max supported input voltage on pca9685 bonnet/shield. Moderators: adafruit_support_bill, adafruit

[knoxvilles_Joker]

https://learn.adafruit.com/adafruit-16- ... ing-servos

I am running into a situation where I can power servos with more than 6 volts but I am not seeing anything that indicates a max board supported voltage and amperage.

https://www.savoxusa.com/collections/se ... size-servo
This one supports 6,7.4, and 8.4 volts DC. The kicker is at stall 10.5V@8.4VDC

I know that you folks typically design the boards to have the power supplies separated, but, with the sandwiches I am using to integrate things with, I can not risk smoking a 100 dollar board setup.

[knoxvilles_Joker]

And I am going through things some more as well:
https://www.powerstream.com/Wire_Size.htm

at 6VDC:
Left arm on my setup would draw 12.3A@6VDC and right arm would draw 15.0A@6VDC.
Each I was hoping to run through an ethernet cable that was about 2-3foot in length.

Ethernet cable gauge is typically around 21 gauge and according to the above link I would max at 9amps for chassis wiring and 1.2amps for power transmission.

so my limiter will be load and distance.

Now throw in the other enemy to all power and electronics: heat. What is heat? the result of resistance to the flow of electrons and an indicator that you have massive power loss issues. heat is the shedding of energy that is not otherwise transmitted through the line. heat is the number one killer of effective electrical circuits that are not designed for a wide range of input/output voltages.

So ethernet cable would be a no go for the power portion. Thermostat cable might be an option at 18 gauge but it is rather inflexible and typically a solid conductor. but a better option? DC barrell connectors and they are very doable at 8 gauge and well designed.

[knoxvilles_Joker]

I have done some more looking. current cap on the servo board I have is 4700micro farads.

For converting uf to f:
https://www.inchcalculator.com/convert/ ... -to-farad/

I was looking at using a 0.5 ish farad capacitor instead which would address the momentary load issues.

The issue with the larger capicitors and big ones called super capacitors, they are much lower voltages at 3, 5, 6 volt ranges.

This unit appears to be a good match:
https://www.newark.com/kemet/fr0h104zf/ ... p/33AC0682

OK so I solve the momentary load. So how do I solve the constant load?
I up the amperage on the power supply.

I use this unit to take the 12@5A max power feed and step it down to 5V@10A.
Stepping down rather than up allows for more efficient battery usage
https://daygreen.com/collections/12-24v ... ler-type-1
(yes it ships from China, but I do not have much of a choice.)

Now the issue comes down to power loss across the conductors which means minimizing cable lengths.

So the two together should help with spike/peak loads and constant loads. Unfortunately, I will have to bench test and then wear test the unit to see if performance improves at all.

The other issue is that I will need to stretch the servo signals to make things play better.

[knoxvilles_Joker]

so tthree main issues to ask on the pca9685 bonnet board:

what is the max supported sized capacitor rating?

What is the max supported voltage as there are now 7.2, 8.4, 6v and even higher voltage servos?

What is the max supported amperage on the board?

[knoxvilles_Joker]

I do not like cross posting but I like to reference the source for a possible answer:
viewtopic.php?f=31&p=530011

Per this thread on the board 8A is the max assuming proper ventilation.

I assume this would be 8A of constant draw.

There is a solution to save things: thermal cut off switch.

My problem is I usually push whatever hardware I am using to its technical limits.

Theoretical max draw on current setup is 10A and that is momentary draw at best. Constant draw by my estimate is about 4A, but I need to test things.

The board I identified to use steps 12v5A to 6v10a. amps is just the pipe size, the conduit or wire will limit the draw on that.

[knoxvilles_Joker]

OK, found the solution:
https://www.digikey.com/htmldatasheets/ ... V4QAvD_BwE

Again I hate external posting. But having the data sheet to reference kills arguments about facts. And it can allow folks to read and make their own determination in case I am wrong and can tell me as much.

The chipset supports 2.3-6VDC.

As stated in prior post 8A is the maximum amperage. 6 times 8 is 48 so the power supply is a 48watt capable board. So on the safe side 40 watts is probably the best.


Adafruit support:
Please update the writeup on the pca9685 boards stating max voltages on servo power/amperage and control pin max amperages.

[adafruit_support_bill]

Servo power is entirely independent of the PCA9685. The chip is powered with 3.3v or 5v depending on the voltage of the processor it is connected to. Servo control pulse voltage will be the same as the chip supply voltage. Servo power is routed directly to the servo headers via the reverse voltage protection MOSFET. It does not touch the PCA9685.

[knoxvilles_Joker]

Servo power is entirely independent of the PCA9685. The chip is powered with 3.3v or 5v depending on the voltage of the processor it is connected to. Servo control pulse voltage will be the same as the chip supply voltage. Servo power is routed directly to the servo headers via the reverse voltage protection MOSFET. It does not touch the PCA9685.

So the way you have servo power setup it is completely separate from the pca9685. The ONLY thing the pca9685 is responsible for is the signal pin on the servo connectors?

So if I have the servo board setup as a bonnet or a hat or shield getting power directly from the arduino or the raspberry pi in this case, the pca9685 has what it needs to function.

Servo power is handled by the servo power connector (ground and positive) and it is completely separated from the pi/arduino and the pca9685. If that is the case are you to then say that the servo power input could exceed 6VDC for the servo power connections?

[adafruit_support_bill]

That is correct. I run servos at 7.2v using the PWM/Servo breakout board - which is essentially the same circuit.

https://learn.adafruit.com/assets/33736

[knoxvilles_Joker]

That is correct. I run servos at 7.2v using the PWM/Servo breakout board - which is essentially the same circuit.

https://learn.adafruit.com/assets/33736

THANK YOU!

I wanted to confirm before I burnt up a 500 dollars in electronics on accident.

The documentation on that is not too clearly stated in the servo shield/bonnet adafruit writeup:
https://learn.adafruit.com/adafruit-16- ... pberry-pi/

Not sure if you want to add something in that regards to the FAQ section or the powering section as a note.

I was also looking at adding a quarter or a half farad capacitor to help with insane momentary stall spikes to help keep temperatures down.

I have been switching to try and power all my projects using my dewalt batteries so I only have on type of battery stuff to deal with and can just keep buying more dewalt batteries for my cordless stuff:
https://www.aliexpress.com/item/4001133 ... 3229%22%7D
(note on this outputs are 12V@5A and 2x5V@2A)
But I would also note all battery brands have a similar unit, I like this unit as it has the belt clip and it is easy to plug and unplug as needed.
And I am pushing on this myself on my stuff so I am more ecologically friendly with batteries and do not run into a case where I am tempted to throw away so many bad batteries that are dead and not rechargable.
https://www.dewalt.com/products/accesso ... d-chargers

I am going to be testing this unit with the battery pack setup to see what I get:
https://daygreen.com/collections/12-24v ... ler-type-1
This is unfortunately the only unit I could find that would possibly work and might be within stated specs.

6v times 8a is 48w
6v times 10a is 60w

I could go with this one as the voltage drop over the wire gauge and length would work but I run into a potential over current possibility on the board:
https://daygreen.com/collections/12-24v ... -regulator

The sad part is the daygreen manufacturer is the only I could easily find. I could find no one domestic that offered a similar product. hint hint.

I am finding that as robotics and automation advances, we will find beefier servos as more of a commonality and we will need to create more power solutions that will allow for off ac power solutions.

[adafruit_support_bill]

How many servos are you running and what is their current draw? Voltage is only half the story. Some of the newer super high-torque servos pull 2A or more during moves. If you have more than a few of those, you should consider running power to them directly so you don't overheat the power bus traces on the board.

[knoxvilles_Joker]

prototype setup

IMG_3289.JPG (961.54 KiB) Viewed 303 times

right arm

IMG_3301.JPG (1018.9 KiB) Viewed 303 times

right arm

IMG_3301.JPG (1018.9 KiB) Viewed 303 times

servos:
35kg-cm, Stall current(at locked): Stall current(at locked): 1.9A @5.0V
Stall current(at locked): 1.9A @5.0V
Stall current(at locked): 2.1 A@6.0V
Stall current(at locked): 2.3A@7.4V

savox sb2290sg Stall Current (at locked): 7500mA @6.0V, 5mA @6.0V
Running Current (no load): 180mA @6.0V, 220mA @7.4V, 220mA @8.4V
Stall Current (at locked): 7500mA @6.0V, 9200mA @7.4V, 10500mA @8.4V
Idle Current: 5mA @6.0V, 5mA @7.4V, 5mA @8.4V


hitech: 5v@2.7A stall
Current Drain - idle (4.8V)
8mA
Current Drain - idle (6.0V)
8.7mA
Current Drain - no-load (4.8V)
230mA
Current Drain - no-load (6V)
285mA


2 savox sb2290g. I am looking at switching back to a 35kg servo on one to lower the amperage draw.
3 hitech hs785hb
2 35kg-cm servos

It is the savox that is causing issues I think. Moving back to just one may solve much of my issues. I have to have one on my right control arm setup as the cannon setup on it just has too much weight to move.

The other issue is due to the way they are mounted on the pack I have about 2-3 feet of wire leads running to the servos. wire gauge is another consideration at play there.

[knoxvilles_Joker]

ref photo on right cannon

fe6d8926290726a422e980ecb428751d.jpg (251.99 KiB) Viewed 303 times

left cannon weathered

IMG_3297.JPG (868.81 KiB) Viewed 303 times

This canon I have 3d prined and it is to movie scale. I had to hollow out room for the servo and two leds. I stayed away from lasers due to liability concerns...

I had gotten some auction photos from this item when I asked very, very, nicely:
https://ukm.propstoreauction.com/m/lot- ... items%3D96

This is the shoulder canon test by amalgamated dynamics incorporated:
https://www.youtube.com/watch?v=dpgznoPCRPk
This is what I am trying to copy.

Now on my upgraded pack I will have the travel arm be moved from underneath and not from the side like in the above video. I had concerns with breakage as I am trying to bullet proof the build for durability and ruggedness and to also account for shock loads if a run or drop/fall event occurs

[adafruit_support_bill]

You have some serious current draw there. I'd recommend running power directly to the servos to avoid any voltage drop or overheating issues on the bonnet. Then you will just need signal and ground connections between the servos and the bonnet.

[knoxvilles_Joker]

You have some serious current draw there. I'd recommend running power directly to the servos to avoid any voltage drop or overheating issues on the bonnet. Then you will just need signal and ground connections between the servos and the bonnet.

Ok let me work through this a bit. I am going a million miles a minute on how to do this with my setup.

Thanks on the pointer on the signal and ground connection requirement. That answered a question I had in regards to that separate power connection. Also not well documented in the writeup for the servo shield/bonnet article on your site.

The main power pigs are the savox units. Those as you suggest I will definately need to split out as that is a stall load of 15amps. Movement I suspect is at 8amps and that is causing most of my issues. Though I am not so sure it is drawing that much at 5 volts as the amp draw at higher voltages indicates it would be around 6500ma or so. I would need to use a 12 gauge conductor for the separate power draw.

The hitec servos that are used for general position holding are set to support 3304oz-in of torque max and are never even close to approach the stall load limit. Those are pulling a theoretical max of 5.4amps

https://www.servocity.com/sg20-series-s ... 98-sec-60/ (2 of these)
https://www.servocity.com/servo-driven- ... gear-rack/

The other two servos pull 1.9 amps at 5vdc. So it is getting close if I pull off the two savox servos.

On the upgraded setup I will be replacing one of the hitecs with a stepper motor and using a separate board with a separate power supply for the left arm in place of the gear rack.
I basically have a lead screw with a 12V stepper motor. I will have to hunt a photo and trim it down.
https://www.servocity.com/3501-series-l ... mm-length/
Plus on that new setup I would have another stepper motor for the blade assembly that I keep dragging my feet on fabricating the housing for.

And after that I have plans to replace the left canon with on from aliens-fx that has a cocking motion built into it. so yay another low power servo.

Given all this I will need a battery rigged power setup and an on AC power setup for bench testing.

So yay, more parts. Let us see how many wrong parts I order on this go around.