[S] ProtoPowerSwitch Boosterpack (formerly AC-powered Relay BoosterPack)

[infinityis 16]

NOTE: this BoosterPack uses high voltage (120V) AC power which is dangerous and can be fatal if not handled/packaged properly

I created my first BoosterPack for the MSP430: an AC-powered AC relay. This BoosterPack takes in 120V AC and uses a transformerless power supply to cheaply/compactly create low DC voltage, which is then used to power the MSP430 Launchpad and drive a 7A relay. There is enough overhead in the currently design to power at least 3 relays, but it only has one designed in at the moment to make for the smallest/cheapest proof of concept. The board provides approximately 30mA at 3.3VDC and 30mA at 24VDC (note: the 3.3V VCC is wired directly to the 24V GND).

A video of it operating is

, pictures and schematic are here.


There are some minor changes/fixes in the works for the next version but I thought I'd go ahead and open it up for comments/feedback.

-The board was set up to accommodate either a 7A or a 15A relay, but the overlapping drill hits required for the relay didn't come out right so this one only works with the 7A relay. Also, if the 7A relay is unpopulated, you can still use the coil's thru-holes as a controllable 24V source (on/off).
-One of the AC connector terminals was bigger than the datasheet stated so it currently leans to the side a bit.
-The plan was to have a 6V+ output which passed through a LDO that provided a steady 3.3VDC. However, the LDO doesn't regulate very well without a minimum load, and MSP430 itself doesn't draw enough power by itself to meet that minimum requirement; it only works right when the top portion of the Launchpad is powered by it as well. I'll probably be changing the LDO in the future to one which does not require as much load current. fixed: it was a just board layout issue
-I am currently drawing 24V and 3.3V separately (without common ground); I have considered with the idea of just drawing the 24V and regulating it down to 3.3V levels to get even more current at 3.3V. This would also allow the same board to be used in both AC-powered and non-AC-powered applications (where all you want is an AC relay board).
-I placed the DC caps on the back of the board; unfortunately right where the crystal oscillator is. They'll be moved the next time around.
-Would something like this work better as its own board (with the MSP430 on the board itself, not as a BoosterPack)?
-There is a tradeoff regarding more current vs. being referenced to neutral; this board uses a bridge rectifier to get extra current. A neutral-referenced version would be needed to work with SCRs/TRIACs instead of relays.
-Is the general preference for 3.3V or 1.8V operation of the MSP430? Are there pins that are recommended or discouraged for use in driving the relays?
-Several "tall" components are used on the board; would it be best to move them out of the Launchpad area entirely to allow other BoosterPacks to be stacked, or is something like this more acceptable to always have as the top of the stack?


If there is any interest in this type of BoosterPack (or a customized variant), please let me know what you'd like to see. I should mention that in general, if/when sold in small quantities, it would be as a kit with the AC line cap and/or relay(s) not installed to minimize liability.

[chibiace 46]

which LDO were you experimenting with to get 3.3v output?

[infinityis 16]

It is NX1117C33Z

[RobG 1,892]

As much as I like it, I would strongly discourage from making transformer-less line voltage booster packs. Imagine what would happen if someone connected hot wire to GND and plugged it to USB.

Even isolated line voltage sitting on top of LP could be dangerous.

Have you considered making it a free standing board with MSP430 socket?

[infinityis 16]

Actually, I considered the very thing (plugging in via USB while powered by AC); while it is clearly not an ideal situation, it can be done in a controlled manner with a laptop running on battery power (or even plugged in--my power brick appears to be galvanically isolated) provided adequate insulation from ground exists. It is certainly not a recommended thing to do, because of exposed metal on the laptop. This particular design uses a bridge rectifier to increase the available current, which comes at the expense of not having digital GND = AC neutral (it varies with the sinusoid +/-30V from AC neutral). I would recommend plugging in only to GFCI outlets as an added safety precaution.

 

Isolated isolated line voltage can indeed be dangerous on top of the Launchpad if adequate protective measures are not taken (such as use of an enclosure); exposed high voltage should always be treated with extreme caution.

 

Giving all of that due consideration, there is a place in the world for AC-powered MSP430 applications. Developers for these applications may find something like this useful for prototyping purposes, and alternative methods for dealing with 120VAC aren't necessarily safer. Taking the MSP430 reflow oven as an example, there is an exposed relay controlling 120VAC out on the counter at around 2:50 in the video. An exposed relay works well enough, but if I had to pick what was safer to work with, I'd opt for the relay to be mounted on a board. And if the plan is to embed the microcontroller in the final application (packaged and all), then this type of AC-powered board makes sense to support prototyping. While the cost of the Launchpad/MSP430 is appealing to hobbyists and DIYers, it is also appealing to those who develop consumer products (a market that TI is counting on for volume sales), and it seems odd to dismiss the needs of those users.

 

I did consider and I am considering making a freestanding board with a MSP430 socket; in fact, my first version of the schematic included all that. However, to prove the concept/a prototype, it seemed better to start with the smaller, simpler BoosterPack, which turned out to be a good thing as I learned a few things to improve upon in the next iteration. While I think a board with the socket is a good idea, I think there is still a place for a BoosterPack variant, perhaps to allow live laptop-connected USB debugging (done with proper safety measures, of course) or support interoperability with other BoosterPacks.

[bluehash 1,581]

It's scary, yes.. but don't let this stop you. I'm more afraid to plug the thing the other way round on an LP.. even on the current boosterpacks. :!!!:

 

keep us updated on your progress. Coupling this with the CC2500 really opens up boundaries..especially into home automation.

[kylej1050 27]

You aren't asking for much current so you can use a small pcb mounted power transformer... Maybe even a Small switch mode supply if you're feeling adventurous.

 

As is this booster pack still has plenty of use though.

[xpg 127]

keep us updated on your progress. Coupling this with the CC2500 really opens up boundaries..especially into home automation.

 

My thoughts exactly. I have considered starting to work on such a project many times, but have postponed it until I have some more experience.

[zeke 693]

This is the reason why opto-isolators were created.

 

I sure hope you're using them.

[infinityis 16]

Bluehash: Is CC2500 something different than Zigbee? While I have seen relay boards that include wireless (for the Arduino), and I have thought about how the wireless capability is nicely coupled with the need for electrical isolation, I haven't yet delved into figuring out which protocols/chips/etc would be the best fit for the MSP430.

 

Also, a home automation project is what got me started on this path to begin with, so yeah, I'm in the same mindset there.

 

kylej1050: I've looked at small PCB mounted transformers, but they all seem a bit overkill for the application. Doing a quick search on Mouser/Digikey (...but I'm partial to Mouser as they're only 20 minutes away, so I can pick up my stuff one hour after I order while also avoiding shipping costs), the cheapest power transformers start in the $3-$4 range and take up a 1"x 1" footprint (fully half of my first 1"x2" board). When compared to a 1uF X2 poly film capacitor that costs just $0.70 each (down to $0.25 each in large quantities) which weighs more and has a 0.5" x 1" footprint, it just doesn't make sense to use a transformer if the end result is going to be packaged properly and safe from finger touch, etc.

 

As mentioned in the initial post, I have considered adding a small switch mode supply, specifically to capitalize on the fact that the transformerless supply is mostly current-limited. Thus, I could draw 30mA or so at 24VDC with a 1uF cap, and buck it down to 5V or 3.3V at 80% efficiency or more, freeing up well over 100mA of current capability at those lower voltages. With a 5uF cap, you can get even more, but then you start to wonder who needs all that current and if there isn't a better way--one of the drawbacks of the basic transformerless power supply is that it consumes the same power whether or not you have a load applied or not. If you don't have something to draw the power, it just gets burned up in the Zener diode(s).

 

Zeke: I'm not sure if your comment is a more general safety one (isolate yourself from high voltages) or actually based on the specific design I put together, but I suspect it is the former. Aside from how I've already used one, I don't see a tangible benefit to using optoisolators for this application. Unless the plan is support connections among multiple microcontroller boards with their own grounding schemes (which I have considered), optoisolators aren't going to help with much of anything. Without separately-referenced supply voltages available to each side of the otpo, you may as well just use a transistor. Everything here is referenced to the AC mains in one way or another. Even with an opto the 120VAC is still present on-board, so from a safety standpoint an opto just doesn't do much to help.

 

That said, I am using an optoisolator....not for safety, but because I am interfacing two parts of the circuit that have ground at different potentials. Specifically, the positive side of the 3.3V is at the same potential as the return/ground side of the 24V (as evidenced by the stacked Zener diodes). For the MSP430 to turn on the relay coil (selected specifically to have low coil current at a higher coil voltage of 24V), an opto was the logical choice to bridge the different ground planes. This stacked approach permitted me to have the full ~30mA of current at both the 24V level (enough to support three 8mA relay coil currents) and 3.3V levels (to drive LEDs, microcontrollers, etc.)

[larsie 121]

The cc2500 is a tranceiver chip. Connect it to an mcu. The mcu implements the protocol, which could be zigbee or a free open sourcealternative like simpliciTI. Zigbee with cc2500 would require you having access to a zigbee stack, which is relative,y heavy. TI have a zigbee stack, but I have not looked into it. I think it is more common to use more specific Zigbee modules. I could find out for you if you need sigbee.

 

More info on TI zigbee here http://www.ti.com/tool/z-stack&DCMP=HPA ... OT+z-stack

 

A fully compliant ZigBee PRO feature set on the CC2530 and MSP+CC2520 platforms

 

Btw, I am trying for a sous vide project (controlled water heater). An isolated supply is required I think. Advice is welcome.

[kenemon 29]

I have learned alot about the politics of zigbee recently. It probably is the best protocol for sensor and control wireless. Using a hub you can integrate these devices into a internet accessable mesh network. It is cheap. It is robust, and a pain in the ass. Ti makes some good radios, and they can be used as standalone modems, sniffers, etc , but have limited code capacity and cant handle the stack. Using an uC you have amazing capacity, and good range if you pick the right module. Z-stack is free, but it is only ported to IAR presently. You can play with it for a month, then it locks down and you have a bunch of useless hardware. SimpliciTI is an option, the newest ver is only ported to IAR, but the old uses CCS. Unfortuantely, the example apps are specific to only a few MSP platforms, so you have to make alterations. Many of the examples do not have the UART set up, so this poses a challenge as well to get the serial connection going. As far as I can tell, TI wants you to use its dev kits, to steer you to a specific chip. For some reason too, they are pushing it down to needing IAR as well, a $1500 investment. Please correct me if I have overlooked something. 8-)

[infinityis 16]

Thank you both for the information regarding Zigbee and cc2500.

 

Regarding your sous vide project, if it is anything like this, I see nothing which would indicate the need for an isolated supply. The heating coils and water pump are powered by AC directly so no isolation needed there. As long as the thermocouple is electrically insulated/sealed/bagged away from the water, there should be no reason why an isolated supply is needed. In fact, it is almost an ideal application for something like the board I made because you only really care about controlling a relay. The only thing less than ideal is that there is presently nothing on board to measure temperature from a thermocouple, like the components shown here. That is something I could look at adding (I was already planning on an RTD or some IC-based temp sensing in the future), but for all I know, maybe someone has already made a thermocouple BoosterPack that could be sandwiched between the LP and the AC relay board....?

[GeekDoc 226]

Question from a guy who is just learning:

 

Aren't thermocouples resistive? I would think that a voltage divider hooked to an ADC input and a bit of code would handle it. Correct?

[kenemon 29]

the reflow oven has a simple thermocouple board in the projects forum