Howtek drum scanner DIY thread: maintenance, troubleshooting, mods

[Fixcinater]

2 week update, since last post I have repeatedly tried the MacUtil FST and could not replicate the single time it ran through all of the tests. It would always lock on test 5 or 6, and just sit there.

So, I've found a local Windows XP box and am setting that up now to run the scanner and will update again once I can run the full Windows FST and see what it says.

I really would prefer to run with a Mac box and maybe I will run Silverfast on it if I can get the scanner running smoothly with the PC to test.

 

[brbo]

What are tests 5 and 6 called?

(sorry, I'm running my scanner with a headless winxp machine and if I want to run FST I need to attach monitor/keyboard)

Are you running FST from OSX? I never managed to do that...

 

[Fixcinater]

Sorry, I should have been more specific, it would fail on pretty much anything I started except for the first few tests. If I manually selected #17 (whatever that one is), for example, it would also fail but wouldn't get there if I selected all to run in a cycle.

I used it on OS 9, it would not run in OSX via Classic.

 

[Fixcinater]

Small update, plugged in the 2906 Adaptec PCI:SCSI card I had been using with the Mac and got it to actually scan something in Silverfast demo mode on the Windows machine. Still had some patchy random jitter in the pre-scan and scanned result but it's at least talking and forming some sort of image compared to the total mess it was when I put up the screen shot before.

Odd thing, it shows up in Silverfast demo mode and scans; in Trident demo mode it shows up and I was able to change to the PC-specific R535 firmware as it had the Mac file loaded; it doesn't show up at all in the FST. Will dig a little deeper again tomorrow.

Also, just wanted to say thank you again to everybody here, especially BRBO and onnect17. I'm super grateful for the time you've spent. Hopefully I can get this thing fully sorted and pass the knowledge forward to keep these things running.

 

[brbo]

Hmm... I never thought it needed a PC-specific firmware. Is there a source of this information? I was running the same firmware on Mac and now on the PC (R813 I think).

I'd try a different SCSI card and cable. I'm sure onnect17 will have a better guess at what else could be your problem...

 

[Fixcinater]

The Howtek file page on Aztek's site shows a Mac file and a PC file for every scanner, and onnect17 warned about the same. I don't know any more than that.

I have seen that the Adaptec 2906 card (which I have) has a rather poor reputation so changing it may indeed bring some results.

 

[onnect17]

Talking about the firmware for the scanner. I have no doubt the resulting binary code in the flash memory in the scanner is the same, via PC or via Mac. However, the internal. format of the source files with the hex code used by the FSTs could be different.

Keep in mind also the scsi cards BIOS could be different. Yes, many have a version for PC y otra for Mac.

The jitter is most likely mechanical.

 

[brbo]

Yes, I am using leds (XML-Color). I was never able to get a noise free scan at 6 microns with a regular lamp. To diffuse the light I use sand paper with a 400 grit or thinner directly on the surface of the led.

Hi!

May I ask what is the reasoning behind going with a RGB led instead of single white? I'll go ahead and try to build a led light source and would just like to understand what I am doing. Why, for example, this wouldn't work:

http://www.cree.com/led-components/products/xlamp-leds-integrated-arrays/xlamp-cxa1304-8def74

Thanks!

 

[onnect17]

Hi!

May I ask what is the reasoning behind going with a RGB led instead of single white? I'll go ahead and try to build a led light source and would just like to understand what I am doing. Why, for example, this wouldn't work:

http://www.cree.com/led-components/products/xlamp-leds-integrated-arrays/xlamp-cxa1304-8def74

Thanks!

Sure. The layers in the color film are "tuned" to be the most sensitive to Red, Green and Blue, so the ideal light source to "read" the info would be a pure RGB, in order to minimize the crossover and get the highest color purity. The white leds actually are blue leds with a phosphorus coating to radiate in the green-to-red area of the spectrum, carrying the unwanted yellow-orange portion.
Once you get the XML-Color based light source working then scan an IT8 target and look at the RGB numbers (using PS or similar), particularly at the pure read, green and blue patches. Then compare it to a any other light source scan of the same target.

 

[brbo]

Thanks! Obviously, I don't know what range of wavelengths the RGB parts in the LED cover. Hopefully they cover all the ranges of the dyes that are used in various films.

Deltas after making a profile from IT8 target with halogen and then LED source might tell us something about that, but some other real world scans of films might still be somewhat "off" when scanned with less than full spectrum, no?

One more thing, do PMTs need to be cleaned (the surface) from time to time?

 

[onnect17]

Thanks! Obviously, I don't know what range of wavelengths the RGB parts in the LED cover. Hopefully they cover all the ranges of the dyes that are used in various films.

Deltas after making a profile from IT8 target with halogen and then LED source might tell us something about that, but some other real world scans of films might still be somewhat "off" when scanned with less than full spectrum, no?

One more thing, do PMTs need to be cleaned (the surface) from time to time?

It does not have to cover the full red, green or blue. Actually, a fully monochromatic RGB light source built with laser would perform even better but some issues related to coherence could be introduced.

If I understand it correctly you do not need a yellow light source to get the yellow out of a flower in the color film. The emulsion recorded it as a combination in the red and green sensitive layers.

The surface of the PMTs should be fine, unless the scanner is being used in a really dusty environment. Humidity, salinity, etc. could be a bigger issue. I remember seeing an aperture wheel with some corrosion in the 30 microns hole, blocking half of the light.

 

[brbo]

If I understand it correctly you do not need a yellow light source to get the yellow out of a flower in the color film. The emulsion recorded it as a combination in the red and green sensitive layers.

Sure, there are blue, green and red sensitive layers in the color slide and negative films, but I've only heard about yellow, magenta and cyan layers present after developing the film. As far as I understand it, the CMY dyes then act like filters. When shining the light through the film magenta dyes block the green spectrum, yellow dyes the blue... and so the original scene is reproduced. Now, in the perfect world containing RGB spectrum (of the original scene) respectively in their designated RGB layers of the film would be 100% perfect. The dyes after developing would be perfect filters (100% blocking of one and 100% transmission of the rest of the spectrum). In this case I can see how a RGB light (with narrow peaks) would be able to recreate the original scene.

In reality, though, I think there is far more "leakage" at original recording of the scene in their respective RGB layers in film, dyes are far from being able to perfectly "cut" the spectrum to desired range of wavelengths. But that's ok. Film manufacturers have developed many different methods to solve those issues and, probably most important for our case, live with them. They actually rely on all those issues or even introduce them so they can create various "characters" in films. For example, they might observe that the blue layer is underrepresented in the exposed film, so they would just use dyes that would block less of that spectrum, but they don't want to totally cancel out the original problem etc...

Now, in those cases a broad visible spectrum is actually needed/prefered to "inspect" the film, because the film was designed (I know, another assumption) to be observed under that kind of light.

(I'm perfectly aware that this all might be total bul****, I'm no expert, just using _my_ common sense.)

edit: Interesting link on this subject...

The surface of the PMTs should be fine, unless the scanner is being used in a really dusty environment. Humidity, salinity, etc. could be a bigger issue. I remember seeing an aperture wheel with some corrosion in the 30 microns hole, blocking half of the light.

Ok, I won't bother with getting to the PMTs for the time being.

The aperture was actually my next question. How to be sure that my 60mic aperture actually is 60mic and not narrower. And what to do about it if that is the case...

 

[onnect17]

Sure, there are blue, green and red sensitive layers in the color slide and negative films, but I've only heard about yellow, magenta and cyan layers present after developing the film. As far as I understand it, the CMY dyes then act like filters. When shining the light through the film magenta dyes block the green spectrum, yellow dyes the blue... and so the original scene is reproduced. Now, in the perfect world containing RGB spectrum (of the original scene) respectively in their designated RGB layers of the film would be 100% perfect. The dyes after developing would be perfect filters (100% blocking of one and 100% transmission of the rest of the spectrum). In this case I can see how a RGB light (with narrow peaks) would be able to recreate the original scene.

In reality, though, I think there is far more "leakage" at original recording of the scene in their respective RGB layers in film, dyes are far from being able to perfectly "cut" the spectrum to desired range of wavelengths. But that's ok. Film manufacturers have developed many different methods to solve those issues and, probably most important for our case, live with them. They actually rely on all those issues or even introduce them so they can create various "characters" in films. For example, they might observe that the blue layer is underrepresented in the exposed film, so they would just use dyes that would block less of that spectrum, but they don't want to totally cancel out the original problem etc...

Now, in those cases a broad visible spectrum is actually needed/prefered to "inspect" the film, because the film was designed (I know, another assumption) to be observed under that kind of light.

(I'm perfectly aware that this all might be total bul****, I'm no expert, just using _my_ common sense.)

edit: Interesting link on this subject...



Ok, I won't bother with getting to the PMTs for the time being.

The aperture was actually my next question. How to be sure that my 60mic aperture actually is 60mic and not narrower. And what to do about it if that is the case...

Based on my short experience and IMHO the biggest sources of color error is not related to different manufacturers. It has to do more with:
- Emulsion condition, i.e. proper storage.
- Light temperature control during shooting.
- Film development quality.

Once I got control of it you could hardly see the difference between E100G and 100F.

Regarding the 60 microns aperture wheel, the automatic gain adjustment in the amps makes it hard to compare it the other apertures. The only way I know is a physical inspection under the microscope. In practical terms, I would no bother with it. The optical box in the 4000/4500 is a lot better built than the one in the 8000 (with the 30 microns aperture).

Now, remember the PMTs pretty much convert photons energy in a current-then-voltage value. Well, because the color bandwidth of each channel is narrower with an RGB light source, the amps operate with a better SNR. I must say I was quite impressed with the image coming out the scanner the first time I used the RGB led. Previously, using the standard lamp the level of noise with the 60 microns aperture was quite high, to the point that I used the 130 microns most of the time.

 

[brbo]

I've ordered "white" and RGB leds. It will take a while to get them. I will then compare and see what works best...

How's it possible that flickering is not a problem with LEDs if you are using the existing power supply for halogen bulbs?

Thanks!

 

[onnect17]

I've ordered "white" and RGB leds. It will take a while to get them. I will then compare and see what works best...

How's it possible that flickering is not a problem with LEDs if you are using the existing power supply for halogen bulbs?

Thanks!

I am using the same 11v DC feeding the halogen lamp. The xml-color has 4 dies r,g,b and white. I don't use the white.

I am sure you can handle basic electronics. Here's the list of components I am using in this design. I am sure other members can come up with a better approach.

- Heat sink + fan. The Howteks are able to compensate for small changes in the light source but it's better to keep a lower and stable temperature. It's also used as a base.

- Arctic Silver™ Thermal Adhesive. To attach the led and the resitors to the main sink.

- Capacitor. 0.1uF is fine. Just to be safe with any noise introduced by the fan.

- 2 x 1 ohm metallic resistor. That will allow around 900 mA into the leds.

- Fiber optic coupler. I make these ones in a minilathe using teflon and then cover with hi-temp metallic paint.

- Molding silicone. To cover the "whole-thing"

Perhaps a picture of a few lamps in the middle of being built ... ;-)

https://1drv.ms/i/s!AugK5NQO79gPgyyqi7XoH464u6Nl

 

[brbo]

So the white led (still waiting for the rgb led) doesn't work any better than the stock halogen bulb. It's true that I don't have a good way to get all the light to the optic fibre, I just place the led close to the optic fibre (the reflector I got for the led doesn't really focus the light to a small enough spot like the original lamp).

I asked about this before... Would you be willing to sell one or two of your rigs?

 

[onnect17]

So the white led (still waiting for the rgb led) doesn't work any better than the stock halogen bulb. It's true that I don't have a good way to get all the light to the optic fibre, I just place the led close to the optic fibre (the reflector I got for the led doesn't really focus the light to a small enough spot like the original lamp).

I asked about this before... Would you be willing to sell one or two of your rigs?

Where are you located?

 

[brbo]

Where are you located?

Slovenia (EU).

 

[onnect17]

Slovenia (EU).

Send me your name and postal address to *******************

 

[brbo]

Hmm... Started getting vertical banding (only in red channel). Swapped lamps, made sure calibration area is clean.

Any ideas?