ANNOUNCEMENT: Introducing the Photon


#166

No idea yet!


#167

Cool - I guess I’ll test once it arrives.


#168

Hey guys!

Sorry for the delay in releasing some updates here!

I haven’t forgotten about you :smile:

We are up to v018 now, but many of these versions are iterations of different options we could go with for the RF switch layout and actual component used. More on this later.

My current issue with kicking out another design to the public repo is that it might not be the one we go with ultimately, and we need to have the new design RF-tuned again so a different layout other than the v010 board will have slightly different values on the Pi filter (most likely).

I can offer some suggestions right now though:

v010 photon_at7020_1.brd can be RF-tuned in this manner: replace C1 with a 3.9nH RF inductor, replace L1 with a 3.9nH RF inductor, replace C14 with a 3.3nH RF inductor.

Some cost can be saved by changing C9, C11, C13, C15 10uF 0402 caps to an 0603 package.

The RGB LED footprint needs to be rotated… this is fixed in the Spark.lbr now (will push that out today)

The RF switch needs 10pF RF dc-blocking capacitors on each of the 3 RF ports in series with the various feedlines. This has been the source of the iterations from v012 to v018 to get 3 more components to fit on the board. If your design has more room than the Photon, you will definitely have more hair remaining than me! :slight_smile:

You should do the calculations on your RF traces to ensure they are already very close to 50 ohm impedance, but you should also ask your board house to perform this impedance matching for you as well.

Here is the recommended PCB stack-up, RF traces should be 0.013" with 0.0075" clearance to GND on top layer.

0.018mm - TOP COPPER (GTL)
0.19mm - 7628M/46 x 1
0.035mm - LAYER 2 COPPER (G2L)
1.09 to 1.10mm - Core
0.035mm - LAYER 3 COPPER (G15L)
0.19mm - 7628M/46 x 1
0.018mm - BOTTOM COPPER (GBL)

The photon can be JTAG programmed with the original JTAG Shield with a number of JTAG programmers including but not limited to the ST-LinkV2, R-Link and J-Link. The new JTAG Shield has an FTDI part on-board to handle everything that the ST-Link would off board.

There’s a firmware repo for the photon, but it’s not public yet. It’s coming!!

cc: @mtnscott @mars please let me know if I missed any questions!


A small question, Photon version, which one i should use?
#169

Man, I think RF-switch is cool, but why do you make this on start of project?
First - make Photon v1 with static RF( on pcb antenna) and send to users.
Second - in parallel time make Photon v2 with RF-Switch.
It is logical. Only a few of as need external antenna.


#170

Isn’t that exactly what they’ve done? This isn’t really the start of the “Spark project” if that’s what you’d like to call it.
The Photon V1 you’re referring to would be the Spark Core in this case, which comes as either on-board or external antenna.
It isn’t als logical to proclaim it up as you claim it to be; for two versions, you’d need two PCBs, two BOMs, two supply chains, two manufacturing processes, two separates logistical tracks, and you’d have to guesstimate how many of each are going to be sold as to not have too much left-over of one type.
This is also more user friendly, since you can now use either, without having to buy a new board. For testing purposes, the on-board antenna mostly suffices, but if you’re turning into something more serious, the additional range and stability might be required.
Also, and I don’t know the numbers, but I feel like you’re greatly underestimating the amount of people that actually need an external antenna. I can imagine that when you’re launching a product that uses the Core, you’re going to want the maximum range and reliability, and that’s what the antenna does.
This combination is the best of both world, and I think it’d be more troublesome, had they chosen to split it up.

Just my two cents.


#171

[quote=“Moors7, post:170, topic:8005”]
The Photon V1 you’re referring to would be the Spark Core in this case, which comes as either on-board or external antenna.[/quote]
Yes,
but with price 20$, not 40$.
but with good wifi, with AP mode, not сс3000
but we can get it early, not in the end of April (will we recieve it??? Or it will be more troubles with rf-switch?)

[quote=“Moors7, post:170, topic:8005”]
I don’t know the numbers, but I feel like you’re greatly underestimating the amount of people that actually need an external antenna.[/quote]
Lets ask SPARK TEAM - what’s share between RFconnector and pcb antenna in Spark Core sales?
When I bay my Spark Core from seeed store - pcb antenna was sold out and RFconnector version was free to bay.


#172

@Godz, i think the idea is that:

1.) The Photon will be FCC-certified. There’s no reason to spend money getting individual versions certified. Tuning the RF circuitry is tricky and doing 2 versions wouldn’t be ideal

2.) Having an RF switch built in will demonstrate the FULL capabilities of the BM-09 since ALL the pins are broken out for development purpose.

3.) You would prefer to focus on your end product than to spend engineering effort on RF tuning should you need to have a 2 antenna option in future. It’s there and it works

4.) The onboard ceramic antenna serves as a back up if your main antenna you chose is an external add-on. That way, your end user will never run into an issue of having to strip the product and fix the antenna :wink:


#173

I’m not saying the Core is a replacement for the Photon, but rather something they have learned from. They had a kit with split options, and chose to combine them this time.

Is getting something early really worth getting something inferior? I’d rather wait a bit longer and have an amazing product, then get something half-finished just for the sake of getting it out there. It would’ve also meant that the team had to spend extra time to create, design and manufacture two separate boards, rather than combining all their efforts on something great. If you want to try things out, you can do so using the Core and replace that by the Photon once it’s here. It should be mostly backward compatible.

I can’t say for sure, but I wholeheartedly believe that if there’s any reason we won’t, they will be honest to us regarding that. They’ve send out the mail explaining why things were delayed, which seemed like valid arguments to me. They’re also going through the unplanned, and extra effort of creating a lot of boards by hand so people with tight deadlines can still get one in time. They didn’t have to do that. They don’t gain anything from it, other than a lot of extra work.
Also, they won’t charge you until they actually ship your products, so it’s not like they’re running off with your money. If you’re unhappy with the delays, you can cancel your order, which is stated in said email. You haven’t lost anything.

I think, and I’m not sure about that, that might be due to the fact that most people who buy them at seeedstudio (for example) aren’t planning on using those units for production. People who need production volumes will go directly to Spark. That might explain the difference in availability of the versions in retail stores. Just a guess though.
I’m not sure if the guys at Spark are willing to share the numbers, but it’d be interesting to see what the share is.


Here’s where the discussion ends on my end. It were just my ideas about the process, and continuing this any further won’t really benefit either of us. The choices have been made, and there’s little we can do to change that. For what it’s worth, I’m glad to wait for my Photon, if that means it’s going to be as awesome as I think it’ll be. If it’s anything like the Core, I’m sure it will be.


#174

Thanks @Moors7 and @kennethlimcp for coming to our rescue :slight_smile: @Godz we felt like having an RF switch on the Photon was a great new feature, and because of FCC certification issues we couldn’t just release an early version without it and push towards a v2 later. The RF switch is partially to provide more user choice (multiple options without buying multiple hardware) and to simplify our supply chain, as @Moors7, which helps us hit the $19 price point.

Definitely appreciate the eagerness for the new product, so thanks for your patience everyone!


#175

Very good reason, why this board must be in one version. Spark is my hobby, but it is very important for commercial propose.


#176

If this is not commercial secret, could you share share? :smile:


#178

Man, I’ll wait another year if that’s what it takes to get things right. I may not have a lot of virtues, but I do have patience.

I have bought far more uFL Spark Core devices than I have chip antenna; I deploy my stuff outdoors far from strong Wi-Fi coverage so the external antenna is a requirement.


#179

Sure thing - I think u.FL Cores are about 10% to 15% of sales. However when people need a u.FL, they really need a u.FL (@naikrovek’s use case above is a great example)


#180

A university guy here. Enterprise support really would be helpful for us while being PITA for you guys!

eduroam is everywhere and not having support for it made my 1st spark project a bit difficult. Android hotspots saved the day however!


#181

great info, thanks @BDub. I’ve now got a v001 build connected by SWD and ready to program, will try out the WICED SDK with it.


#182

Instead of an android hotspot, this (http://www.dx.com/p/xiaomi-w1n-portable-usb-2-0-powered-wi-fi-access-point-adapter-white-322131) might also be interesting. It’s a small usb router which you can use to set up an access point using your current connection.


#183

Any tl;dr version on when photon will be shipped?


#184

In the mail it said the following:

This means that delivery will now be at the end of April.

That’s for those in the first batch though. Not sure about the second batch. @Steph?


#185

ETA for new orders is on the store, which shows May at the top right now.

:smiley:


#186

Bit tricky to get the various parts of the WICED SDK working, but eventually got some code running on my Photon v001 build and the results are very promising. With a simple embedded web server I get 50% of requests completed in < 10ms, 100% in < 25ms. This is something the Spark Core was nowhere near capable of. And 5 concurrent requests without any problem. And Soft AP. Good stuff!