RE: Spectrolino repair
RE: Spectrolino repair
- Subject: RE: Spectrolino repair
- From: Ethan Hansen <email@hidden>
- Date: Sun, 07 Feb 2016 17:14:46 -0800
Peter,
I can't speak as to what Steve and co. do at Chromix, but we (Dry Creek Photo) still use Spectroscans for certain profiles. The only currently available automated instrument offering comparable accuracy and flexibility is the Barbieri Spectro LFP. While the Barbieri is a beautiful instrument, it is on the costly side. Replacing our server racks of Spectroscans with Barbieri kit isn't cost effective for us.
We track each Spectrolino's reflectance performance using trend measurements against a set of ceramic color standards. Each instrument is measured weekly. We typically see minimal (within repeatability Delta-E tolerances) shifts for at least a year at a time.
The normal failure signature for a Spectrolino is the lamp output dimming. This is seen in measurements by darker colors having lower than actual L* values. Once the lamp starts dimming, the Spectrolino is on borrowed time. Fairly frequent re-calibrations are required - figure on getting 50 to 100 measurements of ~2K patches between cal cycles.
We do our own in-house calibration using a larger ceramic standard set and hacking of a diagnostic utility (Windows only) GMB sent us 15 or so years ago to determine the calibration data format. Unfortunately we have not figured out how to make this an easy process - at least a day of tech time per calibration. We also haven't found a source for replacement lamps. We cannibalize a stockpile of instruments to keep the current ones running.
The calibration tiles themselves are remarkably stable. We sent 5 or 6 to a ceramic standard vendor for measurement twice. Neither time had they shifted from the original values. The only time we have seen the cal tiles shift is from external abuse.
As a final data point, we have measured two Spectrolino units that were kept in storage for prolonged times. One for three years, the other for almost five. There was no (within tolerances of our setup) shift in measurement output over the timespan. For those using their Spectroscans on an occasional basis this is the key point. I'm sitting in an airport rather than at the office so I don't have precise numbers handy, however I'd swag the reflectance measurement life of a Spectrolino in the neighborhood of 500 measurements of a 2K patch target before recalibration is necessary.
We did, however, find that Spectrolino units are sensitive to jarring and shaking. Yours truly killed one with an errant coffee cup. Two others died when being attached or detached from the Spectroscan. Baby them, clean out the measurement head of dust, clean the cal tile carefully, and you should get quite a few measurements without needing calibration.
There might be a simple-stupid way to check the most common cause of failure - lamp aging. If there is a widely available dark standard with good repeatability (ColorChecker Passport?) a comparison of measured L* values would be a good indication of whether a particular Spectrolino was in reasonably well calibrated condition. If anyone has good ideas for this, I'd be happy to provide baseline data.
Cheers,
Ethan
----------------
Ethan Hansen
Director
Dry Creek Photo
> Roger,
>
> That is a good question. I sent it in and had it recalibrate when it was still
> Gretag, but I never saw a difference before or after, and it’s still making some
> of the best profiles I’ve every seen. What are places like Dry Creek and
> Chromix doing, as I think they’re still using Spectroscans for at least some of
> their work. The real question is how often do that actually need to be
> recalibrated? How much do that actually drift? I don’t know the answers to
> that, but I’m not going to worry about it as these days I’m probably doing a
> dozen profiles a year.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden