Igor Burger's active timer scare

[CircuitFlyer]

I don't have an Igor timer or a Spin ESC but, this conversation got me thinking so I ran some bench-top tests with a cheap Chinese ESC, my standard timer and a bench-top power supply.  Here are the results of 5 different tests that I did, each one starting with the motor running in flight mode:

1) Positive power plug disconnected from the power supply and then reconnected in less than 0.5 seconds -- the RPM dipped a little and then came back up, the LEDs remained illuminated, timer kept running.  No serious issues.

2) Positive power plug disconnected from the power supply and then reconnected after more than 0.5 seconds -- The motor stopped, the ESC shut down, the BEC output and subsequently the timer also shutdown.  When the plug was reconnected both the ESC and timer did a power-up reset and the motor did not restart.

3) Simulated brownout condition.  The current limiting of the power supply was quickly reduced to only allow approximately 1A of current to flow, voltage unchanged -- The motor stopped, the ESC and BEC remained energized and the timer continued to operate in the flight mode outputting a throttle signal.  When the current limiting was lifted and full power restored the motor did not restart.  The power had to be removed and reapplied to initiate a power on reset to return everything to a normal start condition.

4) PWM signal wire (yellow wire) disconnected from the servo connector and then reconnected in less than 2 seconds -- Motor continued to run, timer continued to run.  No serious issue, operated as expected.  This particular ESC has a built in failsafe feature that will hold the last RPM throttle setting for a few seconds in case of a loss of input signal.

5) PWM signal wire (yellow wire) disconnected from the servo connector and then reconnected after more than 2 seconds -- The motor stopped, the timer continued to operate in flight mode.  After the signal wire was reconnected the motor did not restart but did emit the initialization beep.  The power had to be removed and reapplied to initiate a power on reset to return everything to a normal start condition.

My take away for this experiment is to not be so quick to accuse the timer of screwing up.  How the ESC handles power problems can make all the difference.  Brownouts from partially broken wires, bad solder joints or oxidized connectors are the worst, they can make really weird stuff happen.  The microcontrollers used in the ESC and timer are very reliable and don't just jump to any random location in the code.  Normally if the microcontroller has a hiccup that is not handled correctly it throws an error code and the program just stops, locks up and becomes unresponsive.  A power reset should return everything to normal.

ESC's are designed primarily for RC.  Do you want the motor to restart after an interruption in the battery power supply?  RC guys may say "yes", most control line guys would say "no".  We are stuck with whatever the ESC designers had in mind.

Food for thought: Active timers use a sensor to monitor centrifugal forces.  If during flight mode, the sensor reads 0 for more than 5 seconds or so, it's obvious something must have gone wrong.  Why not automatically shut it down just to be safe?

Bottom line - understand how your equipment operates.

[phil c]

So, we have traded holes in the fuel lines and a loose spray bar for tin whiskers and piezoelectrical discharges. Progress? 

I love this hobby - Ken

One big feature for fuel motors.  They rarely fail- only about 20-25 chunky metal parts that RARELY break.  There might be one fatigue failure or one internal fault in 100,000 hours.

Phil C

[Ken Culbertson]

One big feature for fuel motors.  They rarely fail- only about 20-25 chunky metal parts that RARELY break.  There might be one fatigue failure or one internal fault in 100,000 hours.

Phil C

That would be one failure in every 1,090,090 flights.  Lets say we have 3,000 CL fliers in the country.  Probably high but lets say that on average they put in 4 flights a week also probably high.  That would mean that only one flier would experience a failure about every two years.

Ken

[AMV]

One big feature for fuel motors.  They rarely fail- only about 20-25 chunky metal parts that RARELY break.  There might be one fatigue failure or one internal fault in 100,000 hours.

Phil C

The same low failure rate applies to ESCs and electronic timers in the electrical world.  It's just that the more likely point of failure is in the interconnections between the major components -- the wiring and connectors between the battery pack, the ESC, and the timer/sensor.  Poor connectivity among them is the equivalent of cracked fuel lines, loose spray bars, etc.
Any system is only as reliable as its weakest link.

 
-Andrey

[Paul Walker]

One big feature for fuel motors.  They rarely fail- only about 20-25 chunky metal parts that RARELY break.  There might be one fatigue failure or one internal fault in 100,000 hours.

Phil C

Yup  there's never crap in the fuel,  holes in fuel lines, taters on glow plugs, missed needle settings, loose motor mount bolts  and what else did I miss. If I can't put a good score on the board  it is a failure.  Maybe it's not 1/100,000...

[AMV]

...Here are the results of 5 different tests that I did, each one starting with the motor running in flight mode:...

Paul, I applaud your initiative.  I can't thank you enough.  This is an excellent experiment that gives us insight in the proper direction.

5) PWM signal wire (yellow wire) disconnected from the servo connector and then reconnected after more than 2 seconds -- The motor stopped, the timer continued to operate in flight mode.  After the signal wire was reconnected the motor did not restart but did emit the initialization beep.  The power had to be removed and reapplied to initiate a power on reset to return everything to a normal start condition.

I am still currently betting my rusty $0.02 that this is what happened during Matt's flight.
In my mind this is the MOST DANGEROUS failure mode due to large variation in behavior among the ESCs of various manufacturers when it occurs.  One model may restart the motor, another may not....yet another may wait a while and restart...who knows?  It's an unknown until you actually perform this test on a specific ESC.  I think the safest behavior should be for the ESC to shut down the motor indefinitely if the signal disappears continuously for longer than a certain time.  Require a power reset to return things to normal.  More on this below.

The microcontrollers used in the ESC and timer are very reliable and don't just jump to any random location in the code.  Normally if the microcontroller has a hiccup that is not handled correctly it throws an error code and the program just stops, locks up and becomes unresponsive.  A power reset should return everything to normal.

Thank you!  I cannot stress this enough -- computers in general, and microcontrollers especially are known as "deterministic" devices.  When our timers power up, their program starts running at the beginning.  There is no way for it to start in some random state.  However, from that point they can get into an unexpected state if they encounter an input that the program handles in an erroneous manner.  That's typically called a "bug" and the onus is on the software programmer to fix it.  I can cite a few such stories from the aerospace/defense industry.
On a power reset, microcontrollers restart at the beginning from a known state.  Soooo....I conclude that the symptoms that have been cited in all incidents such as Matt's here must stem from something else.  People tend to stand at a distance and wave off computers as "magical black boxes" that operate on some unknown magical principles.  That is not true.  They are deterministic systems.  They operate according to a set of extremely prejudiced, rigid laws.  Every fault that I have observed with them has always led me to a set of logical steps that make sense only AFTER I put in the effort to get to the root cause.  Blaming the symptoms on random behavior in the computer is just laziness and/or lack of understanding.  Paul here has made good progress toward such understanding.

ESC's are designed primarily for RC.  Do you want the motor to restart after an interruption in the battery power supply?  RC guys may say "yes", most control line guys would say "no".  We are stuck with whatever the ESC designers had in mind.

This is a good point.  Maybe we are not stuck, though.  This behavior can be changed in ESC firmware if we knock on enough doors.  I bet JETI, Castle Creations, etc., would be willing to listen if we were to lay out the issues and stress that this is a safety concern in Control Line.
For starters, why not start with these ideas:

• Interruption in battery power supply --> stop motor indefinitely until power reset.
• Lost throttle signal for at least N seconds --> stop motor indefinitely until power reset, even if signal reappears.

Such options for CL use could easily be added by ESC manufacturers in a firmware update, so there would be no need to run out and buy a new ESC.  For example, Castle already has a "Control Line governor" option under "throttle type"; that page seems like a logical place to add the above options in their CastleLink software.

Food for thought: Active timers use a sensor to monitor centrifugal forces.  If during flight mode, the sensor reads 0 for more than 5 seconds or so, it's obvious something must have gone wrong.  Why not automatically shut it down just to be safe?

Also an easy modification in timer firmware.    Updating the current hardware on the market, however...?? different story.

Bottom line - understand how your equipment operates.

AMEN.

 
-Andrey

[jerry v]

Considering carbon fiber props, high RPM, weight and materials in the modern electric stunters maybe it’s a good safety feature to have a remote electronic device for emergency motor shut off by the pilot or the helper. And the switch should be with that person during the time the model is armed, not on the control handle.

Jerry

[Brett Buck]

I note that this thread has Igor's name all over it, but most of the reported problems would be problems with *anyone's* system, not unique to Igor's system.

    Brett

[AMV]

I note that this thread has Igor's name all over it, but most of the reported problems would be problems with *anyone's* system, not unique to Igor's system.

    Brett

I also feel it is unfair to Igor.  This is not a timer's manufacturer problem.
Also, this thread would be more appropriate in the "Gettin all AMP'ed up!" board. 

 
-Andrey

[Joe Yau]

Igor Burger's active timer scare on August 03, 2022.

I have changed the Target Throttle (TT) value from 123 to 125. This was the only change.
Max. Throttle was and is 180 and Min. Throttle: 66.
Sensitivity: 50
Delay: 35 sec.
Time of flight: 5 minutes 20 seconds

I have used calibration = 001 after changing the TT from 123 to 125 and flew horizontally for 11 laps.
During the 12th. lap the RPM went up for a moment only and the motor stopped.

It is possible the Min throttle  = 66 setting is too low. I believe 80 is the lowest  you want to go,  as it could kick in the full brakes = full stop. (According to Igor)    So basically you have to lower the rpm settings on the jeti 66 /ESC  to a lower rpm setting, so you could increase the target  throttle up to around 150 for the desired lapspeed,  then set Max at 200, and Min at 80.. should be in the ballpark.

[Igor Burger]

I met Matt, but he did not tell me about it, so I am responding, now after my come back home.

Exactly this happened to Samantha in France and Orestes in Poland. In case of Sam, it was faulty XT60 connector, its male pins were soft (probably overheated during flights before, as the spring material is also contact material and that is bad design choice).  Sections of male pin were compressed (missing gap) and connector was very easy to connect and disconnect. Problem was solved by quick replacing.

Orestes had Deans connectors, they are good with nice large flat contact area and separete spring material. So far so good, but this time we flew on dusty place and one piece of sand can convert flat contanct to sigle line contact or even single point contact and vibration during flight can cause this problem.

I recommand power connectors with multisegment contacts wit separate spring. And of course, keep it clean, periodically check and time to time replace all connetros, including signal and do not forget also SWITCH and its leads which are soldered (can break inside insulation) - that needs periodical inspection.

To the security. I know about problem of restarting, but it is hard to solve. The best way is to keep good health of installation, but it is not only that what can cause it. For example if ESC overheates, motor will stop and when it cools down it will start again. The only solution is to have sensor of RPM, real time clock and flash or eeprom, (or mechanical memory like spring powered "memory" activated in such case, or something based on battery voltage) to see if something like that happened. Yes, I am speaking about price of the system.

May be I can invent some "additional" device so in case that we have something reliable, we can "add" it to existing systems. So far I have wifi cut off operable by android which can be hooked between my timer and ESC, it can replace this feature .

............. BUT ....................

It cannot be 100% safe, it is absolutely necessary act with electric models as they can any time start, if the battery is connected or not, human can do any type of mistake (for example that is why I have always model flipped on back, always pointed to safe direction etc. when I am not at it, also without connected battery)

[Igor Burger]

It is possible the Min throttle  = 66 setting is too low. I believe 80 is the lowest  you want to go,  as it could kick in the full brakes = full stop. (According to Igor)    So basically you have to lower the rpm settings on the jeti 66 /ESC  to a lower rpm setting, so you could increase the target  throttle up to around 150 for the desired lapspeed,  then set Max at 200, and Min at 80.. should be in the ballpark.

Theoretically hazard value is 50 (that is 1.25ms) so 66 is very close to that hazard value, but can work. I really recomment minimum 80. But I do not see anything wrong on 100 

[Igor Burger]

I note that this thread has Igor's name all over it, but most of the reported problems would be problems with *anyone's* system, not unique to Igor's system.

    Brett

Thanks guys, however at least it immediatelly attracted my eyes, may be if there is anything about my stuff, it is good to notify me in PM, so I can replay earlier. I was out this time, but I spoke with Matt, he vould get answer immediatelly 

[Paul Walker]

I have had numerous shutdowns over the years. Overheating, bad connections etc, but not one of them was anything other than user caused issues.

As Igor says, check them often!
My stoppage this year at the Nat's was due to the spring clip being missing. I changed the entire system on  Sunday, while talking to someone. The fact that this clip was missing went unnoticed. My bad!  The fact that it ran perfectly Monday through practice flights Friday morning lulled me to sleep on this. Then it just stopped on a critical flight.  Igors sysyem worked perfectly. No power, no run!!

As I said before, 10 years and running strong.

Thanks Igor

[Steve Helmick]

NEVER talk to anyone while they (or you) are doing stuff in the pits, whether fueling a slimer, changing a battery or something more involved.  Steve

[GERALD WIMMER]

NEVER talk to anyone while they (or you) are doing stuff in the pits, whether fueling a slimer, changing a battery or something more involved.  Steve

Good point , when flying with helpers especially flying combat or team race agreed handle signals are useful and everyone helping should know how to disarm or stop a motor/engine. I tend to get distracted by my teenage sons but my poor hearing helps lessen this now.

Regards Gerald

[Matt Piatkowski]

Hello Everybody,
The title of my original message was 'Igor Burger's active timer scare.'
After reading the answers and checking connection by connection and wire by wire in the entire electrical circuit in Big Red, I feel that the timer stand-alone did not contribute to the flight interruptions. Actually, I still do not know WHAT caused the problem.

I have replaced the ribbon cable between the timer and Spin 66 F2B with a brand new ribbon cable with AWG 26 copper wires.
The old ribbon cable does not show any signs of wear and tear and can conduct electricity in each wire bunch.
I twisted, bent, and stretched the old ribbon cable, but it maintained the uninterrupted conductivity.
There was no broken soldered connection anywhere.

I run several ground tests with the new ribbon cable and found no problem. I performed these tests with a 9x5x3 Graupner propeller drawing about 7 Amp. Each of these tests lasted 5 minutes and 20 seconds.

I will fly Big Red next week. Initially, I will fly 5 minutes 20 seconds long horizontal flights with small "jerks" up and down. If the system cooperates, I will try the pattern.

Question for Igor: when a mechanical switch arms the electrical system, the beep is heard. The loudness of this beep varies: sometimes it is loud and can be clearly heard, but sometimes the beep is barely audible. Why is that?
Thank you,
M

[pmackenzie]

Question for Igor: when a mechanical switch arms the electrical system, the beep is heard. The loudness of this beep varies: sometimes it is loud and can be clearly heard, but sometimes the beep is barely audible. Why is that?
Thank you,
M

The source of the beeps is the motor, acting as a very inefficient transducer.
The volume depends on the position of the windings relative to the magnets, which can vary depending on motor position.

[Dwayne Donnelly]

The source of the beeps is the motor, acting as a very inefficient transducer.
The volume depends on the position of the windings relative to the magnets, which can vary depending on motor position.

I've been flying electric cl/rc  over  10 years and didn't know that, you learn something new all the time. 

[Frank Imbriaco]

The only change I suggest for the Igor system would be for the motor to spool up slowly. Not the full blast RPM at take-off.
 I prefer the manner in which Will Hubin's (non-active) system gradually increases RPM. Otherwise, no complaints.

[Igor Burger]

I have replaced the ribbon cable between the timer and Spin 66 F2B with a brand new ribbon cable with AWG 26 copper wires.
The old ribbon cable does not show any signs of wear and tear and can conduct electricity in each wire bunch.
I twisted, bent, and stretched the old ribbon cable, but it maintained the uninterrupted conductivity.
There was no broken soldered connection anywhere.

I spoke mostly about power connectors, but for sure there can be also problem with low voltage lines and also on/off switch.

[Igor Burger]

Question for Igor: when a mechanical switch arms the electrical system, the beep is heard. The loudness of this beep varies: sometimes it is loud and can be clearly heard, but sometimes the beep is barely audible. Why is that?

Pat already answered it - but to be sure - there is no repro or beeper or anything like that on ESC, sound comes from motor and prop, so yes its sound can depend on actual position of rotor to iron core.

[Paul Walker]

The only change I suggest for the Igor system would be for the motor to spool up slowly. Not the full blast RPM at take-off.
 I prefer the manner in which Will Hubin's (non-active) system gradually increases RPM. Otherwise, no complaints.

That can be adjusted in the Jeti ESC settiogs, just like the Hubin timer.

[Igor Burger]

The only change I suggest for the Igor system would be for the motor to spool up slowly. Not the full blast RPM at take-off.
 I prefer the manner in which Will Hubin's (non-active) system gradually increases RPM. Otherwise, no complaints.

Paul already answered, but I will add that the settings of spin does not allow full slow ramping of RPM. We use range of say 8000 to 11000 RPM and there is stair from 0 to 9000, does not matter how slow you go with signal. So timer cannot do anything on this place and therefore also does not try to do it.

However Spin (I mean our C/L non pro version) spools up slowly on begin of run from 0 rpm to 8000, then there is adjustable speed from 8000 to flight RPM as Paul wrote. I send ESCs set to my optimal seetings what is compromis which I prefer. It is not good if motor accelerates too quik, but also not good if it wants to travel slowly at start on grass surface. And FAI rules have also defined minimal and maximal travel of model on start. Siply I found some setting which are OK for me, they are preset as defaults, but you can change it if you prefere.

However spool up from 0 to minimum is fixed property of Spin.

[Howard Rush]

You’d think more people would use TUTs.

[Brett Buck]

Paul already answered, but I will add that the settings of spin does not allow full slow ramping of RPM. We use range of say 8000 to 11000 RPM and there is stair from 0 to 9000, does not matter how slow you go with signal.

   Not to overstep, but I presume that you mean it is a "step function", that is, it goes from 0 to 9000 with no intermediate steps.

   While we are at it, what is the resolution/quantization, once you do have a set range? I presume that it is quantized, maybe 8 bits for the set range?  I feel free asking now, since I have one of the things...

     Brett

[Igor Burger]

Yes, sorry. Exactly said, it has rpm 0 from 1 ms to 1.25ms and 8000 to 11000 from 1.25ms to 2 ms. So there is abrupt change of rpm from 0 to 8000 at 1,25ms.

I am not sure with quantization, it has relatively complicated formula, because there is range of poles, range of RPM, range of gear ratio and if I multiply all, it will come to too large resolution. In any case timer itself uses 8 bits, but it does not mean too much since timer does not keep certain number as constant for a while, it makes something like saw signal and integration of motor and prop inetria will solve the rest and give better resolution.