Epson 7900 from the inside - out

[Eric Gulbransen]

I apologize.  I have hit my head far too many times, nut just my thumbs.  The more I read about Piezoelectricity the less I understand.

[Mark D Segal]

I've also been trying to figure out the Peizo effect as related to the nozzles. I understand the concept but not much on how the nozzles could be damaged.I don't think moving the machine is what caused damage such as this.If it were the case Epson would not be able to ship them at all. They would stand the risk of having to replace a lot of heads under warrantee! Just doesn't add up!  I also agree with your non freeze assessment. Dried ink so deeply embedded is much more likely, and as you say possible? to burn out the nozzles, we may never know.
  The field guide ( http://www.americaninkjetsystems.com/epson_mimaki_roland_printer_rep/pro79_fg.pdf )starting on page 231 has a list of diagnosis for the causes and remedies for clogged nozzles but  it is  somewhat ambiguous to say the least.
 I believe you followed the correct path!

David

On page 233 it says this: "Note: If the same nozzle or nozzles are always missing, this indicates that the Print Head has a problem mechanically or electrically."

Based on what Eric has related, it sounds as if the problem could be a mechanical or electrical problem in the printhead. If so, we still don't know what caused it from the time it was shipped till the time he used it. Eric mentioned the possibility of expired inks. Eric: do you have the precise expiry dates of the cartridges for the offending colours? I have always been given to understand from Epson staff that those dates are pretty conservative (to be safe) and that in practice the inks should be good for much longer than the stated expiry. When you loaded the inks, did you gently shake the cartridges first? They always recommend this.

The next item of interest in that manual, and Epson staff have mentioned to me before, is the entrapment of air in the printhead from the cleaning cycles themselves, which would show up as nozzle pattern breaks - but usually IN DIFFERENT PLACES from one cleaning to the next. To mitigate that risk, Epson staff have recommended running prints between cleanings to try to force the ink through in the printing direction. But Eric's symptom is that the problem is lodged in the same place, so perhaps air entrapment is not the issue.

On reflection it's really hard to fathom how ink could dry-up so severely within such a short period of time to permanently prevent the in-built cleaning routines from freeing the ink and allowing the nozzles to operate normally, let alone permanently damage them. This is the part of the whole story that actually concerns me the most: why did the basic, ordinary consumer-accessible cleaning routines fail to work? Clogging that users can easily clean-up is an occupational hazard one can live with to an extent, but ink flow problems that users can't clean-up at all seem to be deadly.

[chaddro]

Hey Eric!

Here's a crazy idea that just popped into my head...

If these nozzle are still clogged, I wonder is an ultrasonic cleaner would work to break what ever particles are possible preventing the nozzle from firing/working? I googled it and found this:

http://www.youtube.com/watch?v=poEDY4f5_4Y

...maybe hope is not lost? In one of the several vids you can see them checking the nozzle to see if they are clear... you see a stream of fluid coming out each nozzle.

[Eric Gulbransen]

I expect politics may throw up some walls here but I am already actively pushing the limits of standard protocol.  My goal is not to find a source to buy print heads from.  Rather my goal is to develop an understanding of how these heads work, what are their weak points, and how best to care for them - and therefore ultimately our Epson x900s.  My parts contact in SF mentioned to me on Friday that he understands Epson does not make the heads of the x900 printers.  He suggested the heads are made in Germany.  I do not know if this is true.  However there is a US company called Dimatix which specializes particularly in the production of Piezoelectric inkjet printheads and related systems.  I do not know if Dimatix makes Epson's heads.  The more I research this though, the less I believe Epson makes their own heads.  I therefore suspect the most direct route to acquiring more knowledge on Piezoelectric printhead technology and maintenance is likely not through Epson.

My understanding of electricity is pretty basic.  Piezoelectricity even less.  But I do know that sending electric current (electrons) through a substance that will permit this flow (a conductor - like a wire) is always met with some level of resistance.  That resistance generates heat.  Think of a fuse - one with higher resistance stands up to more heat than one with less resistance.  A 2amp fuse has a thin wire with low resistance - it burns out easily.  A 20amp fuse has a thicker wire with a higher resistance - it has a higher resistance to burning out (melting). 

Considering this I am suspicious of the following:

1 - just how thin are the "wires" sending electric current to these microscopic piezoelectric crystals?
2 - what is their ability to resist melting?
*3 - how much of a cooling role does the flow of ink past, or through these piezoelectric crystals, play?

*if the healthy flow of ink past, or through these electrically charged crystals does play a key role in their cooling - then my friends we might now have yet another possible scenario to consider in the explanation of one clog growing into groups of clogs.  "Overheating"


Again, this unfortunately is all just speculation.  But it stands to reason that a nozzle won't get hot until you send a charge to it, so under those circumstances it doesn't need cooling.  But if you print a page that's 90% black for instance, your PK nozzles will be firing full time in back to back sequence of about what - four million continuous firings?  I don't care how big you are, because I imagine it's all relative, if you charge a microscopic line four million times in a row and don't cool it - it's probably going to overheat.  Exactly what happens inside these heads during a power cleaning I wish I knew.  Do all nozzles fire continuously flat out?  Does SS really stand for super sonic CHARGING?  If a nozzle is clogged are they sending it MORE power to help release it?  Does more power mean more heat? 

These are all questions I think we might all be better off knowing the answers to.  I may just be a carpenter but I happen to be one about ten miles from Dimatix.  If I have to I will dress in drag and apply for a job answering phones there...

[Mark D Segal]

The people who design the printheads and the people who manufacture them could well be different, but I recommend you have a look at the following:

http://en.wikipedia.org/wiki/Seiko_Epson, and this: http://www.letsgodigital.org/en/10511/epson_micro_piezo/; and this: http://www.selectsign.hk/ArticleShow.asp?ArticleID=280; http://fotoflock.com/index.php/learn-photography/tips-tricks/29-tips-and-tricks/2958-the-piezoelectric-phenomenon.

I think it's pretty clear whose technology this is.

[Eric Gulbransen]

The people who design the printheads and the people who manufacture them could well be different, but I recommend you have a look at the following:

http://en.wikipedia.org/wiki/Seiko_Epson, and this: http://www.letsgodigital.org/en/10511/epson_micro_piezo/; and this: http://www.selectsign.hk/ArticleShow.asp?ArticleID=280; http://fotoflock.com/index.php/learn-photography/tips-tricks/29-tips-and-tricks/2958-the-piezoelectric-phenomenon.

I think it's pretty clear whose technology this is.

Well there goes that theory.   Thank you Mark.   I can never get the lipstick right anyway

[chaddro]

chaddro,
 Great find, but this system is designed for solvent based inks.

It sure would be nice though if it worked  for a clog,but would not help  totally dead nozzle due to an electrical problem.!

David

Hey David,

Looks like the main site has some interesting info, including using with water based inks.

http://hp9000inks.com/store/catalog/Technical-Support-sp-6.html

One thing to note is the red warnings in the pdf.

[Farmer]

There's absolutely no doubt that Epson design and make their own heads, and they're not manufactured in Germany.

[Mark D Segal]

There's absolutely no doubt that Epson design and make their own heads, and they're not manufactured in Germany.

Phil, this means then that the guy who told Eric the head is made in Germany was wrong:

(ref early on in this thread: "Got a huge deal thanks to Justin, our ebay parts contact in SF.  I couldn't make Justin's 2pm deadline for pickup so I met his father instead.  Pretty cool guy.  We talked for a while.  I gave him my input on what I thought he should stock up on and why.  I even told him about this thread - and David buying one of his damper assemblies just to have on stand-by, most likely due to what we have shared here.  He seemed happy.  Just before we signed off I said, "Don't forget.  Keep your eye out for x900 printer heads.."  He responded with a classic smile - "You are my friend.  ...Come with me."  I followed him through a maze of printer boxes so impressively stacked even a Bloodhound would struggle to get through it.  About mid-way through the maze we arrived at his desk.  He reached from it's cluttered top holding toward my face a small square box in a manner suggesting it's worth was unparallelled.  "It's a 7900 print head.  Did you know they are made in Germany?" )

Until I read this I was under the impression that Epson made these heads in Japan, but then again these days companies make so many things under their supervision in so many parts of the world one never knows, so I thought this could be possible. Thanks for clarifying.

[na goodman]

Eric, you're too funny! At least you can still have a sense of humor about the situation. I think just your creativity and personality would be enough to get you the job, forget about the drag thing. Could you please clarify if you did install a new print head and if you had the service program to run. And thanks guys for being gender sensitive. 

[davidh202]

Well there goes that theory.   Thank you Mark.  I can never get the lipstick right anyway

Eric,
  The manufacturer may be wrong but your  overheat theory could still be valid.
Under normal operating conditions the head may run cooler than other systems, but given a clog it may indeed overheat and cause detrimental issues with the nozzle circuitry. The warnings in the operating manual for the cleaning machine chaddro pointed out, say that it is possible to damage heads by overheating, while cleaning with that machine.
Of course this does not specifically say the Epson heads, but overheating due to clogs, and consequential failure of nozzles, cannot definately be ruled out!
 David

[Wayne Fox]

I paid $2350 new after the rebate including shipping  1 1/2 years ago for the 7900.
Those machines are $2300 and up + accessories ( even if they worked for an Epson head)

The problem that arises is there is just no way todetermine if the nozzles are simply mechanically clogged with dried ink, or dead due to an electrical -mechanical problem, the later of which, would make the cleaning machine useless!
David

actually I think there might be.  after following this thread and the efforts to unclog a nozzle including flushing, removing and soaking etc. the symptom of unrepairable seems pretty clear ... nozzles that will not clear or change at all despite any attempts to clear it most likely means the piezo effect of those nozzles no longer functions.  Despite all efforts (which have been extensive and interesting to follow) not a single missing nozzle cleared, only new ones.  Unfortunately head seems dead.

I'm not clear on the technology.  However it seems possible that they are sensitive, perhaps more sensitive in the new printers, and moving them might be more risky than before. Maybe even static electricity.  Interesting to note that Epson recommends in the 9900 user manual when moving the printer to leave the ink cartridges in place.

I don't think the outdated cartridges are an issue.  There are thousands of users that use outdated cartridges all the time.  I don't think the ink really "settles" out.  Issue with old cartridges has more to do with potential color variations.

[Farmer]

Before you consign a head to the dead pile, may I suggest that you check the ribbon cables?  Both for a secure connection and for possible damage.  Even if you can't see a break, it's possible.

[Eric Gulbransen]

Before you consign a head to the dead pile, may I suggest that you check the ribbon cables?  Both for a secure connection and for possible damage.  Even if you can't see a break, it's possible.

Oh great, two birds with one stone:  My friend in SF with the allegedly bad information about where Epson's heads are made, not only held that spare printhead up to me - but he actually gave it to me.  For free.  Apparently he had sold this head to someone a year and a half ago.  They couldn't get it to work right so he gave them their money back.  The head has sat in a box in his possession ever since.

After extensive testing with our original cleaned head once we fired our machine up Saturday afternoon, the writing was pretty clear on the wall.  Our head is toast.  That's when I started the clock, and the job of swapping printheads.  Like I said it took us 45min start to finish, mechanically.  The next step is "registering" the head.

These heads come with a series of letters and numbers printed on a sticker, glued to one side.  There are 45 characters in all.  In the adjusting program you have to enter those numbers in order for the head to work properly.  The head will work, and the printer will print without registering these specific numbers in the system - but it won't work properly.  I don't know yet what those numbers represent or do, but I expect it has something to do with how each head is calibrated, or aligned, or lord only knows what.  The point is on a nozzle pattern with this replacement head our PK and YW clogs were clear - printing fine.  But many of the other channels needed alignments or had miss-fires. 

The reason we couldn't get this head's numbers registered in the system properly is because one row of them simply wasn't there.  This left us with three missing characters.  End of story.

So I am right there with you Farmer, and already tested it.  Unfortunately I believe our head may be hosed.  This will not, however, deter me from my goal of communicating with Dimatix.  I could be wrong I do this a lot, but I still think the most potent source for getting answers on the characteristics of these piezoelectric printheads is not Epson. 

I say we post an attractive prize, perhaps a box of chocolate in the spirit of na goodman (because I know women far better than I know printing - chocolate fixes everything) offered to the first one who finds answers to these questions:

1 - just how thin are the "wires" sending electric current to these microscopic piezoelectric crystals?
2 - what is their ability to resist melting?
3 - how much of a cooling role does the flow of ink past, or through these piezoelectric crystals, play?
4 - what is the best way to clear a piezoelectric printhead clog?
4 - what is the best way to avoid clogs.


If I win I want ice cream.  I'll even buy it for myself - I just need an excuse to celebrate something

[Ernst Dinkla]

However there is a US company called Dimatix which specializes particularly in the production of Piezoelectric inkjet printheads and related systems.  I do not know if Dimatix makes Epson's heads.  

My understanding of electricity is pretty basic.  etc

There are many companies that manufacture piëzo heads of different designs: Spectra, Xaar, Dimatix, Seiko, Hitachi, Fuji, HP, Konica. They are used in all kinds of inkjet printers from mail address printers to 5 meter wide solvent printers. Epson manufactured heads themselves and licensed the manufacturing to other companies for some designs. For example the 9000 heads are still available and at low prices, I am told that is the result of a continuing license to Chinese manufacturers. In contrast some heads are no longer available or very expensive where Epson could influence the production. See the German pricelist I gave a link for. A tale here going round  in Eindhoven is that Philips made the very first Epson heads possible with their galvanic processes that were developed to fabricate electric shaver heads. Modern manufacturing processes follow another path and piëzo heads are also made with MEMS technology like the large thermal heads are.

I have written in the past that piëzo heads used at high frequencies will heat up despite fresh ink cooling the system, it simply depends on the droplet size squirted , the frequency and the total nozzle area what the heating up and cooling effect will be. Thermal heads will heat up too for obvious reasons but are cooled with the ink as well and then the frequency per nozzle + the nozzle area has to be considered. The print speed of the Canon iPF models, HP Z6200 will be comparable to the x900, 11880 models, their thermal head printers have 2000 nozzles per ink channel, the Epson 360 nozzles. That is roughly 5x the number, if we keep in mind that Canon and HP keep nozzles in reserve for others to fail. Minimum droplet size is about the same; 3.5 versus 4 picoliter, the Canons have only that 4 picoliter size the HPs have 4 and 6 picoliter depending on the ink channel. In the highest resolutions the Epson will not use the larger droplet sizes it can also produce from the same nozzles, it has to squirt at a 5x rate similar droplets said thermal heads have to squirt to cover the same media area in the same time.

Then there is air in the ink lines, the much simpler thermal nozzle channel and its "steam" pump seems to deal better with air in lines than the piëzo heads do, HP docs often refer to that advantage. Piëzo pumps become ineffective when some air just acts as a shock absorber in the nozzle channel when the piëzo element deforms for pumping. What happens to the ink at that point when no ink is really flowing and heat builds up is something you can imagine. What becomes a permanent clogg and worse may have actually started with air in the lines for whatever reason. It will be hard to determine what a nozzle permanently not functioning contains, simply a build up of dried ink, pigment agglomerations or all that plus electronic damage as well. You probably need an electron microscope and slices of the head to get the information. Epson will know for sure, we do not.

Edit: considering the transport of this 7900 and the multiple 4900s in the other thread I wonder whether air getting into the head from the nozzle side when the printer is tilted etc may be the cause in those cases and possibly even with some new ones delivered. Would a process help where the waste tank pump is first activated + the ink carts pressurised to bring ink back into the head before the heads themselves are activated?  On the 5000, 9000s and 10000 I worked on I had for maintenance etc the waste tank tube running outside the waste tank to a vacuum pump with a bottle in between. That way I could reload other inks fast, clean channels and deal better with cloggs without using the heads. All had third party inks then and in the end all went belly up due to pigment settling, fungi or me going too far with DIY approaches.


met vriendelijke groeten, Ernst
Try: http://groups.yahoo.com/group/Wide_Inkjet_Printers/

[Eric Gulbransen]

Fascinating..

[Ernst Dinkla]

The reason we couldn't get this head's numbers registered in the system properly is because one row of them simply wasn't there.  This left us with three missing characters.  End of story.

Would it be possible to ask the SF people to look at other heads from the same production batch and copy the numbers to the spare head you got? There is a possibility that the total sum (or alike) of the numbers has to be correct if it is a code but it could also be that they represent a certain character of the head channels. On the 9000s the number represented the voltage that had to applied to let the piëzos act to a standard pump activity. According a Chinese from HK they changed the voltage themselves that way when the thicker sublimation inks were pumped instead of the dye and later pigment inks used with Epsons. I doubt that today's head numbers are that simply related.

Edit: Google - K T Chan Epson -


met vriendelijke groeten, Ernst
Shareware now:
Dinkla Gallery Canvas Wrap Actions for Photoshop
http://www.pigment-print.com/dinklacanvaswraps/six-canvas-wrap-actions.htm

[multigary]

This thread is starting to read like an international high tech suspense story!  A working guy carpenter in California tries to save a buck by buying a used printer from another carpenter across the country, hoping to promote his passion for printing his photographs, only to be drawn into an all consuming "journey inside the 7900"!  Now even the Chinese in Honk Kong are becoming characters!  Eric, forget about the technical tutorials - write a novel instead! You can call it something like  "The Piezo Code" or  "The  Printer That  Ate Brooklyn". My favorite scenes so far are when you made your way through the maze of boxes and discovered the German made print head; and the color map of the printer's cross country journey, with the  daily average temperatures included.
I do hope this story has a happy ending!

[Alan Goldhammer]

If the insidious Dr. Fu Manchu makes an appearance I will really know all is lost!  On topic, I started doing a patent search yesterday (the US system is on the Internet and pretty easy to search) to see what is disclosed.  Interesting that the ink cartridges for the 3880 and 4900 (thanks to Mark Segal for the 4900 numbers) share the same patent numbers.  These patents only relate to the mechanics of the cartridge (hard outer shell and a polymer bag inside that contains the ink) and not the ink itself.  I have not yet found any patents on the Epson inks and these may be protected by trade secret (I'm still searching so can't be certain this is the case).  I did find one patent on the piezoelectric use to lay down ink but suspect there are more.  This may not be the whole story since it is the implementation of the technology in the move from the x800 to x900 machines that may be the major difference.  I will certainly report further on anything I dig up.

My read of what has gone on to date with the various posts leads me to think this might be head failure but the cause still remains to be determined.

[Mark D Segal]

My read of what has gone on to date with the various posts leads me to think this might be head failure but the cause still remains to be determined.

This is exactly where my mind is at, and I really think it would be in everyone's interest to have an authoritative determination of what happened. So many theories have been floated here, some more plausible than others, that it runs the risk of creating numerous and for the most part probably needless doubts in peoples' minds about the robustness of the technology. Once the cause or the most likely group of causes are known, people will know what to be careful about and this could save both Epson and its customers a lot heartache down the road, as well as preserve confidence in what is, after all, fantastic printing technology when you think of it.