Iqe Share Discussion Threads

[sweenoid] 31/3/2017 16:25

I was going to post the link in Hammers post too:-)
Its interesting that Lumentum ( who will use VCSAEL IP presumably from IQE ;-)) have recently been very bullish , their recent conference transcript is interesting , the answer to this question is interesting and the CEO seems rather bullish

I wonder if you have any comments on sizing the current opportunities that you are pursuing relative to your previous experience in consumer 3-D sensing or gesture recognition. Thanks.

--------------------------------------------------------------------------------

Alan Lowe, Lumentum Holdings Inc - President and CEO [54]

--------------------------------------------------------------------------------

Yes, I think, putting aside the timeframe for when this happens I would say that the market is multiple times bigger than the connect business and could imagine quarters of $100 million plus type of marketplace growing to over $1 billion, I would expect, as we get into multiple mobile devices and multiple customers with mobile devices.

So, I'd say that every indication is that it's a huge growth driver for us. Timing for initial ramp, I think we're still going to remain quiet on that until it actually starts happening. So one, the tidbit that we wanted to leave you with this time was we're actually shipping revenue units and are very comfortable with the progress that we've made and the milestones that we've achieved.

--------------------------------------------------------------------------------

sounds VERY NICE to me - look at thos enumbers he is quoting
S
full transcript here ( worth a read- concentrate on the 3D sensing 'bits',

[poombear] 31/3/2017 16:09

It's not full service wafers, it's full service lasers. So specific to DFB lasers.

Edit
Sorry the above is incorrect, lesson learnt , don't assume the broker understands the solution....

[sheep_herder] 31/3/2017 16:00

max, I think it's just for the bragg grating.

[maxwellsdemon] 31/3/2017 15:53

Sweenoid
Thanks,

I was just wondering whether they might do more of the fabrication steps for other devices in house. If they are "providing FULL SERVICE WAFERS to customers" then it sounds as if they might be doing a bit more than just epitaxy for other devices.

(Or, maybe the term "FULL SERVICE WAFERS" just pertains to DFB lasers and perhaps also VCSELs?)

[jaws6] 31/3/2017 15:48

good talk in MICRON and few other co in CNBC before by Jimm creamer

[hammerd2] 31/3/2017 15:45

Chatter about 3D sensor using VCSEL in next iPhone.

[sweenoid] 31/3/2017 15:29

maxwell- dont have a clue

[aimingupward2] 31/3/2017 14:10

"Just in case anyone is interested"

Yes, definitely, as you'll see by the positive ratings, so thank you.

[maxwellsdemon] 31/3/2017 14:02

Are they going to be doing photo lithography on these "FULL SERVICE WAFERS" for other devices?

[sweenoid] 31/3/2017 12:52

Just in case anyone is interested
IQE are now making noises about providing FULL SERVICE WAFERS to customers and also being a 1 stop shop- what does that mean

Well take the photonics business and in particular DFB LD

Previously IQE provided a base wafer to customers , this wafer would have to be transported to be 'etched' at another facility - this could be across the world and the wafers would be transported by container at great expense and wasting a lot of time. Once etched the wafers were then transported to the customer OR competitor for its final epistaxial process!

That's why IQE bought NIL - nano imprint lithography ( for a song ;-)), IQE now have an in-house capability to do the whole process - no transport needed, time for production minimal) This etching process is highly complex, I am not sure if anyone else offers a full service , I suspect not.

As with other acquisitions, the NIL acquisition was opposed vehemently by people who still post here, this was great use of £ and those guys should eat humble pie.

The outcome will be that the revenue we now derive per wafer will be increase x2-3

So along with an increase in sales we will get more bang for each Buck, and of course a one stop shop is always more attractive to customers so there is a likelihood of taking new customers from competitors.

S

[poombear] 31/3/2017 08:36

Today is a likely day for closing positions and readying cash for ISA accounts next week?

[shanieboy01] 31/3/2017 01:17

Ipavlou it's coming soon

[lpavlou] 30/3/2017 23:14

Expect an RNS tomorrow that the Board are helping themselves to more free shares. Now the SO has more than doubled will they halve their take or still take a handful? Let's see.

[rivaldo] 30/3/2017 22:53

Skyworks are seen as being a beneficiary from the new Samsung Galaxy S8....



"Further, radio-frequency chipmakers Broadcom, Qorvo and Skyworks will see an increased content opportunity from the S8 handsets, Vinh said in a report."

[someuwin] 30/3/2017 09:13

IQE in Shares mag today...

IQE’s earnings upgrade cycle accelerates

Compound semiconductor technology specialist looks perfectly placed

[...Body of article - nothing we don't know...]

SHARES SAYS:

Years of R&D are on the cusp of paying off for IQE. We continue to see value in the long-term investment.

[poombear] 30/3/2017 08:29

I'm not sure they use Vcsels for Iris Scan and facial recognition. It may be IR LED?
I think what the Samsung phone doesn't have is the laser 3D depth scanner that is being touted for the next Apple phone.



Edit.

Defo IR LED for Iris scanner, no mention of 3D depth sensor, so must assume there is none.

[shavian] 30/3/2017 08:22

I'd say that if the G8 has meaningful facial recognition, then it must have 3d sensing.

[sweenoid] 30/3/2017 07:46

All largely Irrelevant to IQE? and tech high-brow

Shouldn't we be discussing whether the Samsung S8 announced yesterday has 3D sensing to facilitate its 'facial recognition '?
If it does then it's HUGELY GOOD news for IQE as it will require lots of VCSEL IP At present VCSELS are used in the camera's for laser autofocus but 3D sensing would add to the IQEBOM ( bill of materials) very nicely indeed. Of course if we are in with facial recognition ALL the other vendors will follow :-)
S

[horneblower] 29/3/2017 23:52

IOT power requirements are in general incredibly low. However, even tiny power usage needs to be replenished over prolonged periods needing batteries that can recharge in non-mains-supplied environments.
In this case ambient power from low-level sunlight or vibrations can be used to harvest sufficient power to recharge batteries. see Ilika Stereax 250 batteries.

[boboty] 29/3/2017 22:02

I don't claim they use a lot of power but you don't need to use a lot of power to drain a battery - especially small ones used in some devices. I like the sleep modes suggested in your note but even there you will need a clock. I think cs will be here quicker than you suggest. The power saving is an order of magnitude less.
Data centres are where big savings will be made (not iot)
Thanks for all the research - very interesting

[sheep_herder] 29/3/2017 20:03

maxwellsdemon - in a word, yes. In more words:

[sheep_herder] 29/3/2017 19:51

boboty - not doing your homework. Naughty.

Your claim that IoT devices will use a lot of power and require CS materials is plainly wrong. Look at a high end phone - it has 100-1000X the power needed in an IoT device and runs for a day on a battery. It doesn't have a fan so I'm really not sure of the relevance of that comment.

Most of the IoT devices that will form the 3B+ a year market are going to be network connected sensors. So the likes of HVAC, light control, temperature/humidity sensor etc that are installed and forgotten about. These are available today and use Cortex-M based CPUs. These MCUs are programmed to sleep 99.9% of the time. They wake up, take a sensor reading, connect to the network, send their data and go back to sleep. They sip energy using current technology nodes and last for years on batteries.

I agree that there may be use cases for CS materials in more complex sensor systems, but currently the majority of the market has no need.

ARM now supply whole systems aimed at people like NXP and Freescale such as the following link.


Nordic are one of ARMs closest partners in the IoT space and have an upcoming product range based around NB-IoT standards:



One of Nordic's highest power MCU's is this series which is designed more for the wearable markets. This is at the high end of the IoT space as it used the M4F which includes DSP instructions and a floating point unit. It's designed for doing some heaving hitting processing that would have traditionally needed a second MCU in addition to the one handling radio tasks.


On the other end of the scale you have the M0 based nRF51822 which is designed for sensor applications. The following blog app note from Nordic shows some current consumption values for the product in what I think is the Bluetooth beacon advertising mode (not too hot on the Bluetooth modes sorry):
Homework #2: work out how long a coin cell powered device would last under those conditions.

So in summary, IoT will not boost the demand for processing power directly. It may boost the need for data centres to process that data and indirectly require faster network technologies that are built upon CS materials. IQE note this in their latest results if you search for IoT and it is the only mention of IoT.

CMOS++ refers to the technology that will allow CS materials to be used for or in conjunction with current IC technologies. When I say technologies here, just to avoid confusion, I'm referring to fabrication technology nodes and it will be a long time before we see any CS only IC's being produced. I think the first ones we'll see will likely be combining the digital side of a baseband IC with some of the filtering/switching/PA side of the RF Front End. That would be a fantastic board space saver and probably offer performance advantages over the current setup. It may also be advantageous for the power switching market by again having the digital logic on die with the power transistors. IoT devices use much older technology nodes due to a number of factors including, lower cost, less performance requirement, better leakage current characteristics, and the fact that flash technology has to match the host CMOS process and doesn't scale down well. The following link shows a current top of the range MCU that has just made it to 40nm: So hopefully you can see there will be a long delay before we even start talking about FinFET never mind CS materials in the IoT space for IC manufacture.

HTH.

[poombear] 29/3/2017 18:51

Macom pdf around their next generation GaN on Si Base stations. Mentions collaboration partner and delivery of 8" (300mm) GaN on Si wafers in 2017.

[maxwellsdemon] 29/3/2017 18:36

I came across the review paper below which describes some GaN based RF filters.

I wonder if the filters being developed by IQE (or IQE in collaboration with others) are similar, in any way, to those discussed in the paper. The possibility of integrating all the components (e.g. filters, power amps, mixers etc) for a RF system in a MMIC is also discussed.


GaN AS AN ELECTROMECHANICAL MATERIAL