FLIPSKY 4.20, A Explanation to the mystery, and how to stop the cutouts entirely!

Schtekarsten · March 21, 2019, 9:04am

Hello good folks of this forum, first post here. kinda nervous Moi and my good friend Gamer43 #thebrain have an explanation to the cutouts and how to fix it. Long long long story short, swap the gate resistors and you wont have the cutout issues anymore.

The reason behind the cutouts of the 4.20 based vesc controllers are something called slew rate (Gamer43 will do a better explanation of that then I ever can, so wait for that if interested), basically when you you pull a shit ton of Amps and do so fast the DRV ship says see ya and cuts out. Sounds familiar… well that Is probable why you threw it in the fuck it bucket at the first place. hopefully you are to lazy to empty the fuck it bucket in the garbage and can still bring it back to life. because it is really quit f***** good when the cutouts are gone, and it can handle continues load surprisingly well, like really well!!! And to combat that you can change the gate resistor or the mosfets used. The cheapest and simplest solution for us garage folks is to change the gate resistors. Change all the 4.7 ohms resistors of the 4.20 plus to 22ohms and on the other 4.20 based versions to 47ohm. (again we will do a video and explain that better down in. y later on…) The 4.20 plus uses “better” mosfets that’s why you need a lower value. Basically we have done maybe 50 hours trying to disprove our theory by messing around with settings and different hardware but yeah cutouts that wont happened even if you decide to strap the vesc to a truck and turn it’s wheels. (well I lied, we have not tried that but you get the point) Well as far as we got before i ran to mamma crying. Was with a 12s4p setup, dual 6374, sensored/hybrid foc, 80a motor max, 60a battery max on each side. doing so at the steepest hill I could find in the area and started from a standstill. And the only rule was: hit the throttle to max as fast as you could possibly could until you went flying up the hill, or the board cutout, or you went tree hugging.

Apart from that we have done, y anything imaginable to have the vesc cut out but nop, not a single cutout as long as you don’t lower the resistor value, and if you do so just a tiny bit it will cut out instantly with just slightly aggressive settings and throttle movement. Yes even the plus version cuts out real easy. (the 6.6 Vesc’s actually has the same issue with cutout but that is at like 80A (alot higher), so you kind of need to strap that to a truck to have it cutout, and very few people do that and we believe that’s why people are not posting and crying about flying into a bush on the 6.6 vescs because of cutouts.)

Oh and I would go as far as to say, stay away from the plus versions their anti-sparks are not the most reliable in the world…

Damn I wanted to keep this short but I guess I messed up… Oh and Gamer43 is really the one you should applaud for this. he is the brain and beauty behind all this.

Flipsky are now shipping all their 4.20 based Vesc’s with the newer resistors. They just have not had time to update their website as they are in the middle of moving their entire factory. Have seen and gotten a few questions regarding it so I thought I might post it here.

If you want to go ahead and make the resistor swap yourself, I have 2-3 posts below and a video covering everything you need to know.

Gamer43 · March 21, 2019, 9:12am

Here is a photo of the resistors in question:

TL:DR, the resistors circled in red need to be replaced with 47 ohm resistors (22 ohm on the plus versions). The result is no more cutouts, even on the settings Schtekarsten mentioned above. The resistors’ original value is 4.7ohm, no other resistors on the VESC 4 has that value.

Basically what was happening was that since the NTMFS5C628NL MOSFET has about 6.5 nC of Miller (gate-drain) charge, this resulted in very high frequency ringing on the switch nodes. Long story short, it caused the DRV8302 to glitch and throw a fault. This would occur when current would change very quickly, so the ringing most likely triggered the VDS monitor on the DRV8302.

I first suspected this was the case when I was lurking on the TI forums, someone reported spurious fault reporting by the DRV8302, and the solutions were about proper routing of PCB gate drive and ground traces.

Will this negatively affect switching losses? Yes. But by my calculations, when switching losses are equal to 2W, conduction losses are equal to 18W, so the effect is negligible, (meaning at the point where temperature becomes a problem, the size of resistors doesn’t really matter). In Fact, the switching losses are still BETTER than 4.12 hardware versions. (The 4.12 uses the IRFS7530, which has an onerous 70nC of Miller charge)

Also, I really only gave Schtekarsten some ideas, his rework and testing is what allowed us to find a solution to this issue.

DeathCookies · March 21, 2019, 9:15am

Great work of you two! I am excited to see the feedback of the users

Schtekarsten · March 21, 2019, 9:19am

Hey thanks, we plan on getting some videos going on how to do the resistor swap your selves.

ervinelin · March 21, 2019, 9:24am

Wow… just wow… much respect!!

Now if only my hands were steady enough to fix this…

By the way why is it that different firmwares seem to suffer these cutouts differently if it’s a hardware issue? Vesctool 0.94 and the latest 1.05 were good for me, 0.95 (which was what came with my dual 4.2 non-pro) would spaz out above 30A.

Gamer43 · March 21, 2019, 9:26am

I think Vedder changed the way the current filter works between those firmwares, and it’s the rapid change in currents that would trigger the faults.

ervinelin · March 21, 2019, 9:55am

Sorry if it’s a wee bit off topic… but do you also know why the voltage readout for the flipskys are not stable? I run telemetry to see my decks voltage but for flipsky escs they have so much jitter in the readout.

pat.speed · March 21, 2019, 10:03am

Definitely following this thread, I’ve been wanting one of these for a while but the cut-outs put me off. Now you are making me spend money…damn it

Schtekarsten · March 21, 2019, 10:51am

For those eager to get going with the resistor swap, i’l just put together a short writeup before i do a more in-depth video thingy: For the dual 4.20 PLUS vesc’s get the 22 ohms resistors in a 603 package For all the regular dual 4.20 and the 4.20 single vesc’s get the 47ohms resistors in the 603 package Get that stuff from arrow FREE WORLDWIDE shipping ( like 1-2 days to Sweden, click around on the shipping options when at check-out Fedex is the quickest for me) For the dual 4.20 plus get the: https://www.arrow.com/en/products/rc0603fr-0722rl/yageo For all the dual 4.20 and the 4.20 single get the: https://www.arrow.com/en/products/rc0603fr-0747rl/yageo

You need 12 of them for a dual and 6 for a single, but F it get like 50 of them they are not even pennies… and you will most likely lose half of them.

You need a ton of flux, get the no clean stuff or have fun cleaning afterwards, some normal lead solder, a good iron with a small tip and preferably a large one too. ((and no if the tip is black and crusty toss it!) I happen to like the “ts100” with the TS-D24 and TS-C4 tips with a Mean Well LRS-100-24 power supply and 20awg cable, might get more into that if interested.) Some solder wick, a pokeee thing or some fine tweezers.

Start by desoldering every resistor with the 4R7ohms mark you can find (se Gamers43 picture above). Use the fat tip and a big blood off solder on the iron. The ones in a tight spots use the small tip to desolder. A big goop of flux on the resistors might help. Use solder wick to get rid of the disgusting lead-free solder on the pads. Squirt out a big glob of flux on the pads and tin the pads with solder. (I have had better luck with tinning both pads before adding the resistors) Carve off some of the plastic on some of the JST connectors to make room for the iron. (you will know what I mean when you get into it) Flux on, heat up the easiest of the two pads and push the resistor into the iron and the iron will kind of move about the resistor. When positioned correctly, stab the resistor with the pokee thing to hold it down on the pad and remove the iron. Try to lightly poke on the resistor to se if it stuck down well enough. place the iron on the other pad and reflow the solder. Important! in order to make sure the resistor has positive contact with the pad try heat up one of the sides/pads with the iron and push it off and then do the same with the other side. If it is still stuck down then that’s all good! Celebrate with a beer or two and jump to the next resistor. When done with all the resistors double check that they are all on correctly by poking each resistor with a iron on each end of the resistors. Remember flux is king! if the joints are grey and dull more, flux my friend get them joints shiny! Clean off the flux with some alcohol if you want to make it look a bit nicer.

ugh shit that got long… a video will be coming in a day or so, grab the resistors from arrow in the meantime and the video will hopefully up by then. (btw I am by no means an expert on this and there are probably better ways to do this with hot air and all that, but I have swapped about a hundred of these little buggers and done so successfully.)

Schtekarsten · March 21, 2019, 10:57am

No worries, both my BT-modules have gotten a sip of the Swedish salty roads so I have not been able able to se if it has to do with this or not, got some new ones on the way though so I will let you know if that is the case or not. but if I remember correctly I don’t think I have ever had that issue. idk

pat.speed · March 21, 2019, 11:08am

Well, by the time we get to the 12th one the solder job is gonna be pretty darn sloppy lol

riverside.rider · March 21, 2019, 11:33am

yes, same here, would love to know why the voltage seems to be rather unstable with the Flippies…any ideas @Schtekarsten? great work on the write up

Rzkyyy · March 21, 2019, 11:43am

Would you guys just please sell me the modified version of dual fsesc 4.20. I’m more than happy to be your guinea pig

High-roller · March 21, 2019, 12:21pm

Thanks again for posting this! I’m not confident enough in my own soldering skills- or my budget- to try this so hats off to you for taking the risk and experimenting and sharing the knowledge. You might want to get a bunch of these to modify and sell before flipsky finds out.

Well, that’s a new one to add to the lexicon.

Riako · March 21, 2019, 1:51pm

Thanks a lot to both of you !!

Hummie · March 21, 2019, 3:52pm

Did u tell flipsky? Assuming it stops the cutouts the info is golden. I’ll link @BarbaraZ and email here Wow. Nice job! Just moved the world two steps closer to cheapness and reliability. Probably saved like ten peoples’ face

Fiori · March 21, 2019, 4:02pm

Thanks for all the hard work and research fellas. This is awesome. This is why i love this place.

808Chopz · March 21, 2019, 4:49pm

Wow! that’s Awesome news! Can’t wait video tutorial how to? I want to save my ugly face

Gamer43 · March 21, 2019, 5:56pm

We’ve contacted Flipsky, they haven’t given us a timeline on making the change though :-/.

I also would like to add that the “proper” solution would’ve been RC snubbers on the switch nodes, but idk why, no one making these things seems to give a damn about transient suppression >:(.

Schtekarsten · March 21, 2019, 6:50pm

You people are to kind To be honest I have no idea why the voltage is jittery and unstable on the 4.20 units. Im too dumb to figure that out, but I did just receive three bt-modules from ray8801 so can do some testing and see if it is still there or not. I also plan on doing the soldering thingee sometime tomorrow and them post the video, just need to figure out how to edit first though I have been successful on both the dual 4.20 and the dual 4.20 plus version with the resistor swap and I have been running the modified plus version for like two weeks without any issues. (and no I dont ride like a granny, the non-modified 4.20 plus cuts out every single time I go from a standstill more or less ) Even though I dont like the anti-spark design on the 4.20 plus, and moi and my friend would not recommend it, I am running It anyways cus im lazy and like to not get my fingers dirty pushing the on/off button, and it was the last vesc I did my testing on and I could not be bothered swapping back.

Back to my point: The 4.20 plus uses a different set of mosfets that are more efficient and in theory would increase the continues amps from something like a 45A per side (on the old NTMFS5C628NL mosfets) to 55-60A (for the new NTMFS5C612N mosfets). For Flipsky it would result in a cost increase of like 8 dollars per dual vesc by using the “better” 612 mosfets. What do you guys think would it be worth the cost increase to get some better mosfets going, should we kindly ask flipsky as a community to use the more expensive mosfets? I for one is just gobsmacked about the performance of the modified 4.20 plus vesc, the board goes like a hot damn and the Swedish winter does not want to let the vesc temperature throttle (well I have ridden it without telemetry up until now, so who the fuck knows really). But I guess the anti-spark might take that away from me sooner or later

We do have a solution for the damn anti-spark to though, but we are not too sure if we want to give them that yet I mean, they are not open source and we have not even heard from them for like a week or so when gave them the answer to the cutouts. And fuck me, now we are basically promoting their products.