EEVblog Electronics Community Forum
Products => Thermal Imaging => Topic started by: Fraser on January 02, 2020, 09:10:34 pm
-
My latest acquisition and my second attempt to own a Bullard Eclipse thermal camera! This was intended to be a personal fire fighting thermal camera issued to every fire fighter on a crew. It therefore had to be compact, convenient and affordable.
The last one that I bought, an Eclipse LD, went back to its rightful owner after I discovered that it was not the property of the eBay seller to sell to me :( Its condition was too good and so I had contacted Bullard USA for a 'lost camera' check. It came back positive so I lost it from my collection :( That really hurt as I have long wanted one of these interesting little thermal cameras.
The Bullard Eclipse interests me due to its compact dimensions and ergonomic styling for fire fighter use. To my eyes, its case styling looks almost organic in nature as it has a front, a rear and a waist between. It looks a little like a Gourd or Monkey nut :)
The LD model that I handed back to its owner had the large 3.5" LCD display and was a QVGA model. The unit that I have just purchased is the standard Eclipse that provides a 2.5" LCD display and only 80 x 60 pixel resolution. I wanted one of these cameras so badly that the low resolution did not concern me and I have some thoughts regarding how Bullard implemented the low resolution. If I was right, I might be able to release the cores full capabilities. If I was wrong, well I still had an Eclipse in my collection.
So what is the back story to this particular Eclipse camera ? Well I was hunting eBay for a Bullard Eclipse located in the USA as that is where they are most common. I found an auction that was asking $1300 for a Bullard Eclipse and a Bullard T320. The seller stated that he would split the cameras if requested. I quite thought that the cameras would be $650 each and way above my budget for one when shipping and importation costs were added. Even so, I asked the seller if I could buy the Eclipse from him. The reply was better than I could have hoped. The camera was mine for $300 and I was dealing with the Deputy Fire Chief of a Fire Brigade that was selling the cameras to convert unwanted kit into useful funds for the station.. I sealed the deal and the Eclipse was mine :)
From the picture of the Eclipse camera in the Auction I had already established that it was most likely a standard model with the limited 80 x 60 pixel resolution. The low resolution model has a plain stainless steel front plate around the lens. The higher resolution model has a checker plate finish instead. The camera also came with its vehicle/desk charging dock, battery and a USB communication cable. That communications lead on its own is $70 on the secondary market ! I know because I bought one for the Eclipse LD :(
The camera arrived on New Years Eve and thankfully had no trouble leaving the USA, unlike an MSA 5600 that I purchased a while back that got inspected twice !
The unit is in excellent condition considering that it is a fire fighters camera. It does not smell of smoke and there is no soot on it. I suspect that it was mostly used in training and saw little, if any, real fire fighting action. The "Powerhouse" charging dock is also in great condition and just requires a 12V power supply. It has been mounted in a vehicle at some time but was clearly looked after. The battery is dated 2014 so the camera is likely a similar age. The communications cable is perfect and looks unused. It is needed to talk to the cameras Bullard designed main board within the camera. Via this link the PC software can access the camera functionality and even upgrade its capabilities. More on that below........
The Eclipse camera was designed to be very versatile, yet affordable. As such the basic Eclipse model was 'just' a simple thermal camera without temperature measurement, hot spot highlighting, thermal throttle or Zoom functions. There was also an option to have a DVR module fitted at the factory. The resolution of the camera was chosen at the time of ordering but Bullard did offer a resolution upgrade service. This required the unit to be returned to them for the resolution increase work. The other 'software' functionality could be added to a basic camera by a user, after purchase, by paying an additional fee to Bullard and unlocking the new feature(s) using the PC software that links to the Bullard upgrade server.
Let us just analyse this upgrade situation for a moment.....
1. The Basic Eclipse Camera can have software based functionality added by unlocking such using the PC app. This is the Bullard Main board firmware interacting with the PC software and upgrade licensing server at Bullard.
2. The DVR option is hardware and can only be added by returning the camera to the factory. That makes sense as placing the separate DVR into every camera and enabling it via software link makes little financial sense. The DVR board is often only needed for training or specialist scene analysis so it is a specialist requirement. Those who need it will normally buy it when they order the camera.
3. The Resolution option is offered as a 'return to Bullard' upgrade if a basic Eclipse camera user ever wishes to improve image detail. When I first read about this upgrade I thought that Bullard possibly remove a 80 x 60 pixel core and fit the more expensive and capable 160 x 120 pixel core. Why did I think this ? Well a 80 x 60 thermal core was available and cheaper than a 160 x 120 pixel core. Why waste a 160 x 120 pixel core in a 80 x 60 pixel Eclipse camera ? If cores were inexpensive then doing so might make sense, but that is not the case. However, the profit margin on these cameras might be adequate to permit a higher specification core to be fitted and windowed down to 80 x 60 pixels. The wasted capability of the core and increased cost might be offset by the use of a single core model across a whole camera range.
Then I studied the 80 x 60 pixel and 160 x 120 pixel models, especially the visible lens design. I noted that the lower resolution model had a different lens design to that of the higher resolution model. I thought that enough to make the use of different cores in the two models very likely.
So what about the Eclipse that I have purchased ? Well it is the basic model with no temperature measurement or other options and it is 80 x 60 pixel resolution. On a 2.5" screen the 80 x 60 pixel resolution is not too obvious. There may be smoothing and/or interpolation being used to improve the image ? Now the great discovery for me..... I looked at the lens front and immediately recognised it ..... my little Eclipse is equipped with a FLIR TAU core :) I powered the camera after charging the battery and my thoughts on the core were further strengthened....there is a FLIR logo on the display. I believe the core to be a TAU2-160 (25um pixels) that provides the lowest resolution available on a TAU core, namely 160 x 120 pixels. The core is fitted with a 5mm lens so it cannot be the higher resolution QVGA+ TAU core as that does not come with a 5mm lens option.
So what cores have been fitted in the Bullard Eclipse over the years ?
In truth I do not know for certain but I can make an educated guess from what I know of the camera.
The camera was released with an Indigo decal on its side. The available Indigo cores were the Omega and Photon. Both are 160 x 120 pixels minimum. The unusual lens that I saw on the lower resolution 80 x 60 pixel core may, in fact, have been an Eclipse that was fitted with the distinctive Omega lens and this is a likely candidate in my opinion.
The Indigo core appears to have been replaced with the later TAU series as the Omega and Photon became obsolete. This would be a logical move as the TAU is a descendant of the Indigo Omega and photon series.
The Bullard Eclipse LD that I briefly had in my care was equipped with a QVGA L3 Nano Core that was a A-Si microbolometer descendant of the L3 4500AS series. Why they moved to an A-Si core from L3 is not known to me.
The latest Bullard Eclipse cameras are equipped with a DRS core.
So what about the upgrade requiring a return to the factory ? Why not just a software function, as with other options ?
Well it is my belief that the software functions are controlled on the Bullard produced main board that communicates with their software via the communications port. It is also my belief that the core in the Indigo and FLIR based cameras is not operating in a configuration slave mode with the Bullard main board. As such there is no direct communications link to the core to access its configuration. As such the camera would need to be returned to the manufacturer, opened and direct access gained to the core I/O port via a configuration cable like the one I already have for my TAU cores. After opening the camera and configuring the core, the camera would need to be serviced, reassembled and re-sealed before being tested and returned to service. This cost around $2K as an upgrade service. $2K would not normally cover the cost of a new core, so that also adds weight to the same core providing both the 80 x 60 and 160 x 120 pixel resolutions.
I will be delving inside the Eclipse camera to gain access to the core. At that point I will have a positive ID on the exact FLIR core model and will be able to access it with my configuration cable and software. With luck I will be able to reconfigure the core to its full 160 x 120 pixel resolution. We shall see :)
I will update this thread when I have more news :)
The pictures of my camera are "as received" and I will clean it during its service. For a fire fighting camera it is very clean and relatively unmarked ! Just how I like my cameras :)
Fraser
-
Pictures continued
-
Some pretty awful screen captures of the LCD screen. Apologies for the rubbish quality but these were just hand held quick 'snaps'. The display looks better than in the pictures ! My cats are in some of the images. Note Oskar's tail is dark as he had just come in from outside and his tail was cold ! The camera easily sees the plaster dabs and central heating pipes behind the plaster board in our lounge. I will produce better pictures later and hopefully some 'Before' and 'After' images of the camera if I can release the full 160 x 120 pixel resolution of the core.
Fraser
-
Just to underscore my thoughts on the Bullard Eclipse 'Organic' case shape.... some pictures of Gourds and Peanuts ;D
Fraser
-
Just some publicity pictures of the Eclipse. Some show the ergonomic grip facilitated by the cameras narrower ‘waist’ mid section. Great for heavily gloved hands :-+
-
More thoughts on the resolution differences between the Eclipse 80 and 160 models.
Whilst I still believe the same TAU core is used in both cameras, I cannot find a function in the TAU configuration that would output an 80 x 60 pixel image without also halving the lens FOV. This is important because the 5mm lens on a TAU2 160 provides just over 40 degrees HFOV. That appears to be the same HFOV as specified for the low resolution Eclipse 80. This means that the simple X2 Zoom is not being used in the TAU core as that would result in a HFOV of around 20 degrees. Unless there is an undocumented/custom pixel combining function in the TAU, the resolution reduction would need to be carried out using quad pixel combining in the video stages of the Bullard main board. I shall have to investigate the Eclipse main board and test the TAU core by driving it from a PC using the FLIR configuration utility.
Fraser
-
I think the 2.5" LCD display and only 80 x 60 pixel resolution is too low for displaying a better image, the 3.5" LCD display (https://www.szmaclight.com/3-5-inch-tft-lcd-module-display.html) and QVGA resolution model would be more better for the camera display.
-
Kathy,
Fair comment. The preferred size of LCD display is 3.5” and that is why Bullard produced the Eclipse LD model. The original Eclipse was intended to be a very compact personal TIC to provide thermal vision capability to fire fighters. It has served well in that role to date. For a fire fighter it is more about identifying sources of heat rather than highly detailed large images. They are often looking for an adult or child lying on a floor having been immobilised by smoke inhalation. They also keep an eye on eachothers location and hot areas behind walls, doors, ceilings and floors indicating the location of hot gases or fire. All is done with care but quickly so the issue of the amount of scene detail is very different to use for building thermal surveys etc. It is worthy of note that the basic Eclipse spec has temperature measurement as an option. I would have thought that function essential for fire fighting but my unit does not have it enabled. I do agree that 80 x 60 pixels is far from ideal and I would consider 160 x 120 a minimum resolution for many tasks that I use a thermal camera for. We will have to see whether I can release my cameras full 160 x 120 pixel capabilities.
Fraser
-
I have not got the spare time to dismantle my Eclipse at the moment, but I can add the pictures I took of the Eclipse LD whilst servicing it before its return to its rightful owner. Or as they say on cooking programs "here is one I did earlier" ;D I will add the pictures of the standard Eclipse once I have time to dismantle it. It will likely be the very similar to the LD version as the LD larger screen is basically just a bolt-on casing and a different core was used. There was a late revision of main board in the LD that incorporated the USB-serial UART bridge rather than it residing in the interface module as in earlier releases. I suspect mine is the earlier main board with external USB-UART bridge.
The core in the Eclipse LD model was a L3 "NanoCore 320" QVGA module. More recent models use a Leonardo DRS core.
To the pictures......
-
Pictures of Eclipse LD continued
-
Pictures of Eclipse LD continued
-
Pictures of Eclipse LD continued
DVR module that sits behind LCD display panel. Known to be a bit flakey and crash prone !
-
Final set of pictures of Eclipse LD disassembly.
I did not get the opportunity to delve deeper into the main board or core as this was just a quick look as part of the service.
I did reverse engineer the communications interface and USB-UART bridge section of the main board however.
Fraser
-
For those interested, I attach the datasheet for the L3 "NanoCore320". It is not a well known core but likely the replacement for the earlier L3 Thermal Eye 4500 series. This datasheet took some finding !
Fraser
-
It is worthy of note that the basic Eclipse spec has temperature measurement as an option. I would have thought that function essential for fire fighting but my unit does not have it enabled. I do agree that 80 x 60 pixels is far from ideal and I would consider 160 x 120 a minimum resolution for many tasks that I use a thermal camera for. We will have to see whether I can release my cameras full 160 x 120 pixel capabilities.
Fraser
There are two schools of thought on spot temperature measurement. The counter-argument is as follows:
In a fire any emissivity calculation is impossible. The temperature will be wrong.
Firefighters have enough to do than think about spot temperatures
Many cameras have a built in bug. They stay in 'high gain-low range' mode and therefore will read out '150°C' when looking at a very hot small object. Some do add a '>' but the readout scale stays put.
This keeps a low noise picture, but the spot temperature is incorrect.
Readout obscures pixels
Spot temperature is likely to be less prevalent in the future, or become an 'extra training' option.
Bill
-
Dear Bill,
That is really interesting to hear. I had mistakenly thought fire fighters needed the temperatures to determine ferocity of fire beyond a barrier or accumulation of hot gases. As you say they have a lot going on whilst in a burning building. Is that why some fire fighting cameras provide a high temperature highlighting mode that displays yellow, then red for any part of the scene above a set temperature ? It would certainly help in spotting the hot spots in a building that is on fire. I suspect fire fighters have specialist scene risk assessment training and almost a sixth sense for where is safe to go in a burning building and where might present a backdraft or high flashover risk.
It is really interesting to hear that thermal cameras are capable of misleading a user in a fire fighting situation. It is always good to know your equipments limitations where safety is involved. As you have told me previously, fire fighting thermal cameras are an aid to the fire fighter but not a licensed life saving equipment.
Fraser
-
Today I was offered a faulty Eclipse LD free of charge :) As I have ‘blown the budget’ of recent, this was a lovely surprise :-+ More on that poorly camera once it arrives. It is the 320 x 240 pixel model so likely contains the L3 NanoCore 320 :)
Fraser
-
Dear Bill,
That is really interesting to hear. I had mistakenly thought fire fighters needed the temperatures to determine ferocity of fire beyond a barrier or accumulation of hot gases. As you say they have a lot going on whilst in a burning building. Is that why some fire fighting cameras provide a high temperature highlighting mode that displays yellow, then red for any part of the scene above a set temperature ? It would certainly help in spotting the hot spots in a building that is on fire. I suspect fire fighters have specialist scene risk assessment training and almost a sixth sense for where is safe to go in a burning building and where might present a backdraft or high flashover risk.
It is really interesting to hear that thermal cameras are capable of misleading a user in a fire fighting situation. It is always good to know your equipments limitations where safety is involved. As you have told me previously, fire fighting thermal cameras are an aid to the fire fighter but not a licensed life saving equipment.
Fraser
Spot temperature can be used (subject to the counter argument above) for checking any gas cylinders reaching critical temperatures, and the yellow/red colours can be helpful in seeing if a backdraft is building up (you see fire gases boiling up towards the ceiling).
Again some cameras have what might be seen as a bug, the red/yellow can appear at two different temperatures, approaching the lower gain switch point as well as getting to 'darn hot'.
Both the problems are down to trying to keep a nice picture as long as possible before resorting to a low gain range to image up to 600 or 1000 °C.
As detectors have got better this is less of a real problem, but some cameras still behave this way.
'Doing it right' has no prizes, such cameras then get panned for having a noisy picture when a small hot object is being measured or put in the scene for a sales argument.
Bill
-
Thanks Bill,
This has been a very useful insight into the challenges facing fire fighter camera designers and also fire fighters who use various brands of camera in the field. Once again it comes down to knowing your tools and their limitations :-+
Cheers
Fraser
-
A comment on the design of the Eclipse camera that surprises me...... this is primarily an affordable, ergonomic fire fighting camera. As such it is robust and water proof. Why then did the designer not include a replaceable lens protection window as found in most other modern fire fighting thermal cameras ? The lens of a thermal imaging core can be coated with Hard Carbon to help preserve it in hostile conditions but a sacrificial layer of physical lens protection is still a good idea for lens longevity. An unusual, but by no means historically unique, decision by the designer and one that I see has been changed in the latest releases of Eclipse camera. The new models equipped with a DRS core have a lens protection window incorporated in the front panel assembly. A good move in my humble opinion :-+
Fraser
-
Humble? Fraser? :-DD
(in friendly jest ;) )
-
:-DD
-
I have produced two new threads to cover the disassembly of the Eclipse 80 and the Eclipse LD that I gifted.
https://www.eevblog.com/forum/thermal-imaging/inside-a-bullard-eclipse-80-fire-fighting-camera-by-fraser/ (https://www.eevblog.com/forum/thermal-imaging/inside-a-bullard-eclipse-80-fire-fighting-camera-by-fraser/)
https://www.eevblog.com/forum/thermal-imaging/inside-a-poorly-bullard-eclipse-ld-320-and-its-l3-thermal-eye-nanocore-by-fraser/ (https://www.eevblog.com/forum/thermal-imaging/inside-a-poorly-bullard-eclipse-ld-320-and-its-l3-thermal-eye-nanocore-by-fraser/)
The Eclipse 80 contains a FLIR Tau core as suspected but the surprise was in its part number. It is supplied by FLIR as an 80 x 60 pixel core and is not just a TAU 160 that Bullard have hobbled. From what I know of the TAU, the lowest resolution manufactured was 160 x 120 pixels so it is likely FLIR have applied pixel combination to reduce the resolution at its output. There is an additional PCB on the rear of the core that is either a cable type adapter or may be a configuration board that sets up the core at boot. I will be taking a much closer look at the core to assess whether it can be reconfigured to 160 x 120 pixel resolution.
The Eclipse LD 320 that was gifted to me has arrived and is in surprisingly good condition considering it has come from active fire service. The materials from which these cameras are made is certainly very durable. The camera contains the L3 Thermal Eye “NanoCore 320” as expected. The lens is hard carbon coated and after a clean is as shiny and scratch free as new :) The camera suffered minor water ingress through ‘abuse’ and not a failing in the cameras design. The small amount of water that got past the seals caused corrosion on the bottom of the control PCB. That side of the PCB is unpopulated. Thankfully the NanoCore appears to have been unharmed and has not come into direct contact with the water :phew: I will work on the control board to determine what repairs are required and attempt to resurrect it from its current inactive state. Power rails have taken the brunt of the damage as you would expect. This Eclipse LD is also equipped with the DVR option which was nice to find :-+
The poorly Eclipse LD is proof that no fire fighting thermal camera is indestructible ! When I say ‘abuse’ caused its damage, I do not mean deliberate action, but rather that the camera was exposed to conditions that fell outside its protection specification and it could not defend itself against. Bad stuff happens to fire service cameras.... it goes with the territory.
Fraser