SMALL - espritmodel.com SMALL - Telemetry SMALL - Radio
Reply
Thread Tools
Old Mar 21, 2012, 06:33 PM
OlliW
Joined Sep 2009
1,583 Posts
The FT232 adapter isn't copmpletly covered in the manual, HolgiB uses the DIY adapter (maybe you can eventually fill/correct these parts ). You need to tell a bit more about details of your FT232R adapter you built
- you inverted the Rx and Tx line?
- you followed the AVRootloader procedure for connecting/flashing? Everything has to be powerless, you then press the connect or program button of AVRootLoader, and only then you power up the GA250.
- you used a 100 Ohms resistor inbetween Tx and Rx? If so, connect in parallel a 1N4148 diode (anode to the Tx side). With the latest GA250 versions the 100 Ohm alone doesn't always work.
- did you take care that the GA250 is indeed powered? Recall that the Ga250 can be powered ONLY through the red gain plug! (have a look at the photo here: http://www.olliw.eu/2011/ga250-gyro-firmware/#hardware)
- did you power the GA250 through the USB? If so, power it independentendly

The last two points are maybe most easily met by pluging both the rudder and gain plug into the reciever and to to power up the heli by connecvting the lipo to the heli's BESC.

My guess is that your GA250 doesn't get power or that the diode is needed.
OlliW is online now Find More Posts by OlliW
Reply With Quote
Sign up now
to remove ads between posts
Old Mar 21, 2012, 06:34 PM
OlliW
Joined Sep 2009
1,583 Posts
too late
OlliW is online now Find More Posts by OlliW
Reply With Quote
Old Mar 22, 2012, 01:38 AM
Registered User
Czech Republic, Olomouc
Joined Sep 2011
126 Posts
Quote:
Originally Posted by OlliW View Post
good question
actually two conditions must be met in order to enter the programming box mode
1) rudder must continously be hold maximally left for 10 sec
2) the gyro sensor signal must continously be zero for 10 sec
so, only if the plane turns extremly slow for more than 10 secs even with the rudd stick maximally left the programming box mode will be entered
if this can happen, then sorry man, you will crash
I think that both condition cant happen in one time. Not that long time. Dont know how high is the trigger for zero gyro sensor signal, but even slow wind initializes gyro compensation. (Imagine what happens when full rudder).
I hope so
By the way I found that ASSAN now doesnt mill chips signs. This could make gyro more reliable because chips dont suffer of vibrations during signs milling.
ablack is offline Find More Posts by ablack
Reply With Quote
Old Mar 22, 2012, 05:31 AM
Registered User
Joined Jan 2011
23 Posts
Quote:
Originally Posted by OlliW View Post
too late
Thanks for the extensive answer anyway. Sorry I posted before I tried the trivial options...
robbancs is offline Find More Posts by robbancs
Reply With Quote
Old Mar 22, 2012, 06:01 AM
OlliW
Joined Sep 2009
1,583 Posts
don't worry, no problem
glad that it works now for you
OlliW is online now Find More Posts by OlliW
Reply With Quote
Old Mar 23, 2012, 01:55 AM
Registered User
Czech Republic, Olomouc
Joined Sep 2011
126 Posts
Does anyone have experienced that sometimes after setting parameters by PC the gyro stops reacting to movements? It react only on stick movements.
I have two GA250 in plane each from different batch. Both sometimes stop to work after programming. When I power it off for longer time (about 2 minutes) it works again.
ablack is offline Find More Posts by ablack
Reply With Quote
Old Mar 23, 2012, 03:11 AM
OlliW
Joined Sep 2009
1,583 Posts
yes... since the 3.0V version I also sometimes have this experience...
I could not figure out what it is, it "feels" as if some charges build up or whatever, maybe the gyro sensor is not perfectly mounted or residuals are not well cleaned (some pins needs very high resistance environment). Maybe someone of you identifies the issue. I have no clue.
I do the same as you, I just have it sitting for some time, and then it's working perfectly again (I at least could not infer any issue so far).
This issue may happen ONLY after a programming step using the USB adapter (with AVRootLoader or AvrConfig, not the progbox)(the type of USB adapter, FT232 or DIY, doesn't seem to make a difference), but it never ever happend in routine operation.

If you look at AvrConfig in the tools menue you find a Diagnose option. If you use it, it will also tell you the gyro sensor mid signal. The above phenomenon was actually one of the reasons for introducing this Diagnose option. What you will see is that if the above happens the gyro sensor mid signal is not at mid but at full range... hence no gyro but only rudder reaction.

BTW: to activate the Diagnose you have to set Diagnose to a value larger than 0. The next time you look into this window you will see the values as determined in the last power on period.
OlliW is online now Find More Posts by OlliW
Reply With Quote
Old Mar 23, 2012, 03:42 AM
Registered User
Czech Republic, Olomouc
Joined Sep 2011
126 Posts
Thanks Olli,
if this issue has been uncover and the solution too, it should be OK. I was worried about sensor failure during flight.
ablack is offline Find More Posts by ablack
Reply With Quote
Old Mar 23, 2012, 06:35 PM
Registered User
Joined Jan 2011
23 Posts
I've done my first round of setting up the gyro values tonight.

My config is hk250gt with Henge MD922 on the tail.

I've followed the idiot's guide and let me tell first of all that I'M ALREADY IMPRESSED! This is far better than anything I've ever had with the original GA250 or the leaderhobby GY520. I'm not questioning that these work very well on bigger helis, or even some people managed to set them up properly on a 250 with a better servo, or even with this servo, but to me this is certainly the most successful attempt so far to make this little b*tch flyable.

Now I'll re-read this thread, and I'll ask my questions if I can't find the answers.
robbancs is offline Find More Posts by robbancs
Reply With Quote
Old Mar 23, 2012, 07:05 PM
Registered User
Joined Jan 2011
23 Posts
Second part of my post...

So I ended up with the following values:

PID I: 3
I'm hesitating between 2.5 and 3. Only in violent stops (if I pretty much release the rudder stick and let it center, which I never do normally while flying) I get some wag when stopping with 3. I can't reproduce it with holding the rudder stick, so I'll probably stick with 3

De-/Acceleration Limit Right: 20
I'm hesitating between 20 and 25. The interesting thing is that what happens in this range (with 20 just a tiny bit, with 25 it's more noticable) is not really a bounce but a quick stop and very slowly drifting to the opposite direction. And then settling down a few degrees (<5) later.
Again: this doesn't happen in normal flying, only violent stops.

De-/Acceleration Limit Left: 0
This will be one of my main questions. I couldn't really get rid of the backbounce in violent stops from this direction. Even with zero, it bounces back 10-30 degrees when I let the rudder stick go. In fact I didn't see much difference anywhere in the range between 0 and 30. Are you sure it's working?
I'm not particularly worried about this either because I don't do that normally, and it's hardly noticable at all if I move the rudder stick back to center "by hand".


Further questions:

I haven't started experimenting with the debounce values yet. How is the new debounce feature related to the deacceleration values? Is it either this one or that one, or I should find the right values for those at the same time? If I have to adjust both at the same time then what method would you recommend?

I don't understand the meaning of a lot of other parameters. Should I worry about anything else than these? For example, my Henge MD922 description says Dead Band <= 3us. Is it something I should set in one of the two dead band parameters? Which one?


This is it for today. So far I've only tested it in the living room with piros and some moderate pitch pumps. Tomorrow I'll take it outside and give it a proper go.
robbancs is offline Find More Posts by robbancs
Reply With Quote
Old Mar 24, 2012, 08:58 AM
OlliW
Joined Sep 2009
1,583 Posts
glad to hear that it seems to do something

may I ask first:
- what's your head speed?
- what transmitter/reciever system you are using?
- may I list you on my webpage as a user?

Next, MOST IMPORTANTLY, in order to test piro stop behavior DO NEVER EVER LET THE RUDD STICK JUST GO!
Do this while you are not flying such that you can look at what your stick does... you will see that it will "oscillate" around center before it stops at center! (that's with my transmitters, DX6i, EFlite RTF, ESky RTF). It should be obvious what this means, the tail would bounce at a stop just because that's that the transmitter tells the tail to do!
I do the piros by putting my finger at the base of the rudder stick.

Now to your questions. I do not own a 250, so I can only share the experience I had for my (two) 450, and what I think this means for a 250.

Quote:
How is the new debounce feature related to the deacceleration values? Is it either this one or that one, or I should find the right values for those at the same time?
That's a good question. The intention is that either this one or that one is used but not both at the same time. However, I left both for compatibility reasons, and to have the opportunity to actually test this out. Maybe it turns out that using both could further improve things, but I couldn't yet do that systematically (and it wouldn't simplify things since these are mutually dependent parameters). So, for the moment, I suggest to use either the debounce or the deacceleration feature.

For completeness:
the debounce feature is so to say inactivated with the parameters
Debounce CCW = 30
Debounce CW = 0
the deacceleration feature is so to say inactivated with the parameters
De-/Acceleration Limit Right = 30
De-/Acceleration Limit Left = 0

Quote:
The interesting thing is that what happens ... is not really a bounce but a quick stop and very slowly drifting to the opposite direction. And then settling down a few degrees (<5) later.
This suggest that your PID I is too low. I do see exactly this behavior with max deacc (=30) then the PID I is too small. That's BTW also "predicted" by theory.

What you should see (with both debounce and deacc deactivated) when starting at a low value and increasing PID I is:
- the tail starts to hold the heading in hover, but after a piro stop the tail slowly drifts back some degrees
- with increasing PID I the "back drift" after a piro stop gets shorter and shorter, until it produces a short quick backbounce
- with further increasing PID I the tail oscillates few times after a piro stop

To actually see this behavior your gain should not be too large (too large gain also produces oscillations!). So, in principle, for each new PID I value the GAIN needs to be readjusted. In practice I just change, for each new PID I, the GAIN a few % up and down and watch if an oscillations goes away or not.

The goal is to find the PID I for which you get one as short as possible back bounce without oscillations.

This point is actually more quickly found than it may sound now, in particular as you don't have to overdo it, it really doesn't matter if PID I is 8 or 9. That's more matter of taste. If you absolute bounce free is your goal be a bit more conservative (lower PID I) if you rather like to be on the grippy end be more aggressive (higher PID I).

Once you have found this one strong back bounce you got it. BTW the back bounce can be easily as much as 30 degrees! That's no point to worry. It should be one, and it should be as short as possible. The amplitude is not any issue. (it tells how fast your tail servo is, but that's not a variable)

If, however, the amplitude of the back bounce is very different for a CW and CCW piro stop, i.e., if there is a pronounced asymmetry, then you should investigate that! For my two 450 I found this to depend on quality issues such as the tail servo (you really see the performance of the tail servo here), as well as the mechanical setup. If you get significantly different amplitudes of the back bounce at this stage you really might wish to work this out first (the gyro can't totally correct the weaknesses of a setup!). It seems that for the TRex type of helis the stop after a CCW piro is generally nice, while the stop after a CW piro causes the troubles. But with "playing" with the mechanical setup I found it possible to significantly improve this. I in fact find it impressive how much can be achieved by just a little mechanics.

Having set PID I now (and worked out drastic mechanical asymmetries), you may go on and get rid of the backbounce either via the debounce or deaccel feature.

BTW: deacc = 20 I find very large... on my 450 I have something like 8-10... it is interesting that you can note a difference between 20 and 25, however mestre also reported large deacc values... interesting...

Quote:
I couldn't really get rid of the backbounce in violent stops from this direction. ... I'm not particularly worried about this either because I don't do that normally
but I DO worry... this should be better... and I know it can be better... and we will get this better

Quote:
In fact I didn't see much difference anywhere in the range between 0 and 30. Are you sure it's working?
hm... I tested this extensively for the v010 version (the first firmware release) but since then actually never used different deacc for left an right. I won't rule out that I made a mistake then integrating it into the v016 version. (v010 to v015 are so to say the "old" versions and technically identical). I will inspect this, thanks for pointing out!

Quote:
I don't understand the meaning of a lot of other parameters. Should I worry about anything else than these?
no, don't worry about them. Only the "main" parameters (hopefully) do matter.

I have them (still) exposed just for the case. Remember, this project is still at a very early stage and not much experience has been gained. However, that the project progressed you can see form the fact that these parameters are more "hidden" now than in the previous versions!

If you have a "good" transmitter though you may lower to Rudd Dead Band = 5, that's what I use with my DX6i.
OlliW is online now Find More Posts by OlliW
Last edited by OlliW; Mar 24, 2012 at 09:15 AM.
Reply With Quote
Old Mar 24, 2012, 10:48 AM
OlliW
Joined Sep 2009
1,583 Posts
Hey Rob,

Quote:
The only thing I notice now, which is barely noticeable, is while hover very still I can see the tail doing extremely small wags. ... you can't see it unless you are hovering very still, like in the garage, and look closely.
I looked today at my heli. In hover my tail also wanders around. It's not at a regular frequency or amplitude (as you would get for too large gain), it's kind of an irregular hopping around. It also seems to depend a bit on wind. Does this sound like what you observe?

Anyhow, I do have this since I started with a CP two years ago... i.e., I did have this with all other gyros and configurations I had before... and still have that... on both my 450er... so, I haven't actually noticed this explicitely... I just got used to that and accepted it as normal LOL.

I don't know what it is. I never spend anytime to investigate this because it didn't ever bother me. Since I have that since "ever" I think however that it's probably not related to the gyro (i.e. the alternative firmware). I am thinking that it may be related to the BESC... both my helis "jump" up and down a bit even with absolutely no throttle stick movement and no wind (in particular the cheaper heli does this). But as I said, I have no idea. Maybe someone of you know the real reason.

PS: I didn't yet find time to investigate your 60% wagging... we had bad wheather for quite some time so that I prefer to use the better temps now to fly
OlliW is online now Find More Posts by OlliW
Reply With Quote
Old Mar 24, 2012, 11:49 AM
Rob C.
United States, FL, Citrus Springs
Joined Nov 2007
40 Posts
Quote:
I looked today at my heli. In hover my tail also wanders around. It's not at a regular frequency or amplitude (as you would get for too large gain), it's kind of an irregular hopping around. It also seems to depend a bit on wind. Does this sound like what you observe?
Yes that is exactly what I am experiencing. I probably didn't really notice it before because I never looked so closely at what the tail is doing till I started testing your firmware.

On another note, I had the Power HD 3688HB on 2 of my 450's and I DO NOT recommend them. I had the brushes on the servo motor fail in fast forward flight on my HK 450 GT and I couldn't save it. I have it rebuilt and I am waiting on a new tail servos for both 450s. My neighbor just started getting into the helis while watching me. He built a HK 450 pro and had the same tail servo. His servo motor brushes gave out after about 10 flights of nothing buy hovering. Thankfully he saved it. He is using the GA 250 gyro with stock firmware.

If figured I would throw that out there because I mentioned that I use this tail servo during testing. Mine lasted longer than my neighbors, but the most abuse my servo took was just testing piros.

Rob
mestre is offline Find More Posts by mestre
Reply With Quote
Old Mar 24, 2012, 12:21 PM
Registered User
Joined Jan 2011
23 Posts
Thanks for the answers, it's most appreciated!

Quote:
Originally Posted by OlliW View Post
- what's your head speed?
- what transmitter/reciever system you are using?
- may I list you on my webpage as a user?
Head speed 4000
DX6i with orange receiver
Please feel free to list me on your web page


Quote:
Originally Posted by OlliW View Post
Next, MOST IMPORTANTLY, in order to test piro stop behavior DO NEVER EVER LET THE RUDD STICK JUST GO!
Do this while you are not flying such that you can look at what your stick does... you will see that it will "oscillate" around center before it stops at center! (that's with my transmitters, DX6i, EFlite RTF, ESky RTF). It should be obvious what this means, the tail would bounce at a stop just because that's that the transmitter tells the tail to do!
I do the piros by putting my finger at the base of the rudder stick.
Understood. I'll redo the whole thing without letting the rudder stick go


Quote:
Originally Posted by OlliW View Post
For completeness:
the debounce feature is so to say inactivated with the parameters
Debounce CCW = 30
Debounce CW = 0
the deacceleration feature is so to say inactivated with the parameters
De-/Acceleration Limit Right = 30
De-/Acceleration Limit Left = 0
Good to know. I've misunderstood, and put all 4 to 30 to start with. Then what I had in one direction was probably conflict between debounce and deacceleration. I'll redo tonight with the right starting values.


Quote:
Originally Posted by OlliW View Post
If you have a "good" transmitter though you may lower to Rudd Dead Band = 5, that's what I use with my DX6i.
Thanks, I'll try with 5. One very annoying thing with the leaderhobby gy520 was that my rudder stick didn't do anything in a few degrees range in the center. Was it this value that they got too high on that one?


I'll let you know how the second round of my setup will have gone.
robbancs is offline Find More Posts by robbancs
Reply With Quote
Old Mar 24, 2012, 12:39 PM
OlliW
Joined Sep 2009
1,583 Posts
@rob
thanks for the info on the Power HD 3688HB, much appreciated

@robbancs
Quote:
I've misunderstood, and put all 4 to 30 to start with.
oh, this would do the same! A zero (0) in the second parameter for each set (i.e. CW or Left) means that the value given for the first parameter is used.
So, e.g.
30 , 0 means in fact 30 , 30.
18 , 0 means in fact 18, 18
10 , 0 means in fact 10 , 10
but
30, 25 means 30 , 25
and so on

Quote:
One very annoying thing with the leaderhobby gy520 was that my rudder stick didn't do anything in a few degrees range in the center. Was it this value that they got too high on that one?
well, 10 seems to be what many gyros are using. So 5 gives you a bit quicker reaction than usual. It might well be that the LH gy520 uses more than 10... I will test that

for your next tuning tests you might wish to try debounce... (with deacc inactivated)
OlliW is online now Find More Posts by OlliW
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion A Free firmware project for the GA250 Gyro cruzado Mini Helis 344 Apr 29, 2014 04:19 AM
Discussion GA250 gyro issues rangamatchstic Electric Heli Talk 29 Jan 20, 2012 09:33 PM
Sold Assan MEMS GA250 gyro Jonathan Ott Aircraft - Electric - Helis (FS/W) 0 Jan 12, 2012 07:21 AM
Sold assan ga250 mems gyro $15 shipped ohsnap Aircraft - Electric - Helis (FS/W) 2 Oct 27, 2011 09:57 AM
Found ga250 gyro tcb22185 Aircraft - Electric - Helis (FS/W) 1 Sep 21, 2011 06:44 PM