ColorHug 2.0

[Richard Hughes]

Hi all,

So the number of pre-orders for the ColorHug+ device are trickling in,
but nowhere near as fast as we hoped. Until we get anywhere close to
150 (break even point), I'm thinking about doing a hardware refresh on
the original ColorHug device. I've been looking at the MaZET TRUECOLOR
sensors. They are still colorimeters, but matched to CIEXYZ, which
means less people would need a CCMX file to get an acceptable result.
They are *much* more accurate.

The catch is the sensor+driver chip are a lot more expensive than the
TAOS chip I use at the moment. In volume the TAOS chip is about £1,
but the MaZET chip is £25. This would mean the RRP of the ColorHug
would rise from £60 to £85.

One large advantage to the MaZET chip is that it would fit in the
existing enclosure with the sensor in the same location, so we could
provide an upgraded PCB for existing customers to install at home.
This wouldn't be cheap, perhaps £40 (rough estimate) but significantly
less expensive than a new device.

Feedback would be great, thanks.

Richard.

[Yi Ding]

Would there be a way to produce CCMXes with the upgraded colorhug?  So that those of us who bought the original version can "free ride" with the upgraded hardware?

I'm asking because my current colorhug has basically spent the last year gathering dust because I can't get a non-redshifted profile out of it.

Richard.

--
You received this message because you are subscribed to the Google Groups "colorhug-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to colorhug-user...@googlegroups.com.
Visit this group at http://groups.google.com/group/colorhug-users.
For more options, visit https://groups.google.com/groups/opt_out.

[Richard Hughes]

On 23 January 2014 21:28, Yi Ding <yi.s...@gmail.com> wrote:
> Would there be a way to produce CCMXes with the upgraded colorhug? So that
> those of us who bought the original version can "free ride" with the
> upgraded hardware?

A good question. I suppose you could, although I don't know how
accurate it would be.

Richard.

[Yi Ding]

That would be excellent.  I would probably even buy one just to contribute some CCMX files for the displays I own/have access to.

Richard.

[Graeme Gill]

Richard Hughes wrote:

> the original ColorHug device. I've been looking at the MaZET TRUECOLOR
> sensors. They are still colorimeters, but matched to CIEXYZ, which
> means less people would need a CCMX file to get an acceptable result.
> They are *much* more accurate.

Just beware that they are probably not as good as they look -
Gerhard Fuernkranz pointed out that they have "filtered" those response curves
to a 27nm bandwidth. Characterized to more normal (smaller) bandwidths
and their filters are much bumpier. Many displays have fairly
bumpy spectra, so in practice they will still interact with this
sensor in odd ways. How much would need to be determined.

See <http://pdf.directindustry.com/pdf/mazet/data-sheet-mtcsics/15658-46042-_4.html>

Graeme Gill.

[Graeme Gill]

Yi Ding wrote:
> Would there be a way to produce CCMXes with the upgraded colorhug? So that
> those of us who bought the original version can "free ride" with the
> upgraded hardware?

Evan an improved sensor would need calibration - there will be
unit to unit variation in the filters. The easiest calibration
would be much like Richard currently does - against certain
displays. Ideally the calibration would be by characterizing
the spectral sensitivities, but this would need fairly specialized
measurement gear.

Graeme Gill.

[Chris Lilley]

Hello Richard,

Thursday, January 23, 2014, 10:23:06 PM, you wrote:

> Hi all,

> So the number of pre-orders for the ColorHug+ device are trickling in,
> but nowhere near as fast as we hoped.

I must admit that the reported issues with the original ColorHug, plus
a certain amount of "blame the user" (its supposed to look pink,
honest) made me wary of chipping in for this one.

My ColorHug sits in the spares drawer and my ColorMunki spectrometer
also doesn't get used for screen, only for reflective, because of the
well known accuracy issues with spectros on dark colours.That is a
second reason to not jump on the ColorHug+, for me, because screen
calibration and screen measurement is my most important use case.

Due to the above, I use the i1D3 (with Argyll, not the proprietary
software) for screen calibration and it works well on the three
wide-gamut screens I have available.

Given the price point, a comparative table showing how this device is
better than the very similarly priced ColorMunki spectro (and ideally, how
it compares with the i1 Pro2) might go a long way towards encouraging
interest.

Have you considered a Kickstarter or Indiegogo campaign?

> Until we get anywhere close to
> 150 (break even point), I'm thinking about doing a hardware refresh on
> the original ColorHug device. I've been looking at the MaZET TRUECOLOR
> sensors. They are still colorimeters, but matched to CIEXYZ, which
> means less people would need a CCMX file to get an acceptable result.
> They are *much* more accurate.

I was looking at them as well, due to the recent mention on the list,
and they do seem to be really nice devices assuming the curves on
their website are the actual measured curves and not just the ideal
XYZ curves they pasted in.

> The catch is the sensor+driver chip are a lot more expensive than the
> TAOS chip I use at the moment. In volume the TAOS chip is about £1,
> but the MaZET chip is £25. This would mean the RRP of the ColorHug
> would rise from £60 to £85.

Yes, but that seems a small price to pay considering that the results
would be more reliable, more consistent from screen to screen, and not
require either owning a spectro or happening across some CCCMX files
that seem to sort of make it better maybe.

Does that price include the instrumentation amps which the datasheet
suggests as well or are they not actually needed?

> One large advantage to the MaZET chip is that it would fit in the
> existing enclosure with the sensor in the same location, so we could
> provide an upgraded PCB for existing customers to install at home.
> This wouldn't be cheap, perhaps £40 (rough estimate) but significantly
> less expensive than a new device.

I notice that they require an acceptance angle of less than 10 degrees
so some sort of optical baffling or even a lens might be needed there.
Does the existing grommet hole geometry provide that?

> Feedback would be great, thanks.

I would be interested in a MaZET-based ColorHug 2.0.


--
Best regards,
Chris mailto:ch...@w3.org

[Richard Hughes]

On 24 January 2014 08:56, Chris Lilley <ch...@w3.org> wrote:
> My ColorHug sits in the spares drawer and my ColorMunki spectrometer
> also doesn't get used for screen, only for reflective, because of the
> well known accuracy issues with spectros on dark colours.

Yes, agreed, that is another concern.

> I was looking at them as well, due to the recent mention on the list,
> and they do seem to be really nice devices assuming the curves on
> their website are the actual measured curves and not just the ideal
> XYZ curves they pasted in.

Right. I'm going to order one and try it out for myself.

> Yes, but that seems a small price to pay considering that the results
> would be more reliable, more consistent from screen to screen, and not
> require either owning a spectro or happening across some CCCMX files
> that seem to sort of make it better maybe.

Agreed; but as Graeme pointed out spiky illuminants like LEDs don't do
very well with even idealized curves.

> Does that price include the instrumentation amps which the datasheet
> suggests as well or are they not actually needed?

Yes it does, and I very much think they are needed.

> Does the existing grommet hole geometry provide that?

Yes, I think so.

Richard.

[syrou]

Have you considered a Kickstarter or Indiegogo campaign?

I think that is the key to making ColorHug+ succeed. Not sure why, but I am quite sure it would! 

[Nikolaus Waxweiler]

I'd be interested in an upgrade :) Would the new sensor work better with high-gamut screens, OLEDs like in Samsung's Galaxy smartphones and the like?

[Pascal de Bruijn]

On Sat, Jan 25, 2014 at 7:53 PM, Nikolaus Waxweiler <madi...@gmail.com> wrote:
> I'd be interested in an upgrade :) Would the new sensor work better with
> high-gamut screens, OLEDs like in Samsung's Galaxy smartphones and the like?

It's hard to say up front. Having a better sensor, still doesn't get
us spectral data, like the ColorHug+ would.

Regards,
Pascal de Bruijn

[B.S.]

CDN$0.02

With no offense intended.

Granted, I have not had a chance to get back to things - life busy elsewhere.

My ColourHug experience was so disappointing (red shift, multiple monitors/computers, and unable to get the laptop and external monitor to look remotely similar), that I would not do the upgrade. I needed / expected a push here / magic completed, experience, and did not get that. (And I'm a computer guy. The CD burn / distros / etc bothered me not in the least.) I am not displeased with the expense however, being happy to have supported Open Hardware/Software some.

My conclusion at the end of the experience was that I should not have bought a colorimeter product, instead spending more for ColourHug 2.0 type hardware.

If / when / should I get back to this, and can afford it, I would be inclined to pursue a ColorHug 2.0. I would not be inclined to upgrade.

FWIW.

[Enzo Cappa]

Hi Richard,

I can't afford a ColorHug+ right now, but a Colorhug 2.0 will be more
"reachable" for me.

Anyway, I'm a little bit disappointed with the limitations of the
ColorHug. The red shift workaround is just that, a workaround. I had
the hope of having a good amount of CCMX while the community
increases, but that never happened.

So, if the Colorhug 2.0 is actually going to work properly with the
available CCMXes, I'll go for it. If not, I prefer to pay for some
kind of CCMX creation service that can potentially leads me to a fully
calibrated monitor.

Thanks,
Enzo

[christian pellegrin]

Honestly I would think twice before buying another colorimeter (colorhug 2.0 or others). I learned a lot by studying why it doesn't work and this makes me happy. But from a practical POV it's not an useful instrument. And I have a pretty standard setup, a 200-250€ range IPS display, not something exotic.

colorhug plus may worth it's price if  it can be used for other things, for example amateur spectroscopy (but I really don't have any idea if I'm saying something really stupid: I never studied how much sensitivity is needed for example). Otherwise, if only good color response is needed, perhaps the money for colorhug plus could be better invested in a better/professional display.

[B.S.]

> Honestly I would think twice before buying another colorimeter (colorhug 2.0 or others). I learned a lot by studying why it doesn't work and this makes me happy.

Care to share why / what you've learned?

> But from a practical POV it's not an useful instrument.

Could you explain what makes you say that?

> colorhug plus may worth it's price if  it can be used for other things, for example amateur spectroscopy (but I really don't have any idea if I'm saying something really stupid: I never studied how much sensitivity is needed for example). Otherwise, if only good color response is needed, perhaps the money for colorhug plus could be better invested in a better/professional display.

Except ... how do you know the accuracy of that new display without such a device as ColorHug (1.0 or 2.0), as it degrades over time, even? Or the monitor you purchase after that, or in addition to? Or that your left monitor matches your right monitor, or middle, or ...? One ColorHug, multiple devices throughout your lifetime. No?

[B.S.]

"as it degrades over time, even" - ach! The monitor, not the ColorHug. (He said, hopefully?) (-:

[christian pellegrin]

Care to share why / what you've learned?

the best reference I found are, IMHO of course, the articles on pcode.nl blog because they are short and focused and linked to relevant wikipedia articles for more advanced topics. Apart from theoretical knowledge, there is not much documentation around about how it's implemented in Linux (and there are many ways to do it that differ between distros and even different desktop environments on the same distro). It was nice to understand this aspect starting from the mentioned articles and the posts in this list. Perhaps this sounds pretty basic to you experts, but for a beginner it's a rather difficult stuff and one doesn't know where to start. Naively one could look at Colorhug like a plug'n'play tool, but it's not so.

Could you explain what makes you say that?

Calibrated instrument are useful because they eliminate, as far as possible, the subjective aspect of measurement. If I have to judge, by eye, the results (is there some redshift left?) I don't see much value as a characterization tool for performing calibration. From a practical point of view it enhanced a bit what I see on the monitor. As far as I understand (but maybe I'm mistaken, correct me if you think so, I can provide you any information you need to judge this) it's more because it fixed the gamma correction than the color balance.
 

Except ... how do you know the accuracy of that new display without such a device as ColorHug (1.0 or 2.0), as it degrades over time, even? Or the monitor you purchase after that, or in addition to? Or that your left monitor matches your right monitor, or middle, or ...? One ColorHug, multiple devices throughout your lifetime. No?

After the troubles (reported non only by me, but by many people around the world) I decided that the best thing to do is to print some reference images with different print services an check them against what I see on the monitor. It's cheap. Another thing that I learned is to stay away from Adobe RGB color space if you don't have a wide gamut monitor and a print service that you trust not to ignore (or otherwise misinterpret) the color space specification. A pro would laugh at this, but without an expensive wide gamut monitor and a well though workflow from camera to print, not using sRGB just means looking for troubles (and pics put on the web should be in sRGB by specification AFAIK). Nowadays even prosumer cameras like the Nikon D5100 have an excellent .jpg sRGB output that I use most of the time (I have raw alongside if I have to save an underexposed important shot or if I have the time and the inspiration to go for an elaborate editing of some shots).

On the other hand a pro would value peace of mind more than hacking joy, so perhaps he will bring the (expensive) monitor to the service every couple of years or so to check if it's calibrated. I guess it won't cost much more than 350€ planned for the spectrometer during the monitor life-time. Having some color output on many devices could be an attractive application but, as it's mentioned on the FAQ, it makes little sense because you have to accept the quality of the worst device. And, if you read the articles mentioned above, "So practically speaking, if you need a reliable color managed setup, you should probably avoid dual screen setups altogether." It's hacking playground, not something working out of the box. Anyway, if I can afford a multiple wide gamut monitor setup, laptop with wide gamut monitor and a TV with wide gamut in the living-room, the cost of Colorhug plus won't be a problem. :-)

Don't get me wrong: I don't have nothing against this project, I'm just answering the question (and I think I motivated the answer). When I got my Colorhug 1.0 at first I was disappointed but then I said: stop, this is an Open Source project, not a boring tool, let's hack on it and enjoy. The lmit is, as usual unfortunately, the available time. If I will read some good reports about Colorhug 2.0 of course I will buy it (I'm short of time to be an "early adopter" unfortunately). But a tool such Colorhug plus just for color correction (assumed that it works, the 1.0 version showed that color science is not an easy field) makes little sense to me: a 350€ tool to calibrate a 250€ monitor. This is the reason I mentioned other possible applications could make it more attractive as a hacking tool.

This is just my 2 €urocent from my experience as an amateur photographer, of course the opinion from experts are much much more valuable

 

[B.S.]

On Monday, 3 February 2014 02:42:57 UTC-5, christian pellegrin wrote:

This is just my 2 €urocent from my experience as an amateur photographer, of course the opinion from experts are much much more valuable 

Thank you! A most excellent and reasoned post.

One thing I will point out though ... personally, I concluded that I am no longer going to worry about print. (For the few times, if ever, I'm going to, I'll worry about it then. And accept that there may be some hoops to jump through at the time. Every time.) Therefore, my viewers and myself will only ever view via web / monitors, and since there is no way to control the colour calibration of every one else's monitors, I'll live with sRGB and non-wide gamut monitors.

So, within those constraints, perhaps ColorHug 2.0 class hardware, applied like you say, across all monitors, TVs, smartphones, et al, makes some amount of sense?

I'll know, for myself and my devices, blues and reds will be 'sufficiently' close, and black and white details apparent, to the extent I care at the time / on that image. So when I have e-mail up on one screen, and photo editing the other, and slide something to the other monitor for the moment I won't be completely exasperated at the sudden non-trivial colour shift between the two?

[christian pellegrin]

On Monday, February 3, 2014 10:24:44 PM UTC+1, B.S. wrote:

So, within those constraints, perhaps ColorHug 2.0 class hardware, applied like you say, across all monitors, TVs, smartphones, et al, makes some amount of sense?

I really don't know if this makes sense because the gamut between the devices you cited have a very different gamut. I was hopping at least to be sure to have a good white balance between my PC monitor (that I use for editing pictures) and the TV (that I use for showing them to a larger audience). But I haven't tried yet because I'm still not satisfied with the results on my working PC's monitor. Let's see if the problems with 1.0 can be fixed and how 2.0 will look like.
 

I'll know, for myself and my devices, blues and reds will be 'sufficiently' close, and black and white details apparent, to the extent I care at the time / on that image. So when I have e-mail up on one screen, and photo editing the other, and slide something to the other monitor for the moment I won't be completely exasperated at the sudden non-trivial colour shift between the two?

I guess that the part of color calibration done by the X11 server will applied easely (I don't see any problems with loading different ICCs to different displays at startup). The part specific to the application is much harder to get right because there isn't a per-display indication of the used ICC. And what are you saying is even harder: there has to be an event notifying the application that ICC profile has changed (and applications must be written to do this). AFAIK there isn't something like that available.
 

[Pascal de Bruijn]

There is, and it's called colord-gtk IIRC (and Richard wrote it)...

But few applications actually use it in reality. Darktable is one of those few.

Regards,
Pascal de Bruijn

[edwin]

On 02/04/2014 07:50 PM, christian pellegrin wrote:
> I guess that the part of color calibration done by the X11 server
> will applied easely (I don't see any problems with loading different
> ICCs to different displays at startup). The part specific to the
> application is much harder to get right because there isn't a
> per-display indication of the used ICC.

There is an _ICC_PROFILE atom in Xrandr, per display, and you can ask for change notifications.

http://www.freedesktop.org/software/colord/faq.html#application-support
http://www.oyranos.org/wiki/index.php?title=ICC_Profiles_in_X_Specification_0.4

I only use one monitor so I don't know how well this works in practice.

Best regards,
--Edwin

[tilew...@gmail.com]

Any chance we'll see some Windows native calibration software in addition to the Linux stuff?
Nothing against it, it's just that while you can calibrate in Linux and import the ICC profile over, the different variables from the AMD/ Nvidia/ Intel display drivers can mess with the calibrated results as far as I know.

[Richard Hughes]

On 3 November 2014 16:22, <tilew...@gmail.com> wrote:
> Any chance we'll see some Windows native calibration software in addition to
> the Linux stuff?

From me, very unlikely. There are a couple of thorny legal issues I
can't really go into, and aside from that, it's been over a decade
since I did any UIs for Windows. You probably don't want me dusting
off my copy of Visual Studio 6.

Richard

[Florian Höch]

Am 03.11.2014 um 17:22 schrieb tilew...@gmail.com:
> Any chance we'll see some Windows native calibration software in
> addition to the Linux stuff?

When Argyll CMS supports the ColorHug 2, then dispcalGUI will too.

Florian.

[Richard Hughes]

On 3 November 2014 20:58, Florian Höch <lists+colo...@hoech.org> wrote:
> When Argyll CMS supports the ColorHug 2, then dispcalGUI will too.

It's supported in the ArgyllCMS dev-src tarball right now. I probably
ought to get one of the prototypes to you too, dispcalGUI is used by
tons of people with a ColorHug. If send me your address off list I'll
send you one of the prototypes.

Richard

[Florian Höch]

That's great, I really appreciate the offer. I'll send you my address.

Florian.

[tilew...@gmail.com]

You have any more of those prototypes to give out? :P hahaha
But being serious,
I wasn't aware of dispcalGUI, it sounds great!
Any idea when you guys are going to release 2.0? If it's soon enough I'll just wait for this refresh to buy.

[Richard Hughes]

On 4 November 2014 05:43, <tilew...@gmail.com> wrote:
> You have any more of those prototypes to give out?

I only made 4, each of which costing a great deal more than £60 due to
the small batch.

> I wasn't aware of dispcalGUI, it sounds great!

It's a power-user tool for people that know what they are doing.

> Any idea when you guys are going to release 2.0? If it's soon enough I'll
> just wait for this refresh to buy.

That's a good question. My tentative plan was to sell the remaining
ColorHugs in stock, and then have a few months break with no stock.
First and foremost, Ania deserves a break as she's the one packaging
these devices up and taking them to the post office 3 times a week. I
also need some time to actually sit down and make ColorHug2 actually
work properly, and I can't really do that and make the current devices
at the same time. Also, and more importantly, I don't want someone to
buy a ColorHug and then I announce ColorHug2 the very next day. I'd
feel gutted if that was me.

I'd say I have about 6-7 weeks worth of ColorHug stock, then maybe
take a 2-3 month break, so fingers crossed start selling ColorHug2's
about March 2015 time. None of that is set in stone, of course.

Richard

[robert....@gmail.com]

On Friday, January 24, 2014 at 10:23:51 AM UTC+1, Richard Hughes wrote:

> Yes, but that seems a small price to pay considering that the results
> would be more reliable, more consistent from screen to screen, and not
> require either owning a spectro or happening across some CCCMX files
> that seem to sort of make it better maybe.

Agreed; but as Graeme pointed out spiky illuminants like LEDs don't do
very well with even idealized curves.
 

I am curious about that "bumpiness" in the spactra. Is my understanding right, that a tristimulus sensor with perfect XYZ curves would give accurate colors independently of the spectral power distribution of the LEDs (or whatever illuminant)? Ive got the feeling, that my eye is pretty happy with its LMS-tristimulus curves ;). Designing a sensor with equal spectral response (ignoring technical difficulties for the moment) should give accurate colors every time or am I wrong? Is there a difference between a sensor with perfect LMS response and a sensor with perfect XYZ response in this regard? I mean a sensor with equal LMS response to the human eye should always give the correct colors. Can the same be said of a perfect XYZ sensor? By the way: a Colorhug 2 with improved color response would be much appreciated. I had my share of disagreements with the original design when matched to professional printing services. A photospectrometer aka ColorHug+ might be nice as well. But I honestly fail to see why a tristimulus should not be sufficient given the workings of the human eye.

cheers Robert

[Richard Hughes]

On 11 January 2015 at 11:17, <robert....@gmail.com> wrote:
> By the way: a Colorhug 2 with improved color response would be much
> appreciated.

That's what we've done :) http://www.mazet.de/en/products/jencolor --
I'll be launching CH2 in March/April if everything comes together in
time.

> A photospectrometer aka ColorHug+ might be nice as well.

Still the plan, and I'm in talks with Hamamatsu that I can't really go
into yet :)

Richard.

[Graeme Gill]

robert....@gmail.com wrote:

> I am curious about that "bumpiness" in the spactra. Is my understanding
> right, that a tristimulus sensor with perfect XYZ curves would give
> accurate colors independently of the spectral power distribution of the
> LEDs (or whatever illuminant)?

Yes, within the limits of XYZ matching.

> Ive got the feeling, that my eye is pretty
> happy with its LMS-tristimulus curves ;). Designing a sensor with equal
> spectral response (ignoring technical difficulties for the moment) should
> give accurate colors every time or am I wrong?

You seem to be saying the same thing. Putting aside some technicalities
(Vision specialists are often interested in cone sensitivity curves with
other optical elements like the lens etc. treated separately), XYZ
is a 3x3 matrix transform of the eye's LMS spectral sensitivity curves.
The matrix was chosen specifically so that the Y curve is the same as the
luminous efficiency function, and X and Z curves are positive.

> Is there a difference
> between a sensor with perfect LMS response and a sensor with perfect XYZ
> response in this regard?

Not in the sense you are using.

> But I honestly fail
> to see why a tristimulus should not be sufficient given the workings of the
> human eye.

In detail, things are a lot more complicated than that. XYZ only works perfectly
under identical viewing conditions with two observers who perfectly match the
standard observer, observing patches of identical angular size.
As soon as you vary any of these conditions, various adjustments have to be made.
That's why there are the 2 and 10 degree observers, why there is a parametrized
version of the standard observer that takes aging into account, why there
are color appearance models, and why there is a good deal of research at
the present into observer variability modeling.

Graeme Gill.

[michael...@gmail.com]

On Sunday, January 11, 2015 at 10:30:17 PM UTC+1, Richard Hughes wrote:

That's what we've done :) http://www.mazet.de/en/products/jencolor --
I'll be launching CH2 in March/April if everything comes together in
time.

Could you please give us an update on CH2? I was thinking about buying a CH but reading about the CH2 I got the feeling that waiting a few weeks now will be worth it. Did you run into any complications or will the device be announced soon?

Cheers,
Michal

[Richard Hughes]

On 28 March 2015 at 08:55, <michael...@gmail.com> wrote:
> Could you please give us an update on CH2? I was thinking about buying a CH
> but reading about the CH2 I got the feeling that waiting a few weeks now
> will be worth it. Did you run into any complications or will the device be
> announced soon?

Well, a two part answer. The reason for the delay is that we're having
now the roof replaced on our house at the end of the month, so I've
been super busy with that. Also ColorHugALS; we've sold 40 or so in
the last few weeks but it slowing down now and I have stock to spare.

The second more exciting part is that this week I'm going to sell the
first batch if anyone wants to be an early adopter. If you're not
familiar compiling code and using the command line the original
ColorHug is the best bet for quite a few months, but I know at least 3
people that really want to get their hands on the new device to start
playing with the high speed acquisition mode for latency measurements,
and making a few available now will hopefully buy me some time for
more testing and getting the firmware up to snuff for general users.

I've asked Ania to sanity check the page and I'll write a mail to this
list in a few days.

Richard

[michael...@gmail.com]

On Wednesday, April 1, 2015 at 9:57:46 PM UTC+2, Richard Hughes wrote:

The second more exciting part is that this week I'm going to sell the
first batch if anyone wants to be an early adopter.

Good news! Count me in.

Cheers
Michal

[germano....@gmail.com]

Il giorno mercoledì 1 aprile 2015 21:57:46 UTC+2, Richard Hughes ha scritto:

The second more exciting part is that this week I'm going to sell the
first batch if anyone wants to be an early adopter.

Will an upgraded PCB for existing customers be available?

[Richard Hughes]

On 3 April 2015 at 10:59, <michael...@gmail.com> wrote:
> Good news! Count me in.

I've got to spend a few hours tonight fixing up the shipping program
to recognize the new device and then I should be ready to go.

Richard.

[Richard Hughes]

On 11 April 2015 at 18:53, <germano....@gmail.com> wrote:
> Will an upgraded PCB for existing customers be available?

That's an interesting question I'm not sure I know the answer to. The
PCB is 100% different from the old one, and the new rubber assembly
and mounting pad has the appropriate cutouts for the sensor IC that
has to sit very close to the sensor. There's a certain knack to
mounting the PCB on the pad, and once you've done 2000 odd it's pretty
easy. If you get it wrong, you've got to sit there for a few mins with
a bottle of isoproply alcohol scrubbing bits of the old pad off from
both parts as when it's stuck you've got to rip it apart to get the
PCB off.

Also, from a cost point of view, the PCB must be like 70% of the
complete cost, so it's not like I could offer the PCB for much cheaper
than the fully built (and tested) unit with all the accessories. I've
decided to only send the CH2s with signed-for shipping as I've spent
so much time on each one if one got lost I would probably be quite
upset!

One person also asked me if I'd consider shipping the unpopulated PCB;
this doesn't make much sense as you either need a reflow oven or a
hot-air soldering station for the IC and sensor. So I'm not sure it
makes sense. Ideas welcome.

Richard.

[B. S.]

One thing I have to say ...

Thank you so much, Richard.

Not many would take the time to offer such a well explained and reasoned
out response.

What you say makes sense, and probably not intuitive to an average
retail buyer (like me). Superficially, one (like me) might intuit that
significant cost savings should be effected, while those in the know
would not.

Thanks for putting us all in the know.

-- Bill