stunthanger.com

Electric Stunt => Gettin all AMP'ed up! => Topic started by: Howard Rush on December 29, 2021, 03:27:48 PM

Title: Timer Signal Optoisolator
Post by: Howard Rush on December 29, 2021, 03:27:48 PM
What's a good optical coupler that will pass a PWM servo signal unadulterated?  Thanks.
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on December 29, 2021, 04:07:40 PM
What's a good optical coupler that will pass a PWM servo signal unadulterated?  Thanks.

What are you doing now?  Are you sure your ESC isn't already opto-isolated?

If you can find just about any optoisolator made for RC use, it'll work.  If you're looking for chips, just about any will do.
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on December 29, 2021, 06:09:05 PM
What are you doing now?  Are you sure your ESC isn't already opto-isolated?

If you can find just about any optoisolator made for RC use, it'll work.  If you're looking for chips, just about any will do.

I'm gonna guess that some little voice mighta suggested having the throttle signal in flight data might be a good thing to have in order to understand more better how timer control out put may be acting.
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on December 29, 2021, 06:58:22 PM
On reflection, I left out detail.

The hobby/cheap electronics market seems to be stuck in the 1970's in a lot of ways.  If someone is using the same opto-isolators that are called out in some article in their favorite 1980's Popular Electronics project, then you may get an isolator that isolates just fine, but that loads the ESC down more than necessary.  I'd expect an opto that's made using those older components without much thought would load down an ESC quite a bit.  If you were doing something like driving two optos with a 'Y' cable, it may not be able to work at all.

If you're going to shop the hobby market for RC opto isolators (I looked, there's a few but I didn't recognize the brands) then check how much current they pull on the signal line.

If you don't like what you find, check with me & I'll jot down a few schematics.  this part (https://www.digikey.com/en/products/detail/fairchild-semiconductor/FOD0720R2/13500326) looks pretty easy to use electrically, but it's surface mount.
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on December 29, 2021, 07:54:09 PM
On reflection, I left out detail.

The hobby/cheap electronics market seems to be stuck in the 1970's in a lot of ways.  If someone is using the same opto-isolators that are called out in some article in their favorite 1980's Popular Electronics project, then you may get an isolator that isolates just fine, but that loads the ESC down more than necessary.  I'd expect an opto that's made using those older components without much thought would load down an ESC quite a bit.  If you were doing something like driving two optos with a 'Y' cable, it may not be able to work at all.

If you're going to shop the hobby market for RC opto isolators (I looked, there's a few but I didn't recognize the brands) then check how much current they pull on the signal line.

If you don't like what you find, check with me & I'll jot down a few schematics.  this part (https://www.digikey.com/en/products/detail/fairchild-semiconductor/FOD0720R2/13500326) looks pretty easy to use electrically, but it's surface mount.

Well, I'm planning on using a ADA Feather 32u4 with onboard SD card for collecting data. It uses 3.3V logic that needs to be dropped down. I was just going to use a voltage divider to do that. Something like 3k total resistance. But then looked at the logic level shifters. I dunno exactly what I need  to do. But my package as I think Howard is leaning towards will be a pair of strain gages, a 6 DOF sensor and an input from the timer. The concern I have is using a separate power source and differing logic levels. The Feather has a USB max of 5V and the ESC outputs 5.5v which presents another problem for the non EE to solve. In my perspective the preferred power source would be the ESC BEC.
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on December 30, 2021, 12:02:32 AM
I'm gonna guess that some little voice mighta suggested having the throttle signal in flight data might be a good thing to have in order to understand more better how timer control out put may be acting.

Nah, this is for something more mundane.

this part (https://www.digikey.com/en/products/detail/fairchild-semiconductor/FOD0720R2/13500326) looks pretty easy to use electrically, but it's surface mount.

I'm a a through-hole guy.

How about https://docs.broadcom.com/doc/AV02-0876EN or https://www.sparkfun.com/products/9118 ?  The Broadcom part appears to the layman to have fancy stuff making it less picky about input and output, but the Sparkfun board is easy.

Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on December 30, 2021, 10:06:11 AM
How about https://docs.broadcom.com/doc/AV02-0876EN or https://www.sparkfun.com/products/9118 ?  The Broadcom part appears to the layman to have fancy stuff making it less picky about input and output, but the Sparkfun board is easy.

The Sparkfun part would draw around 16mA per input; 32mA is way more than any typical timer will drive (and way, way more than a TUT -- there's a story there).

The Broadcom part should work.
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on December 30, 2021, 10:09:47 AM
I'm a a through-hole guy.

We need to fix that.  It's so very 1980.
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on December 30, 2021, 10:45:08 AM
The Sparkfun part would draw around 16mA per input; 32mA is way more than any typical timer will drive (and way, way more than a TUT -- there's a story there).

The Broadcom part should work.

What about these PC817X Series ?

chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/viewer.html?pdfurl=https%3A%2F%2Fwww.farnell.com%2Fdatasheets%2F73758.pdf&clen=287234&chunk=true

Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on December 30, 2021, 11:15:19 AM
It's so very 1980.

It was for me.
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on December 30, 2021, 01:09:13 PM
What about these PC817X Series ?

chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/viewer.html?pdfurl=https%3A%2F%2Fwww.farnell.com%2Fdatasheets%2F73758.pdf&clen=287234&chunk=true

Here's a (hopefully) better link (https://www.digikey.com/en/products/detail/american-bright-optoelectronics-corporation/BPC-817-B-BIN/9678142).

Those are LED input, so they'll either need a lot of current or they'll need a driver.

The ones I'm recommending have an internal driver with an input that looks like a logic gate.  The ones that are most common just have an LED -- it's your responsibility to drive it adequately (which usually means a logic output that can sink or source at least 10mA, and a resistor to limit LED current).

Think of the logic-input ones as having power steering for the LED.
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on December 30, 2021, 02:48:00 PM
OK, I'll get me some purple boards made with two Broadcom isolaters each and no buffers.  Thanks.
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on December 30, 2021, 04:46:21 PM
OK, I'll get me some purple boards made with two Broadcom isolaters each and no buffers.  Thanks.

Don't forget the bypass caps. 100nF across each power supply, right at the chip.
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on January 21, 2022, 06:08:27 PM
Here’s the board.  Is it still static sensitive with the condensers in place?
 
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on January 21, 2022, 07:36:39 PM
Here’s the board.  Is it still static sensitive with the condensers in place?

That is very cool. What does it do?
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on January 21, 2022, 07:44:36 PM
It lets one timer drive two ESCs with no electrical connection between them.
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on January 21, 2022, 07:59:24 PM
It lets one timer drive two ESCs with no electrical connection between them.

Why's it got an antenna?
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on January 21, 2022, 10:48:09 PM
There’s no antenna.  Ends look funny because input and output sides are configured so both halves of connectors can be held to the board with tape or heat shrink.
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on January 22, 2022, 09:09:23 AM
Here’s the board.  Is it still static sensitive with the condensers in place?

Electronics are always static sensitive, it's only a question of how bad.  Should be no worse than the timers we all use.  I recommended the bypass caps because that's just what you do, like putting on your skivvies before you pull on your pants.  Bypass caps have more to do with protecting circuit operation when there's long wires between the circuit and the power rail.
Title: Re: Timer Signal Optoisolator
Post by: pmackenzie on January 22, 2022, 04:33:19 PM
Yep - bypass caps are there to keep the IC power stable.
If you want static protection then bleed resistors on all the inputs and outputs can help.

But just don't rub your cat with the PCB indoors in the winter, and you will probably be fine :)
Title: Re: Timer Signal Optoisolator
Post by: Rick_Huff on January 22, 2022, 06:15:10 PM
Howard,
Are you going to try a twin (with an Igor system)?
Sounds exciting.

Rick
Title: Re: Timer Signal Optoisolator
Post by: pmackenzie on January 22, 2022, 06:22:02 PM
Twin, or Contra?
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on January 23, 2022, 12:01:37 AM
Are you going to try a twin (with an Igor system)?

One of these days, I hope.  The purpose of making the isolator now is to put in Jimby's twin so he can determine if a problem he saw is electrical or something else.

Twin, or Contra?

I'd make a twin.
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on January 23, 2022, 06:05:31 AM
One of these days, I hope.  The purpose of making the isolator now is to put in Jimby's twin so he can determine if a problem he saw is electrical or something else.

I'd make a twin.

There's lots of opportunity there for noise. A twin is in my future but I gotta finish the road I'm currently on first.
Title: Re: Timer Signal Optoisolator
Post by: TDM on February 01, 2022, 03:29:57 PM
What about a MKR Zero board with built in SD. Use the 5v from BEC to power up the Zero
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on February 02, 2022, 09:41:23 AM
For what would one use the Zero?

(Although I’m an analog guy, I can still talk digital.)
Title: Re: Timer Signal Optoisolator
Post by: Mark wood on February 02, 2022, 09:58:02 AM
For what would one use the Zero?

(Although I’m an analog guy, I can still talk digital.)

It's a datalogger similar to the one I'm working with. Cept the one I have is 3.3V.

Quote
What about a MKR Zero board with built in SD. Use the 5v from BEC to power up the Zero

The isolator is for isolation from the two ESCs being run by one timer.

Title: Re: Timer Signal Optoisolator
Post by: Igor Burger on February 02, 2022, 10:42:50 AM
HI Howard, what is exactly the problem? We have here also twins and also contraprops with 2 escs and they work without problems.

BTW Do you know BEC on Spin can be feeded backwards from the other ESC?
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on February 12, 2022, 02:43:27 PM
HI Howard, what is exactly the problem? We have here also twins and also contraprops with 2 escs and they work without problems.

Sorry to take so long to respond.  I think Jim told you about the problem.  As I understand it, one Igor timer will drive both motors well with sensitivity zero.  With nonzero sensitivity, motors make strange noises.  I do not know how he wired them.  He also tried two independent systems: each Spin 44 driven by its own Igor timer with sensor packages close together, but electrically isolated. That worked with both zero and nonzero sensitivity.   
Title: Re: Timer Signal Optoisolator
Post by: Tim Wescott on February 12, 2022, 03:10:51 PM
Sorry to take so long to respond.  I think Jim told you about the problem.  As I understand it, one Igor timer will drive both motors well with sensitivity zero.  With nonzero sensitivity, motors make strange noises.  I do not know how he wired them.  He also tried two independent systems: each Spin 44 driven by its own Igor timer with sensor packages close together, but electrically isolated. That worked with both zero and nonzero sensitivity.   

Many great minds in Stunt have been puzzled by this.  I was even brought in as an electronics consultant, and added the weight of my very own consultant's shrug to the general shrug-ness of the responses Jim's been getting.
Title: Re: Timer Signal Optoisolator
Post by: pmackenzie on February 12, 2022, 05:06:28 PM
HI Howard, what is exactly the problem? We have here also twins and also contraprops with 2 escs and they work without problems.

BTW Do you know BEC on Spin can be feeded backwards from the other ESC?

I got the impression the system is running two motors and two batteries.
So if the only ground connection is at the timer there might be some noise?
Title: Re: Timer Signal Optoisolator
Post by: Howard Rush on February 12, 2022, 06:50:06 PM
Two motors and one battery, I think, so not so independent.
Title: Re: Timer Signal Optoisolator
Post by: Igor Burger on February 13, 2022, 02:11:36 AM
Yes, I remember something like that. I thought it was solved afterwards. I do not see reason why sensitivity off or on can make this problem, beside one - if you hit certain PWM % then motor starts to make "some" noises. It happens to me at certain temperature when I adjust RPM to keep lap time. But it is only "other" sound, not any screaming like with free bearing or so. ALex calls it "muses in model  ;D It can happen if you change sensitivity and accidentally hit that PWM %.

I did not hear of see what is exactly happening there, so hard to suggest something, but for example when I started with development, I had model which was improperly trimmed and acceleration of motor caused yaw of fuselage and that confused sensor installed in nose, so at some sensitivity it started wildly oscillate. That is why I had sensor installed back of CG in first tests.

Regarding connections in such configuration - I give very thin (it is "light" grade, it is much thinner than "standard") cables with timers, so maybe you can try standard or thick R/C cables for larger servos. If you think it is problem of grounding, but since it was one battery, it cannot be this problem. On the other hand it can be problem of too long leads, if motor gets command strongly accelerate, it can make voltage drop on long leads and it will push governor to draw even more and it will make even larger voltage drop.

And the last think I have in mind, one of BECs should be off and cable disconnected from connector, or simly cut, with switch off. That switch only makes BEC on or off, ESC is powered independently and all the time under power if battery is connected. If you do not disconnect that red cable, other ESC will feed it backwards by Y cable. If they are both on, it can make teoretically some strange thing (they are switched BECs) but when I tested them they worked normally also both on, however it is not so "technical"  ;D.     

Heres is the link to my FB page with explanation and that is used on several models (mostly contraprops, so one battery) without problem. I think Jim saw it and I think that is how he wired all together.

https://www.facebook.com/permalink.php?story_fbid=1346574585520642&id=363984487112995