RCMF Donations

Enjoy using RCMF? How about a wee donation to help us keep you in the style to which you've become accustomed?

Welcome to RCMF. Please login or sign up.

December 05, 2020, 11:20:43 am

Login with username, password and session length

FrSky Neuron Speed Controllers (mini review)

Started by PDR, June 25, 2020, 10:44:41 am

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

PDR

Having recently changed my main Tx to a FrSky Horus I looked at these speed controllers because they have extremely useful embedded telemetry. For example if you plug the S-Port connector into a FrSky Rx you will get Voltages (flight battery, Rx Battery, BEC), Currents (motor and BEC), Power(mAh) consumed, ESC Temperature, motor RPM. They also have a huge range of selectable/adjustable settings (including BEC voltage which can be set anywhere from 5 - 8.4v). The particularly useful one of these is the Power Consumption. Pack voltage is a poor indication of remaining charge in a large lipo, and so this allows me to set a fairly accurate electric fuel gauge with an alarm at (say) 4,000mAh consumed from a 5,000mAh pack.

These ESCs come in two types and a range of sizes. The standard (black) ones are about double the size and weight of the newer "S"(red) versions, but the S-types only come in 40 and 60 amp versions while the originals had an 80A (120 peak), and have robust machined aluminium cases for both strength and heat-sinking. They are all the same size - that is to say the standard ones are the same size and weight for the 40, 60 and 80A versions and the two S-types are the same size and weight as each other (about half that of the standard ones). There also isn't that much of a price difference between them (the cheapest 40 amp one is £45 while the most expensive 80amp one is £63, and the difference between the 40S and 60S is about a fiver).

So I got one of the 80Amp ones for my AcroWot - I actually needed the extra weight, and couldn't see any point in getting the lower rating for the price. It arrived in a nice plastic box but with no connectors fitted. One of the first things to grab my attention was that it doesn't have a lead to go to the Rx - it has a pair of servo-lead *sockets* (Rx and Smart Port) instead. So you may need a couple of plug-to-plug servo leads - I feel these could have been supplied in the box! But that aside the impression is of a very sold, well-made unit with a decent build quality.

I mentioned the plethora of adjustable settings. Unlike other ESCs these are not set with throttle movements or a programming card, but are set from a piece of software called "BLHeli_32" that you have to download. To use it you also need a USB-linker" device (available from T9 for £6.60 - cheaper alternatives are available but you need to know what you're doing to choose one). This adjusts everything other than the BEC voltage, which is adjusted via a LAU script in an OpenTx transmitter. You can just plug it in and use it, and it will work (including the telemetry), but it will probably have the prop-brake set to on (fine of you're a gliderist with a folding prop or a multicopter flyer, less-so for normal people).

Also in the box is the manual (two sides of A4) which, like most FrSky documentation, pretty hopeless - especially on how to use the USB-linker. This took me some research and head-scratching (because some aspects of it are unexpected and counter-intuitive) so at the bottom of this post I've included step-by-step instructions so that they're written down where people can find them!

Once you have it connected you get this delicious set of things to play with:


BLHeliSuite32ESC Setup_200625_1.png

The full details of what they all do can be found in the manual (in a PDF under the "BLHeli_32 Info" menu. Things to note include that the range of options presented will depend on what ESC is connected, and that most of these ESCs seem to be shipped with the prop brake or "damping" (this setting only appears when certain ESCs are connected) set to "on". So if you're a normal fixed-wing user with a non-folding prop you'll want to change this. I believe they default it to on because a lot of FrSky customers are multicopterists, and they like to have "throttle damping" because it slows motors down quicker and thus gives them better control response. In some cases this menu item appears as "Regenerative braking/damping" which seems to imply that the ESC can actually put energy back into the batteries - an interesting idea that I have no way of testing. Lots of these settings can be ignored by the typical sport user - they're included so competition flyers can fully optimise their setups and for the inveterate, compulsive tinkerers who tend to be attracted to FrSky kit!

Anyway - my overall conclusion after one evening of head-scratching followed by success is that this is an excellent piece of kit whose telemetry functions offset the slightly higher price and will allow be to operate my larger electric models with a much better understanding of how much duration remains in the battery!

How to use the USB Linker with BLHeli_32
I appreciate that to some who speak arduino the following is a "well DUH!", but the information isn't in the manuals so I thought I'd add it as a summary in case anyone else comes at it for the first time and searches for help.
The Process goes like this:

Before you start plug the USB Linker into your PC and confirm that there's a driver for it. When you plug it in Windows should recognise it and announce it exists. If Windows doesn't seem to recognise it, or it it doesn't seem to work when you follow the steps below, you may need to manually install the driver. Go into Windows Device Manager and look down the list for Ports - expand this branch and you should see something like "Silicon Labs CP210 UART bridge" with a Com Port number. If you don't then ask Mr Google for a "Silicon Labs CP210 UART bridge driver" - choose the option from the Silicon Labs website and download/install the driver in accordance with the instructions.

1. The S-Port and Rx leads must BOTH be disconnected from the ESC

2. The programming lead from the USB linker goes into the PWM port (ie the normal one from the Rx, not the telemetry port)

3. The programming lead MUST have the +ve pin (the middle one) disconnected

4. Plug the USB linker into your PC and fire up BLHeli32 (suggest you maximise it so that bits of the window aren't hidden beyond the bottom of the screen)

5. After firing up BLHeli32 you must find the top Windows menu bar (above the tabs) and open the "Select BLHeli_32 Interface" menu and select "BLHeli32 Bootloader (USB/Com)"

6. At the very bottom of the screen there's a box labelled "Port" - click the down-arrow and select the one from the list which includes the words "Silicon Labs CP210 UART bridge" or similar (ignore the "baud"  box - this is automatic).

7. WITH THE ESC DISCONNECTED FROM THE FLIGHT BATTERY connect the programming lead from the USB Linker to the PWM port of the ESC.

8. In BLHeli32 click on the "connect" button (at the bottom of the screen next to the "Port" and "Baud" boxes. The screen will ask you to power-up the ESC - do this by connecting a flight battery. If you're too slow it will stop trying - just press connect again and this time connect the flight battery more quickly! It will announce success and the "connect" button will change to a "disconnect" button.

9. Click on the "Read Setup" button (above the Ports box at the bottom of the screen). It should load all the data from the ESC including its type and settings.

You can now make whatever settings changes you wish. You can also give the ESC a specific name (a good idea if you have more than one). You can save settings in files on the PC using the top menu "ESC Setup" and load them again later. There are other options for working with multiple ESCs - these are for multicopter use, Instructions on what all the settings do can be found in the "BLHeli_32 Info" menu which links to a PDF copy of the manual.


10. When you have made your settings changes click on the "Write Setup" button (next to the Read Setup button at the bottom of the screen). Do not use the "Flash BLHeli" button - I assume this is for updating the ESC firmware itself, which is a very different thing.

11. Now click on "Disconnect", and when it has done so (indicated by the button changing back to "connect" you can unplug the flight battery, then unplug the programming lead and then reconnect the PWM lead from the Rx (and the S-Port lead if you're using it).

I've managed to make that sound very involved - it isn't. The second time you do it it's obvious - it's just a series of very simple steps. I wasted a lot of time doing an Eric Morecombe - I was doing the right steps, but not necessarily in the right order...

PDR
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

RobC

It's worth looking at Rampup Power and Low RPM Power Protect too as some users have found this limits max power available in certain setups.  The first time I closed the throttle on my 1300W UltraStick I didn't know about Undamped Mode and the damn thing stopped so quickly I thought it would rip the nose off!
I think the smaller Neuron S versions come with the USB linker.  Unfortunately they don't come with a power lead either so you need suitable length heavy duty male to male cables handy.
flying's easy - it's getting it back down in one piece that's the hard part

PDR

I tried some power tests today and found that it seemed to be limiting at around 33 amps, so I suspect that I need to look at the ramp-up power parameter. Too warm to play with it today - I'll try it over the weekend.

PDR
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

PDR

New runs today with a freshly charged battery. I turned off the Low RPM Power Protect option and set the Rampup Power to 100%. I also turned on the data logging (couple to the engine kill switch, so it's logging whenever the throttle is live). This time it didn't mess about - serious grunt all the way to full chat, smooth running and barely breaking a sweat (temperature less than 45degC throughout).

So the Axi4120/18 on a 6s pack (two turnigy 40-50C 3s3000 packs in series) turns a 14-8.5APCe at 9,200rpm and around 58A which (by my sums) works out to around 1300watts (or 1.75BHP if you prefer). That should be enough to be going on with.

PDR
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

FlyinBrian

It would be a good idea if this post was saved in the "Tech Articles" section
Basic Research is what I do - when I don't know what I'm doing!.

PDR

One other thing I have learned about these controllers is that apparently it is recommended that the S-port connection (the telemetry lead) should be a single wire to the signal pin - no positive (red) or negative (black) wire. Apparently there have been reports of problems with earth looping. I haven't experienced any problem, but can see that it could well be "best practice", so I've changed by wiring accordingly. The S-port socket has no centre (positive) pin anyway, and removing the contact for the negative pin makes the connection a bit floppy, so cut and folded back both positive and negative wires right nect to the connector. I put some heatshrink over each folded wire to avoid shorts, and then put a larger piece of heatshrink over all the wires and the end of the plastic connector body to give it support. The unconnected negative contact grabs nicely and makes a much more stable connection.

PDR
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

PDR

So I've now been using the Neuron-80 in the AcroWot for a few weeks and I'm pretty happy with it. The telemetry data is giving me lots of insights with things like plotting BEC current against servo positions and throttle settings. I've learned that my aileron servos barely run above tickover, but the rudder servo is maxed out in flicks (so I may change it for a better one).

BHut I have been perplexed by minor discrepancies in the current consumption sensor. At the end of each flight I note the charge consumed from each battery (as reported by the ESC) and then compare it to the charge put back in by the charger. There was always a small amount more put back in than was taken out. I initially assumed this was differences in the chargers, and maybe some calibration error, but the amounts weren't random and some analyses suggested there was a trend. The discrepancy was small, but it annoyed me enough to trigger my OCD*.

Then I had a thought. What if the current and current consumption sensors were only measuring the current going to the motor, and didn't include the current consumed by the BEC? A few fag-packet calcs suggested that the discrepancy was about that sort of size. That would explain it, but I wanted prove this before I accepted it as an answer.

The Neuron ESCs have telemetry on BEC current but not consumption. But of course this is an OpenTx system and OpenTx allows you to do all sorts of clever stuff. My first approximation was to create a new virtual sensor which used the BEC current sensor as an input and integrated (a posh way of saying "added up") the current to give me BEC current consumption in mAh. I then created a second virtual sensor which had the main current consumption sensor and the BEC current consumption sensor as inputs and adds them to give a total current consumption. I then switched my "electric fuel gauge" on the Tx display - job done.

Except that it wasn't - the numbers were now over-reading. But I was sure I was on the right track, but I was missing something. I now had the data in the telemetry logs so I worked the numbers backwards. If my theory that the discrepancy was the BEC current was correct the correction seemed to be too much by a factor of around four. But why? I didn't want to apply a correction without a reason - that would just mean I didn't trust the instrumentation.

Then while driving back from Yeovil yesterday the penny dropped. The BEC was set to 6v, and the measured BEC current was the output current at 6v. But the current was being drawn from the 6s battery at ~24v and then chopped to drop it down to 6v. If I took the data and multiplied the BEC current by the ratio of the BEC voltage to the battery voltage before integrating it the numbers matched up to within a few mAh.

So I've now added a further virtual sensor that multiplies the BEC current by BEC/Battery voltage ratio in real time (so it doesn't get corrupted by the variations in battery voltage). Next time I fly I'll test this!

PDR

* I prefer to call it CDO because it's neater if the letters are in alphabetical order
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

RobC

I must admit that I hadn't noticed that discrepancy (not that I've flown much this year) but now you mention it I wonder why they did it that way?  In-line sensors would work on the total current draw from the battery, perhaps the Neuron reported ESC values should more accurately be described as Motor Current and Motor Consumption which are fine if you are using them for energy limited competition class but not much use for the average Joe in the street.
I'll be interested to see how your testing pans out, I may need to steal your ideas and start integrating total consumption on my Neuron setups.
flying's easy - it's getting it back down in one piece that's the hard part

Dave_S

I haven't used my Neuron ESC yet (it's earmarked for the Acrowot) but I'll read this thread again when I'm ready to start setting it up.
Dave S in West Oxfordshire

PDR

Quote from: RobC on October 23, 2020, 14:34:44 pmI must admit that I hadn't noticed that discrepancy (not that I've flown much this year) but now you mention it I wonder why they did it that way?

I guess you could argue the pros and cons of this either way. The in-line sensors can only ever report total current (unless they are installed downstream of an SBEC) where the Neuron has the ability to report both separately. If you're looking to do motor power testing then the motor-only current number is useful to have. Whichever way they did it you could extract all these numbers because the Esc has a separate current sensor on the BEC.

But it dawned on me this morning that I already know this (or should have, if only I'd paid attention). The EscA and EscC sensors have a resolution of 10mA(h), but the logs and the displays show zero unless the motor is running. The servos, Rx and ESC are drawing more than 10mA so if the EscA sensor was recording total current it should show up as a "hotel load" in this number. It's always obvious in hind-sight...

PDR
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

RobC

Yes, I have noticed on other ESCs the consumption clocking up even when setting up models on the bench but it completely passed me by that nothing changed when setting up with a Neuron.  I can't remember which version of OpenTX added BecV and BecA but I'm pretty sure my first Neuron (in a H9 UltraStick) preceeded it.
flying's easy - it's getting it back down in one piece that's the hard part