Propeller Powered From Non-Isolated 120VAC Power Input - Prop Plug Acting Sketchy/Not Working Right

Mahonroy

Hey guys,
I have a propeller device that is powered from a non isolated 120VAC to 12VDC power converter module:
http://rohmfs.rohm.com/en/products/databook/datasheet/module/power_module/acdc_converter/bp5067-12.pdf

I then am using a 3.3VDC voltage regulator (RECOM R-78E3.3-0.5):
http://www.mouser.com/ds/2/468/R-78Exx-0.5-21170.pdf

Here is an image of the power rail schematic for my circuit:


I have a feeling that this has to do with using the neutral line as the ground in my schematic. This is also the same ground that the prop plug uses. I notice that with the device powered down, plugging in the prop plug, that the prop plug led stays constantly lit. Sometimes I can detect the propeller chip, and sometimes the serial terminal works and receives messages from the device. Other times the prop terminal and the propeller tool will freeze and I have to unplug and re-plug in the USB prop plug to get it to work again. Using the prop plug on other propeller devices work just fine, so I have a feeling its something to do with this specific design.

Am I missing something? Is there something that I should do differently with my schematic so that I can use the prop plug while this thing is connected to 120VAC?

If I power only the 3.3v rail of my device externally with a wall wort (leaving the device unplugged from 120VAC), the prop plug works perfectly fine (and the device works perfectly fine without the prop plug plugged in)... it only malfunctions when I actually plug the device in with the prop plug.

So I think it has to do with the neutral line. If someone could give me an explanation and possibly a fix that would be great, as I do not fully understand the problem. Any help is greatly appreciated, thanks again!

Peter Jakacki

Yikes, non-isolated power and you you are connecting to another computer!!?? This type of suplly is not intended for designs that connect to other devices that then connect to mains as well. Fine if the computer is a laptop and it is running off batteries but if not then you are fortunate that this has not blown the USB port etc... Neutral is not ground and also cannot be relied upon to always be Neutral plus even though Neutral is grounded somewhere you can have many volts depending upon load currents etc.

kwinn

What Peter said, plus the lack of filtering also provides a path for all the noise spikes on the AC line to the DC output.

Mahonroy

Ok thats making sense. Under normal operating circumstances the circuitry is completely inaccessible and safe. Its just the step where I need to flash these things (and they are flashed through a bed of nails jig... the jig also tests them as well).

So is there a way I can have this powered from the 120VAC and still connect it via USB - only for the flashing step? Or will I be required to power it externally in order to flash? I was hoping I could get serial communication while its running for debugging purposes.

Do I just need to create an isolation somewhere for it to work? E.g. I'm almost wondering if there is a 1:1 power transformer that I could use. Or will I have to make a circuit that goes between the Prop Plug and the device that uses for example optocouplers to get the isolation?

Thanks again for the help and sorry for my newbie questions!

Mahonroy

And if that is the case, would just a USB to USB isolator be a quick fix?
This for example?:
http://www.ebay.com/itm/USB-isolator-1500v-isolator-ADUM4160-USB-to-USB-/181880309284?hash=item2a58e94624:g:C~4AAOSw~bFWL42y

Peter Jakacki

Simply run your laptop on battery and there is no problem although you should avoid connecting anything else to it or touch the laptop's ground as you may get a bite but normally this is fine. The USB isolator will certainly work although it is overkill and you simply need to isolate the serial and reset channels.

jmg

Mahonroy wrote: »

Its just the step where I need to flash these things (and they are flashed through a bed of nails jig... the jig also tests them as well).

So is there a way I can have this powered from the 120VAC and still connect it via USB - only for the flashing step?

You have a few choices :
* Provide power via your bed of nails jig.
This avoids any high voltages, but may limit some tests

* use a small 120VAC transformer (faraday shielded)
Depends on how much power the unit sucks

* use isolated USB
Protects the PC, but you still have a 'hot' test jig.

If you do need 120 VAC for all tests, I'd favour the isolated transformer and add an Earth Leakage breaker too.

Phil Pilgrim (PhiPi)

jmg wrote:

... Earth Leakage breaker ...

IOW, a ground-fault interrupter (GFI), as they're called over here.

-Phil

Mahonroy

Thanks again for the input! This helps a lot.

This is another power supply I'm using in another design:
http://www.mouser.com/ds/2/260/IRM-20-spec-794574.pdf

I am unable to determine if this is isolated so that I can connect a USB up to it or not. It mentions isolation class II?

MikeDYur

Phil Pilgrim (PhiPi) wrote: »

jmg wrote:

... Earth Leakage breaker ...

IOW, a ground-fault interrupter (GFI), as they're called over here.

-Phil

Technically isn't it GFCI ?

_mike

https://simple.m.wikipedia.org/wiki/GFCI

Mahonroy

Or better yet, is there a way I can plug it in and test it before actually plugging a USB in & potentially damaging something? Can I just use the ground of my computer and the power rail on the device through a voltmeter? And if I get the correct power value as opposed to an inflated value then I'm good?

EDIT:
Now that I think about it... if its isolated, then I should not get any voltage, and if its not isolated then I will get a voltage reading, is that correct?

jmg

Mahonroy wrote: »

Or better yet, is there a way I can plug it in and test it before actually plugging a USB in & potentially damaging something? Can I just use the ground of my computer and the power rail on the device through a voltmeter? And if I get the correct power value as opposed to an inflated value then I'm good?

EDIT:
Now that I think about it... if its isolated, then I should not get any voltage, and if its not isolated then I will get a voltage reading, is that correct?

Broadly correct.
You need to measure AC, but keep in mind a Voltmeter high impedance's, so 'isolated' can still give some AC voltage.
(just like your fingers give some AC reading on a Voltmeter)
Properly connected AC isolating transformer with a faraday shield, should measure a few volts max.

You still need care with the hot side, even with an isolate transformer, and in many cases you can first-test AC systems at much lower and safer voltages.
How low can you go and still test ?

Peter Jakacki

If you power your device through the USB while it is disconnected from the mains then there is no problem. Although if you forget then it could very well be a problem.

The block diagram of the IRM20 indicates that this is an isolated supply but sometimes they connect a cap and resistor across from the neutral to the output ground and this can upset some circuits and even give you a sharp little zap. But certainly this is a much better option than the non-isolated version which I would not recommend for general designs simply for safety reasons.

Just as a general note on switch-mode supplies is that I prefer switching down to 5V and then using an LDO to deliver nice smooth 3.3V to my logic and sometimes the 5V goes off to interface logic and displays. This is because SMPS always have some switching ripple and the response time is limited by the switching rate whereas a linear reg will look after that side of things much better although you can add a post filter to the SMPS to remove the ripple. SMPS for the conversion efficiency, linear for the smooth and responsive regulation.