About the eshell upgrade kit

[Thierry Garrel]

Hello,
I've just finished a week using the new tungsten/blue led flat lamp upgrade kit supplied by Sheliak.
It replaces the Led lamp inside the calibration unit. The upgrade is well documented from Sheliak and easy to do. The latest version of Audela and ISIS allow the used of both tungsten and led lamp for order detection and geometric corrections, and flat field division.
The first thing was to found an exposure time of the tung/led lamp for order detection, from order 31, 7300A to 52, H gamma. This simply calling Tungsten series. The order detection came from the 2700K tungsten lamp, so enough exposure has to be set for deep blue detection. 3 images of 30s were able to detect all the order with an ST10XME camera.
The second thing was to set an exposure time for Led series. In this series both lamp are on, tungsten and blue led. This series is use for flat division and geometric corrections. It was delicate to find a good exposure time cause blue led are very intense and saturate sensor quickly as tungsten lamp need longer exposure to be at mid saturation. So the best compromise was to do a ten exposures of 5s.
The third thing was to go to the sky !
All image are processed with Audela, eshell module and Instrumental response was done with Vspec. I present here a 1B merged spectra of Altair. Note the absence of waves and ripples. A very clean continuum without high frequency artifact prove the very good geometric correction done by the eshell pipeline.

Alatir 1B

altair_1b_20111011_884.png (8.53 KiB) Viewed 6473 times

Only the overall shape have to be corrected by an instrumental response calculation done under Vspec. The result is a smooth instrumental response and 1c merged spectra proving the efficiency of the correction.

altair 1c

altair_1c_20111011_884.png (8.33 KiB) Viewed 6473 times

instrumental response

instrumental_response_20111011_884.png (5.13 KiB) Viewed 6473 times

So first conclusion is that the new kit is very effective for working on the full range of the eshell spectral domain. Flat, order detection and geometric correction allow the user to do easily merged spectrum from the different orders. That was a real difficulty before with only the led lamp. I think that we should be able to extend spectral domain to near UV and farther in the Nir.
More results will come here further.
Cheers
thierry

[Bernard Heathcote]

Thierry,

Postage takes a little longer to Australia so I only received my kit yesterday, which I installed the same day. The only problem I found was that the diffuser disc in my lamp unit was at the bottom of the 25mm extension tube, not at the outside end. Obviously I had to correct this in order to fit the long tungsten bulb.

I am puzzled by the different exposures suggested by Christian (relatively long, to achieve around 80% of dynamic range) and Olivier (relatively short).
What are the average maximum ADU you are obtaining for your 30s tungsten and 5s LED+tungsten exposures? Also, what exposures are you using for ThAr?
With a 4 sec LED+Tungsten exposure my ST-8ME gives a max of 5500 ADU in the blue (LED) and 4000 ADU in the red (tung) regions.

I've not yet had the opportunity to obtain a spectra with the new setup.

Cheers,
Bernard

[Thierry Garrel]

Hello Bernard,
It was very delicate to find good exposure. I made dozens of images to test for order detection first. On a single tungsten exposure, 30s leads to 35000 ADU max in the red part of the orders and allow a precise detection of the blue orders.
And dozens more test for tungsten+led, 5s leads to 35000 ADU (from led) in the blue part, and 6000 ADU on the red part (tungsten). 10s was too much.
Max ADU for a ST10XME is 55000 so 35000 is pretty good.
Individual ThAr time exposures is 8s, i used to take two or three of them before each star capture. Everything works well with eshell module under Audela i did not test with ISIS but it is the same pipeline.
I have to implemented new lines of thor arg in the config files to extend domain of eshell.
Cheers
T

[Stephane Charbonnel]

Hello Thierry,

I have tested too the new Eshel upgrade kit.
The number of ADU on one order depends on which orders you make focalization. In order to keep a homogenous resolution between orders 34 to 48, I have made focalization on order 39. So, my resolution is a little less on order 34 (as if I have made focalization on order 34), but my RMS is now greater on order 42 and resolution is always correct on order 48.
So, with a focalization on order 39 :
Typically, with my ST10XME :
• TUNGSTEN : about 48’000 ADU on order 39 with an exposure of 9 seconds. 35’000 ADU on order 34. A poor 4’000 ADU on order 48.
• LED : With a 20sec exposure, about 45’000 ADU on order 48. I will test if these values are constant during the time.

Stéphane

[Thierry Garrel]

Stephane,
it seems that it is the reverse than me. 30s for tung to me against 9 for you and 5s on led-tung for me 20 for you.
Surpise !

[Stephane Charbonnel]

Thierry,

Nevertheless, I can confirm these values for me.
In fact, I have a real in the blue with LEDs ... LED Orders 53 and 54 are so faint as order 48 for example. So, ISIS has some normal problem to obtain order 53 and 54.