New VESC from China. Seems like not only Maytech is making VESC's now

ESK8 Electronics
356

Ah bugger.

I also decided to try out CAN instead of split PPM. I reconfigured both ESC’s and everything was working but when I tried to check via Ackmaniac’s App, no BT module seen. No lights on the module. Try a second module, same thing. Connected the one at a time to an Adruino Uno and bang, powers up, lights on and I can connect to it. Both these modules used to be connected each to an FSESC and worked fine before.

I have been speaking with them about the split PPM issue and powering up both ESC’s. Just need clarification on some thing they want me to try.See what they say about the UART port as well.

357

Blew the DRV on my 6.6 esc today. Was running great at 12s 140A phase on foc but decided to try bldc mode today with the same current. Blew in about 5 minutes :frowning:

I bought some better mosfets to replace the existing ones so I will try and get those soldered in while I do the DRV.

1 Like
358

I sorted the UART problem. Picnic error plugging a 6 pin JST into a 7 pin socket badly.

CAN is working fine etc

359

Oh damn. Did it over heat or just pop?

360

ESC was cold. It just died randomly, no fires or anything dramatic.

Some good news though. The DRV repair and mosfet upgrade went well.

2988×5312

5312×2988

3 Likes
361

Hi All,

As promised, data from testing two FlipSky v6.6 ESCs …

ambient temperature in test room - between 22 and 25 degrees c airflow - nil cooling - case with heatsink only constant input - 48v / 30a (conditioned power) test load - 48v (1.5kW) BLDC sensored motor duration of 30 amp test - 37 minutes maximum temperature ESC 1 - 63.2 degrees c maximum temperature ESC 2 - 68.4 degrees c

Immediately prior to this we’d done 1 hour at 15 amps AND 36 minutes at 25 amps. So total test time was well over 2 hours. The testing was contiguous. This is the reason the graph starts at 50 degrees c.

flipskydata25062018

This is a huge step up from the VESC 4.12 even with a heatsink. We are really impressed.

I can see there may be a few component issues based on @district9prawn experience, but in our testing of constant throughput of over 1200w it was really impressive.

FlipSky seem to be on the right track. We’ve found them responsive and they appear competent and willing to take feedback and incorporate it.

They also seem prepared to sell them at a reasonable price, so if you can help them it may well just help everyone.

Seems to me that they are close to getting this sorted and they can mass produce.

Worth supporting with feedback / input and may end up a good commercial option.

Cheers

4 Likes
362

Testing indoors so effectively zero air flow? Impressive that there is no overheating.

1 Like
363

I went for a ride yesterday after work along some mountain bike trails etc. Here is the log of my ride - https://drive.google.com/open?id=1ccv4J5gTz9E-4b-9FYA0GsTlYhx3sx1K

Almost an hour long ride with short breaks during a hot evening +26c ambient. 80/-80 motor 40/-10 battery running sensorless FOC. As you can see I was not able to push the ESC’s due to the terrain but the ESC’s survived the constant rattling and worked flawlessly.

Flipsky have been in communication with me about the spilt PPM issue, I have a few things to try tonight and will report back.

Overall I’m happy with how these ESC’s have performed so far. Time will tell with regards to overall reliability. I quickly designed & printed a bracket to mount the ESC’s on top of each other IMG-20180623-WA0009

It’s going to get a total redesign with a BT module holder integrated on the side, a bottom bracket screwed to the old nowind MAX6 mount and then I will try to figure out a way to clamp the top ESC down to the bottom ESC (right now I’m just using cable ties).

It is currently like this - IMG-20180625-WA0000

So it needs sorting out :wink:

6 Likes
364

@district9prawn … in a lab. Definitely no airflow :slight_smile: It’s really really good. Are you able to detail your DRV issue, changes made to it and MOSFETS with exact component changes and why? Thanks.

365

The mosfets are NVMFS5C604NL. They have half the rdson as the stock ones. So low now that with drv8301 overcurrent at lowest setting, current is reported as 50-60A when tripped :joy:

The drv death was pretty unremarkable. I had the esc in trap mode and gave it full throttle. Motor cut out and esc threw drv error.

140A phase seems to be the limit for how far I can push. Setting higher in sensorless foc causes the rotor position to be lost at full throttle. I don’t think it is saturation as I have an identical motor with higher turn count that also has the same problem at the same current setting.

4 Likes
366

I just want to point out that Rdson isn’t the only electrical characteristic of MOSFETs that generates losses.

The stock MOSFETs also have less than half the gate switching charge, which means less than half the switching losses. The NTMFS5C628NL has 10nC of switching charge, the NVMFS5C604NL has ~25 nC of switching charge, resulting in a total of 40nC and 100nC respectively). With four in parallel on a 44V system at a switching frequency of 25kHz, Switching losses dominate until around 40A with the NTMFS5C628NL with a power dissipation of 1W (total dissipation 2W), with the NVMFS5C604NL, switching losses dominate… well, always (150A). The formula for calculating switching losses is (GC * SW * Vcc* I)/Igs where GC is Gate Switching Charge (Gate Drain Charge plus Gate Source Charge minus Gate Threshold Charge), SW is switching frequency, Vcc is system voltage, I is current, and Igs is gate drive current.

For reference, the stock MOSFETs have a total dissipation of 1.8 W at 40A, while the NVMFS5C628NL (four in parallel) have a total dissipation of 3W at 40A (most of that is from switching losses). Both at 25 kHz switching frequency. MOSFETs have this attribute called “FOM or Figure of Merit”, which is switching charge times Rdson. When paralleling MOSFETs, this becomes very important, as in this case, where there are four in parallel on each phase (well, two on the top and bottom). Switching loses affect both sides of the half-bridge since both the top and bottom need to turn on and off, whereas conduction losses only affect one side of the half-bridge at a time. In fact, in switched mode power supplies, switching losses is the main source of power loss, not conduction losses, so Rdson becomes even less relevant in those applications. Trying to exclusively minimize Rdson in a switching application is like trying to overclock a CPU while reducing SDRAM read/write speed.

Gate switching charge is what makes Internal Rectifier’s high performance MOSFETs such trash when compared to any other MOSFET in that performance range. They literally have sometimes up to 4 times the switching charge as a MOSFET from a different manufacturer with the same Rdson. Compare IRFS7530 with IPB014N06. (The latter is also cheaper too). (Well, Technically Infineon owns IR now, but optimos and IRFET are different technologies. The IRFETs released under Infineon are getting better with the switching charge, but are still inferior when compared to the competition).

Has anyone here checked out the ODrive? it’s another Open Source ESC that can drive up to two motors and an auxiliary DC motor. It works with (relatively cheap) shaft encoders to provide a cheap but powerful servo solution with BLDC motors.

The 48V variant costs $149. I would get one myself, but I can’t really install shaft encoders on 8" hub motors, maybe it can be rigged to run without encoders.

5 Likes
367

Looks nice - any idea if it can be an alternative for the VESC? Does it accept PPM remote control ?

368

It looks like it can accept PPM as well as many other forms of communication. I looks really promising as an alternative to the VESC; however, it appears to still be in beta testing and currently has stocking issues. Oh I’m a fool, it has a sensorless mode. I’m going to get one to replace the two VESC 4.12s I have once the 48v variants are back in stock.

369

I does accept PPM. In short, the STM32F405RTG6 chip used in the VESC can be coded to run two DRV8301, which is what they done.

Here’s a direct link to their schematic of their most recent version. https://github.com/madcowswe/ODriveHardware/blob/master/v3/v3.4docs/schematic_v3.4.pdf

370

@district9prawn They’ve got some additional detail now on their website which outlines the changes they’ve made and some key components. I had a message from them and they’ll be changing their MOSFETs based on your suggestions and testing… like I said, looks like they want to get this to work. https://flipsky.net/products/flipsky-fsesc-6-6-based-upon-vesc®-6-heat-sink?variant=8739261743164 … and they definitely appear to have some engineering ability at the hardware level which is refreshing as most of the Chinese VESC derivative (and copy) suppliers we’ve dealt with seem not to …

If I can just soapbox for 1 minute …

there is an incredibly high level of electronics design skill on this forum and tons of experience. If these people on this forum can help FlipSky make this product bullet proof, this is the chance to finally stop the bins full of dead VESCs that are posted here as a badge of honour. They aren’t; they are a complete waste of money if they don’t work, but awesome value if they were commercial quality motor controllers. A high spec ebike controller with bluetooth is $50 … the bar is high.

Mostly, the group buys rely on one person’s design with no ability to do much ‘looping’ to incorporate changes. Understandable of course as you aren’t dealing with a company, but not ideal. More power to them, but it’s not the path to mass production and eventually lost cost, reliable, robust e-skate (and other stuff) for all. Then the various e-skate suppliers on here could just focus on making the best boards and know the controllers aren’t going to fail. So could all the builders …

FlipSky happen to be the ones we’ve found most willing to assist and they seem willing to incorporate suggestions / feedback and take real world data and use it to make the product better. We don’t ultimately care who does it, but if FlipSky ends up being the ones to bring bullet proof ESCs to the world with VESC lineage then we certainly won’t be complaining and we’re doing what we can to help. We just need reliable product at a reasonable price. Right now, it’s not.

disclosure: We are looking at FlipSky as a potential supplier. We have not made a decision yet but hope their ESC can meet our requirements. The no.1 issue for us currently is reliability … it cannot fail and you don’t have to look far through these forums to find them still failing. We are trying to get their ESC to work in our product/s and hope that it can be sorted in time for us to use it. Our feedback and testing is also looping back into the project through FlipSky. We’ve found them to be responsive and willing to undertake continuous improvement.

The collective experience and knowledge of a subset of users on these forums could fix the VESC derivative issues … if they want to. This isn’t supposed to be an electronics hardware design forum as far as I can understand as it’s over the head of most of the users.

There’s a huge amount of time dedicated to hardware design and re-design because of continued failures. It’s been years since BV released the first designs and the improvements on the v4.12. It seems like a good time to stop continually looping through the hardware design and get it rock solid. Sure the firmware and software improvements/iterations and accessories (like @rpasichnyk 's work) should continue and will but it has to be on the back of hardware that does not fail any more. It’s got to be smart enough to ‘trip’ before it fails.

So, if you think this is all reasonable, get behind them, get in touch, give them feedback and suggestions … and help stop the dead VESC problem. Help them make their VESC bullet proof. At least that’d be one commercially available at a reasonable price that doesn’t fail.

It’s time to move up a layer so the hardware does not fail and it’s all about firmware and software.

Cheers

7 Likes
371

My flipsky vesc hw v4.12 arrived today. Exactly 2 weeks since the date I ordered on ebay. How do I get things started? What to avoid doing so I don’t fry them? Any help is appreciated! Thank you! image

2 Likes
372

nice, i hope mine will come this week.

373

Did you order it when it was $60? Today I checked and price went up to $80 lol

374

no still the same

375

Wow what app is that? Looks cool. Edit: think I found it. Is it this one?

1 Like