New EYEMAGIC Scanners EMS7000

[EyeMagic]

Hello Everyone,

Just to help things up, we would like to note a few points :

At speeds of 60Kpps and angles of 10deg, every minute detail counts.

In our test setup we use a +/-28V linear power supply using a toroidal transformer and a Pangolin QM2000 card inside a desktop computer (PCI).
The above have the following advantages as accuracy is concerned :

The linear PS is lower noise and supplies peak currents much faster and better. It is no accident that the best audio amplifiers still use linear PSUs.
The QM2000 card is a high quality output device and by sitting inside the computer it has the advantage of being referenced to a real ground. On top of that we use an extra wire from the computer chassis to the center tap of our linear power supply.

We use the above setup just to make sure that the scanners are tested in a ''neutral'' environment with no external issues.

What we see in Solarfire's setup is that the X axis has a tendency to ''pull'' on the left. This could be from the switching power supply not being able to deliver equal current on each side, or from the output card not having a steady ground etc..

What is important to know is that when the scanners are displaying that circle, they are essentially ''flying'' way out of their commanded position and the slightest disturbance in the system is immediately visible in this ''flight''.
Remember that we are speaking for 60K and 10deg. This is more than twice the speed of a 30K at 8 deg...

Also, if the PS used by Solarfire is lower voltage (for example +/-24V) then this makes a small difference in the ILDA circle. We estimate that with a +/-24V, the ILDA should be displayed correctly at around 8-9 deg @ 60K.

What is important though is that the ACTUAL performance of the scanners is EXACTLY the same when using a 24 or a 28V power supply. The extra voltage just helps the servo overcome the back emf produced by the motor when it is making that sine wave (the circle). The step times and the settling accuracy are the same. The scanners, when displaying real graphics, are never commanded to display a circle like the one in the ILDA test.
So, unless your customers logo is the ILDA test frame itself (!), you are OK with a +/-24V PS. It also produces less heat on the amps.

All the best,

Tom Kamaras
EyeMagic

[JoJo]

Sorry, but the scan pictures I see here look completely awful.
Is the main target of all your tests bringing the circle to the square or making a good looking scan?
I know many "fast" chinese scanners fitting the circle to the square at speeds of 45K or more. But what is the advantage to have a high speed when the lines look like hooks and waves?
This all remembers me many "fast" scanners of some companies, where speed and not picture quality was the only sales argument.
I don't mean specially the EMS-galvos, which I don't know, but I mean the global practice selling and classifying galvos by possible slew rate only.

It is not that easy to tune the scanner to high speed and then "make the lines looking good later".
Such "small" problems like mirror resonances, uncentered mirrors (unbalanced load), noise on power and signals, wobbling bearings all together make a bad scan picture and are not easy to eliminate just by driver fine tuning.

Joachim

[JohnYayas]

Hello Everyone,

Just to help things up, we would like to note a few points :

At speeds of 60Kpps and angles of 10deg, every minute detail counts.

In our test setup we use a +/-28V linear power supply using a toroidal transformer and a Pangolin QM2000 card inside a desktop computer (PCI).
The above have the following advantages as accuracy is concerned :

The linear PS is lower noise and supplies peak currents much faster and better. It is no accident that the best audio amplifiers still use linear PSUs.
The QM2000 card is a high quality output device and by sitting inside the computer it has the advantage of being referenced to a real ground. On top of that we use an extra wire from the computer chassis to the center tap of our linear power supply.

We use the above setup just to make sure that the scanners are tested in a ''neutral'' environment with no external issues.

What we see in Solarfire's setup is that the X axis has a tendency to ''pull'' on the left. This could be from the switching power supply not being able to deliver equal current on each side, or from the output card not having a steady ground etc..

What is important to know is that when the scanners are displaying that circle, they are essentially ''flying'' way out of their commanded position and the slightest disturbance in the system is immediately visible in this ''flight''.
Remember that we are speaking for 60K and 10deg. This is more than twice the speed of a 30K at 8 deg...

Also, if the PS used by Solarfire is lower voltage (for example +/-24V) then this makes a small difference in the ILDA circle. We estimate that with a +/-24V, the ILDA should be displayed correctly at around 8-9 deg @ 60K.

What is important though is that the ACTUAL performance of the scanners is EXACTLY the same when using a 24 or a 28V power supply. The extra voltage just helps the servo overcome the back emf produced by the motor when it is making that sine wave (the circle). The step times and the settling accuracy are the same. The scanners, when displaying real graphics, are never commanded to display a circle like the one in the ILDA test.
So, unless your customers logo is the ILDA test frame itself (!), you are OK with a +/-24V PS. It also produces less heat on the amps.

All the best,

Tom Kamaras
EyeMagic

How hard is it to create a video of a test frame along with distances and a shot of the computer screen showing the scan angle? You should be able to prove your case in no more than 5 minutes with a smartphone and a YouTube account. I really want to believe you and for you to prove that you actually have a 60K 10degree scanner because frankly, I might buy one, but at this point I am left with doubt. I know that PhotoLexicon is a just a tiny part of your market place so this thread really doesn't matter all that much but you have spent far more time going back and forth with Bill and other forum members than it would take to show some proof with a video. Just make sure you do it in a manner than takes out any doubt that you might be faking it..

[planters]

JohnYayas is exactly right. Ignore the conflict and continue with the approach in your last post. That information was USEFUL. Now add a video as he described and you've accomplished a lot to restore some confidence in these scanners. I'll be watching.

[EyeMagic]

Mr. "JoJo'' please find the factory tuning in post 289. Also, the image size in this photo is at 10-11 deg which is a little big for the +/-24V used.

As for the conflict that someone else referred to, it is not our style. It was another person's idea and realization. We just protect our work.

As for making a video, it is not a bad idea at all. Our problem is that if some customers do NOT want to believe, then they will just say that this is a fake video.
And the posting will go on, and on, and on...
We informed the forum that we asked some of our big companies (that use both our and CT scanners) to do a side by side EMS7000 - CT6215 test, with Y cables. This cannot be faked as one scanset is reference for the other and vice-versa. The response we received was not positive. Why ?
So yes, if someone does not want to believe, then there is little we can do...

Al the best,

Tom Kamaras
EyeMagic

[-bart-]

As for making a video, it is not a bad idea at all. Our problem is that if some customers do NOT want to believe, then they will just say that this is a fake video.
And the posting will go on, and on, and on...

You're right, if you argue with a fool, chances are, he's doing just the same.

You'll never convince a person that doesn't want to be convinced.

However, the world isn't divided in fanboys and haters. Most (of us) are just healthy sceptics, who will surely be convince by some plausible proof.

[planters]

That is a shame that the response to your request was not positive. I might guess that these companies do not want to take on the responsibility for testing and the implied liability if the results do not PLEASE someone. They have nothing to gain. But, you do.

Why not preform this same test yourselves? I do not know your R and D budget. If it is very limited you might find a volunteer on this SMALL AND OH SO INSIGNIFICANT FORUM who would lend a CT scanner for the purpose. At the least you could do a side by side comparison with one of your own 4000 series that many contributors already own and are familiar with. I really think you need to do this.

[andythemechanic]

Mr. "JoJo'' please find the factory tuning in post 289. Also, the image size in this photo is at 10-11 deg which is a little big for the +/-24V used.

I think Tom is right here. Although the Circle is a bit open on the image in post 289 (which could still at least partly be a wrong blank shift) in this original factory tuning the image looks quite good, straight lines straight etc.
The retuned images look really ugly. Solarfire, contrary to your belief this shows exactly that the rest of the ILDA frame does matter as well. In many scanner, amp combinations it is easy to tune the scanners much higher than the speed were a nice ILDA frame is achieved when only judging circle in square, but than you see exactly what is seen in your retuned images. There will be oscillations and all kinds of distortions. There was for sure a reason the scanner was factory tuned like this. It is also dangerous to tune the scanners higher and higher as they could overheat and get damaged when there is much distortion.
I have several scanners, CT6800, CT6210, Raytracks, Tera Scans all of them could do more speed at the expense of image quality and lifetime.

Andreas

[Solarfire]

I think Tom is right here. Although the Circle is a bit open on the image in post 289 (which could still at least partly be a wrong blank shift) in this original factory tuning the image looks quite good, straight lines straight etc.
The retuned images look really ugly. Solarfire, contrary to your belief this shows exactly that the rest of the ILDA frame does matter as well. In many scanner, amp combinations it is easy to tune the scanners much higher than the speed were a nice ILDA frame is achieved when only judging circle in square, but than you see exactly what is seen in your retuned images. There will be oscillations and all kinds of distortions. There was for sure a reason the scanner was factory tuned like this. It is also dangerous to tune the scanners higher and higher as they could overheat and get damaged when there is much distortion.
I have several scanners, CT6800, CT6210, Raytracks, Tera Scans all of them could do more speed at the expense of image quality and lifetime.

Andreas

Andreas, here is my current result, after looking at the picture I see there’s still a little room for improvement. This is the best overall compromise I could get so far. This is 62kpps @ 10.5° showed @ a distance of 4.6m.
Once again to the ILDA frame, for speed checking mainly the circle in the box is of relevance, which is defined by the circle being round, closed, undistorted touching all 4 insides of the box. This occurs @ a certain frequency which in this case was @ 64kpps (if there wouldn’t have been a slight offset), this has nothing to do with how the rest of the image looks. As you see yourself strictly tuning to the ILDA frame specs doesn’t work in this case. I’m also aware of what scanners can handle and what not, but if you’re going to test something then you also have to scratch on the boundaries. At no point did these scanners show any signs of stress even @ much larger angles, they didn’t really even break a sweat. The scanner block never went over 29.8°C without TEC cooling activated. The scan amps reached up to 60°C and here I will improve cooling. The output is excellent and razor sharp, I’ll get some pictures posted maybe tomorrow.

[Pangolin]

Bill, Do the laws of physics indicate that anyone wins a flame war? Is this discussion a good idea when you don't have a product on the market yet? You're smart - I knew that years ago - but this is ego - you're not winning customers doing this...

Brian Wirthlin

Hi Brian, great to see you here on PL.

First let me say that it is not my *intention* to flame -- only to determine the FACTS.

To me the main question is -- are there exaggerations on Tom's datasheet yes or no. This should not cause flames. It should have a simple binary outcome. Either everything on the datasheet is factual, or it isn't. If everything isn't then we come to a question -- how many non-factual things are there on the datasheet. If there's only one non-factual thing, perhaps we could chalk it up to a mistake. (Happens to everyone.) But if there is a lot of non-factual things on the datasheet, then the question is -- why is this?

Hmmm. It is entirely possible that the power supply at FLEM couldn't deliver enough current. That would explain the lack of performance.

It's a good point and I thought of this too John, but the power supply capability would only matter with patterns that require a high RMS demand, for example raster images or abstracts with a lot of squarewave content. None of that was used during the scanner tests. Also, the initial tests we saw (tuned to 45K) parallel tests done at Pangolin with a 30 Amp linear supply.

Hmm… It’s suspiciously quiet on this thread? I would have expected the Pango/CT disciples to be crawling out of the woodwork by now.

Hehe. Don't forget I have a real job ;-) As I have said, many times on PL, I don't visit this forum very often. Certainly not every day. So if you're looking for instant responses from me every second, you're nog going to see this.

Funny thing that just now someone is starting to question the setup and results of the FLEM.

Yes, and this is entirely fair.

Commenting further -- are the EMS7000 scanners that I have representative or not? I don't know. And I'm certainly up for Tom sending another pair to Adam or someone else for independent testing. That seems like a good idea to me. What might be an even better idea is for Tom to send me a replacement pair -- if he is highly confident that the pair I have is defective.

HOWEVER, there's still the bigger question of whether or not there are exaggerated claims on Tom's datasheet. To me this is a separate issue, and it's the main thing that stirred me.

Do people here always take what Mr. Benner says as gospel? I guess Pangolin is like Apple, they have followers just for the sake of being a part of the status symbol. Well Mr. Benner I’m afraid you’re going to have to give back your merit badge as the ultimate scanner expert.

Hehe, well it's certainly fair to question anything and everything that I say. I encourage it!!

Nevertheless, to give some credentials, I currently have 14 issued patents (most related to scanning and scanning technologies) and many pending patents. Also, I've not made any factually incorrect statements. Moreover, scanner companies far and wide wouldn't come to me for advice if I didn't have advice worth listening to!

In any event, sure, absolutely, question everything I say. No problem at all. And ask good questions and I'll give you good answers.

************************************************** ************************************************

However, let me ask you a pointed question too, and it was asked indirectly by AndyF. Instead of telling us the number of degrees, tell us the distance to the screen and the width of the projection.

One thing I have learned over and over and over in this business is that "people don't know from angle". And your quick snotty response to AndyF ("don't you know trigonometry") is not an answer. The problem is -- the most simple trig is for right-triangles. In order to correctly calculate the angle, you'd either have to use law of sines, or split the projection into two right-triangles, and then double the angular result.

Just like I encourage everyone to question what I say, don't be surprised if I question what anybody else says. Even if someone in my office -- or Greg Makhov -- someone who I respect, and have worked with for the past 20 years told me "it's 10 degrees", I'd want to confirm it for myself. So please supply the distance and width instead of the angle.

The reason I even got into this discussion was due to the fact that this thread was turned into a Pangolin propaganda thread

I certainly don't see it that way. My comments have been trying to drill down on the factual inaccuracies portrayed in the datasheet and to provide some clarity.

In the next post, I'll provide pictures and more specifics about what specifically bothers me. Then perhaps you'll understand where I'm coming from. Again-- I'm only after the FACTS.