Mµ Thermal Imager - real or fake?

[EEVblog]

The cheapest Redpine modules looks to be $27
http://www.futureelectronics.com/en/technologies/semiconductors/wireless-rf/rf-modules-solutions/80211-wlan/Pages/2003041-RS9110-N-11-02.aspx?IM=0&IT=False
No QTY price breakdown though.

[BravoV]

Just wondering if any backer ever asked them to make a video instead of just photos the update of the "secured" prototype, its easy, just wrap the camera unit with cardboard and duct tape it, make a hole at the card board big enough for the lens to work, and run a short "video" demo to show if they really have a working unit no matter how messy it is now.

They should have at least the proof of concept by now even only one right ?

[MacAttak]

(Marcus is the person handling all refunds so far - not sure what his actual role with mu optics is)

Marcus Deely said 4 days ago
Kris,

We do have a thermal sensor. It’s in an open prototype; I can see one from where I’m sitting!

Also, the target price of this thing is $400. The $150 price on IGG was planned to be at a loss.

Which is odd really. I can't recall any other crowdfunded projects where they planned to operate at a 50% loss for all units. Often there will be a very limited number of "early bird" slots at a very steep discount, but those fill up quickly and the bulk of support comes from people who are paying retail price or very near retail.

But anyhow, when trying to figure out the financial angle, you need to budget with a $400 retail price instead of $150.

[Rasz]

But anyhow, when trying to figure out the financial angle, you need to budget with a $400 retail price instead of $150.

Why? They didnt get 400, they got something like $250/unit  from the campain. You arent suggesting they run a charity?
Hmm, maybe the angle is to scam some big company into buying them out?

[rougeaux]

But anyhow, when trying to figure out the financial angle, you need to budget with a $400 retail price instead of $150.

Why? They didnt get 400, they got something like $250/unit  from the campain. You arent suggesting they run a charity?
Hmm, maybe the angle is to scam some big company into buying them out?

Maybe they're channeling Sir Clive Sinclair; lose money hand over fist selling a brilliant product at a ludicrous price, then get Alan Sugar to swoop in and take over.

Of course, all that is predicated on the thing actually being a product at all, and not just some files on a dude's desktop and a couple of e-mails from potential suppliers.

[MFX]

Re. the money numbers :-

"$125 ....... Shipipng not included. This represents $200 off the MSRP."
"$150 ....... Shipping not included, This represents $175 discount off the MSRP."
"$300 ....... (BRANDED camera)Shipping is included only in the U.S, This is a $100 discount to what we expect will be the MSRP."

So any budget calcs need to be done assuming a retail price of $325.

You need to assume that even if this was legit they are not looking to make any profit on the early units. There can be good reasons for this, I.E. to just get them out there, look at how many units Leap motion have given away free.

Martin.

[EEVblog]

What about the "H.264 Video recording" capability?
Do they expect the Atmel ARM to do that at the claimed 30fps?
Add a H.264 encoder chip to the BOM

[tom66]

I'm guessing H.264 will be done on the phone, but if they intend to do it on the Atmel, a 160x120 image should be doable (at very low bitrates and possibly 15fps max), but they need to add a $5 payment to MPEG-LA for the patent rights. I've seen 320x240 software MPEG decoders run on the Nintendo DS, which has a clock of only 33MHz and a basic ARM processor.

[EEVblog]

I'm guessing H.264 will be done on the phone

Doh, of course.

[firewalker]

They will transmit the video in a lossless format?

Image sensors usually spits out raw uncompressed image data?

Maybe their sensor produces H.264 data?

There are many phones with no H.264 coding capability.

Can an iPhone encode a video stream with H.264? Yes, it can play H.264, and record H.264. But if the camera module produces the H.264 it may not have an H264 encoder for anything else.

Alexander.

[CanadianAvenger]

Crowdfunded projects usually operate at a loss. Sometimes by design, and sometimes due to a lack of proper planning or knowledge.

The question is how much of a loss?  IF you can break even on your hard costs [raw materials/parts & external contractors] you're doing good. Internal costs like labour are essentially free at this stage. These projects are usually single persons with an idea, or small teams. In both cases they donate their time to the project to develop it, with the expectation that as a "founder" they get a payoff after the retail launch. [quite often these projects are worked on only on evenings and weekends, the founders still keep their day jobs to pay the bills] It is highly unlikely for a campaign to purposely run below cost, as no sane person wants to pay people to take their product.

So if we assume that Mu is operating with free labour and a designed operating loss of zero profit, we can look at the $125-$150 figure to be their raw costs for manufacture of the camera. So the question now is CAN we make it for that price?  My gut says no, the sensor array is likely going to cost more than that, not to mention the optics.

[Rasz]

Re. the money numbers :-

"$125 ....... Shipipng not included. This represents $200 off the MSRP."
"$150 ....... Shipping not included, This represents $175 discount off the MSRP."
"$300 ....... (BRANDED camera)Shipping is included only in the U.S, This is a $100 discount to what we expect will be the MSRP."

So any budget calcs need to be done assuming a retail price of $325.

I can tell you are not an accountant You cant budget for expenditures you have no resources for (unless you are a government).
 They dont have 400, they dont have 325, they have ~$150/unit to make this happen.
280K / 1850 units. Not even $250 I wrote previously !
No, they dont have >$460K of their own money to cover losses. If they had even a fraction of that this campain would be pointless - they could just develop 50 prototypes with own money (at $400 per prototype like you suggest), put on ebay at $800 a pop and sell overnight.
(Ironically 400 is the bottom minimum parts cost this is doable in very large scale with current image sensors).

Campain goal was 200K, with > 5000 discount slots. This is not a discount. This is like supermarket windows with permanent "going out of business prices!! promotion!!11one" signs. Basically this is a lie some like to call marketing.

Crowdfunded projects usually operate at a loss. Sometimes by design

Please tell me more, any examples of deliberately giving out money? Any campains that set their goals deliberately so they will have to fund > half of the parts cost (not to mention labor)?

[MFX]

Did you read the rest of my post? Many of the earlier posts doing budget calculations were assuming they were looking at making a profit on these early sales so assuming they are willing to supply them at cost just to get the money in/the units out there then they have $150 for component cost/development with the profit (if any) to come from future retail sales.

Martin.

[Corporate666]

Labor essentially is the assembly cost. Ok, so you have some testing, but this qty could be done in-house by them.

The often quoted figure for small player prototypes is Retail = 2.5 x BOM cost (minimum). To make it worthwhile.

Don't forget these guys are planning to sell into retail, where as a small independent approaching the huge national chains, they will need to support keystone pricing.  So for a $200 product, they will need to be able to sell it for $80-100 and make a profit... and they will have co-marketing costs, return costs (about 5% of sales), and will likely be paying for shelf space and warehousing.   So they need to be able to sell for $100 and make a profit, even after allocating about 20% of that $100 for all the ancillary BS.   Actually, given that, your 2.5x (small player to consumer) and my 10x number (small-medium player to consumer through distribution) are pretty close.  Bet Mu is nowhere near those margins though.

Don't forget the battery. Or NRE for the case. NRE for PCB and assembly gets easily absorbed into 2000 boards, but I doubt that's the same for the case.

!!!  Didn't even think about the battery, or the Wifi module!  Good catch!  Wifi modules are slim pickings at $10/ea - and these guys won't have the volumes to get the price down much.

Yes, that is very strange. If the sensor has been on the market for years, someone would have used it already in a low cost product.

But ultimately you are right, they can't possibly meet this price point and make a profit for sub 2000qty unless the sensor + lens is going to cost maybe $20 tops. And if a 160x120 thermal sensor exists for that price (and has been for some time) the market would have already gone apeshit with it.
We haven't factored in the several employees they have either.
The numbers don't add up.

This is the most confusing part of all.

It starts with "we have a really amazing new technology we've discovered and we're going to bring it to market with a cheap TIR camera... but it's so new and super-secret that we can't tell you about it or even show you our PCB because it will give it all away... we're relying on secrecy and 'first to market' to make it work!"

Then...

"Well that all fell through... but not to worry, there is another company that has had a product on the market for years and is well tested and tried-and-true, so we're going to use that instead.  Oh, but we still can't tell you anything about it because... uhhh... well, we just can't". 

The only way this makes any kind of sense is if they are repurposing an existing device for thermal.  That has been discussed quite a bit on here, but if it was easy or simple, the legions of engineers at Fluke or FLIR would have done it and be selling low end cameras as you say.  Not to mention that even if there was a technology we've all missed that would allow them to make a $150 TI camera... they have set themselves up for failure by showing images from a 'real' thermal camera, and setting expectations (in their backers mind) of getting the performance of a $2,000 device for only 6% of the cost of the real thing.

[Corporate666]

Did you read the rest of my post? Many of the earlier posts doing budget calculations were assuming they were looking at making a profit on these early sales so assuming they are willing to supply them at cost just to get the money in/the units out there then they have $150 for component cost/development with the profit (if any) to come from future retail sales.

Martin.

I disagree they have $150 for component costs.  Not even close.

Firstly, they sold a bunch at $125, but let's go with the $150 number and let's assume that is the sum total of all expenses to make a run of 1,500 imagers. 

Since that price includes all costs, it would include packaging, assembly, shipping costs, printed materials, accessories, labor, etc.  Those costs will be much higher than I bet most people assume. 

The other big one is the NRE costs - NRE for PCB's is small, but NRE for injection molded housings is huge.  Add in the one-off parts they've bought, prototype PCB's, software and hardware and such, and they must have spent at least a few tens of thousands of $$$'s on this - assuming none of them are taking a salary.  Since they needed funding to make this idea a reality, and since all those NRE and R&D costs must be paid up front, it must be subtracted from the amount they raised.   Even if they were miraculously able to spend only $30k on all of that... that leaves them $250k to make 1850 imagers.  Considering all the other costs I mentioned, that would be a max of maybe $100 in parts per unit.

Add up the cost of the parts we already know about - and it leaves them virtually nothing to spend on the most important parts - the imager and the lens.

The reality is more likely that R&D and NRE costs on such a device would cost $50k-100k easily and all their other costs will easily add 20% to the parts price.  Which means they have $75 or so to spend on parts, in total. 


Coming at it from the other side - as a small independent who approaches Best Buy and such to sell their product, they will be paying for shelf space, marketing, returns, shipping, promotions/sales, samples, etc.  Then there is the cost of sales and much more.  If they want to retail it at $400, it better have $40 or less in parts in it, or they're screwed.

Here's a decent intro to all the costs associated with distribution:

https://www.nuvonium.com/blog/view/how-to-price-your-product-for-retail-distributor-and-direct-to-consumer-sal


And we never even considered testing/approvals (FCC, and whatever equivalents around the world).

[envisionelec]

Here's a decent intro to all the costs associated with distribution:

https://www.nuvonium.com/blog/view/how-to-price-your-product-for-retail-distributor-and-direct-to-consumer-sal


And we never even considered testing/approvals (FCC, and whatever equivalents around the world).

 
Warranty repairs can be the source of big, big potential losses. My personal rule of thumb for first-run productions is 8-10% fully assembled replacements for DOA and warranty exchanges. Higher if the application is going into an abusive environment (thermal, shock, etc). Spare parts are essential, but whole assemblies are far cheaper to expedite repairs and keep customer happy.

[araugh]

On the hardware end, I’m happy to be able to say that we’ve settled on a microcontroller from Atmel, the Sam3x...

We’re narrowing down our assembly options(there have been a lot of potential manufacturers) and for our electronics and pcb manufacturing, our current top contenders are “Advanced Circuits” and “American Standard Circuits”...

Just now figuring out the MCU and assembly house?  They also seem to think that it's all downhill from here.  I've used advanced circuits for assembly runs before.  They're a good company and do good work, but they're not cheap and they're not fast.  Expedited assembled boards could cost a third of the $125 base price before you even get into the components.

Ridiculous.

[tom66]

I remain sceptical until an actual board is shown, with an actual thermal imager on it.  I'm also struggling to see how they're going to fit the MCU, sensor, wifi and presumably battery in that small camera size they show (unless it's a multi-board design or the case isn't real.)

It does have a battery right? Given that it uses wifi? Have they stated how long it will last on that?

[CanadianAvenger]

It does have a battery.  I don't see any issues with size, and the given functionality.

[WattSekunde]

After reading nearly all 22 pages here and looking the MuOptics Video Art now I couldn't resist. 

That's an old update, here is the latest:

Hello All,

I thought I’d share with you a few pieces of hardware and manufacturing news.

On the hardware end, I’m happy to be able to say that we’ve settled on a microcontroller from Atmel, the Sam3x. This 84MHz MCU is the brains of the imager, it drives the image sensor and the communications equipment. We’ve been using this chip since early in the development cycle and trust it to be able to give great performance in the camera.
We also have a wireless chipset from RedPine Signals that enables realtime wireless video from the  imager to your smartphone or computer. It’s been very fun adding wireless and we think that we have a few more tweaks we can implement to get some added functionality out of it.

- They said before that they having a lot of trouble with WiFi and that's now all fun? 
- So many problems on all parts and now it's time to add more functionality?    And more bugs and problems... 

We’re narrowing down our assembly options(there have been a lot of potential manufacturers) and for our electronics and pcb manufacturing, our current top contenders are “Advanced Circuits” and “American Standard Circuits”. Both great companies.

What a great BS bingo.
"Money doesn't care. Only the best for our customers".  Pffff.

As we work to get the case and body of the imager finalized, we have come up with a few slight modifications to the body that we are currently considering. If those ideas go anywhere, I’ll try to post pictures of the possible changes down the road.

Showing changes without any state before? 

As always, Thanks for all of your support. We can’t wait to get the first imagers into your hands and see what uses you come up with. (I personally really want to see someone mount one to a quadcopter drone).

Cheers,
-Charles and the MuOptics Team.

It is so bold. 

Once again, they talk about their development cycle and infer prototypes, but continue to still not show them.

I am very sure they have nothing. At no time! 
The only thing that interests me: Have they written the story before?

I never baked anything because there are enough really nice DIY and open hardware projects out there where you can see every detail before you buy, talk a bit with the developer and so on. You know.
Look for example at the WP 34S project (http://wp34s.sourceforge.net), at http://mutable-instruments.net, at www.midibox.org and others. The difference in every detail to MuOptics and most of the other crowd funding nerd touched vaporware couldn't bigger.

 

[Corporate666]

On the hardware end, I’m happy to be able to say that we’ve settled on a microcontroller from Atmel, the Sam3x...

We’re narrowing down our assembly options(there have been a lot of potential manufacturers) and for our electronics and pcb manufacturing, our current top contenders are “Advanced Circuits” and “American Standard Circuits”...

Just now figuring out the MCU and assembly house?  They also seem to think that it's all downhill from here.  I've used advanced circuits for assembly runs before.  They're a good company and do good work, but they're not cheap and they're not fast.  Expedited assembled boards could cost a third of the $125 base price before you even get into the components.

Ridiculous.

The part about the assembly house really bugs me.

I've spent the last 10+ years doing product design and manufacturing, and I think I have gotten pretty good at it, at least as far as independent small manufacturers go (I'm not on the same plane as Apple, for example!).

But in all that time, and dozens of products brought to market, I have never EVER even given the slightest thought to the "assembly house" when I didn't even have a device working.  Not only that, they make it sound like it's some kind of competition with Advanced Circuits (AC) and American Standard Circuits (ASC) being the two leading horses.  I'd wager neither of those companies have even heard from the Mu Optics guys - but even if they had, what in the hell did Mu say to them?  If I were someone working at AC or ASC, I'd laugh if some jackass called me up and said they were "considering" using me for their upcoming project and wanted me to justify myself to them.  And yet, Mu makes it sound like they have undertaken some deep competitive analysis and AC/ASC were rated best.

Aside from that, while I never have used ASC, I have used AC and they are not cheap - not at all.  I can't understand why Mu would choose these two as their suppliers?  Why do they even need to 'choose a supplier' when they are still changing WiFi modules and imaging sensors?  It's sort of like going to Ford dealers and telling them you "might" be considering a Ford vehicle and you want the dealer to justify why you should consider giving them your oil change business.

What a joke!

[EEVblog]

Aside from that, while I never have used ASC, I have used AC and they are not cheap - not at all.  I can't understand why Mu would choose these two as their suppliers?

Maybe because they rank high on a local google search?

Why do they even need to 'choose a supplier' when they are still changing WiFi modules and imaging sensors?

They don't.
And to anyone with any electronics product design and manufacturing knowledge, most of what they are saying seems just plain silly.
At best they seem to be completely cueless.
I'll be dumbfounded if they manage to ship this year, or at all before the money runs out.

[EEVblog]

It does have a battery.  I don't see any issues with size, and the given functionality.

Neither do I, based on the relatively chunky case shown in the promo video.

[CanadianAvenger]

Though they may want to check their operating temperature range with all that "hot air" they're putting in that case

[Towger]

After 13 pages of conjecture, how about contacting them and interviewing them on the Amp Hour?