Transferring data over induced coupling

Printable View

- 27th January 2010, 07:40
ardhuru

2 Attachment(s)

Transferring data over induced coupling

I built a propeller clock (one of those POV things) a few years back, and am always looking for ways to improve the design.

My version uses a brushless fan motor, with a hand wound pair of coils to form an air-core transformer to transfer power to the rotating clock PCB.

The 'primary' is driven by a 555 oscillating at about 32 Khz; this I found to be somewhere near the optimal value. The pic on the PCB does everything, including driving the display, reading the RTC etc.

Now, I'd like to move the RTC to the base unit, to make the rotating PCB significantly lighter (the RTC chip, battery, holder). So, I need a way to transfer the time information to the rotating PCB; one way of course, would be to use IR.

But, I was wondering if I could inject this information into the transformer coil itself? Suppose I use another pic in the base to read the RTC, and then output the time serially, how could I incorporate that in the 32 KHz signal required by the transformer primary? Would it make sense to interrupt the modulation from the 555 to inject the time signal every few 100s of ms? I suspect the power transfer can tolerate this. My worry is, how well would the raw serial output travel over the induced coupling? And also, how easy would it be to filter out that signal from the rotating 'secondary'?

Yet another idea would be to get rid of the 555, and use the base pic itself to generate the 32 Khz, and periodically stop that to give out the time serially.

Do hope this doesnt sound too confusing.

Any suggestions, folks?

Regards,

Anand
- 27th January 2010, 11:11
mackrackit

I have not tried sending data over a coupling like that, but are you sure this will help?
You will still have hardware on the arm for the receiving PIC and you probably need/want that for a counter balance.

The easy way to cut down on the weight would be surface mounts.

But the induction idea is intriguing. Of the top I wonder if the same or similar protocol used for IR would work...
- 27th January 2010, 12:37
Bruce

1 Attachment(s)

You definitely can communicate with coils. Look at the data sheet for the MCP2030. You
can use the MCP2030 with any PIC type.

Or use the PIC16F639 with the built-in MCP2030. I used the 16F639 to build the ID badge
in the attached photo. The board on the right is the initiator, which communicates with the
ID badge up to about 6 foot away with 125kHz pulsed communications.

Look for AN1024 PKE System Design Using the PIC16F639 for schematics.

I built this for a customer, so I can't share all the code, but I can give you enough to get
started if you need it.

Here's some more good info: http://ww1.microchip.com/downloads/e...Doc/51637a.pdf
- 27th January 2010, 13:02
ardhuru

Quote:

Originally Posted by mackrackit

You will still have hardware on the arm for the receiving PIC and you probably need/want that for a counter balance.

The easy way to cut down on the weight would be surface mounts.

But the induction idea is intriguing. Of the top I wonder if the same or similar protocol used for IR would work...

Thanks for the insight.

Getting rid of the CR2032+its holder itself would reduce the weight significantly. Of course, looking at the RTC consumption I could use a much smaller battery with solder tabs. I agree, I would need to redesign the PCB, but a slimmer, longer board with the pic right at the very opposite end should be enough of a balance, I think.

SMDs? Sure, but I just havent mustered the guts to make the transition yet :(

And of course, IR would be the easiest way to go, but there would be issues with it interfering with the index pulses. Also, if I dont use IR, I could also get rid of the 3 pin module I am currently using. Plus, I find the concept of being able to achieve this thru' the existing hardware (nothing extra specifically for that) too tempting!

Regards,

Anand
- 27th January 2010, 13:33
mackrackit

SMDs are not that bad. One or two now and then a loupe and a "pointy" iron will do. More? A toaster oven works fine.

But I am going to look at Bruce's idea. Six feet! WOW! That would be great for many things.
- 27th January 2010, 14:04
ardhuru

Bruce, thanks for the pointer; it made for some very interesting reading.

Like Dave says SIX FEET? Wow!

My requirement is to transmit just a few mm, so I'm hoping I could get away without using any extra hardware at all. The question I'm wondering about, considering the *very* modest range requirement, would I need to modulate the Debugout signal, or just feed it to the MOSFET driving the coil every few 100s of ms (the rest of the time the coil wouldkeep getting its 32 KHz for power transfer).

BTW, I had no idea of the existence of the 16F639. Sure opens up loads of possibilities!

Regards,

Anand
- 27th January 2010, 15:59
Bruce

1 Attachment(s)

It took a bit of experimentation to get the 6+' distance. It's all about placement of the
small RFID LF antennas on the receiving unit. Once you get that dialed-in you get a
whopping range. The first pictures were the initial prototypes. These had around 3+'
range.

This picture is the final version that works at 6+' range. If you compare placement of the
small RFID antennas on this one to the one above you'll know how/where to put them for
excellent range. You can walk up to the initiator with the ID badge in your back pocket, and
it still gets solid comms.

The final ID board is about 1/3 the size of a credit card, and modified to fit into a keychain
type plastic enclosure.

SMD caps & resistors are all 0402 size. Boy were those fun to hand-solder on the last
prototype. Phew...:)

Quote:

would I need to modulate the Debugout signal

I used a TC4422A high-speed MOSFET driver to drive the large LF antenna, and PWM to send
data like this;

Code:

ID_Init: T1CON.0 = 0 ' Shut-down Timer1 during wake-up routine ' Wakeup sequence. 4mS ON, 500uS gap, 2mS ON, 2mS OFF. Then ' follow up with header & data. CCP1CON = 12 ' PWM ON PAUSEUS AGC_STABLE CCP1CON = 0 ' PWM OFF PAUSEUS GAP CCP1CON = 12 PAUSEUS 2210 ' 2mS ON CCP1CON = 0 PAUSEUS 2210 ' 2mS OFF ' Wakeup sequence complete CCP1CON = 12 ' Send header after initial wakeup PAUSE 1 ' 1mS high on receiver LFDATA pin CCP1CON = 0 PAUSEUS 500 ' header complete ' Transponder should now be waiting for data payload FOR LOOPS = 0 TO 7 ' loop for all 8-bits CCP1CON = 12 IF DAT.0[LOOPS] THEN PAUSEUS 500 ELSE PAUSEUS 250 ENDIF CCP1CON = 0 PAUSEUS 500 ' this pause allows 500uS for ID badge to respond NEXT LOOPS ' after challenge code is sent. CCP1CON = 0 ' turn off PWM ' Data payload complete PAUSE 5 ' Pause a bit so interrupt flag-bit can be set when ID badge responds TMR1L = 0 TMR1H = 0 ' Clear Timer1 count before re-enabling T1CON.0 = 1 ' Re-enable Timer1 on exit RETURN ' Notes: After wakeup the receivers 16F639 AFE inactivity ' timer will expire after 16mS. Any gap in transmision that ' exceeds 16mS will require another wakeup sequence before data ' can be sent.

The receiving end just uses PULSIN to grab the inbound data.
- 28th January 2010, 03:04
Byte_Butcher

Quote:

Originally Posted by Bruce

The final ID board is about 1/3 the size of a credit card, and modified to fit into a keychain
type plastic enclosure.

SMD caps & resistors are all 0402 size. Boy were those fun to hand-solder on the last
prototype. Phew...:)

Neat project! But 0402? :eek:
Do you have REALLY good eyesight or do you have really good lighting and magnification?
I'm pretty comfortable with 0603 but it's sure easy to lose those little suckers on the bench...

I'd love to see a thread with "Tips & Tricks" for dealing with SMD parts, along with suggestions and reviews of really handy tools, lighting, magnifying device, etc to make the job easier. More and more "neat new parts" are available as SMD only, and a lot of the "old familiar" through-hole parts are disappearing. SMD is a way of life now. I'd love to see a "SMD Help" category added to the forum...

steve
- 28th January 2010, 03:53
Bruce

2 Attachment(s)

Quote:

But 0402?

Stuff just wouldn't fit otherwise, and we needed the free space for the 418MHz UHF & 125kHz RFID parts. There's a UHF antenna for the Linx RF transmitter on the back of the board just behind the Primeport logo.

These RF parts need x amount of clearance from board components, so we had to go with 0402 to squeeze everything into a keyfob enclosure on a board 1/3 the size of a credit card. The CR2032 battery holder & UHF antenna sucked-up 50% of the
board space.

The board photo looks big, but it's definitely not.

We do all SMD placement under a lab microscope, and beyond the initial prototype, we do all soldering with a RoHS comlpiant reflow oven, and lead-free solderpaste screened boards.

We use the Gold-Flow GF-B-HT reflow oven and a zoom stereo trinocular microscope mounted on a boom stand for SMD placement. Photos attached.

RF design is a whole new world, and if it's not spot-on, you ain't getting that FCC stamp of approval. So you allow for whatever clearance is needed for the RF stuff.
- 28th January 2010, 04:55
Byte_Butcher

Quote:

Originally Posted by Bruce

These RF parts need x amount of clearance from board components, so we had to go with 0402 to squeeze everything into a keyfob enclosure on a board 1/3 the size of a credit card. The CR2032 battery holder & UHF antenna sucked-up 50% of the
board space.

The board photo looks big, but it's definitely not.

You don't need to sell me on the NEED for small precisely placed components! I understand why small SMD parts are desirable. :)

Now I'm interested in good tools and techniques to make SMD more accessible for us mere mortals.
I wanna know how us wannabees can do it most efficiently, without a HUGE expense. :)

Quote:

We do all SMD placement under a lab microscope,

Excellent! Does your microscope have a model number? Are there things you particularly like or dislike about it? Is it... horribly expensive? :eek:
These are the kind of tools I seek, but it's nice to have good reviews from folks who actually USE them, rather than blindly whipping out the credit card.

I love SMD components! But they can definitely push the limits of my eyesight and dexterity, so I'm in search of good tools and techniques to make things easier.

steve
- 28th January 2010, 05:23
mackrackit

I do not do stuff as small as Bruce and I am not building for certification....

I have a set of loupe and one of those headband magnifyers like jewelers use.
I also, do not laugh, use one of these.
http://www.intel.com/support/intelplay/qx3/

pretty much everything I do is custom, so all I have to be able to do is a good prototype.

For reflow I use a toaster oven from walmart.

That is how this newbee does it....
- 28th January 2010, 06:22
ardhuru

Bruce, thank you so much for sharing the code. But the reason I'm hoping to use the hardware UART for data transfer is that the receiver would then be able to receive data in the background, making the the dot-pattern display servicing much easier.

Intend to try out that; will keep you guys posted on the results!

Anand
- 28th January 2010, 15:47
Bruce

1 Attachment(s)

Steve,

A couple pair of SMD tweezers, a toaster oven, and a pair of the headband magnifiers like
mackrackit uses is all you really need. The microscope I've had for about 10 years now, and
the nameplate was peeled off a long time ago, so I don't know the manufacturer, but it
was around $400.00 new.

You can probably get one on eBay for a fraction of that if you want one. Get one with the
ring light if you do. I love the one I have. Really comes in handy when you're working with
very small SMD parts. The batch reflow oven is awesome, but it's pricey at >$5K. If you
have deep pockets a reflow oven is the way to go, and there are some good ones a lot less
expensive than the one we have. You probably don't need the high-temp lead-free type.

Anand,

Couldn't you use IR? A surface mount IR module like in the photo would work if mounted
dead center on the board, and you can connect it to the UART.
- 28th January 2010, 18:06
ardhuru

Quote:

Originally Posted by Bruce

Anand,

Couldn't you use IR? A surface mount IR module like in the photo would work if mounted
dead center on the board, and you can connect it to the UART.

Absolutely; the only concern is it would interfere with the IR index pulses I currently use to sync the display frame. I could perhaps change the index sensor to a Hall effect device.

Regards,

Anand
- 28th January 2010, 18:16
Bruce

Cool looking clock. I've always wanted to build one myself, but keep getting interrupted by
things I have to do VS things I want to do....;o)

Do you have a video of your clock in action?
- 28th January 2010, 18:16
Byte_Butcher

Dave & Bruce,
Thanks for the SMD tips. I've already got a toaster oven (with homemade controller) to do reflow. That parts covered. :)
Got decent tweezers so I'm OK there.

I just need better eyes!

I've tried the headband magnifiers and haven't liked them much. It seems like the depth of focus wasn't very good and unless my head was just the right distance from the work then it wasn't focused. Maybe I just need better ones?
Right now I do my work under one of the magnifying lamps with the fluorescent tube around it, but that's not really ideal either. It's in the way a lot.

I think a good lab microscope may be my next adventure...

Dave, who's laughing at your Intel Play microscope? Not me! I already bought a Aven "Mighty Scope" for inspection. Up to 200x magnification...
http://www.weirdstuffwemake.com/swee...scope_0863.jpg

It gets in pretty close. The resistor in the lower right is 0603.
http://www.weirdstuffwemake.com/swee...closeup_01.jpg
Great for inspection, but it doesn't really respond quickly enough to work under "in real time", so I still need...something... else.

Anand,

I'm sorry I highjacked your thread to ask about SMD. But since Bruce brought up 0402 sized parts I just *had* to ask. :)

Quote:

The question I'm wondering about, considering the *very* modest range requirement, would I need to modulate the Debugout signal, or just feed it to the MOSFET driving the coil every few 100s of ms (the rest of the time the coil wouldkeep getting its 32 KHz for power transfer).

I doubt you'll need to modulate the signal. I'd just stuff the signal direct into the MOSFET and I'll bet it works fine.
Certainly you could do IR too, but I don't see any reason why you can't couple the data through your transformer.
I'm looking forward to photos of the finished product in action. :)

steve
- 28th January 2010, 18:56
Nicmus

Just in case someone needs this information here is a good place to find stereo microscopes:
http://www.carton.ca/p1micsstereoboom-dsz44.html
I’m using the one with the light for inspection purposes and I’m very happy with it.

HTH

Nick

I’m using the one with the light for inspection purposes and I’m very happy with it.

HTH

Nick
- 28th January 2010, 19:21
ardhuru

3 Attachment(s)

Quote:

Originally Posted by Bruce

Cool looking clock.

Do you have a video of your clock in action?

Sure do, and love to show off! But the size is a couple MB, and I dont see .avi or .mov as a permitted valid attachment. Maybe I'll upload to YouTube.

Till then, here are some steady shots;

And Bruce, thanks for the compliment; coming from you I really feel flatterred.

Steve, no problems at all; I benefited from the SMD advices too.

And your opinion on the serial transfer-over-the-coil is reassuring. Intend to try out at the earliest!

Regards,

Anand
- 28th January 2010, 20:28
Bruce

WAY COOL....:)

Thanks for the pictures. The green LED on the outside edge is really a nice touch.

Quote:

coming from you I really feel flatterred

Err .. OK, and thanks, but I don't think I could have built that sweet looking clock...:)
- 30th January 2010, 01:35
Byte_Butcher

Anand,
I keep going back and looking at those photos and thinking I need to build one of those someday! (when I get about hundred and eleventyseven other projects finished :( )

You did a great job on it! Thanks for sharing.
--

Nick,
Thanks for the link to the microscopes. "Very happy with it" is the kind of reviews I like! I may quiz you about yours a little more when I get ready to shop for one. Still working on getting a decent 'scope first.

steve