Electronic level/horizon indicator

[Eric Fletcher S.O.C.]

Steve, why is there a 30 second warm up time?

It's for error correction, the sensor is reading the jitter and then masking that out. Presumably the sensors accuracy should get better the longer its on.

 

What's more troubling to me is the +/- .25 degree error that is baseline. If you are using the one degree setting that error renders that resolution all but useless

[Jens Piotrowski SOC]

just checked the XCS PDL specs:

 

PERFORMANCE SPECIFICATIONS
Range: Visual goal posts +/- 1, 2, 4, 8, 16 degrees.
Range: Limits +/- 90 degrees.
Nominal Accuracy: At 1 degree 0.03 degrees.

 

It would mean it's 8 times more accurate!

 

 

Why is there such a difference in accuracy? Or/And does it matter...

[Eric Fletcher S.O.C.]

just checked the XCS PDL specs:

 

PERFORMANCE SPECIFICATIONS

Range: Visual goal posts +/- 1, 2, 4, 8, 16 degrees.

Range: Limits +/- 90 degrees.

Nominal Accuracy: At 1 degree 0.03 degrees.

 

It would mean it's 8 times more accurate!

 

 

Why is there such a difference in accuracy? Or/And does it matter...

A few reasons why. PRICE and profit margin being two of them. Also I have a pretty good idea what Tiffen is using for their sensor and its just not as accurate as the sensor that XCS is using. Yes you will see the difference in accuracy with the Tiffen unit

[Jerry Holway]

Some clarity here: (This info comes from the genius behind the device). The +/- 0.25 degree specification is for absolute repeatability, over temperature, which, of course, is different from accuracy or sensor resolution and short term drift.

 

The viewable short term resolution is limited by the bargraph. Therefore, on the +/- 2 degree setting, over 53 segments you get about .07 degrees per LED 'pixel' segment. With the pixel smoothing it's visually about half that, or .035 degrees!

 

Of course, the sensor itself has a much higher resolution, in the order of micro degrees.

 

Regardless, once the unit is zeroed with the zero button and at a constant temperature, the display doesn't move more than about 1 segment or +/- .035 degrees due to sensor noise induced drift.

 

Of course, the real joy of the device is in its acceleration compensation, i.e., how it works while you are moving around. For this – for how good this unit is, I think a little video is in order.

 

This is a prototype of the "low-grade" version. Production ones work a tad better!

level quick.mov

[chris bangma]

The viewable short term resolution is limited by the bargraph. Therefore, on the +/- 2 degree setting, over 53 segments you get about .07 degrees per LED 'pixel' segment. With the pixel smoothing it's visually about half that, or .035 degrees!

If you have 53 leds, and each one corresponds to .075 degrees, how do you "smooth" it to see half of one led. In pixel smoothing, you usually blend to the adjacent "pixels" or in this case leds, so you would have one full brightness led, and 1 on either side at half brightness. This means you are visually seeing about 1.5 leds, giving you essentially .1125 degree resolution, and not .035.

[Chris Van Campen]

Looks pretty cool. Any idea when the BG is shipping??

[Victor Lazaro]

 

The viewable short term resolution is limited by the bargraph. Therefore, on the +/- 2 degree setting, over 53 segments you get about .07 degrees per LED 'pixel' segment. With the pixel smoothing it's visually about half that, or .035 degrees!

If you have 53 leds, and each one corresponds to .075 degrees, how do you "smooth" it to see half of one led. In pixel smoothing, you usually blend to the adjacent "pixels" or in this case leds, so you would have one full brightness led, and 1 on either side at half brightness. This means you are visually seeing about 1.5 leds, giving you essentially .1125 degree resolution, and not .035.

 

If you look at his video you get the answer, the indicator is about 5 led wide and the leds on the side are dimming to get in. Then you get the color coding for the out of balance sides. It gives the indicator a more fluid look and make it easier to see from afar at a single glance.

PS: I don't have the device no do I have any intention of buying it so far, just making observations.

[Eric Fletcher S.O.C.]

 

 

The viewable short term resolution is limited by the bargraph. Therefore, on the +/- 2 degree setting, over 53 segments you get about .07 degrees per LED 'pixel' segment. With the pixel smoothing it's visually about half that, or .035 degrees!

 

If you have 53 leds, and each one corresponds to .075 degrees, how do you "smooth" it to see half of one led. In pixel smoothing, you usually blend to the adjacent "pixels" or in this case leds, so you would have one full brightness led, and 1 on either side at half brightness. This means you are visually seeing about 1.5 leds, giving you essentially .1125 degree resolution, and not .035.

If you look at his video you get the answer, the indicator is about 5 led wide and the leds on the side are dimming to get in. Then you get the color coding for the out of balance sides. It gives the indicator a more fluid look and make it easier to see from afar at a single glance.

PS: I don't have the device no do I have any intention of buying it so far, just making observations.

And being 5 wide degrades that resolution even more

 

Smoke and mirrors....

[Jerry Holway]

Geez Eric, can't you see the display is always 4.5 wide?

 

So each full jump is .075 degrees at max resolution (i.e., 2 degrees full scale).

 

And it is acceleration compensated amazingly well, which is the key to its usefulness.

 

Not at all smoke and mirrors, but just plain good at an astonishing price.

[Chris Van Campen]

Practically speaking, I'd like to know how often one would be expected to have to zero this during use. If it's on a scale of hours, no problem. But if it's every 15 minutes, that's a lot less handy.

[RonBaldwin]

I had a chance to play with this level for a few hrs around cinegear time. The function was great...the form, not so much (admittedly to me). It worked very well but having to look away from the frame to see the display is a deal breaker (again, to me). I am an old dog and have been using the XCS level since '97 so the knight rider display drove me crazy. It would help a lot to change some of the colors/narrow the target area (it was hard to see what was what with one's peripheral vision). I will put up with a little delay to have my graphics on screen -- though sadly nobody seems to be able to do this in hd (Greg? Chris?) with the sensor at the post.

 

Here comes the dead horse...I was supremely disappointed with the radio shack power connector -- yeah, that'll last. I know it's soldered to the board...but it's YOUR board! I can't believe it wasn't at least a connector that can be found on Tiffen rigs.

 

All the bitching aside I assumed it was an early model/work in progress I was looking at and it will be improved before release into the wild. It's better than sticking a bubble on your screen for sure!

 

As far as resetting goes...I imagine that won't be necessary more than a time or two per day -- unless going from warm int to cold ext/visa versa

 

rb

[Chris Van Campen]

Thanks Rob ...

 

Also really underwhelmed with the power connector. Would be nice if that got upgraded.

 

I'm still running SD on a Transvideo/Zephyr, so looking for a way to get away from the liquid bubble - which only works when I'm moving exactly the same speed as the earth's rotation right under my shoes ... ;-)

[RonBaldwin]

Thanks Rob ...

 

Also really underwhelmed with the power connector. Would be nice if that got upgraded.

 

I'm still running SD on a Transvideo/Zephyr, so looking for a way to get away from the liquid bubble - which only works when I'm moving exactly the same speed as the earth's rotation right under my shoes ... ;-)

brief thread-jack:

for sd one cannot beat the xcs pdl...it is built really well and is just plain badass. Nebtek tried to make one as good for sd/hd a few years ago but as far as I know failed miserably (anyone compare it to the pdl?).

[Jerry Holway]

Ron-

 

to your questions and concerns: the connector is being changed, I believe to a Lemo.

 

Knight rider is just one of the display modes, see the video I posted earlier today.

 

I'd also love to see an alternate form factor to get the display closer to the image, and of course, eventually get it overlaid on the picture like all the old SD ones.

 

Yes, reset is not often unless the temperature changes a lot.

 

The sensor in the post is irrelevant with this degree of acceleration compensation. A sensor in the post is not really valuable for any spatial move, as that acceleration/deceleration will show up in the display. There are some pans where we don't move the sled much (staying in place), and a sensor in the post will work better, but the real ticket is all axis acceleration compensation, which this unit has in spades.

 

Is it perfect? Could it be better? Your concerns and suggestions are welcome.

 

Jerry

[RonBaldwin]

thanks Jerry -- good to know. I had just looked on the Tiffen site before posting and didn't see lemo listed -- it's good that it will be addressed.

 

I did see the video (this is how I had it set up when I played with it) and the change from green to yellow is barely noticeable...again, to me...especially in my peripheral (watch the video while looking an inch below the led display). Could be I am blind and don't know it :-( .

 

It is definitely a step in the right direction especially since no one else is doing a stand alone version for hd now.