Electric Raven update

[gjestico]

My Raven electric conversion has gone through some changes since it appeared in this thread:
http://www.rcgroups.com/forums/showthread.php?t=266753

After seeing Goodrums success with Hackers in his Raptor conversion, I got a C50-15XL and (2)TP 6000 5s3p Packs from him to try out. Gearing is at 19.53 to 1. I have also got an eagle tree data recorder that has been lots of fun to use.
My real point with this thread, is that I hooked up the RPM recording feature for some eye-opening results. I have configured the scaling so that the rpm's read out as headspeed on the heli. As many others do I use KC's great spreadsheet to work out power combos. I was interested how the real life RPM numbers work out compared to the computer predictions.
Below I have posted 3 screen shots from the FDR results. The first one shows the RPM's and Volts graph. Second is watts, and last is amps.
Interesting to note that rpm generally averages around 2000, but drops to a low of 1637 and drops to right around that level at any point of high load.
so thats a (Huge!) RPM drop of about 360 , not including the many peaks that are above 2100. The spreadsheet predicts a 140 rpm drop for this combo, but I supect drooping battery voltage under load is the culprit that the spreadsheet cant simulate. Speaking of drooping battery voltage notice how the volts and RPM sag exactly the same way.The heli does have more power than it did with the Tango moter, and doesnt "seem" to bog anywhere in flight. It will take Tic-Tocs all day long. Imagine if it only drooped to 1800 RPM. Plus sitting on the ground unloaded headspeed flares to 2100 plus, and that makes for those impressive pop-up launches that look so impressive on videos. However the headspeed decays fast.
Being a converting Glow heli guy, this wildly varying RPM is a little unsettling as I am used to governed .91 glow motors that will hold the head at +/- 10 RPM or so unless you really bog it. So my next test will involve trying out a governor on the machine and slightly upping the gearing a little to allow the gov some headroom to work and also to lift the low rpm droop a bit. Yes I am aware this may be hard on the ESC as it will be doing a lot of throttling, so first few flights i will check it often
Greg

[GGoodrum]

Very interesting, Greg. How did you setup the rpm sensor? I haven't found a convenient rotating spot that I can mount the magnets that would allow the sensor to get close enough.

I've found that the real strength of the SS is for relative comparisons of different setups. It is hard to get the "real world" numbers to match, but the differences do seem to track. I have noticed, for instance, that when I'm videoing a flight I can tell the rpm drop in the view finder because at 1800 rpm the blades appear stopped (due to the roughly 30 frames per second frame rate...) As the rpm goes up, the blades appear to start "walking" in a clockwise direction. The higher the rpm goes above 1800, the faster the blades "walk". If the rpm goes below 1800, the blades start going counterclockwise. Anyway, I'm getting so I tell what it is within about 50-100 rpm, and I'm also seeing higher-than predicted drops, but usually in the 200-250 range, not as high as what you have recorded, but then my "calibration" might be off a bit.

In any case, I think you will find the 10s2p PL pack holds the voltage up better than your 10s3p g2-lite setup which means that rpm drop you are seeing will be less. BTW, did you ever do a capacity test on those 5s3p packs? Your voltage curves seem a bit lower than the 10s3p setups we've tested.

Also, this looks like you are using TCs, is that right? This might be a great technique to document just how well a typical governor mode does in holding the rpm consant, and compare that with an "external governor, like the Throttle Jockey.

-- Gary

[RC Man]

It’s very interesting that the battery performs better later in the flight than it does at the beginning. It looks like the LiPo’s you are using must have lower IR as they heat up – that probably explains why LiPo performance is so poor in cold weather.

What ESC are you using?

Have you tried using a throttle curve to get a more constant headspeed?

[gjestico]

On the Raven, due to its two-stage reduction there is a convenient spot on the primary driven gear (55tooth) The molded gear has 8 "corners" molded around its rim that are ideal for gluing in 2 magnets spaced 180 apart.
Then I was able to simply glue the sensor to the motor mount facing the magnets. ESC is Hacker 77-O.
I havnt checked the capacity of these g2 lites. Perhaps that is something I should do. Yesterday was an ambient temp of about 60F, and the batts started out there. End of flight temp was 110F and the batts perform better at that temp.. Nice thing about that is the power seems very consistent throughout the flight, not the usual electric powered strong at the start and weak at the end.
Yep I am using a simple throttle curve of 100-90-85-90-100 on this. I mounted the governor last night, I can turn it on and off at will so if the rain stays away I may be able to try it later today.
Greg

[GGoodrum]

I've seen the same thing, re: pack temps. They really need to get up to about 90-95F to get the most out of them.

I'm definitely going to make another attempt at installing this sensor. I also have a Throttle Jockey that I may put on as well. I think it has an optical sensor, which might be easier to get setup. I'm VERY curious about how the 32.55's govornor mode compares to the TJ. I know the 77-O's governor mode doesn't work worth a damn.

--Gary

[gjestico]

Well it wasnt raining too hard so I rushed home and got a flight in the quickly fading daylight. TJ givernor was installed, I flew the first part of the flight with it off, then landed and took off again with it engaged. Looks like it works very well, you can clearly see where it is active, holding the headspeed right at 1860~.
I did some hovering and then worked it abit with some tic tocs, then at time index 400 I deliberatly entered a full negative descent, which usually results in a huge RPM flare. The gov allowed it to come up a bit but much better than without it. Ended the flight with a minute or 2 of hovering as I figured this would heat the speed control up the most, as the machine hovers at 2050 without the gov, so thats about 200 RPM being knocked off by the governor due to throttling back the ESC. Measured temp on the ESC at 92 degrees. I figgered it would be hotter !. So this begs more testing and when the weather co-operates I will do.
I think I will try the governor trick on my E-pred as well. Among many reasons, one nice thing about having a constant head speed is that you can maximise the gyro gain for good tail holding without having the gyro freak out when the RPM flares by a few hundred.
Gary, One nice thing about the Data recorder RPM sensor is that is uses the same type of sensor as the Throttle Jockey governor does. So once you have found a spot to mount the magnets they can do double duty on RPM recording and gov control. I have both sensors installed so both systems can work at once.

[GGoodrum]

Wow, this is pretty amazing stuff! TheTJ definitely looks like it's doing a great job of holding the RPM. Now we just need to see what a typical governor mode "curve" looks like. I think I have the sensor/magnet issue solved,and it is great news that they share the same sensor. I think mine came with an optical sensor. Do you know if thereare separate versions, or do you think this one will also work with the magnetic sensor?

Once I get both the FDR rpm, and the TJ operational, I'll try getting the same sort of data for the 32.55's governor mode, vs. the TJ. In my present setup (boxed 1515-2Y, 86T/9T. 10s2p PL...), the 100% headspeed is about 2250. The governor is dialed down to about 2050. This will be interesting.

-- Gary

[RC Man]

How about some plotos.

The CSM governor is quickly becoming more popular than the TJ – why don’t you test one of them also.

[GGoodrum]

Well, I have the TJ in hand, but if it turns out it works significantly better than the internal Schulze governor, it would be good to give the CSM version a try as well. I'm not familiar with this one. Is there something different about how it works?

[RC Man]

The CSM RL10 is suppose to be cheaper, more features, hold RPM better, and easier to setup – but that’s what they all say. It’s also very small and light weight. http://www.rchelisplus.com/CSMRL10displaypage.htm

[misskimo]

hey I wonder if you can put a heat prob on there to see what the esc temps are like


Tony

[GGoodrum]

Yes, the FDR has two temp channels. I'll try adding one on there, although the 32.55 runs fairly cool. After an 8 or 9 minute flight the other day, the 2Y, the 32.55 and the pack were all about 110-120F, about the same as a half-hour old cup of coffee. The only time I've seen the 32.55 get hot was the first time we tried the 1.5Y with the 10s2p pack in my XeroG-based Raptor with the 95T/8T gearing, which spins the motor at 50,000 rpm. This was the "autoland" flight, however, where the receiver died completely and the motor wouldn't shut down. It ran at like 1/4 throttle for a couple of minutes, so I don't know for sure if it got hot because of the 50,000 rpm, orif it was the partial throttle run-on at the end. We flew it again last weekend but the flight was cut short when the tail pushrod let go (it was evidently weakened/damaged in the autoland "incident"...). The flight was only about 4 minutes, but neither the 32.55 or the 1515-1.5Y were any warmer than normal.

Actually, whether the ESC's internal governor mode is used, or one of the external varieties are added, the ESC will still be working partial throttle, so the heat, if there is any, is going to be the same. Since these setups are mechanically optimized, the amount of bogging that needs to be "governed" should only be around a couple hundred rpm.

-- Gary

[misskimo]

I can tell you one thing , that Graph sure tells a lot of good info , I bought one last week to put on my fury E , I been running it on gov 100% though , with 13s , the headspeed didnt change that much ,

[gjestico]

One note : the Throttle Jockey and FDR use the same sensor, IE the same actual electrical component, but even though both units' sensors use a 3 wire plug that is identical to the average 3 wire servo plug, they are wired differently. Thus one cannot use the same sensor interchangeably between the units without some sort of adapter harness. My solution was to mount the respective sensors on each side of the gear, but they read the same spinning magnets.
Greg

[GGoodrum]

Ah... good to know. Thanks! I'll do the same.