Question about Castle data logging

[longballlumber]

Hello fellas,

With the recent interest in the conversion of the Castle (AIR) controllers, I have a different question for those who have some experience in the logging feature of these controllers.

I will try and set up the situation without getting too long winded.

Earlier in the year, I converted one of the new Castle Edge (Phoenix) Lite 200 controllers to water cooling. The original application was for a rigger, so space was tight. In the process of using this controller I really enjoyed the data logging information that I was receiving. Because of the information I was getting, I decided to run the controller in my P-Limited Sport Hydro at the Nats. At that time I was seeing 80A-85A average in a heat. Oh, and the controller would get around 150-160 degrees in temp (200 amp version)

Fast forward - because I liked the setup so well, I decided to purchase a few of the 100 amp controllers and convert those. I didn't change anything in the boat. Same prop/motor/batteries… yada yada yada… Well, the 100 amp controller is logging 10A (plus) more in a run. The lower rated controller would also get hotter as well; around 190-200 degrees (at its peak)

What would cause such a drastic difference between controllers? Certainty the IR of the controllers itself wouldn't change the actual amp draw that much would it? Logic would tell me the difference in heat could be a factor. I am not sure. I am no EE by any standard, but I would still like to know why such a drastic difference in Amp draw numbers.

There has also been some discussion about the castle stuff not having adequate caps for our boating applications. Again, logic tells me the additional caps are to help the highly variable amp draws (spikes). I would expect an airplane to have a much smother discharge curve. Now that I think about it a helicopter being tossed around like the extreme 3D guys fly you could see similar types of discharge curves. I don’t know if the caps make a difference in the amp draw logging, I am just grasping at straws now…

Thoughts? Input? I am all ears.

Later,
Ball

[kfxguy]

I'm no expert, but I'll give my opinion or thoughts should I say. Maybe the lower amp version has:
Smaller wires? That's a factor there
Lower amp circuits..that don't conduct enough current (think smaller wires. To that effect)
Then your running it to its limits more? Seems like that would heat it more for sure. That's just some ideas I have as to why.....

I once tried an esc that was a mere 25 amps lower in its rating than the usual esc I was using....the lower rated esc heat up much faster and ran a good bit hotter. So I've experienced the same thing myself.

[longballlumber]

The heat increase isn't the main concern. Actually, it's something I would expect. I more interested in understanding why there is such a big difference in the amp draw values.

Later,
Ball

[kfxguy]

Does it have smaller wires? I think that would be a factor....

[RaceMechaniX]

Mike,

I run both a Castle Ice 200 with cooling mods and a ICE 100 for P-LTD. Your statement about the caps is correct although for a L-LTD application the ripple current is pretty low unless you are throttling it a bunch. The IR does contribute to heating as does the square area available for heat transfer to the air or water.

I just checked a datalog from the FE nats in P-LTD sport. I ran on average 85A and my controller temp was 289F at the end of the run. I am only using a small fan to cool the ICE 100, but it works. If you are measuring 200F from the datalogging, this is fine.

You can see the difference in quality, quantity and location of caps by looking at the ripple voltage in the log file.

TG

[longballlumber]

Mike,

I run both a Castle Ice 200 with cooling mods and a ICE 100 for P-LTD. Your statement about the caps is correct although for a L-LTD application the ripple current is pretty low unless you are throttling it a bunch. The IR does contribute to heating as does the square area available for heat transfer to the air or water.

I just checked a datalog from the FE nats in P-LTD sport. I ran on average 85A and my controller temp was 289F at the end of the run. I am only using a small fan to cool the ICE 100, but it works. If you are measuring 200F from the datalogging, this is fine.

You can see the difference in quality, quantity and location of caps by looking at the ripple voltage in the log file.

TG

Thanks for the reply Tyler, I was hoping you, Brian, or Jay would chime in...

What I still don't understand is the difference in amp draw number between the 2 different controllers... I like the 85 number using the 200 amp controller however, the 95+ amps I am seeing with the 100amp controller is what is confusing to me. Why are they so much different just by swapping controllers? What amp draw value is more realistic or closer to the "real" value. In P-Limited classes the difference between those to number could be the difference between cooking something or not cooking something...

Oh by the way, Castle uses the same gauge wire in all of their controllers of that particular line (Phoenix Edge Lite)

Later,
Ball

[Whitey]

Mike,

The difference in current readings between the two controllers may not be as bad as you think.

The actual current can be 90A. If it is, the 100A controller would be reading +5% (of range) and the 200A controller would be reading -2.5% (of range).

+/- 5% is not really all that bad.

The data loggers on these controllers are not pieces of precision test equipment. It may be possible to make the accuracy higher on the loggers but you can be sure that the cost would go up along with that accuracy.

Has anyone ever seen the current reading accuracy of the data logging function of these controllers specified? Not on the Castle website. Nowhere that I can find.

Has anyone ever seen the current reading accuracy of an Eagle Tree data logger specified? On the Eagle Tree website or anywhere? I haven't found any spec. There might be somewhere but I can't find it.

The point is that these are good comparison tools but, they are not precision test instruments. And, the manufacturers are not trying to present them as being precise. Hence, no accuracy specified.

This is also why 10 people, testing 10 motors, with 10 different data loggers (as on the P-Ltd replacement motor testing), is not going to prove much. 10 Motors tested on one single data logger will provide useful comparison information.

[Fluid]

IIRC the amp reading is NOT a direct reading, but rather is calculated from voltage drop. For me the
logging is more for comparisons rather than absolute values. That said when I compare the average amp draws and the amp hours used they usually compare quite favorably to the average amp draw seen when charging the packs after the run. This tends to validate the calculated amp draw as reasonably accurate.

More FETs means less resistance (superficially at least) so it does make sense that the 200 is more efficient. But the actual reason us...unknown.




.

[raptor347]

Mike,
There may be a setting that is different between the two controllers. The heat is expected. The current may be something as simple as the default settings being different between the two generations of firmware.

It would be worth looking at all the settings.

Is there a difference in motor temp between the the two esc's. You would expect an increase with the rise in current draw.

[kfxguy]

Is the timing set the same? Just throwing that out there.

[longballlumber]

Thanks again for the replies guys...

Mike,

The difference in current readings between the two controllers may not be as bad as you think.

The actual current can be 90A. If it is, the 100A controller would be reading +5% (of range) and the 200A controller would be reading -2.5% (of range).

+/- 5% is not really all that bad.

The data loggers on these controllers are not pieces of precision test equipment. It may be possible to make the accuracy higher on the loggers but you can be sure that the cost would go up along with that accuracy.

Has anyone ever seen the current reading accuracy of the data logging function of these controllers specified? Not on the Castle website. Nowhere that I can find.

Has anyone ever seen the current reading accuracy of an Eagle Tree data logger specified? On the Eagle Tree website or anywhere? I haven't found any spec. There might be somewhere but I can't find it.

The point is that these are good comparison tools but, they are not precision test instruments. And, the manufacturers are not trying to present them as being precise. Hence, no accuracy specified.

This is also why 10 people, testing 10 motors, with 10 different data loggers (as on the P-Ltd replacement motor testing), is not going to prove much. 10 Motors tested on one single data logger will provide useful comparison information.

Whitey - your absolutely right about the "assumed" accuracy/precision. I never thought to reach out to the company or check their websites for specifications of the data logging feature. I do have an eagle tree unit. I should try and run the eagle tree with each of the controllers and compare those logs and see how closely they match. Assuming your correct on the +/-5% for the error between, a delta of 10% still seems high to me, but I understand what your saying. At best these things help establish a baseline showing the effects of changes in the test eviroment assuming you keep the same logging device (the controller in this case.)

IIRC the amp reading is NOT a direct reading, but rather is calculated from voltage drop. For me the
logging is more for comparisons rather than absolute values. That said when I compare the average amp draws and the amp hours used they usually compare quite favorably to the average amp draw seen when charging the packs after the run. This tends to validate the calculated amp draw as reasonably accurate.

More FETs means less resistance (superficially at least) so it does make sense that the 200 is more efficient. But the actual reason us...unknown.

Your right Jay, In the back of my mind I knew the controller was actually measuring something else other than actual amp draw. It also makes sense they are using the battery voltage to do this. That is the same way many of the chargers on the market give the user (what I would call) "user friendly information". Taking that measurement and converting it to something usable for the operator. Where I went wrong is assuming the accuracy of said calculation method would be more consistent regardless of the controller rating. You've also made a good suggestion by paying more attention to the number of mAh's going back into the batteries. Paying more attention to those values would also help validate some/all of the logged information.

Mike,
There may be a setting that is different between the two controllers. The heat is expected. The current may be something as simple as the default settings being different between the two generations of firmware.

It would be worth looking at all the settings.

Is there a difference in motor temp between the the two esc's. You would expect an increase with the rise in current draw.

Hi Brian,

I would need to double check the controllers again, but I think I created a profile for my P-Limited boats and saved it. Then I could/can load the profile knowing each controller has the same profile. I would need to double check it though. I don't recall a huge difference in the heat of the motor between the two controllers. I agree with you, I would expect a large heat difference between the low 80's amp draw compared to a high 90's amp draw. I also wouldn't expect those two curves to be parallel. The heat should start spiking relative to the amp draw (if it was actually plotted on a graph).

Is the timing set the same? Just throwing that out there.

See my comment above about creating a profile. When you load these profiles, they save all of the controller settings into one file. That allows you to load the particular profile without going through each option.


All of this dialog reminds me of a recent conversation that I had with Mr. Newland. I was trying to be very aware of my amp draw numbers, by using data logging. He preferred to put more weight on the temperature of the components, which he nicely stated in his P-Limited motor blog. It's getting to the point where I might need to come up with a "test report" if you will. Using ALL or more of the suggested items in this thread to come up with a more complete conclusion.

moral of the story - While its a good start and it shouldn't be ignored, data logging isn't telling the "whole" story. There are still many variables that need to be considered.

Thanks for the help guys.

Later,
Ball

[kendt]

As I recall GRS (Gil Sonsino) has posted in the past that he found the castle data logging to be unreliable..Maybe you could shoot him a PM.