Xicoy Electronic C.G. Balancer
07-24-2015, 09:57 PM
#201
I may have discovered the cause of my weight sensors giving random readings even immediately after calibration. It seems that due to a tiny error in design calculation or manufacture, the top case may be rubbing against the base or the pcb and thus affecting its readings. It also explains the huge difference in readings caused by moving the wheel slightly across the top of the sensor.
If I am right, then a tiny change of less than 1mm in the position of the bolt hole in the base is needed.
I really like this device and the angle sensors so I don’t want to return it for a refund, I want to make it work. I spent a lot of time last night carefully calibrating the sensors but one of them never quite weighed the correct amount by a few grams, and another sensor would over weigh by several hundred grams then not return to zero by a long way, then tare it and it would under weigh by several hundred grams then not return to zero by a long way, tare it and it would over weigh by several hundred grams, and so it just repeated that cycle.
Inside the unit is the measuring beam which is bolted to the base at one end and bolted to the top cover at its other end. Therefore the position of the holes in the beam, base and top cover determine the position of the cover relative to the base. If you turn the sensor upside down and look at the narrow gap between base and cover, on mine the gap is not even all the way around, at the end where the top cover is bolted to the beam the gap seems to go to zero. I can’t slide a piece of paper around the narrowest point of the gap so the gap is less than the thickness of paper and may go to zero, and I can also feel the paper snagging something. That something turns out to be the pcb inside the unit which not only sits ever so slightly proud of the base unit meaning it has more chance to rub against the cover, but at both ends of the pcb is a very small spike sticking out which gives it another chance to rub against the cover.
I very carefully sanded the edge of the pcb flush with the base including getting rid of the tiny spikes, then applied pressure along one end of the beam to try and push it away from the closed end of the gap as I re-tightened the base bolt, but it is manufactured very precisely so if it moved it was fractions of fractions of a mm. But it was enough! I can just about get a piece of paper to slide through the gap and now the sensor weighs exactly the calibration weight, returns exactly to zero, and does it over and over again.
This would also explain why some of my sensors give a huge difference in weight just by moving the wheel a couple of mm. The top cover is bolted at one side to the beam with a washer in between which means the position of a weight on the cover can rock the cover ever so slightly and thus affect the gap at the narrow side and thereby change the pressure of the interference between cover and base/pcb.
I will see if I can mill or file the hole in the base by about 0.5mm to make it a slot so that the beam and therefore the top cover change position by just enough to ensure a gap all the way around, and I will remove the proud edge and spikes of the pcbs to make sure they don’t touch and interfere with the top case.
If I am right, then a tiny change of less than 1mm in the position of the bolt hole in the base is needed.
I really like this device and the angle sensors so I don’t want to return it for a refund, I want to make it work. I spent a lot of time last night carefully calibrating the sensors but one of them never quite weighed the correct amount by a few grams, and another sensor would over weigh by several hundred grams then not return to zero by a long way, then tare it and it would under weigh by several hundred grams then not return to zero by a long way, tare it and it would over weigh by several hundred grams, and so it just repeated that cycle.
Inside the unit is the measuring beam which is bolted to the base at one end and bolted to the top cover at its other end. Therefore the position of the holes in the beam, base and top cover determine the position of the cover relative to the base. If you turn the sensor upside down and look at the narrow gap between base and cover, on mine the gap is not even all the way around, at the end where the top cover is bolted to the beam the gap seems to go to zero. I can’t slide a piece of paper around the narrowest point of the gap so the gap is less than the thickness of paper and may go to zero, and I can also feel the paper snagging something. That something turns out to be the pcb inside the unit which not only sits ever so slightly proud of the base unit meaning it has more chance to rub against the cover, but at both ends of the pcb is a very small spike sticking out which gives it another chance to rub against the cover.
I very carefully sanded the edge of the pcb flush with the base including getting rid of the tiny spikes, then applied pressure along one end of the beam to try and push it away from the closed end of the gap as I re-tightened the base bolt, but it is manufactured very precisely so if it moved it was fractions of fractions of a mm. But it was enough! I can just about get a piece of paper to slide through the gap and now the sensor weighs exactly the calibration weight, returns exactly to zero, and does it over and over again.
This would also explain why some of my sensors give a huge difference in weight just by moving the wheel a couple of mm. The top cover is bolted at one side to the beam with a washer in between which means the position of a weight on the cover can rock the cover ever so slightly and thus affect the gap at the narrow side and thereby change the pressure of the interference between cover and base/pcb.
I will see if I can mill or file the hole in the base by about 0.5mm to make it a slot so that the beam and therefore the top cover change position by just enough to ensure a gap all the way around, and I will remove the proud edge and spikes of the pcbs to make sure they don’t touch and interfere with the top case.
07-25-2015, 02:26 AM
#202
I was fortunate enough to get a set of scales that required no additional fiddling but just out of curiosity, is it as simple as removing the hex headed screw and lifting the cover off to access the scales internals? I tried removing the screw but the cover didn't want to come off easily so I just stopped before I ruined something.
07-25-2015, 09:34 PM
#203
Yes, removing the hex screw allows the top to come off.
I finally got my Dolphin on the scales. I guessed that it weighed about 38 lbs. The scales measured 37.48 lbs.
The CG was measured at 1mm from my hung CG. It's very important to level the plane as 1 degree from 0 changed the CG by 7mm.
The hardest thing is to get accurate measurements. Overall I'm happy that it works as advertised once the scales are working properly. Mine are still not perfect but are within a tolerable range.
I finally got my Dolphin on the scales. I guessed that it weighed about 38 lbs. The scales measured 37.48 lbs.
The CG was measured at 1mm from my hung CG. It's very important to level the plane as 1 degree from 0 changed the CG by 7mm.
The hardest thing is to get accurate measurements. Overall I'm happy that it works as advertised once the scales are working properly. Mine are still not perfect but are within a tolerable range.
07-25-2015, 10:44 PM
#204
My Feedback: (1)
Join Date: Jun 2005
Location: Hong Kong
Posts: 167
Likes: 0
Received 0 Likes
on
0 Posts
I finally managed to find some calibration weights. 2 of my scales seem ok. But the last scale is quite erractic. I placed 200g on it.
The highest I saw it drift to was 228g. So that's a 14% error. Not happy with that.
The highest I saw it drift to was 228g. So that's a 14% error. Not happy with that.
07-26-2015, 09:22 AM
#205
Just a question, by the time you tested the 3rd scale, had the unit been on for a while? Just wondering if the battery was perhaps getting a little low. I got the best results with a freshly charged battery powering the unit and when it started to get low I did notice some drift.
07-26-2015, 06:32 PM
#206
My Feedback: (1)
Join Date: Jun 2005
Location: Hong Kong
Posts: 167
Likes: 0
Received 0 Likes
on
0 Posts
Hi Wayne,
It had admittedly been on for a while. I will fresh charge my battery now and get back with the latest findings, but I'm pretty sure the battery was good, as I rechecked all the scales and it was only that 1 particular scale that 'wandered'.
I'll let you know...
Also, the other scales did drift slightly, but that said, it was only in the range of a couple of grams (2-4%).
Cheers,
Morten
It had admittedly been on for a while. I will fresh charge my battery now and get back with the latest findings, but I'm pretty sure the battery was good, as I rechecked all the scales and it was only that 1 particular scale that 'wandered'.
I'll let you know...
Also, the other scales did drift slightly, but that said, it was only in the range of a couple of grams (2-4%).
Cheers,
Morten
07-27-2015, 07:05 AM
#208
07-28-2015, 12:30 AM
#209
Attempted numerous times to save into the uSD card for backup of models and found it not working properly. I will not be able to find my 'saved' models anymore and remains so even when I remove the uSD card (while still powered). Once I cycle the power, my onboard models' data come back due to them being on the box memory.
I am on firmware V1.3
Does anyone else face this issue?
I am on firmware V1.3
Does anyone else face this issue?
07-28-2015, 10:55 AM
#210
My Feedback: (58)
Join Date: Dec 2009
Location: peterborough, ON, CANADA
Posts: 92
Likes: 0
Received 0 Likes
on
0 Posts
Your understanding of step 2 is spot on!
Step 3 is asking you to tell the machine where you are going to add or remove weight to get the aircraft to balance. This number is measured from the mains but is not very important; an error on this measure will cause an error on the calculated correction weight, but not an error on the calculation of the real CG location. In other words, move equipment or add/remove weight till you get the CG spot on.
Just a quick note; the gear on an A10, when raised will shift the CG forward so I would suggest also checking the CG with the gear up using the cones from Ultimate Jets or something similar.
Step 3 is asking you to tell the machine where you are going to add or remove weight to get the aircraft to balance. This number is measured from the mains but is not very important; an error on this measure will cause an error on the calculated correction weight, but not an error on the calculation of the real CG location. In other words, move equipment or add/remove weight till you get the CG spot on.
Just a quick note; the gear on an A10, when raised will shift the CG forward so I would suggest also checking the CG with the gear up using the cones from Ultimate Jets or something similar.
Thank you very much Cris. worked perfectly to your description and maiden was a success...
08-23-2015, 03:01 AM
#213
Member
Join Date: Oct 2005
Location: Abu Dhabi, UNITED ARAB EMIRATES
Posts: 41
Likes: 0
Received 0 Likes
on
0 Posts
measures
Hi,
Question regarding the measurements made from nose to mains, shall I measure at the edge of mains and nose or center of the nose leg to the center of main legs?
Question regarding the measurements made from nose to mains, shall I measure at the edge of mains and nose or center of the nose leg to the center of main legs?
11-04-2015, 02:53 PM
#216
Hey guys, just got my Xicoy CG balancer with angle sensors. One of my 3 scales was WAY off upon delivery - reading 9.45lbs with no load. I re-calibrated it, but it's still off by 200+ grams from the other scales and not always by the same amount.
Anything I can do to potentially fix it, or does it need to be exchanged? I noticed that the other two scales are assembled differently as well. The two working and consistent scales, have a washer between the top /con-caved resting plate where the hex bolt screws in. The wonky scale doesn't have a washer, and it seems like the white glue had been flattened down. By comparison with the other two, the white glue or silicone (whatever it is) on those haven't been flattened at all.
That shouldn't be this way on a brand new item... for a product that's all about precision, that's pretty sad.
Anything I can do to potentially fix it, or does it need to be exchanged? I noticed that the other two scales are assembled differently as well. The two working and consistent scales, have a washer between the top /con-caved resting plate where the hex bolt screws in. The wonky scale doesn't have a washer, and it seems like the white glue had been flattened down. By comparison with the other two, the white glue or silicone (whatever it is) on those haven't been flattened at all.
That shouldn't be this way on a brand new item... for a product that's all about precision, that's pretty sad.
11-05-2015, 01:23 AM
#217
My Feedback: (1)
Join Date: Feb 2002
Location: private, UNITED KINGDOM
Posts: 3,672
Likes: 0
Received 26 Likes
on
16 Posts
Hey guys, just got my Xicoy CG balancer with angle sensors. One of my 3 scales was WAY off upon delivery - reading 9.45lbs with no load. I re-calibrated it, but it's still off by 200+ grams from the other scales and not always by the same amount.
Anything I can do to potentially fix it, or does it need to be exchanged? I noticed that the other two scales are assembled differently as well. The two working and consistent scales, have a washer between the top /con-caved resting plate where the hex bolt screws in. The wonky scale doesn't have a washer, and it seems like the white glue had been flattened down. By comparison with the other two, the white glue or silicone (whatever it is) on those haven't been flattened at all.
That shouldn't be this way on a brand new item... for a product that's all about precision, that's pretty sad.
Anything I can do to potentially fix it, or does it need to be exchanged? I noticed that the other two scales are assembled differently as well. The two working and consistent scales, have a washer between the top /con-caved resting plate where the hex bolt screws in. The wonky scale doesn't have a washer, and it seems like the white glue had been flattened down. By comparison with the other two, the white glue or silicone (whatever it is) on those haven't been flattened at all.
That shouldn't be this way on a brand new item... for a product that's all about precision, that's pretty sad.
Next, see my post #194 on this thread for things to do to make a sensor work reliably and repetitively. The main cause of the problem is the top cover rubbing against the pcb and against the bottom case because the bolt holes are slightly in the wrong place so the top cover and base are not concentric but are slightly offset with too small a gap between them. The slot for the cable connector is also too small and can interfere against the plug.
Great concept but not well executed, however once you sort the problems it does work well and thankfully they are easy and quick to put right. If xicoy tweaked some of the cnc dimensions by just 0.5mm or less, the problems would not arise. I would rather have the device and spend the few minutes to sort the problems than not have the device, though I agree the problems should not be there in the first place. The angle sensors are brilliant for getting side to side travels exactly matching so that you don't get uncommanded roll e.g. elevons on a Eurosport, tailplanes, flaps etc.
11-05-2015, 07:08 AM
#218
HarryC,
Thanks for the reply. I did come to realize this - I've enlarged the opening on the top cover that surrounds the servo-wire plug lead with a file on all 3. I installed a washer in the "bum" scale, and between the two mods it's officially "fixed". The only problem I have now is moving around the object/wheel tends to offset the reading by a few grams.
Edited: now fixed it and it's setup almost 100% perfectly to the gram consistently across all 3 scales. Weighed my iPhone 6, and a JR Xport Duo ... each rated as weighing 129grams, and 57grams respectively. All 3 scales read the same values for both - iPhone 6 (64GB) came in at 128grams, and the Xport Duo to my surprise came in at exactly 57grams.
Think I'm officially happy with it
I also opted to use a 2s LiFE battery, regulated to 6v steady so there's no shift on the voltage to the units even after running for a good while.
Thanks for the reply. I did come to realize this - I've enlarged the opening on the top cover that surrounds the servo-wire plug lead with a file on all 3. I installed a washer in the "bum" scale, and between the two mods it's officially "fixed". The only problem I have now is moving around the object/wheel tends to offset the reading by a few grams.
Edited: now fixed it and it's setup almost 100% perfectly to the gram consistently across all 3 scales. Weighed my iPhone 6, and a JR Xport Duo ... each rated as weighing 129grams, and 57grams respectively. All 3 scales read the same values for both - iPhone 6 (64GB) came in at 128grams, and the Xport Duo to my surprise came in at exactly 57grams.
Think I'm officially happy with it
I also opted to use a 2s LiFE battery, regulated to 6v steady so there's no shift on the voltage to the units even after running for a good while.
Last edited by chorner; 11-05-2015 at 05:51 PM.
11-06-2015, 01:18 AM
#219
My Feedback: (1)
Join Date: Feb 2002
Location: private, UNITED KINGDOM
Posts: 3,672
Likes: 0
Received 26 Likes
on
16 Posts
chorner, try calibrating with a much heavier weight, I used a spare small bench vice weighing several kilos!
Before you edited your post you mentioned trying to build up the sealant on top of the sensor, I advise against doing that. I don't fully understand what the sensor is, but because the beam is bolted to the base at one end and the top case at its other end, I guess the sensor measures stress/stretching/bending at the middle of the beam. It will not measure a weight on it, and any weight transferred through it to the beam will not cause a stress/stretching/bending at that point for it to measure. I am guessing that is why the washers are added, the groove in the inside of the top case is not deep enough to give clearance for the sensor so the washer lifts the top case clear of the sensor.
Before you edited your post you mentioned trying to build up the sealant on top of the sensor, I advise against doing that. I don't fully understand what the sensor is, but because the beam is bolted to the base at one end and the top case at its other end, I guess the sensor measures stress/stretching/bending at the middle of the beam. It will not measure a weight on it, and any weight transferred through it to the beam will not cause a stress/stretching/bending at that point for it to measure. I am guessing that is why the washers are added, the groove in the inside of the top case is not deep enough to give clearance for the sensor so the washer lifts the top case clear of the sensor.
11-06-2015, 06:53 AM
#220
HarryC, I will try with a heavier weight as soon as I can get an accurate one. The only reason I used those two things to test is that generally speaking, I've found Apple and JR to be pretty precise in their listed weights. I calibrated using 1,5Kg object.The amount of weight you calibrate with does not matter, and if all sensors are reading equally that is the most important part. If the sensors are each off by 2 lbs, it still doesn't matter in terms of finding out CG, and whether it is tail or nose heavy. Those calculations are all reliant on equally accurate scales - and if they are all reading the same between 3-5 different objects you test, you're good to go. Be it 1kg, or 20kg. I do notice however that each scale is calibrated with a 20Kg weight to begin with from the factory (it's etched on the side of the beam).
Yes, before I edited the post I mentioned building up the sealant on top. I took this part out of the post for the very reason that I figured out how the sensor works, and that glob of silicone is simply there to stop the wires being crushed, or the top case meeting with the rest of the beam otherwise it throws off the results. Edit: do your wires run through the silicone on the top, or under/through the beam?
So how it works is akin to how a see-saw works. There's a bolt that comes up through the bottom to secure the mechanism in place, and hold the beam to that spot. This is the measurement/hinge point. It's in that same location that the sensor is mounted, and it detects the amount of pressure put on that "hinge line". The focal point or pressure point of the weight needs to be directly over that spot where the bolt goes through and secures the top cover. Make sure to loctite and ensure the cover cannot at all be shifted by hand or otherwise, as this can affect the reading. The gob of silicone is simply there I presume, so that in the case of a certain amount of weight applied, it won't warp the top cover as it's only being held in by that single bolt at one point. I'd imagine repeatedly heavy weights could eventually slightly warp that dish - I noticed the aluminum is very soft. I was able to file off half a mil with one or two single passes with my hand file.
If you trace the wires - you'll see where they connect to a sensor inside the centre rail, on the hole opposite to the inside hole that mounts the cover. So again, it relies on a sort of "see saw" action, and for that pressure point to not change for repeatable results. Get it setup right, the scale itself is very accurate. It's most impressive that I can actually read accurately across all scales with a very light object, down to 7grams I had available - much more so than a heavier weight.
Yes, before I edited the post I mentioned building up the sealant on top. I took this part out of the post for the very reason that I figured out how the sensor works, and that glob of silicone is simply there to stop the wires being crushed, or the top case meeting with the rest of the beam otherwise it throws off the results. Edit: do your wires run through the silicone on the top, or under/through the beam?
So how it works is akin to how a see-saw works. There's a bolt that comes up through the bottom to secure the mechanism in place, and hold the beam to that spot. This is the measurement/hinge point. It's in that same location that the sensor is mounted, and it detects the amount of pressure put on that "hinge line". The focal point or pressure point of the weight needs to be directly over that spot where the bolt goes through and secures the top cover. Make sure to loctite and ensure the cover cannot at all be shifted by hand or otherwise, as this can affect the reading. The gob of silicone is simply there I presume, so that in the case of a certain amount of weight applied, it won't warp the top cover as it's only being held in by that single bolt at one point. I'd imagine repeatedly heavy weights could eventually slightly warp that dish - I noticed the aluminum is very soft. I was able to file off half a mil with one or two single passes with my hand file.
If you trace the wires - you'll see where they connect to a sensor inside the centre rail, on the hole opposite to the inside hole that mounts the cover. So again, it relies on a sort of "see saw" action, and for that pressure point to not change for repeatable results. Get it setup right, the scale itself is very accurate. It's most impressive that I can actually read accurately across all scales with a very light object, down to 7grams I had available - much more so than a heavier weight.
Last edited by chorner; 11-06-2015 at 06:56 AM.
11-11-2015, 03:38 AM
#221
We have been applying sound engineering principles with these machines since day one and we had only one customer returning one scale due to a hard sensor failure over 100+ scales sold.
01-09-2016, 09:06 AM
#222
Senior Member
Join Date: Aug 2002
Location: Broken Arrow,
OK
Posts: 280
Likes: 0
Received 0 Likes
on
0 Posts
Sorry, but I don't think it's a good idea. In spite of appearance, lateral balancing is a lot more complex than it looks, and the fix might be worst than the problem.
Why? Because for 10 grams (for example) of unbalance, the fix should be applied at the wing panel lateral center of gravity, not at the wingtip, while on the scale, 10 grams are 10 gra.
let's say that
Why? Because for 10 grams (for example) of unbalance, the fix should be applied at the wing panel lateral center of gravity, not at the wingtip, while on the scale, 10 grams are 10 gra.
let's say that
01-09-2016, 09:15 AM
#223
Senior Member
Join Date: Aug 2002
Location: Broken Arrow,
OK
Posts: 280
Likes: 0
Received 0 Likes
on
0 Posts
Gs turnSorry, but I don't think it's a good idea. In spite of appearance, lateral balancing is a lot more complex than it looks, and the fix might be worst than the problem.
Why? Because for 10 grams (for example) of unbalance, the fix should be applied at the wing panel lateral center of gravity, not at the wingtip.
Let's say that, for a given wing panel, the center of gravity (which might be at the same position than mean cord, but not always) is 1 unit (or whatever) from the symmetry line (or thrust line), and the wing tip is 4 units away (3 more) from the center line. 10 grams on the wing tip, on the scale can look like the problem was solved, the plane is balance laterally. BUT, in flight, under load (say, 9 Gs turn), 10 grams will be multiplied by 3 (moment), multiplied by 9! In other words, you will end up with 270 grams at the wing tip!
In that example, you better fly with 10 grams lateral unbalance.....(or you correct it properly, but then good luck...)
Bernard
Why? Because for 10 grams (for example) of unbalance, the fix should be applied at the wing panel lateral center of gravity, not at the wingtip.
Let's say that, for a given wing panel, the center of gravity (which might be at the same position than mean cord, but not always) is 1 unit (or whatever) from the symmetry line (or thrust line), and the wing tip is 4 units away (3 more) from the center line. 10 grams on the wing tip, on the scale can look like the problem was solved, the plane is balance laterally. BUT, in flight, under load (say, 9 Gs turn), 10 grams will be multiplied by 3 (moment), multiplied by 9! In other words, you will end up with 270 grams at the wing tip!
In that example, you better fly with 10 grams lateral unbalance.....(or you correct it properly, but then good luck...)
Bernard
