Author Topic: signal ground between two isolated voltage circuits  (Read 2439 times)

0 Members and 1 Guest are viewing this topic.

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
signal ground between two isolated voltage circuits
« on: April 16, 2019, 03:05:35 pm »
I have a design where there are two separate circuits. One circuit operates on 15 volts DC. (Isolated electronic motor control) The other operates on 5.0 volts DC.
The 15 volt electronic motor control has a 5.5 volt output that I do not want to use for isolation reasons.
The 5.0 volt circuit needs to send a low voltage signal (PWM) to the electronic motor control.  Will a single signal wire to the motor control (no ground) work or will the signal wire require a signal ground… And, what measures (precautions) should be taken to avoid ground loops and EMF issues?
I know that this is a bit like asking how long is your well rope due to not knowing anything about the motor control schematic ( and I have no way of knowing), however any comments are welcome.
 

Offline helius

  • Super Contributor
  • ***
  • Posts: 3688
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #1 on: April 16, 2019, 03:09:19 pm »
Welcome to the site.
A single wire between two points is not an electrical circuit and cannot serve any purpose. All electric circuits must have a loop for current to flow.
If you need to galvanically isolate two circuits while sending a signal from one to the other, there are means to achieve this, in particular, optoisolators and signal transformers.
In your case, I suggest you investigate optoisolators, since they will handle PWM signals. They are used in most switching power supplies in this fashion, to control the primary power from the secondary side, while maintaining galvanic isolation between primary and secondary.
For controlling a motor, you also need some kind of feedback from shaft speed or position and some means to prevent damage if the motor is stalled.
« Last Edit: April 16, 2019, 03:15:37 pm by helius »
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #2 on: April 16, 2019, 03:29:45 pm »
Thank you, helius:

I will check out the optoisolators, thanks again.
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #3 on: April 16, 2019, 04:39:34 pm »
As usual the devil is in the details.  The opto's appear to be just the ticket, however, since I am a newbe, I have more issues.  The Motor control is also a BEC or battery elimination circuit designed to use the same battery that powers the motor (15v) and the other electronic circuits (5.5v).  I will not be using this 5.5v BEC power supply from the motor control, I need the two circuits isolated.

There are three wires coming out of the motor control (BEC line) White (signal), Red (5.5v+) and Black (5.5v-).  As I see it, the signal (white wire) is an input. Then there is the issue that you brought up regarding feedback.  Since this situation uses PWM is the feedback issue necessary? 

Also, would grounding the signal ground (isolated 5v circuit) through the opto going to work or cause some other issues?

I will try and send some schematic details;  These will be snippits of DWG files.  Please do not be too harsh regarding this schematic, It is a work in progress.

 
 

Offline pwlps

  • Frequent Contributor
  • **
  • Posts: 372
  • Country: fr
Re: signal ground between two isolated voltage circuits
« Reply #4 on: April 16, 2019, 05:45:26 pm »
As usual the devil is in the details.  The opto's appear to be just the ticket, however, since I am a newbe, I have more issues.  The Motor control is also a BEC or battery elimination circuit designed to use the same battery that powers the motor (15v) and the other electronic circuits (5.5v).  I will not be using this 5.5v BEC power supply from the motor control, I need the two circuits isolated.

In a similar project with a DC motor, where the control and power circuits were working at different PS voltages and also needed separate and isolated grounds,  I have been using the HCPL2200 optocoupler.  What is nice about this IC is that it has an integrated amplifier so it can translate your 5.5V PWM pulses to 15V PWM pulses without any additional components, just a transistor for the current gain. 
« Last Edit: April 16, 2019, 06:03:16 pm by pwlps »
 

Online rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #5 on: April 16, 2019, 05:55:02 pm »
I need the two circuits isolated.
Why?  It would be far easier to just tie the common wires together and send the digital PWM straight to the controller.  I have done galvanic isolation with a high speed opto but it was because the speed controller wanted a potentiometer input and I was using PWM and I needed to convert to 24 VDC.

It is not a 'given' that you need isolation of the common return wire.
Quote

There are three wires coming out of the motor control (BEC line) White (signal), Red (5.5v+) and Black (5.5v-).  As I see it, the signal (white wire) is an input. Then there is the issue that you brought up regarding feedback.  Since this situation uses PWM is the feedback issue necessary?
Whether you need feedback depends on your goals.  The feedback mentioned was to prevent a stalled motor burnout.  So will a fuse...  However, if your application includes PID control for speed regulation then, sure, you need some kind of feedback.  Hall effect sensor, disk with slots and opto sensor, DC generator, or any of a number of schemes.
Quote

Also, would grounding the signal ground (isolated 5v circuit) through the opto going to work or cause some other issues?
That isn't how the opto works.  On the isolated side, all you have is a transistor.  It is up to you to add a collector resistor from the local power source and connect the emitter to the ground of the local source.  There is no direct connection between the power supplies or across the opto.

https://www.sparkfun.com/products/9118

Google has a lot of hits to 'opto isolator for motor control'
Quote
I will try and send some schematic details;  These will be snippits of DWG files.  Please do not be too harsh regarding this schematic, It is a work in progress.
Which will limit their utility to just those who can read the files.  Better to print them off as a .pdf (using Microsoft Print To PDF or equivalent) and post them in that format.

The .png are probably ok as well.
« Last Edit: April 16, 2019, 06:01:47 pm by rstofer »
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #6 on: April 16, 2019, 06:31:44 pm »
Why?  It would be far easier to just tie the common wires together and send the digital PWM straight to the controller.  I have done galvanic isolation with a high speed opto but it was because the speed controller wanted a potentiometer input and I was using PWM and I needed to convert to 24 VDC.

It is not a 'given' that you need isolation of the common return wire.


That is preferable, I just read about so many horror stories regarding ground loops with these expensive micro circuits.  The 4s. lipo is rated @ 40c. on the 15v side.  That is a lot current dump potential.  And...since I am new to all this, I take heed perhaps too cautiously.  I will resend some of the details on a PDF to help in the communication process.  It may take me awhile.

When these images are more readable, I guess you will laugh at all the overkill but, since its my first attempt at building one of these things, I choose to error on the redundant side.
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #7 on: April 17, 2019, 03:41:43 pm »
SILLY ME!

I just looked up the specs on the motor control.  I completely forgot about this:

Specification:
Output:  Continuous 65A, burst 70A up to 10 seconds.
Input Voltage:  2-6 cells lithium battery or 5-18 cells NIMH battery.   
BEC:  Linear 4A @ 5V
Control Signal Transmission: Optically coupled system.
Max Speed:     
   2 Pole: 210,000rpm
   6 Pole: 70,000rpm
   12 Pole: 35,000rpm
Size:  70mm (L) * 34mm (W) * 17mm (H).
Weight:  73g.


So, if i cut the positive out of the three wire connector and leave only the signal and ground, that should be isolated.
 

Online rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #8 on: April 17, 2019, 06:03:19 pm »
OK, how about the rest of the information on the ESC?  Do they show example circuits?  Are you sure relatively high frequency PWM is what they want?  Are you sure it isn't some variation on RC servo pulses (a very low speed form of PWM)?  Is there a publicly available datasheet?  Manufacturer and model?

That it is optically isolated is good.  We can assume that the white wire goes through a resistor to the anode of the LED inside the opto coupler and that the black goes to the cathode of the LED.  But we're just guessing!  How do we know the resistor is internal?  How much current do we have to pump through the LED?  Can the PWM source even drive that LED?

If the drive is made for RC, I would almost bet on the signal input wanting to be an RC servo pulse stream.
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #9 on: April 19, 2019, 07:31:32 pm »
If the drive is made for RC, I would almost bet on the signal input wanting to be an RC servo pulse stream.

Thanks for responding, The signal is coming from a Pixhawk Mini Auto-Pilot and should be compatible with the ESC.  There are no readily available data on the ESC as far as I can determine.  It is some time ago that I purchased the ESC, and that is why I have had time to forget about this issue.  I intended to buy an ESC w/o the BEC  feature and there were none available where I shopped at the time.

I agree that the internal circuitry is sketchy at best.  The reason I do not want to use the power from the ESC/BEC is that ESC/BEC's are nortorious for being failure points.  In this application where I will be flying long range autonomous missions, I want to know that in an event of loss of power due to the ESC/BEC  I still have a chance to safely land the vehicle due to no loss of avionics and video.

I guess I could crack open the ESC and have a look, however I wouldn't know what I was looking at anyway especially with the micro surface mounted components.  Not to say that photos won't be useful for others.

 

Online rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #10 on: April 19, 2019, 08:08:41 pm »
In the Arduino IDE installation, there is an example called Knob.  This reads a potentiometer, does some scaling and produces an RC servo pulse.  It would be quite easy to load this into an Arduino Uno and generate 5V pulses.  What we don't know is whether there is an internal resistor on the LED portion of the opto.

I suppose you could put a 220 Ohm resistor between the BEC +5V and the white wire while measuring the current with a DMM in series.  You can then do a little arithmetic to figure out whether there is an internal resistor.  It would be useful to also measure the voltage drop across the 220 Ohm resistor.

Resistor voltage = BEC voltage - VF of the opto  This is 5.0 - 1.5 for a 4N35 opto coupler.  So, we have around 3.5V dropped across the resistors.  The maximum allowable current for the 4N35 is 50 mA but, in use, it usually runs around 10 mA.  So, Total R = 3.5 / 0.01 or 350 Ohms.  We can start with a 220 Ohm resistor and see how much current flows.  From that, we can work out the internal resistance.  If it seems like this internal resistor will limit the current to a couple of mA then we don't need an external resistor.

What I'm trying to avoid is just twisting the red and white wires together and dumping multiple amps into the opto LED.  It's best to test at a low current and try to rationalize whether there is an internal resistor.  An 'oopsie' will destroy the opto.

Look at "Forward Voltage" spec on page 154  = VF = 1.3..1.5V at IF = 10 mA

We almost certainly don't want the motor connected during these tests.

You'll know right away because if you use a 220 Ohm resistor, about 16 mA will flow if there is NO internal resistor and something much less if there is.  It's a quick test!

Take some measurements and post back.
 

Offline grizewald

  • Frequent Contributor
  • **
  • Posts: 612
  • Country: ua
Re: signal ground between two isolated voltage circuits
« Reply #11 on: April 19, 2019, 08:17:17 pm »
The vast majority of Pixhawk based craft run a common ground. Years ago, Engine Speed Controllers could be bought with an opto-isolator built in, but these days, when an ESC is described as an "opto" ESC, it normally means that it does not incorporate a Battery Eliminator Circuit and not that it actually uses an opto-isolated PWM input.

Is there some special reason why you must have the motor power completely isolated from the Pixhawk's power?

You don't tell us exactly what parts you are using here, but I'll put money on the black, red and white wires from your ESC being:

Red: 5.5V
Black: ground
White: signal

There is nothing in your craft which uses -5.5V.

The reason why the BEC outputs +5.5V is to account for the protection diode on the Pixhawk's 5V power input. The diode will drop around 0.5V and therefore the output from the BEC is a little higher than 5V so that the Pixhawk actually get 5V power and not 4.5V

Unless there's a special reason for isolating the two supplies; then don't!

« Last Edit: April 19, 2019, 08:19:33 pm by grizewald »
  Lord of Sealand
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #12 on: April 20, 2019, 07:03:04 pm »
Thanks again:

I can't work on the project until Monday.  It is in the lab at my tech school.  I am a 71 year old student.  Now that you know a little about me, I hope you won't find that too amusing.  Yes I went back to school after retirement.  It's the best thing for me now.  It keeps me sharp as if that is still possible.  I'm not a complete novice at electronics, however I may as well be with the advent of miniaturization.

It is good to be around younger people and be able to share some life's work experiences.  I singed up for Cad Classes about five years ago.  Now I help out the instructors and do substituting for some of them as needed.  And.. I get to play with all the "toys"...3-D Printers and such.  As a plus, My instructor has let me build this project in our lab.  He wants today's millennials to get more interested in  hands on stuff' as he puts it.

So, thanks again for your help.  I have not been able to save the wiring schematic as a PDF as of this note, however I do have some snipits of my years long solid works rendition of the project sans wires.  I will post it just so it may be of some help and interest.

Hopefully, I will have the schematics converted on Monday and some results to share.
 

Offline jberookc@gmail.comTopic starter

  • Newbie
  • Posts: 8
  • Country: us
Re: signal ground between two isolated voltage circuits
« Reply #13 on: April 20, 2019, 07:18:51 pm »
Here are two examples of the project in Solid Works.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf