Thread Tools
Sep 11, 2015, 02:33 PM
Registered User
Sure send it to me and I'll flash it for you and bring it along October 11th at the next race.
Sign up now
to remove ads between posts
Sep 14, 2015, 02:00 AM
Registered User
Don't forget a controller that is not running full throttle heats up more than controller in part-load (as your governed controllers are doing). When running part-load it's recommend to keep at least 20% headroom in the current rating, to deal with the extra heat this generates.

The full throttle correct is the leading factor, not the governed current. While reducing power to stay at the right rpm, the controller is still switching the full throttle current, just very fast to get to a lower average power.

Especially when using a BEC it would be prudent do leave some headroom, a burnt controller would mean a crashed model.

Sep 14, 2015, 01:27 PM
Registered User
One of the reasons I am happy to do the software is that I want to try some algorithms to reduce the part load power, especially at high throttle percentages such as in the regulation region of E2K, or in the last part of a ramp into base A for F5B. This will require testing with a temperature measurement in defined circumstances.

I was told in Bad Bruckenau that it was worthwhile to use a 550gm battery rather than a 450 gm one (minimum allowed) because the 550 gm one had more spring at the end of the ramp. This could be emulated by a very slight third ramp if the efficiency was high enough.

Sep 18, 2015, 12:47 PM
Registered User
So the red bricks arrived yesterday and as I "unpacked" the ESC, I noticed a couple of things:
To start with, the ESC is quite bulky and as soon as the heath shrink came of it shows why.
First of all there is quite a big red heatsink on the FET side. Secondly, on the other side is a small alu plate which also acts as a heat sink for the LM2596 step down converter. This must be the device that powers the receiver. Strange thing is that this IC is specced for 3A max while the outside label of the Brick says 5A BEC... really? The alu plate is tacked to the board with a little bit of heath glue. The notch of glue is covering on of the corners of the controller chip, which is an Atmega 8A (I was happy to see that). So I put my nail under the notch of glue and it came right off so giving me access to the full controller chip. I wonder if I can't do without this alu plate? After all output of the chip is 5V and input with 4S is max 16,8V while the LM2596 can handle upto 45V max. So do I really need this additional heatsink? I am going to try without it (I will do a static test first running it on the ground for 90 seconds). Also the heatsink on the FET side seems quite bulky, but for now I will leave it.
Finally some pictures to illustrate all this talk:

Last edited by ragbak; Sep 22, 2015 at 08:35 AM. Reason: Heatsink foto added
Sep 18, 2015, 02:29 PM
Registered User
Nice! I like pictures.
As you already concluded, if the stepdown voltage is from 14 to 5 in stead of 45to 5V the dissipation is probably less. Perhaps that is also why they can spec a 5A bec, because 6S is only 25.2V max in stead of 45V.
A picture of the FET / heatsink would be nice too.
I stripped the heatsink from the CC Talon90 too. For size/ weight. And it ran fine this season, but not very surprizing at just 40A.
The heatsink is below the heatschrink right? Then it won,t be very effective I guess.

still have to order the 60 yep....

Sep 19, 2015, 01:15 AM
Registered User
I have not received my order yet.
It looks as if it has two regulators, a switching one at the top for the BEC and a linear one on the bottom left for internal use. No need to cool either of these. It seems to have only a voltage doubler.
I don't think cooling will be required on the FET side either, but that is what we want to try.
I have a similar (but different) one which has the flat plate moved to the FET side.
Have you seen the thread?
This is yet another board, but it has a number, is similar, but different from the 2P0603A1 in that thread.
As much of the layout is the same as the 2P0603A1 there is a strong possibility that it will use the same software.
But you should try with a lot of care. HambaWena may have been lucky or careful in trying all these firmwares! You need a current limiting supply or a series car bulb to limit the current, in case an upper and lower FET get switched on at the same time.
If it works, however, it will save a lot of time in tracing the circuit!
Sep 21, 2015, 01:19 PM
Registered User
Actually I am not quite sure about the LM2596S device. I took a look at the datasheet and this is a TO-263 package (indicated by the last letter 'S'). The thermal resistance is specified 50 degrees C per Watt in worst case (if mounted on a small PCB surface, like it is on the ESC). There is also a section about thermal considerations, it shows this picture:

Note that these measurements are valid when the device is mounted on top of 9 square inches of solid copper plated PCB! So this is in no way a worst case scenario plotted here. But we will see, I can simulate extra current draw at 5V in a static situation and measure the temperature.

B.t.w. I do have a power supply that I can limit in current so that should be OK.

Sep 22, 2015, 02:34 AM
Registered User
Thread OP
I would not worry about that regulator too much. Its only for the ESC itself. There is a separate buck converter using the large inductor for the BEC. Power loss in the controlling FET is minimal.
Sep 22, 2015, 04:35 PM
Registered User
I got my red brick at the weekend.
Very nice and neat once the heat sinks are removed.
We don't need them. We run the three board version at 200A without them, though I met someone at the weekend who had a fire at 300 - 400 amps!

The connections are indeed the same as the 2P0603A1 and the software ran. But the timing is retarded at 15000 rpm on the 14 pole I test on. There are three small capacitors in the phase voltage circuit. Normally these should be removed for simonK software and my software. They are not there in my 2 board version, but I can't remember if that was original or if I removed them myself. They are not good as they cause the timing to vary according to the rpm. Presumably they have been put there to reduce the effect of spikes causing misfiring. Simon invented a digital filter, the zcfilter (zero crossing filter) to do this without affecting the timing which I also use.

I could probably adjust the timing to be correct at 15000 rpm. But the capacitors are very easy to remove if you have two small soldering irons. Just one on each side and lift it off. Have a look and see what you think.

I have got the rpm limiting software working on the mega8 so we are nearly there.

Sep 23, 2015, 12:10 PM
Registered User
So if I understand correctly that zcfilter is solved in the FW code itself? Neat! Considering the fact that we allow 10 and 14 pole motors in the E2K class I don't think it is wise to leave in the capacitors, because then you would need to optimize the timing for each flavour of FW (10 or 14 pole). Also if somebody wants to have the rpm limitation removed then again you probably need to tweak it again for the FW version without rpm limitation. So I will take them off. Just to be sure, are we talking about these?:

The disadvantage is of course that this is not something that everyone can easily do, it just goes a step beyond flashing the thing. On the other hand: my goal is to make sure we have second source controller CPU with suitable FW for E2K racing. So it does not has to be the RedBrick necessarily but it could be an alternative ESC running on the ATMega8A and I know now that there are other ESCs out there with the same chip and without capacitors. So in theory those could be flashed too using the same FW. So we can proceed as far as I am concerned.

One request though: the FW you did for the YEP 80A ESC is limited to an average of 15232 rpm. To keep the racing honest I would like to aim for the same (average) limited rpm.

Sep 23, 2015, 12:15 PM
Registered User
By the way: I flashed two YEP 60A ESCs for Tjarko yesterday, and all is working fine. Tjarko will do testing to see if the 60A ESC can handle the E2K abuse (flying multiple packs back to back and see if the ESC survives... hey Tjarko, maybe an external BEC would be a thing to consider?)

So if that works out well, we already have a cheaper alternative then the 80A ESC. But still not a second source controller of course.

Hopefully Tjarko wil post his findings here once he get's his hands on the ESC (which won't be before Oct. 11th because then we have our next E2K meeting).

Sep 23, 2015, 03:25 PM
Registered User
No! No! Not these.
There is a group of 12 components, 9 resistors and 3 capacitors which do the voltage sensing.
I'll remove mine first and check operation and send a photo.
Attached is a photo of the other board with the capacitors removed.
Bit big, sorry.
Here they are all grouped together.
Sep 24, 2015, 09:34 AM
Registered User
Well here you have it!
The photo shows the capacitors removed, the black spaces next to the "183" resistors.
The four files are as their names suggest for 10-pole and 14-pole motors, forward or reverse.
Some of the code is the latest, you will hear that, but I have tried to keep most things the same, especially the rpm regulation values, so I hope it gives the same results.
Please try and let me know.
Good flying, George.
Sep 24, 2015, 12:10 PM
Registered User
I notice that you have the forward and reverse files. Does this mean that you can't reverse the motor in the usual way by changing two motor connections?
Sep 24, 2015, 12:17 PM
Registered User
Yes, reversing two connections reverses the motor. But sometimes, especially in F5B, space is tight and it is easier to connect up with the battery connections one way, so it is interesting to be able to reverse in software.

Also if it is all connected up and goes the wrong way it may be easier just to reload the reverse software, depending on how the software is reloaded.

Quick Reply

Thread Tools