I would not say, that it is sensor issue, but maybe, I do not see something what some other ones see. But the TM1000 use 3,3V and read it without problem, therefore I do not suspect any problem with it.
Believe me, I tried different needle settings, but nothing helped. Now, the engine is little more on rich side, because the governor overspeed it (I have more experience with gas engine tuning and Walbro carbs, where is general rule - higher RPM, more opened needles), therefore I opened the needle little more. I will not risk the engine because of governor tuning.
Because of rich side, the engine might boggy, but not overspeed. And as you can see - when I switch to V curve, the engine is not overspeeded, like it would be when is too lean.

Regarding the pitch - I fly about 11,5 or 12 of pitch, but it does not play any role for overspeed problem when I go to hover – I am able to reach better rotor speed control with V curve then with governor.
I tested throttle servo as well (without running engine), but it seems to be good. I will test the throttle servo without blades and e.g. at 14k RPM to be sure, that it is OK, but when it works with V curve, I do not suspect that it could be faulty…

Last idea is to install Futaba GV-1 governor and use it. I only would have to lie about gear ratio, because GV-1 is constructed to 2100 RPM on the head. I am asking about 2500 – 2600. There could be decreased RPM stability, but I think that it would be better in every case. Problem is, that I have to switch back to receiver and it means a half day of work (wiring, bidding, settings Spirit, radio, next tests fly, etc…)

Regarding the temperature – I placed the sensor out of main air flow stream, but as front as possible.
General rule is that the engine should have around 90 – 95C after landing – it means that about 100 – 110C during fly. But I do not tune it for temperature – I used power test like I use for gassers.
I use high output fan, it is quite cold, therefore the temperature do not have to be the primary parameter.

Well,
Because I was not able to tune Spirit GOV, I have decided to roll back to standard received and install Futaba GV-1 Governor (in standard settings - no modification in hidden menu).
I only lie with gear ratio to be able get 2600 at rotor (governor thing that there is 2100/maximum RPM for GV-1/, but real RPM is 2600).

The result was GREAT. See here (it was very windy):





I did not changed nothing (like throttle servo, linkage, throttle curve, needle settings /btw. during next fly, I have to more open the H needle, because the weather has significantly changed from last test (sunny and 17C versus, snowing/raining and 1C now)/). During this fly, the engine was on the edge between lean and ideal condition, I would say.

Based on these test, I have to say, that Spirit Governor (FW 2.0.1) just DOES NOT WORK for OS MAX 37 (in GAUI NX4).
I do not know how about 50/90 size helis, but here, Spirit Governor is not the way.

I am sad with it, because I wanted to save some space (receiver and governor), because I am going to upgrade to 56 gas engine and I will need some space for ignition and next battery, but there is no other way - Futaba is the way for me.

Maybe, it makes sense to teach the governor MAX, IDLE and STOP, like GV-1 needs...

@ZeXx86, if you want to cooperate on this issue, let me know. In other hand, it would be good to let know, which engines/helis were tested with the governor...

Sorry for delayed response and thank you for the videos!

I have checked it and I think I know where is the problem.

The governor itself is actually working probably in the latest video, from time it looked like there was no active governor - it was actually working now.
All other attempts are not working properly, because governor is in a mid-state.
Anytime when governor is really On, you should see the event - Governor was engaged in any transition from disabled state (TC is under 50% or disabled in a bank).
If you will not see this, governor was not prepared yet for some reason.

So it is actually not governing the RPM, but only trying to achieve requested RPM. This is reason why response there is very slow and you can't see too much of difference with different values.

So definitively in this case governor can't work and will not work until it is really engaged.

So the problem is very probably somewhere in the sensor - each governor is working better with a different sensor configuration and RPM range.
So while it can work with GV-1 it might not work with other for some reason (e.g. mentioned noise). GV-1 is designed for combustion motors, so there is well defined how RPM data are processed.
On the other hand configuration that is working very well might not work at all with GV-1.
Can you tell me exactly which Spektrum sensor do you use?

Did you tried to play with RPM sensor filter values?

_________________
Spirit System developer

I did not play with the sensor filtering.
At the beginning, I was connected to Spirit via phone and saw, that it reads RPMs well.
Once, I started engine with connected BT unit and achieved about 16k RPM (on the ground) and saw, that Spirit reads RPM well.
I do not use Spektrum RPM senson, but Align RPM sensor (from Align RCE-G600). I read that Align RPM senson should work with Spirit.
The Align sensor is powered from Spektrum telemetry (about 3,3V).

Align sensor works well with spectrum telemetry and is able to work with Futaba GV-1 (even though that GV-1 is showing “detecting magnet” when the magnet is not under sensor and showing “magnet is not present” when the magnet is under sensor /it is reversed/, but it works).

Frankly speaking, I saw that some settings in Spirit has some effect on rotor speed. E.G. When I was increasing Holding performance, there was bigger and bigger overspeed when I went from a maneuver to hover. When I was decreasing gov response, I saw that RPM went down during tictics, when I was increasing gov response, it holds better during maneuevers. Based on it and logs from Spirit, I decided that governor was enabled and “worked”.

I made some PRT SCS after one of the test from fly – see attached. There is mentioned that governor was activated and no other information about noise from sensor etc.…


Frankly speaking, I was in touch with a guy from Helifreak and he told me, that Spirit Gov did not work either in his GAUI NX4 (unfortunately, I got this information when I already switched to Spirit GOV – but I tried to test and share the results at least).

As I mentioned before, I would say, that it might work something like:
Gov response: 7 or 8
Holding performance: about “-2”
- or somehow teach the Spirit Gov, when is idle, when stop and when max. Maybe set there, that it is glow/gas and not electric and use different algorithm.

Based on the results, logs, videos, feedback from different users and workable Futaba GOV (and Align gov worked as well - but it overspeed it little as well - therefore I decided to change it) I still have to claim that the Spirit Gov does not work for OS 37 and GAUI NX4... It is able to hold RPM during maneuvers, but it overspeed rotor when you exit from maneuver.
Depends, if you want/are you interested to deal with it for, lets say, smaller group of users...

The later model Futaba governor shows the sensor signal strength when you are setting up the unit. does the GV-1 have this function - with an Align sensor ?

Yes, it does and I already checked it.
When the magnet IS NOT under the sensor, it shows me about 68% (from user manual for GV-1: 60% is minimum - but it shows when the magnet is not present, therefore I do not suspect that it would go under 60%). When the magnet IS under the sensor, it shows about 1% or 0%.
As I mentioned - the interpretation is exactly opposite between Align and Futaba RPM sensors - but it works for my purpose.
Align sensor works as it should with Spektrum telemetry (if magnet is under sensor, the LED lights red on TM1000)

Are you sure that the wiring is really correct?
Align sensor has black wire with +3.3V and red wire is GND (just in case you are powering it from the unit - which is not the current case).
From point of view of the electricity, it is not ideal as the current must go through additional devices and then to the unit.
In the unit the signal could be very different as it can pick much more noise during the "route".

_________________
Spirit System developer

Are you sure that the wiring is really correct?
Align sensor has black wire with +3.3V and red wire is GND (just in case you are powering it from the unit - which is not the current case).
From point of view of the electricity, it is not ideal as the current must go through additional devices and then to the unit.
In the unit the signal could be very different as it can pick much more noise during the "route".

I am sure about wiring and I know this fact about Align sensor. I found it on the internet and I measured it by multimeter for sure (to check that there is only 3,3V and not 6,6V from my LiFe calls).
I know that it do not have to be ideal, therefore I checked whether Spirit reads RPM and it seems that it was OK.
After each fly, when I changed the parameters, I checked the logs and I found information like Governor enabled (see my example of screen shots).
There was no information about noise, vibration, low voltage etc...
From the behaviour, I was sure, that the gov "worked", because when I set very high RPM in IDLE1 (gov off) and switched to IDLE2 (gov on), the gov slowed down the rotor after few seconds.
As I mentioned before, the governor was able to hold RPM in a hover (not is some configurations - there RPM were not stable in hover - they were "pulsating") and during maneuvers, but it significantly overspeeds rotor when I went to hover from power maneuever. I would say, that it was too slow.

Based on the tests:
- I was able to fix asked RPM on the rotor temporary;
- I saw RPM reagings from Spirit (only two tests were made - one during idle, second without blades at about 15 or 16k on the crank )),
- Spektrum telemetry reads RPM without any problems;
- no errors in Spirit logs - no message about noisy signal from sensor - therefore I did not deal with any filters;
- the governor showed behaviour according user manual;
- Futaba GV-1 works with the sensor even though works with oposite signal orientation;
- a next user confirmed me, taht was not able to tune governor for NX4.

...but only I could not go more down with holding performance (according description in user manual, I should decrease it more based on my helis response - see in my videos)...
Based on it, I am still convinced that the GOV in Spirit does not work for nitro/gas (at least for OS MAX 37 in NX4). I use Spirit GOV in MiniProtos and there works great (and there, I think I set holding performance to limit as well: 1 /but not sure now and it is different chapter - do not deal with it here/).

However, if you still think that it is in bad sensor, I could make a test for you. I do not want to rebuild the heli to satelites and Spirit governor /at least for now/ (I spend a half day of build and rebuild, 4 afternoons with testing, burned about 2 liters of fuel,...).
But I have a next free Spirit from heli, which I sold without FBL - I can install it in the heli, connect it via Bluetooth with phone/PC and measure the RPMs during fly. There will be 2 Spirits - one will be only for measuring of the RPM from the sensor. I could record it and share it here if you need it (to exclude a possibility of bad sensor)...

Thank you for the info.

It is possible the sensor is working well, but it might not in the whole range.
Sensor Noise detection is working only when the governor was engaged, so before it can't be visible in the log.

Sensor Filter is preconfigured to be nearly at the max value, so if there is no noise, then it is definitively better to decrease it.
Lower values will improve governor response speed.

I am certain it will work fine with your or any other motor when we will find the problem.
Also, we can decrease the gain in the firmware for testing purposes so that value of 1 will work as for example 1/5.

_________________
Spirit System developer

Sorry to chime in here. As nitro still seams to be a challenge in some configurations and both of you life in Czech, it might help if you meet to fix it. This might also allow to adjust conclusions presented on helifreak as well.

_________________
Oxy4max SpiritPro + HW // SOXOS 550, SpiritPro, HW120A V4, 920KV // SOXOS 600, SpiritPro, HW160A, Pyro650 // SOXOS Strike7, SpiritPro, HW160A, Pyro750 // Skywing Spirit Aero // FrSky Horus X10S + Taranis X7 /