Howtek drum scanner DIY thread: maintenance, troubleshooting, mods

[onnect17]

Is there any way to adjust the analyser lens? Can't see it mentioned in the manual.

If I get good centering on the aperture wheel but off-centered image in viewport this could also be because of flip-up mirror? This is non-issue, but if the cause is in the analyser lens misalignment...

Could you explain how to check if the analyser lens is positioned correctly? Unfortunately, I don't understand how one would use the cross-point laser for that. Thanks!

The analyzer lens is quite similar to a microscope objective, even share the same old thread. You can remove it easily with the help of a rubber band. Imagine if you make a similar piece and instead of a couple of achromats elements you insert a small laser pointing inwards. That will show the sweet spot of the analyzer lens.

The analyzer lens is mounted on a linear stage that can be adjusted.

 

[brbo]

The laser thing will have to wait as I don't have it and don't really understand what I'd have to do...

Anyway, I swapped reflective and transmissive lamps and cleaned analyser lens. No noticeable improvement. Either there is a problem somewhere else or this is "par for the course". Maybe Howtek 4500 just can't be compared to a good CCD scanner (I also have Minolta 5400) when scanning negatives. I'm a bit hesitant to touch the mirror and filters inside the optical box to clean the dust off. Is this something that can safely be done with cotton swabs and lens cleaning solution?

For "grainy/discolored image/colored streaks" further suggestions from manual include replacing PMTs, VAA, VDS, pre amplifier board... So I'd probably be out of luck if that is the source of my problems. My scans of slides look good to me, though. If I push shadows in PP there's some noise, but that's probably normal. Can field service test program give more answers? Can the FST4500.exe be run directly from within WinXP?

I managed to put together an old PC and get Trident 3 going in demo mode. Scans of negatives have considerably less color noise (still look a fair bit "harsher" than CCD scans, but this was all dry mounted on a crazed drum). I guess Trident is using internal 12bits to only get the actual range of the negative. I guess DPL Pro would be roughly the same then so I have a kind of DPL demo now 😉

 

[onnect17]

Maybe Howtek 4500 just can't be compared to a good CCD scanner (I also have Minolta 5400) when scanning negatives. I'm a bit hesitant to touch the mirror and filters inside the optical box to clean the dust off. Is this something that can safely be done with cotton swabs and lens cleaning solution?

It's really hard to compare scanners especially if are using different technologies. A 4500 in good shape produces an excellent scan, with clean color channels than no CCD based can get even close. I owned a few of them, including a Nikon V ED and an Imacon 343. Sold them. Never got used to the idea of having lower resolution in the borders, or uneven focus, flare, etc.

It's fine to clean the optics but be careful with the aperture wheel. You can inspect the holes using a microscope.

For "grainy/discolored image/colored streaks" further suggestions from manual include replacing PMTs, VAA, VDS, pre amplifier board... So I'd probably be out of luck if that is the source of my problems. My scans of slides look good to me, though. If I push shadows in PP there's some noise, but that's probably normal. Can field service test program give more answers? Can the FST4500.exe be run directly from within WinXP?

IMHO the "achilles heel" of the Howteks is the underpower light source. The alignment procedure would get the best of the original hardware but you can also improve the output installing a more efficient lamp, like RGB LED or at least placing a filter UV-IR at the entrance of the fiber. Also replacing the PMT with other models with better performance, like the R3896, R928, etc. With this changes I use the 6 microns aperture all the time, without issues.

I run the FST4500 from a floppy with DOS or directly from a DOS HD partition, using a PC or Notebook. Not sure if I can run it from XP. The section to test the PMTs could give you an idea of their condition.

I managed to put together an old PC and get Trident 3 going in demo mode. Scans of negatives have considerably less color noise (still look a fair bit "harsher" than CCD scans, but this was all dry mounted on a crazed drum). I guess Trident is using internal 12bits to only get the actual range of the negative. I guess DPL Pro would be roughly the same then so I have a kind of DPL demo now 😉

I think Trident uses the Log mode by default, so the scans of C41 should look a little better.

 

[brbo]

Yes, flare in CCD scanners ruins it for me. Film flatness, too. I had a Microtek 120tf scanner that was AMAZING (for a CCD scanner) in regard to those two issues, though. The flare was more than what I get with my Howtek, but very managable. Any Nikon Coolscan would look like a joke in a scenario when you scan with sprocket holes included...






My Minolta 5400 with multisampling can match DMax of my Howtek (but not DRange). My Epson 4990 can match the size of largest film that can be scanned.... I'm sure a drum scanner in good condition can outdo them all in every characteristic, so I'm trying to get to the bottom of my problem with scanning negatives.


I run FST4500.exe, everything seem to be "OK". But this tests would probably need a person that is able to interpret the numbers to get past the "OK/Fail" level of troubleshooting in order to get to the part of the scanner that needs to be checked/improved. So here are screens of tests that I thought might be of relevance with the issue:

The test was done before I changed the lamp for a new one (I don't see any change in scans) and before I installed an uv/ir cut filter that I had at hand (I can see a small change):

I'll run the test again and see how uv/ir filter affects the numbers.

BTW, I'm curious about the reason behind the lamp shining into the optic fibre at an angle.


Oh, one more thing. What is the purpose of the battery in this scanner? Mine is dead as the scanner boots with "Checksum error" and then loads defaults. Could/would any of the pre-scanning routines be skipped if the battery was working?

 

[onnect17]

The test I use the most is the PMTs. It should be close to the max, 4095. The rest is just a guess. Some basic info if you press F1 for help. Here are some pics of my tests here for comparison:

https://1drv.ms/f/s!AugK5NQO79gPgxBihXVwLijEBICW

You should replace the motherboard battery in any case because some users reported stability issues with it.

If I understand correctly, the boot code is in EPROM on the motherboard. In some point loads an updated version from flash memory located in one of the cards. I made copy of all the firmware, just in case.

I am curious about the CHECKSUM error. It would be helpful if you take a video showing it.

 

[brbo]

Thanks for screenshots. My numbers are not that far off. Green channel is lagging in some tests, but doubtful that it should make much difference.

I'll make a video of boot process when I get home. From memor it goes like "Starting scanner" -> "Checksum error. Loading defaults" -> "Soft reboot" -> "Ready to scan. Firmware R813".

 

[brbo]

Here is the link to the video of the scanner boot sequence.

I also ran the FST with UV/IR cut filter installed. The PMTs test shows lower values in green channel. In repeated tests values range from 3900 to 4090.

Found upt that the closer I move the filter to the fiber the higher the value. I've now removed the filter.

 

[onnect17]

Here is the link to the video of the scanner boot sequence.

I also ran the FST with UV/IR cut filter installed. The PMTs test shows lower values in green channel. In repeated tests values range from 3900 to 4090.

Found upt that the closer I move the filter to the fiber the higher the value. I've now removed the filter.

The Flash version seems to be loading fine. Perhaps the battery is dragging down the 5v in the MB during the initial seconds, which is not a good thing anyways. Where are you located? I remember buying a pack of 5 batteries a few years back for not much money. Also make sure the scanner is connected to the same outlet used by the PC.

The UV/IR filter should be hold against the entrance of the fiber. Check for ebay item number 281437567116.

 

[brbo]

I'm from Slovenia (EU).

The scanner and the PC are on the same outlet and have the same surge protector between. Although the manual recomends an UPS. Wonder if that an actual requirement or just over-the-top suggestion.

If the scanner is shut down and powered back on within a minute or so there is no checksum error.

 

[onnect17]

I'm from Slovenia (EU).

The scanner and the PC are on the same outlet and have the same surge protector between. Although the manual recomends an UPS. Wonder if that an actual requirement or just over-the-top suggestion.

If the scanner is shut down and powered back on within a minute or so there is no checksum error.

I am sure you are familiar with how clean the power is in your area. If you are located in a residential area then should be fine. However, if you have a few business around using heavy machinery then protection is recommended. UPS is not a must.

If you do not get an error within a minute then it seems like a MB battery issue. Easy to replace. eBay item 351792398716 should do the job.

 

[brbo]

Thanks! Ordered the battery. I hope there won't be soldering involved when replacing...

Anyway, I've done/checked what I was able by myself at this moment. I'm probably one aperture away from what I would like to have at scanning negative film. I'd be totally happy even with 13m for fine grained neg film, but with the most negs I've tried so far I needed 19m to keep noise in check.

So... I broke down and ordered DPL7. I hope it will help with negative film. Slide film is slowly (but much faster than I would want) going away and Kodak Vision3 stuff is great and not so expensive (even dirt cheap if you buy short-ends).

 

[onnect17]

Thanks! Ordered the battery. I hope there won't be soldering involved when replacing...

Unfortunatelly yes. And be prepare to spend a lot of time removing all the cards and connectors to reach the motherboard. Take pictures before you start!

Anyway, I've done/checked what I was able by myself at this moment. I'm probably one aperture away from what I would like to have at scanning negative film. I'd be totally happy even with 13m for fine grained neg film, but with the most negs I've tried so far I needed 19m to keep noise in check.

"one aperture away" and "one aperture at a time" 🙂

So... I broke down and ordered DPL7. I hope it will help with negative film. Slide film is slowly (but much faster than I would want) going away and Kodak Vision3 stuff is great and not so expensive (even dirt cheap if you buy short-ends).

It should help but don't expect miracles. IMHO, it takes some of the hardware changes previously mentioned to really get down to 6 microns.
Good luck!

 

[onnect17]

A couple of links to videos showing a basic test of the balance in the 7500. If operating properly, the liquid inside a small container on top of the scanner should not show any sign of vibrations, at both, slow and fast speed.

https://1drv.ms/v/s!AugK5NQO79gPgxdyS6e_KmOkGe7j

https://1drv.ms/v/s!AugK5NQO79gPgxgd-RqNCXZdBpuP

 

[Fixcinater]

Over the past months, I've been working on getting a D4000 setup running. It came with two drums, I've been using the worse condition one (few larger scratches on it, haven't tried polishing it at all) to get things started. The other nicer condition drum is safely packed away until I can put it to better use.

Current setup:

Mirror Drive Doors Mac G4 tower with OS 9.2.2 and 10.4.11
Adaptec 2906 SCSI PCI card
SCSI cable that came with the scanner
D4000 scanner on R535 firmware

Luckily, the MDD tower is one that allows booting directly into OS9, the last version with built in firewire 800 ports does not allow it.

So far, I've swapped bulbs from reflective to trans side which helped with some dimming during scanning leading to lines. The lead screw looks dry, should I lube it or not? I've seen both ways of thinking.

Issue 1: I have just updated the firmware on the Howtek to R535 (was 531), and now I cannot move the carriage through the Howtek's panel. It seems that this should not have been lost in the firmware update, but can anyone confirm that R535 does not allow manual carriage movement and AF and aperture choice? I was able to update the FW through the small app provided on Aztek's website along with the FW file itself. I only have tried this once, would it be worth trying again?

Issue 2: Silverfast 6.6.2r5 (demo version until I decide if I should invest in DPL or Silverfast) will talk to the scanner and make scans in OS 10.4.11. None of the demo software downloaded from Aztek's website work in OSX through Classic, which is disappointing but somewhat as expected. When booting directly into OS 9, Aurora will see the scanner but will not scan anything. Trident v3.5.5 and v4.0 will not see it at all, and Polaris will scan but none of the produced files are viewable, they turn into .exe files when moved to OS X or another Mac computer. Should I just quit wasting time trying to get those demo programs to work and buy Silverfast? The scanner came with an early version of Aurora, but no dongle so I can't access full version anyway.

Issue 3: The scanner came with two Howtek labeled drum mounting stations, but both of the waxy rubber pressure spools have literally melted off the spindles. They were kept in a garage in southern California and the door was left open, leaving a pool of waxy rubber on the floor. I've seen some pipe insulation wraps used on DIY mounting stations, are there any other suggestions?

Issue 4: Seems to be a large amount of vibration compared to the D7500 video posted just above. Should the D4000 be held to the same standard of smoothness? Maybe related: Scanner had a cracked old belt, I replaced with one that came with it (soft and supple, never installed before) as well as a new o-ring on the mounting spindle. How do I know what proper belt tension is? I can see the allen heads for adjusting the motor fore/aft on the carriage to set tension. Seems a bit loose now but I don't want to burn any bearings up.

Lastly, I just want to say thank you to all those that have posted their problems and solutions and those that have helped diagnose things, I never would have been able to get it going as far as I have without the existing discussions being able to be referenced.

 

[onnect17]

The lead screw in the 4000 has a solid coating of teflon (PTFE). It's save to use a lubricant with a perfluoropolyether (PFPE) base and PTFE as a thickener. Check some thread in the large format forum about it. I use it in all the howteks because all of them use a plastic anti-backlash nut system.

Issue 1 - I have to go to the basement and check, but if I remember correctly, just follow the procedure to focus manually. That should allow you to move the carriage left and right. Be aware, there is a time out that returns the carriage to home after a while.

Issue 2 - Regarding Trident, I think you also need the scanner driver (version 2.2)

Issue 3 - Check the Hi-End scanner yahoo group for solutions related to the rollers in the mounting stations.

Issue 4 - The driving belt should not be issue here with the vibration. The o-ring in the mounting spindle and the pads at the other end of the drum could be the source. I use 6 or 8 pads instead of 3. Also the bearings could be the source.

Good luck and let us know how it goes.

 

[Fixcinater]

Thanks for the quick reply!

I rebooted scanner by pressing all three buttons and the firmware options became available again. Not sure what was happening there as I couldn't access any of those menus but after the reset I can move the carriage and change aperture, etc with the new firmware. So, I cleaned up underneath the carriage home position and called it a night.

Will update on the others as I make progress. Again, thank you for taking the time to help out!

 

[onnect17]

You are welcome!

 

[Fixcinater]

Minor update after trying a few things to reduce the lines I have been seeing some lines along drum rotation axis. I noted that I only saw it when set to 2000DPI or higher.

-Lubed the lead screw (was totally dry), scanning is quieter - no change.
-Changed the bulb out for yet another one that came with the scanner that was labeled "still usable" - no change.
-Confirmed bulb is shining on end of fiber pack - no adjustment necessary, no change
-Couldn't tell where the drum pads were supposed to be located, checked the other drum that came with the scanner (has been in the factory pouch in safe-keeping), saw where the three pads are and felt rather stupid.

Next step is confirming that having ANY pads helps, which I can only assume it will...and ordering more pads so I can go 6 or 9 pads on each drum.

 

[brbo]

Describe or show us the lines you are getting in the scans. Is the calibration part of the drum perfectly clear/clean?

 

[Fixcinater]

Attached are two files from small scans I just did. Ambient room temp is fairly cold, scanner has been turned on for ~2hrs at this point.

All settings between files are same, I resized with CS3 to same pixel width.

EDIT: This is with a transformer style power supply for the scanner. I can still clearly see variation in the lamp brightness, sometimes seems there is a pattern to the variance and sometimes it just dims quickly and then back to norm.