Amsterdam Scientific Instruments (ASI)

Luna outputs

Ítalo Machuca Flores

Software · 6 months ago

Hi everyone, I have some doubts about the variables that Luna provides for post-processing. When they refer to 'cTOA' and 'cTOT', it means that it is the maximum/minimum value of the Clusters? or it is an average of TOA and TOT? Besides, I want to ask for 'tnc', what is the meaning of 'not correct TOA' ?

For the other side, the TDC events are for external signal, but I want to ask more specific details. What time are you measuring when you save the data to 'Timestamp'?

3 Comments | 431 Views

ASI Support · 6 months ago

Dear Italo,

the cTOT and cTOA are the TOT and TOA of the centroids of those clusters. To achieve the maximum performance of the timing of the clusters, timewalk correction has to be made, and then applied. This is done by LUNA. LUNA then outputs the raw TOA values, dTOA (which is the amount of correction for that TOA reading), and a corrected_toa if the correction is applied to the list of TOA values with LUNA. The amount of correction can also be applied with Python, by subtracting the amount of correction (dTOA) from the original arrival time (TOA).

About TDC, I'm afraid I do not quite understand the question. Could you please reformulate so that I can try to help you better?

Hope this helps!

P.S. here is some further information that might be helpful.

I advise you get a program on your PC (or the ASI PC should already have it installed) that allows you to browse the HDF5 file freely.

HDF Compass - open and browse the HDF5 file comfortably:

https://support.hdfgroup.org/projects/compass/

This is one way to read an hdf5 file using Python.

HDF5 to DataFrame - pd.read_hdf()

https://pandas.pydata.org/docs/reference/api/pandas.read_hdf.html

Under Clusters you'll find:

Clusters

ID à Identifier

Size à Size in pixel numbers

Min TOA à Minimum TOA in cluster

Max TOA à Maximum TOA in cluster

Ctoa à Centroid TOA

Corrected
ct à corrected centroid TOA

Sum TOT à sum of TOT in cluster

Ctot à Centroid TOT

X average à avg position (in pixel coords) of
cluster

X centroid à centroid (in pixels coords) of
cluster

PixelHits

PixelHits

cid à Cluster ID

toa à Time of Arrival (TOA)

corrected_toa
à TOA with correction from timewalk
already applied to the TOA

dtoa à amount of correction on TOA

TOT à Time over Threshold

X à coordinates of pixel hit in x

Y à coordinates of the pixel hit in y

ExposureTimeBoundaries

Ignore

Timewalk

Lookup
table. The lookup table is what is generated when performing the timewalk
correction. The amount of correction on TOA is governed by how much TOT is
recorded. The higher the values of TOT are, the less correction on TOA is required.

Mean,
Sem, Std, SumSquareDiff are extra figures we've made accessible.

Andrei Nomerotski · 6 months ago

Dear ASI Support - could you explain what it means: "cTOT and cTOA are the TOT and TOA of the centroids of those clusters"? what exactly is cTOA? Normally I'd take TOA of the pixel with largest TOT in the cluster, then apply to it the TOT correction and use the corrected TOA as the arrival time estimate for this cluster. from the variables in the list I think min_TOA would produce similar results and could be used as the arrival time estimate but I am not sure about cTOA..

Ciaran · 6 months ago

I can share with you how I'm computing the cluster. Basically clusters are built incrementally. Pixels are added to clusters on-the-fly and the cluster stats are updated.

```rust
// Insert a new PixelData into the cluster and update its fields
pub fn add_pixel_to_cluster(&mut self, pixel: &PixelHit, pixel_indices: (usize, usize)) {
// Update size
self.size += 1;

self.pixel_indices.push_back(pixel_indices);

// update the maximum tot value of the cluster
let pixels_tot_is_largest_we_have_seen_so_far = pixel.tot > self.ctot;
if pixels_tot_is_largest_we_have_seen_so_far {
self.ctot = pixel.tot;
// if we observe a tot value which is greater than the current maximum,
// update the centroid ctoa, which is defined as the toa of the pixel which
// has the greatest energy.
self.ctoa = pixel.toa;
self.x = pixel.x;
self.y = pixel.y;
}
self.min_toa = self.min_toa.min(pixel.toa);
self.max_toa = self.max_toa.max(pixel.toa);
self.min_x = self.min_x.min(pixel.x);
self.min_y = self.min_y.min(pixel.y);
self.max_x = self.max_x.max(pixel.x);
self.max_y = self.max_y.max(pixel.y);

// Update average_x
// running average formula:
// new average = old average + (new value - old value) / n + 1
self.x_average += (pixel.x as f64 - self.x_average) / self.size as f64;
self.y_average += (pixel.y as f64 - self.y_average) / self.size as f64;

// Update sum_tot
self.sum_tot += pixel.tot as u64;

// Update accumulated weighted sums
self.weighted_sum_x += pixel.x as f64 * pixel.tot as f64;
self.weighted_sum_y += pixel.y as f64 * pixel.tot as f64;

// Update centroid coordinates
self.x_centroid = self.weighted_sum_x / self.sum_tot as f64;
self.y_centroid = self.weighted_sum_y / self.sum_tot as f64;
}
```

For the other questions:

> Besides, I want to ask for 'tnc', what is the meaning of 'not correct TOA' ?
I'm not sure I follow the 'tnc' acronym but I think you are referring to the fact that there is ToA and corrected ToA? Basically ToA is what is computed by combining the 25ns and 1.5625ns clocks and the corrected ToA refers to the implementation of timewalk correction used inside luna. This is a statistical method created by Peter Svihra who helped me implement it.

> For the other side, the TDC events are for external signal, but I want to ask more specific details. What time are you measuring when you save the data to 'Timestamp'?
I think this is more of a serval question, but I found this on our internal documentation:

The TDCs offer 35 bits of 3.125ns resolution, thus a dynamic range of 107.37s

I hope it answers your question (but if not, please let me know).

Are you sure you want to
delete this reply?

Hi Andrei, many thanks for your message. I've been on holiday for 2 weeks so I do apologise for the delay in my response. The "c" in cToA and cToT does indeed represent the "clusters" ToA or ToT which is indeed the pixel with the largest ToT in the cluster. I will be sure to document this better! I can share with you how I'm computing the cluster. Basically clusters are built incrementally. Pixels are added to clusters on-the-fly and the cluster stats are updated. ```rust // Insert a new PixelData into the cluster and update its fields pub fn add_pixel_to_cluster(&mut self, pixel: &PixelHit, pixel_indices: (usize, usize)) { // Update size self.size += 1; self.pixel_indices.push_back(pixel_indices); // update the maximum tot value of the cluster let pixels_tot_is_largest_we_have_seen_so_far = pixel.tot > self.ctot; if pixels_tot_is_largest_we_have_seen_so_far { self.ctot = pixel.tot; // if we observe a tot value which is greater than the current maximum, // update the centroid ctoa, which is defined as the toa of the pixel which // has the greatest energy. self.ctoa = pixel.toa; self.x = pixel.x; self.y = pixel.y; } self.min_toa = self.min_toa.min(pixel.toa); self.max_toa = self.max_toa.max(pixel.toa); self.min_x = self.min_x.min(pixel.x); self.min_y = self.min_y.min(pixel.y); self.max_x = self.max_x.max(pixel.x); self.max_y = self.max_y.max(pixel.y); // Update average_x // running average formula: // new average = old average + (new value - old value) / n + 1 self.x_average += (pixel.x as f64 - self.x_average) / self.size as f64; self.y_average += (pixel.y as f64 - self.y_average) / self.size as f64; // Update sum_tot self.sum_tot += pixel.tot as u64; // Update accumulated weighted sums self.weighted_sum_x += pixel.x as f64 * pixel.tot as f64; self.weighted_sum_y += pixel.y as f64 * pixel.tot as f64; // Update centroid coordinates self.x_centroid = self.weighted_sum_x / self.sum_tot as f64; self.y_centroid = self.weighted_sum_y / self.sum_tot as f64; } ``` For the other questions: > Besides, I want to ask for 'tnc', what is the meaning of 'not correct TOA' ? I'm not sure I follow the 'tnc' acronym but I think you are referring to the fact that there is ToA and corrected ToA? Basically ToA is what is computed by combining the 25ns and 1.5625ns clocks and the corrected ToA refers to the implementation of timewalk correction used inside luna. This is a statistical method created by Peter Svihra who helped me implement it. > For the other side, the TDC events are for external signal, but I want to ask more specific details. What time are you measuring when you save the data to 'Timestamp'? I think this is more of a serval question, but I found this on our internal documentation: The TDCs offer 35 bits of 3.125ns resolution, thus a dynamic range of 107.37s I hope it answers your question (but if not, please let me know).

Are you sure you want to
delete this comment?

Dear Italo,

About TDC, I'm afraid I do not quite understand the question. Could you please reformulate so that I can try to help you better?

Hope this helps!

P.S. here is some further information that might be helpful.

I advise you get a program on your PC (or the ASI PC should already have it installed) that allows you to browse the HDF5 file freely.

HDF Compass - open and browse the HDF5 file comfortably:

https://support.hdfgroup.org/projects/compass/

This is one way to read an hdf5 file using Python.

HDF5 to DataFrame - pd.read_hdf()

https://pandas.pydata.org/docs/reference/api/pandas.read_hdf.html

Under Clusters you'll find:

Clusters

ID à Identifier

Size à Size in pixel numbers

Min TOA à Minimum TOA in cluster

Max TOA à Maximum TOA in cluster

Ctoa à Centroid TOA

Corrected ct à corrected centroid TOA

Sum TOT à sum of TOT in cluster

Ctot à Centroid TOT

X average à avg position (in pixel coords) of cluster

X centroid à centroid (in pixels coords) of cluster

PixelHits

PixelHits

cid à Cluster ID

toa à Time of Arrival (TOA)

corrected_toa à TOA with correction from timewalk already applied to the TOA

dtoa à amount of correction on TOA

TOT à Time over Threshold

X à coordinates of pixel hit in x

Y à coordinates of the pixel hit in y

ExposureTimeBoundaries

Ignore

Timewalk

Lookup table. The lookup table is what is generated when performing the timewalk correction. The amount of correction on TOA is governed by how much TOT is recorded. The higher the values of TOT are, the less correction on TOA is required.

Mean, Sem, Std, SumSquareDiff are extra figures we've made accessible.

All

All questions and topics

Hardware

(4)

All topics on hardware

Firmware

Firmware-related questions

Software

(2)

Software-related questions

Application

Application-related questions

Other

(2)