PWM voltage?

[hogster]

Hi all

I've just got 9 new Hobbywing Platinum 30A Opto ESCs for my X8 (8 and a spare!), and before connecting them to my AQ, I decided to put a scope on the AQ PWM outputs to double check they were nice and smooth. They are indeed smooth, and cycle at 400Hz, but I'm rather puzzled why the PWM is only ~3.3V peak to peak? I thought it should be 5V?

Many thanks David

[chschmid]

AQ (the STM32F407 32bit Cortex M4 ) runs @ 3.3V and so are the PWM signal outputs.

Cheers
Christof

[hogster]

Hi Christof,

Ah ok thanks for your reply. Presumably ESCs are designed (or should be designed) to operate on this lower voltage? I've only heard of PWM as 5v pk-pk ... but I guess no-one using AQ would have had any success otherwise

Many thanks David

[chschmid]

Hi David
You can always use a logic level converter if needed. I am not the HW crack but as far as I understand, it is unhealthy to use a 5V logig level for a 3.3V device if the device is not protected (as an Arduino Due for example).

Cheers
Christof

[hogster]

Hi Christof,

Thanks for your reply. Well the ESCs would normally run on 5V, so device protection isn't an issue ... I'm just scratching around to find out why my previous set of ESCs suffered such an appalling failure rate (4 before I even managed a first flight!), and wanted to check that a lower PWM voltage wouldn't have adversely affected the ESCs ...

Thanks! David

[JussiH]

If you poke around with the scope on some RX´s, you will find that 3.3V is very common. All spektrym RXs do 3.3V, so does most of the FrSkys and futabas, far as I know. They all run 3.3V internally because they are commonly supplied from a 5V source to a 3.3V regulator. If they ran 5V internally, they would either have a voltage drop over the regulator, or rely on direct power from a UBEC or SBEC somewhere in the setup.

Most 5V Microcontrollers (such as a AVR on a SImonK ESC) will detect over 1.8V as a high. So it does not cause problems and is very unlikely to have anything to do with the ESC failure.

Id be willing to bet that those ESC´s are borderline for 6S and that is the real cause of the problem.

[kinderkram]

Wait!
You soldered the 5V bridge on the motor outputs?
These are only needed if you want to power the ESC32s - for regular ESCs your middle pins should look like this:



No bridge on the pad! Can you confirm?

[LPR]

David

I know that your ESC are optos but I would still not run power from the AQ to the ESC with the signal wires. The opto ESCs that I've used do not even have the red signal wire connected in the ESC. Your ESC may be different and having the red wire connected may have caused your problems.

The optos I've used have a small 5v regulator to power the Atmega on the ESC.

Larry

[hogster]

Oh ... you might be onto something there Norbert.

Originally I understood from CarbonCore that I would not need to supply the ESCs with a separate 5V feed (eg. from a separate regulator or the like) as the flight controller would do that for me. I thereby assumed that the ESCs needed to be fed 5V from the flight controller in order to operate – something which I've seen on many other ESCs before.

I then tried and failed to get the motors to start when connected to the AQ, which I assumed was because of the gap in the middle track as shown in your diagram, so I bridged them all ... but the motors still didn't run. I then discovered that no PWM would be sent to the motors until the AQ was 'armed' by holding the rudder stick to the right. All the motors then worked fine and I assumed that was that.

However I bought a servo tester over the weekend, and lo-and-behold, with the red wire of one of the original ESCs DISCONNECTED, the ESC works fine, meaning I've been pushing 5V into the ESCs the whole time, without realising they didn't need it I had even measured between the ESC red wire and ground to see if it had an internal BEC, and it read 0V, so I assumed I was doing the right thing by giving it 5V.

Do you think this would have caused so many of the ESCs to fail? If 5V isn't coming out of the red wire, I'm wondering whether it's actually connected internally to the ESC at all ...? I've attached a photo of the ESC that smoked during my first take-off attempts showing one of the FETs with lots of charring around it. I guess it's hard to say why the additional 5V would have caused the FETs to fail?

If this was the cause of the ESCs failing, I'm now feeling slightly stupid for buying a whole new set of ESCs

Many thanks for the guidance guys David

[kinderkram]

Regular ESCs don´t need 5V input at all. They have (except for Optos) 5V BECs on board which output 5V to power the FC if needed.

So yes - this might have been the cause for so many failures. Sorry I didn´t ask earlier...
Either remove the bridges or cut/remove the 5V power lines on the ESCs and test again.