Mµ Thermal Imager - real or fake?

[Corporate666]

See I don't understand why the confidence trickster needs to maintain confidence. He's got the money. Refunds are entirely optional on his part. He could basically walk away tomorrow and never reply.

Two possible reasons, at least.

1) They are still luring other investors and need to maintain confidence (they have openly spoken about having other investors)

2) If they just take the money and run, it was obviously a scam.  If they pretend to be trying and giving refunds to disaffected backers, they can fall back on the "well, we tried, but just could not make it happen" excuse.

If #2 was true, then of course they would not want to spend money on "prototype" stuff such as PCB's and sensors.  They are graphics guys - so it would be comparatively easy to make some one-off screenshots they could use to give the appearance of progress while not actually spending any real money on the project.

Even if they gave back 20% of the money in refunds, it's still something like $200k in their pockets net/net.

[amspire]

Bernie Madoff most certainly gave refunds.  It's part of the documented history against the guy.

Are you talking about refunds, or the promised investment returns. The investment returns payment were all part of the scam to get even more money invested - he had to religiously pay the returns he had promised. Money has been recovered but I thought that was only money recovered under legal threat or by the liquidator and from assets seized by prosecutors. Only half of 17.3 billion dollars has been recovered. He fabricated 65 Billion dollars of profits. Now that is a scam.

In the case of Mu Optics, the fundraising campaign is complete. If PayPal is retaining some money, we do not know how much. It sounds odd that PayPal would retain money - I do not understand that bit at all.

[Corporate666]

Bernie Madoff most certainly gave refunds.  It's part of the documented history against the guy.

Are you talking about refunds, or the promised investment returns. The investment returns payment were all part of the scam to get even more money invested - he had to religiously pay the returns he had promised. Money has been recovered but I thought that was only money recovered under legal threat or by the liquidator and from assets seized by prosecutors. Only half of 17.3 billion dollars has been recovered. He fabricated 65 Billion dollars of profits. Now that is a scam.

In the case of Mu Optics, the fundraising campaign is complete. If PayPal is retaining some money, we do not know how much. It sounds odd that PayPal would retain money - I do not understand that bit at all.

I know personally three people who were invested with Madoff.  Two of them started to feel it was too good to be true, and asked to "cash in their chips" (i.e. get their money back) and got it.  The third had reservations but ignored them for the returns and ended up getting swindled. 

There was heavy coverage on the news about this, because the courts were trying to claw back much of the money that had been paid out as refunds to some - because it was considered to be the money of others and they wanted to spread the losses equally. 

As for PayPal, I do not believe the claim that they are holding money - especially in light of the prior post RE: Paypal and their crowdfunding policies.  But I am very sure that the claims of the Mu guys that PayPal is monitoring the situation and "keeping them honest" is an outright fabrication.  I find that claim to be perhaps the most worrying one of all they have made, because they are claiming financial oversight of their business by the people controlling the purse strings - and by the people controlling the funding campaign (IGG).  They overstate the monitoring role of the latter, and I believe they are outright fabricating the role of the former. 

I can see no reason to fabricate the role and actions of PayPal other than to lull backers into a false sense of security about the safety of their money.  And the only reason they would do that is to prevent people demanding refunds or thinking it's a sham.

[amspire]

I can see no reason to fabricate the role and actions of PayPal other than to lull backers into a false sense of security about the safety of their money.  And the only reason they would do that is to prevent people demanding refunds or thinking it's a sham.

My only experience with Paypal refunds is when I paid $5 for a service manual PDF with schematics, and I got a link  to the freely available manual available from the Agilent site with the service section cut out. As far as Paypal were concerned, I got something, and they said didn't care less if it wasn't what was advertised. They finished the message to me with a "Case closed. We will not reopen it" statement. I would hope they have improved from 5 years ago.

Apparently it does look like Paypal has started holding onto Indiegogo campaign money:

http://www.develop-online.net/news/43946/Paypal-witholds-Skullgirls-funding

It sounds like PayPal went to the company involved and ask if they were good to refund all the money if there were callbacks. They said no - that is why they needed the money from the campaign - they don't have the money. Paypal then froze the account until a Consumer Financial Complaint Bureau complaint was filed against them.

[amspire]

Concerning the cost of the sensors, Ulis is targeting the Pedestrian Avoidance market for the automotive industry, and for that, they would have to be able to make sensors at a reasonable costs. They just invested in a new plant for volume production and they claim to offer a good price for volume purchases. I couldn't find any $ figures.

Looks like the Nano160P sensor has an I²C bus interface, so it would be easy to use. In fact, it would be dead easy to grab one of the many development boards with a fast processor and color LCD screen, connect the sensor, and get a picture. You would need a lens as well of course. Don't know how a pinhole lens would go - probably not enough sensitivity.

Edit: Just thinking about it, of course a pinhole lens would be no good - the intensity of the light from the target would be way less then the environmental IR around the sensor. You obviously need a big enough lens to make the light from the target at least an order of magnitude greater then environmental IR.

[Corporate666]

Apparently it does look like Paypal has started holding onto Indiegogo campaign money:

http://www.develop-online.net/news/43946/Paypal-witholds-Skullgirls-funding

It sounds like PayPal went to the company involved and ask if they were good to refund all the money if there were callbacks. They said no - that is why they needed the money from the campaign - they don't have the money. Paypal then froze the account until a Consumer Financial Complaint Bureau complaint was filed against them.

The guys from Mu said

"Please keep in mind that PayPal continues to hold back a considerable amount of money until such time as we can show them that all the processes and parts and orders, and POs, etc. are in place to begin delivering the product."

I believe that statement to be an outright fabrication.  As I said in an earlier post, PayPal's only concern is chargebacks and making sure that they get paid.  They do not achieve this by inspecting companies and making judgements about the likelihood of delivery of product.  The Mu guys are trying to reassure the backers that the people they gave the money to - PayPal - are taking some sort of oversight role and acting as trustees of their funds and making sure those funds are disbursed only when PayPal is satisfied that the project is "real".  I think that is a flat out lie.  There is no way in hell that PayPal would ever place themselves in any sort of oversight role.  To do so would create liability for them - and that is precisely what they are trying to avoid.

It may very well be true that PayPal is holding some of the funds.  But I believe the characterization given to the backers regarding the nature of that withholding to be an egregious hack-job of the truth.

I believe a lot of what the Mu guys say to be a hack-job of the truth.  For example, their claims that they had a case, now saying they don't.  And the claims that they can't get an image without the WiFi working - which I am sure we all recognize as being not only untrue but the complete opposite of what the truth would be.

There are many other issues - the claims that they can't show *their* design due to NDA's (when do NDA's from a supplier prevent you showing your own work?)... the fake excuses given in the video that got pulled/edited/re-uploaded... the initial TIR images being from a commercial TIR camera and not theirs, etc, etc. 

These guys play fast and loose with the truth.  I cannot conceive of a scenario where things are as these guys have presented them are true... can anyone here?

[sleemanj]

"Please keep in mind that PayPal continues to hold back a considerable amount of money until such time as we can show them that all the processes and parts and orders, and POs, etc. are in place to begin delivering the product."

I believe that statement to be an outright fabrication.

+1, their statement is complete and utter bullshit.

Pay Pal do not hold funds in any way other than to lock an account while it's under investigation, or of course in standard Auth-Capture process, of which this is not.

The only way it could be "true" is if it's indiegogo which is holding the funds and they are "confused".  I don't know if that's how igg works, but I didn't think it was.

Or they are "embellishing" the paypal buyer protection through which people can dispute a transaction within a certain period of time (err 45 days?) and paypal will usually side with the buyer if the seller can't produce evidence of shipping.

They are spinning this out if you ask me, the longer they spin it, the less chance that pay pal will reverse transactions when people start opening disputes.

[amspire]

I believe that statement to be an outright fabrication.  As I said in an earlier post, PayPal's only concern is chargebacks and making sure that they get paid.  They do not achieve this by inspecting companies and making judgements about the likelihood of delivery of product.

I agree it does not sound like a frank explanation of the situation with PayPal. As with many of their other statements, it sounds like Mu Optics have chosen the wording to convey the situation in a favourable way. One reason PayPal would get worried would be that chargebacks are occurring, and they are really worried they will continue to occur at an accelerated rate.

[mikeselectricstuff]

If they just showed a blank prototype PCB, even for an old sensor, it would tell us a LOT about whether or not they are serious with the hardware.
...but they haven't.

It would also tell everyone a LOT about the design. 

Not necessarily - anything critical could easily be blanked out

When we see a PCB, we can work out a lot, and a big problem these guys have is if they can make an IR camera cheaply, there will be a lot of people who can do it 500% better. The moment they post a picture of the PCB, people on this list will be working out a lot about the design. From the pin count, visible traces from pins, clock connections, etc, it may even be possible to work out the micro family, the WiFi chip, the IR sensor. They may as well post the circuit and parts list. You seriously think they could post the PCB photo and not have it analyzed online here in the forum?

If it was as simple as working it out from a picture of the PCB, you wouldn;t need to see the PCB - the MCU, Wifi chipset etc is a total non-issue. It's all about the sensor.

Isn't it the truth that if they show a picture of a PCB that shows no details, they will be slammed in this forum for not showing a decent HiRes picture? It could be a picture of any PCB.

Not as much as they are being slammed for showing nothing.

If they actually succeed, how long do you think they can sell the camera's for until they are undercut by Chinese-made cameras?

If cheap sensors were available, the Chinese would already be selling cameras based on them.

If amazingly they have actually used an original concept in this design. they most definitely shouldn't give away even the smallest clues until they start to ship the product.

On the contrary, their credibility is now zero, and has severey damaged any chance they may have had on getting any more investment or being taken seriously by suppliers. If they had shown some evidence they had anything, it might have still been credible.

They have been accused here of being scammers, liars, frauds - you name it. I don't see why they should add "unbelievably stupid" to the list by showing their PCB at this time or at they time they actually have a genuine first revision prototype PCB.

They could easily show evidence that they are not lying without giving anything away.  They haven't.
My bet is  the guy is just delusional and now into self-denial. I don;t thinl it will be long before the whole thing implodes and he'll start whining on about how teh world is ganging up on him.
Hopefully after that an insider will spill the beams on what actually happenned.
 

[Keef Wivanef]

Bernie Madoff most certainly gave refunds.  It's part of the documented history against the guy.

He had to religiously pay the returns he had promised.

But he didn't.
He PRINTED statements to be sent to shareholders telling them that he had (once again, miraculously) earned 10% interest.

No problem to pay out the odd investor who wished to withdraw.

Heavens to Betsy
Read the book!

No not mine, his.... http://lp.wileypub.com/Markopolos/

Nice website......
Wait a minute.... I have a cunning plan 

[Fraser]

For those wondering about the focus issue (not that important compared to the TI detector IMHO) you may be interested to hear that all of my Fire Brigade TIC's are fixed focus 1m to infinity using a 25mm multi element Germanium lens structure. They are no good for looking at PCB's unless you fit a lens in front of them like I detailed earlier. Focus can be a manually adjusted affair anyway....but you do need a lens of some description, and you do need to bandpass filter the energy coming through the lens if that is not inherent in the lens material used.

People have mentioned NDA agreements on the TI detectors....it goes further than that. When I asked for some details of my EEV Argus 2 circuits I was advised that it was a federal offence to disclose ANY circuit details for a TIC that contained a thermal detector covered under the US DoD technology export controls. That even included the power supply module ! The very friendly and helpful chief engineer has provided me with the full schematics for my other two EEV TICs because they use a Pyroelectric Vidicon tube that is not covered by export rules. I am in the UK yet the USA does have a great deal of influence in this area, even for UK manufacturers like EEV(E2V) supplying to UK buyers. Its weapons control at work. If you ask FLIR for circuit details of one of their older cameras, like my 570, they treat you like a terrorist ! Absolutely no information on their cameras circuits is released into the public domain under penalty of prosecution by the US DoD and associated legal agencies 

The rules have been changed recently with regard to low update, low resolution TIC's, but even then restrictions apply. As a manufacturer, you buy a TIC 'Engine' complete, and fit it into your chassis with your data processing chip set. No schematic detail is released on the TIC engine by the manufacturer, just I/O requirements.

In summary, I don't see the Mu project being allowed to use a high performance thermal imager chip in their design as the US DoD would stamp on them pretty quickly. If they were to buy a TIC engine, they will not be able to produce a product at the suggested price.

This project comes under the heading of blue sky thinking and vapourware.

[Lexy]

Despite the crowd funding ‘problems’, Mu thermal imager raises a technical intriguing question. Is it possible to make a thermal imager in the sub $400  range? I do not know, but a possible path is de DLP + single sensor option. A cheap DLP+controller like the DLP3000+DLPC300 costs under the $150. This DLP array scans the area and sends the pixels one by one to a cheap IR sensor like the MLX90614. Add a FPGA and maybe a microcontroller and you are ‘good to go’. But I think there would be a lot of prototyping and research effort needed to get this working, especially the optical construction. For prototyping sake, you even could replace the IR sensor with a visible light one, just to test the principle.

I’m eager to test this, but I do not have the time/money to make a prototype, but I think it is a feasible path.

My opinion on the Mu Thermal Imager project itself: I think they underestimated the project, and are covering things up, with a unneeded features, vague vendor issue’s and some fake software snippets. They do not have a working prototype, just a (good) idea. They are as far in the project as I am...

[Jay_Diddy_B]

Hi guys,

In response to a request from amspire, I am posting some real thermal images taken with my Ti9 Thermal imager. It gives you an idea of what the images look like. This imager has a 120 x 160 resolution.



In this first image the storm doors are closed. For those of you in warmer climates, the storm door is a glass door in front of the wooden door to protect the wooden door from snow and ice.
You can see that the glass door is not transparent to LWIR.




In the second image I have wedged the storm door open to reveal the wooden door. This is a panelled door, made of cedar, with a glass insert.





In this image I manually focused the imager to get the best possible image of the door.



I now stepped back without adjusting the focus. You can see how the image degrades. This was a crude attempt to show the effect of a fixed focus imager.



The final image shows a different section of the building.


Notice how all these images have a reasonable amount of Gaussian noise.

You can compare these images with other images published on the web.

The ambient temperature was about 4C. Some of the temperature variation shown in the images comes from variations in emissivity, not necessarily variations in temperature.

Jay_Diddy_B

[Jay_Diddy_B]

To Lexy,

The MLX90614 sensor is fine for a IR thermometer. Its response time is not fast enough in my opinion for an imager. The datasheet suggests the fast settling time is 0.04 seconds.

To build a 120 x 160 image (19,200 pixels)

19,200 x 0.04 = 768 seconds (12.8 minutes)

The other challenge is that if you arrange the optics so that the sensor was looking at the entire DLP array and you only turned on one pixel worth of mirrors, you would only get 1/19200 of the thermal energy image hitting the sensor. The rest of the sensor would be measuring whatever the other mirrors were pointing at.

Jay_Diddy_B

[mikeselectricstuff]

To Lexy,

The MLX90614 sensor is fine for a IR thermometer. Its response time is not fast enough in my opinion for an imager. The datasheet suggests the fast settling time is 0.04 seconds.

To build a 120 x 160 image (19,200 pixels)

19,200 x 0.04 = 768 seconds (12.8 minutes)

The other challenge is that if you arrange the optics so that the sensor was looking at the entire DLP array and you only turned on one pixel worth of mirrors, you would only get 1/19200 of the thermal energy image hitting the sensor. The rest of the sensor would be measuring whatever the other mirrors were pointing at.

Jay_Diddy_B

And don't forget you'd need a DLP device with a germanium window.

[tom66]

DLP devices are just arrays of mirrors. In fact they have no filter at all and this is one of the reasons for the dreaded "white/black dots" problem (stuck mirrors) in some older projectors and DLP TVs. Many cheap manufacturer specified and aftermarket bulbs have no UV or IR filters, which in time kills the DLP array (bombards the control transistor and sticks it in one position.)

I think it would be possible but remember how a DLP chip works. It uses PWM to vary the light intensity. A mirror can go to the -15 deg or +15 deg position only, variable intensity is achieved by using high frequency PWM. You'd have to find a way to force them fully on (I have a feeling the DLP chip would have a maximum duty cycle to limit output intensity.)  Plus DLP devices are only supposed to be used with the DLP chipset, which is designed for colour video data, so you'd have to do quite a bit of hacking around to get it to work. The actual DLP pin out is only available to customers spending many millions who want to use their own board (ex. Former LG/Samsung), otherwise, you use the TI appnote & chipset.
 
You could do the random diffusion pattern which was linked to previously, which is able to capture images with greater sensitivity.

[Fraser]

Here are some thermal pictures I took of the microprocessor board inside my Agema 570 using a 360x240 pixel FLIR 25 TIC

I was looking for any components that were either too warm or not warm enough ! On the last picture you will see a crystal oscillator that has a hot spot nearest the supply pin. These thermal cameras can locate the heat source inside a chip very well indeed. I used the cameras X2 digital zoom in many of the pictures due to the minimum focus distance of around 0.5m.

I also include some wildlife pictures just for Dave to enjoy as I know he likes wildlife....can you tell what it is yet ?     

I somehow don't think the Mu would ever have provided this quality of image, even if they had a working prototype. 1st generation products very rarely get it right from the outset. FLIR and AGEMA have been working out the bugs in TIC's for decades.

[mikeselectricstuff]

For those wondering about the focus issue (not that important compared to the TI detector IMHO) you may be interested to hear that all of my Fire Brigade TIC's are fixed focus 1m to infinity using a 25mm multi element Germanium lens structure. They are no good for looking at PCB's unless you fit a lens in front of them like I detailed earlier.

If you can adjust the lens-to-sensor distance you can focus closer - I managed to hack my FireFlir to do this and it works well on PCB-level things, although depth of field is limited.

[MFX]

Some interesting reading here :-
https://developer.paypal.com/webapps/developer/docs/classic/lifecycle/crowdfunding/

Particularly the bit about "Crowdfunding Platform Best Practices". I wonder how long before PayPal really starts clamping down particularly on things like vetting requirements.

Martin.

[Fraser]

Sadly my focus is locked off inside the waterproof housing of the ARGUS series of cameras. The outer lens is not adjustable. My little ZnSe lens trick works fine though. Even the FLIR 25 could not reliably focus closer than 0.5m

[dustout]

Here's an email update!

*oops, copy/pasted wrong section! removed; see next post*

[EEVblog]

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.

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.

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.

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.

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

[Corporate666]

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.

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.

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.

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.

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

What a bunch of bullshit.  They are just giving a tiny bit of meaningless detail so that people will think something is actually being done.  I bet it's just to stave off a rush of refund requests.

They sold these things for $125 (early birds) to $150 (the rest).   Depending on which version of the Sam3x chip they are using, that represents around 10% of the ENTIRE cost of the camera, JUST FOR THAT ONE CHIP!

I bet the PCB will be at least $5-10.  The housing will be another $5-10.  Assembly will be $10-15 easy, or more.  All the passives and connectors and associated simpleparts on the board will be another $10-15.  So you're looking at $35 to $50 before you've spent a penny on the imager or lens, or packaging, or *labor*, or shipping.

I do a ton of product design and manufacturing.  To turn a profit, your parts cost should be about 1/10th of your retail price.  If you are a big company with economies of scale, maybe you can get to 20-30% of your MSRP being your parts cost.  If you are a small player and don't need to turn a profit, then maybe 25% could be your parts cost. 

That means the Mu guys have between $13 and $40 to spend on parts.  If they are already into this for a $10 MCU, $5 PCB, $10 assembly, $5 housing, $10 worth of connectors and passives - there is no money left for the imager or lens!




And I can't be the only one thinking "wait, if they got the price down because some new imager technology had come out that they had to sign an NDA for ... how can they still do the product claiming that company is no longer going to be a supplier, but ANOTHER company has a similar cheap imager that has been on the market for years?".

If it's been out years, why would they be under NDA on disclosing it?  And why not disclose who the original supplier is/was?

The stench of BS is getting unbearable.

[Rasz]

redpine wifi modules starts at $25 (I dont think the basic one can even pair with iphone). More advanced ones are $44, so 1/3 of the total cost 

I wonder what blow off will they use. Airplane crash killing main designer? and he didnt document things because SECRUT? That might work or National security concerns

[EEVblog]

They sold these things for $125 (early birds) to $150 (the rest).   Depending on which version of the Sam3x chip they are using, that represents around 10% of the ENTIRE cost of the camera, JUST FOR THAT ONE CHIP!

$5-$8 2000qty from Digikey.

I bet the PCB will be at least $5-10.  The housing will be another $5-10.  Assembly will be $10-15 easy, or more.  All the passives and connectors and associated simpleparts on the board will be another $10-15.  So you're looking at $35 to $50 before you've spent a penny on the imager or lens, or packaging, or *labor*, or shipping.

Labor essentially is the assembly cost. Ok, so you have some testing, but this qty could be done in-house by them.
I think your lower figures are doable.
Passives/connectors/misc might only be $5.
The Wifi chipset won't be cheap? could even be $10+?

But still, yes, it all adds up.

I do a ton of product design and manufacturing.  To turn a profit, your parts cost should be about 1/10th of your retail price. 

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

That means the Mu guys have between $13 and $40 to spend on parts.  If they are already into this for a $10 MCU, $5 PCB, $10 assembly, $5 housing, $10 worth of connectors and passives - there is no money left for the imager or lens!

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.

And I can't be the only one thinking "wait, if they got the price down because some new imager technology had come out that they had to sign an NDA for ... how can they still do the product claiming that company is no longer going to be a supplier, but ANOTHER company has a similar cheap imager that has been on the market for years?".

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.