. The sad part is I canā€™t buy any of the resistors I chose from Vishay on Mouser or Digikey for the next 6 to 8 months. Apparently itā€™s Vishayā€™s new line of automotive resistors for 2018 so it means waiting. I did make a version for the 0603 resistor since they are in stock but I still want to try my version at redline. I also have to make a case for this.

As for naming this. Iā€™m not branding it like the ESCape, VESC, VESCX, FOCbox, or any other combination of letters to VESC. Itā€™s just going to be the generic term ā€œESCā€. Other languages the pass such as tank in German is Panzer today versus schutzengrabenvernichtungpanzerkraftwagendepending (roughly translates to trench annihilation assault vehicle) depending when it was used. Source for nameing tank in German.

My version will be my mark with no name.

8 Likes

Check out allpcb.com, used them a couple times already, super fast, great price, great quality

Just my 2 cents : I think the cap are on the wrong side as it will be more difficult to desing a thin case and they will also bother the heatsink for the DirectFet. BTW, are those SMD cap the same rating as the through-hole model used in the original VESC 6 ? And, finally, using only ā€œESCā€ is not a good approach IMHO. We see a lot of derivatives of VESC 6 now. And for the GPL licence, you need to mention ā€œbased on VESCĀ®ā€ or something. :wink:

What the purpose of TVS on servo port and CAN ? Do they protect against GND loop as we can see in dual setup using Y-split PPM cable ?

This is a great improvement, congrats for your design !!

2 Likes

a Chuttney ESC!

Iā€™m going to call it ā€œTAAVESCā€: The Trench Annihilation Assault Vehicle ESC!

Ah!, no VESC allowed, I forgot, OK how about thisā€“

TAA8ESC (Pronounced ā€œTASKā€): Trench Annihilation Assault Sk8 ESC!

1 Like

I donā€™t know how to get the 3D rendering to show the capacitors the way I want it on KiCAD.Itā€™s just a placeholder for now. Capacitors are at the same voltage or greater, the capacitance is the same as listed on the BOMā€¦

I thought the transient voltage was the issue. Looks like itā€™s back to the drawing board to prevent the ground loop. There were a few cases the CAN transceiver malfunctioned, I canā€™t do testing on this as I donā€™t have another working VESC controller. Iā€™ll just assume the best way to protect it from literature posted on the internet for the best thing to use.

1 Like

I developed SESC of VESC6 clone. It used a DRV8323 device including three shunt resistor.

11 Likes

Nice work!

That was the next thing I was going to do. Use DRV8323RSRGZR chip to get rid of the current sensing amplifiers and make a smaller version more applicable to the drones, airplane, and water RC people, but I couldnā€™t code the stuff.

Do you have any plans to release your versionā€™s code for implementing the DRV8323?

Update 6/26/2018 : These are the files to my version. Itā€™s two layers and soldered all components. Firmware and bootloader upload through ST-Link work well, but MCU will not give a green light. I post this to the community here for someone to check what exactly I did wrong.

https://github.com/YBMat/Speed-Controler/blob/master/VESC6.1.3_0603%20version.zip

2 Likes

Hi Chuttney, Not a real electro man here (more mechanical), but I got myself interested in VESC pcbs too. I looked at your files and have to say that I couldnā€™t find a hard flaw directly. It was also difficult as not all schemeā€™s are in the zipped folder. I assume you checked for solder balls first and coulnā€™t find any abnormalities.

I do think however you designed a very small power supply line to the DRV8301. See picture attached. This line (0.254mm) feeds 3 SPUPPLY pins, which in turn via the 5V output, and via the LM3671, supplies the power for all ICā€™s on the board. I think it is too small. small_supply_line And then there is a huge VCC track from P7 to U3.

Moreover; there are 4 identical CAPS next to eachother. But I do believe they are more to be placed near each corner of the STM32F045. close_caps

C31 doesnā€™t seem connected? not_connected By the way, my Kicad says there are 16 unconnected nets. Most however look connected, but it might be a good thing to recheck these. And solve themā€¦

Maybe try to get some external / bypased power to the +5V/VCC volt pins of the headers? That way you power the ICā€™s with a bit more power.

Hope you can find the problem! Would be a waist of ā€˜lesson learnedā€™ (and money :wink:)

2 Likes

Thanks for the feedback. I should state the design I posted only adds the D9 diode to put an Oring Diode as stated by one of the commenters in a different thread. Never heard of an Oring Diode until it was stated. I had to change the footprint for C32 because I was too small when I got the PCB made which was the reason it was unconnected.

I increased the track width of the supply line to 0.75mm going to the DRV8301 for the 5V output. It should be enough.

I fixed the row of capacitors in the schematic and PCB file next to each VDD pin as expected. It really is confusing given Vedder made that note and more experienced people like him know how to interpret it. I believe all the VDD pins can be daisy chained to a bank of capacitors by way of a singlecommon power trace, because of the intended operation of placing individual capacitor next to each VDD pin.Iā€™ll probbaly have to add one to the 3.3V/VCC rail.

I fixed C31 by adding traces so it connects to C32.

I fixed the polarity direction of D9 to the correct way after seeing in one of your posted pics. It is the 5V rail for all the connectors and supposedly the ā€œORingā€ diode. Sadly, USB power of MCU cannot happen so battery power must be applied when wanting to upload firmware to the MCU.

1 Like

No. Those caps need to be as close to the VDD pins as possible. They are acting as filters to clean up the power entering the MCU. Look at the 4.12 pcb.

Google ā€˜decoupling capacitorā€™

1 Like

Chuttney, Were you able to solve it? I just came across this post and remembered you didnā€™t have power too: http://vedder.se/forums/viewtopic.php?t=195 Maybe it helps. On a board I did I had the same problems too. Couldnā€™t find any solderjoints, even after checking all pins next to eachother with an ohm meter. But by luck I did see at 3 pins next to each other that I could move when gently pressing them with the tips of the meter :face_with_raised_eyebrow:. It was a sloppy soldering job but I resoldered them (first created a new solder bridge again :confounded:). Solved my problem.