The test sled. @hoytskate 5” tires - these things are awesome.
3 Likes
This is great testing, and really shows how far sensorless startup has come, especially with no tuning. Thanks for taking the time to put it together.
Hopefully more people can experiment with it and post results in this thread.
It’s worth pointing out some details about these ESCs. They are from the first and second batches of VESC 6s released back in 2017 and do not have phase filters. It wasn’t until quite a while later that phase filters were added to the hardware in the MK5 series and above.
If anyone has an ESC with both phase filters and phase shunts it would be really interesting to compare results and see if it resolves some of the issues shown here.
Here is a list of ESCs that have both. There may be more out there.
-
Spintend Ubox 75v single and dual, all versions. Not the 100v. Spintend states on their website that Silent HFI seems to be working well on these controllers.
-
Trampa VESC 6 MK5, MK6, HP, 75v, and EDU
-
Trampa VESC 75/300 R2 and above
A list of ESCs with at least phase shunts (like was used in this testing) can be found here:
4 Likes
The silent HFI looks like it had less kick than screaming HFI?
I think it’s just the way i was throttling in the videos but i can check later and test with that in mind.
1 Like
Update - I tried switching back to HFI to see what it feels like on fw6.0. I heard that HFI was upgraded so I was curious.
I did several flat ground full throttles and every single one resulted in a motor fault where I had to jump off the board.
So from my very short test, it seems that sHFI is way better than original HFI in high current bursts.
3 Likes
I made a discovery tonight while playing with HFI settings on my stormcore 60d+.
I got all the new modes of sHFI to work! They are not silent, but quieter than normal HFI. These new modes also handle more amps than normal HFI so they should improve sensorless performance.
To get sHFI to work with motors reliably turning in the desired direction, I reduced the default “HFI gain” from 0.300 to 0.100
4 Likes
let’s do a drag race! i’m onnnnnn 5.1? 5.2? on HFI on my thane board
2 Likes
Set one side as HFI and the other as VSS and see which side pulls harder
1 Like
I can’t recall if i have temp sensors connected on this board. Vss requires them, right?
Also, it’s not really which one kicks harder. They should kick the same. The thing is VSS will likely fault out at lower amps than 45 deg or coupled v0v7.
rafe me on wednesday
After work - I’m in. I’m more likely to be on the true sHFI dremelseed tho. And that thing does not kick ass at 10s. It’s fun but not winning any drag races.
1 Like
I thought silent HFI was called ASS then got changed to VSS
ASS did get changed to VSS, but silent HFI is new to fw6.0.
Coupled v0v7 and 45deg v0v7 are new and can handle more amps. 45 deg moreso according to vedder but coupled is quite good in my testing.
What is the oldest hardware that runs it?
1 Like
The newer VESC6 MKV and MKVI have the phase filters. They make all the difference on the very bottom end of the RPM range. However, even without these filters it is quite impressive what sHFI can do. Single processor dual ESCs will always generate a bit of noise, regardless of phase shunt or low side layout. You have to split the sampling frequency and therefore you end up with 12.5khz instead of 25khz. 12.5khz is audible for most people.
The sHFI tuning will probably see some automation in future, as it should be possible to calculate the voltage parameters from the motor detection parameters and the battery parameters.
Some motors also work much better compared to others. Ld-Lq difference is one of the key factors.
A motor designed for sHFI operation should feature a high value here. Weak magnets and a big airbag reduce this value for the typical 14 pole outrunner. It also depends on other design aspects.
When everything is well optimised, then sHFI is a lot better compared to Hall sensors. But it can also be less good if the ESC isn’t optimal or the motor or the tuning. So we expect some mixed feedback because of the HW variety out there.
In general the entire sensorless startup is much better now and more quiet. No matter what HW is used, there are huge benefits when switching to FW6.
2 Likes
Not from where i’m sitting.
5 Likes
I have a Unity on my hubs running VSS and dual FocBox 1.7s on my off road running standard HFI, do you think it’s gonna be worth the effort to crack them open?
I’ve had a decent play around with all the silent and non-silent modes now. Some work better than others, and if I was just accelerating it may be passable, but it’s a bit of a mess on a onewheel. Can start ok, but it struggles with low speed producing cogging leading to either nosedives or tail drags. Current and duty cycle modes work ok in the VESC Tool, speed control mode will usually get it spinning in a random direction, and position control just causes shuddering with little actual movement.
Haven’t really had much luck tuning the HFI voltages, I seem to get quite a lot of noise. Data noise that is, not necessarily sound noise although there is that too. This is the foc_plot_hfi_en 1 (enabled in VESC Terminal) experiment data chart, normal ranges should be between -3 and +3 
. Any tips much appreciated.
Trampa VESC 6 mk5 with an APS 80100 50kv motor.
4 Likes
Have a look at this value. The higher, the better. If this too low, then the non sHFI modes might work better for you. Motors with a tight airgap and strong magnets perform better here.
2 Likes
VSS is not a good method for these controllers. They don’t feature phase shunts. VSS really wants phase shunts and phase filters. The non silent HFI modes are preferred for the designs you mentioned.
1 Like