First Post/Fun with a FLIR TAU2

[manic515]

TL;DR: Looking to downgrade firmware on a FLIR Tau2 640 60Hz core but can't find the .hex or .pcf files for version 2.4. I doubt FLIR Customer Service would be willing to help with a Frankenstein-ed thermal optic.

.......................

If you have the time, I have an entertaining project.

First off, I'm amazed by the amount of knowledge on this forum. There are a lot of answers to questions, but just plain curiosity as well. I love the hardware tear-downs, particularly. I'm in the middle of such a project. I don't want to burden anyone's bandwidth, but I'd be glad to post pics of it should anyone be interested. So let's get to it.

I recently came into possession of a FLIR (Armasight) Apollo Pro MR 50 MM thermal clip-on that was non-functioning. To be clear, I wanted to build a handheld thermal device on a shoestring budget for wildlife OBSERVATION at night. The plan is to locate wildlife in foliage with the thermal and then do the actual observing with night vision. I've already repaired/restored a few sets of PVS-7 and PVS-14 devices (fun to share). I know analog night vision isn't thermal (although you could argue an infrared connection), but I'd love to discuss that elsewhere on the forum at some point. So no interest in hunting -- never really liked it. Just went for the Apollo because it was dirt cheap and ruggedized for the outdoors.

Anyway, the Apollo would only boot to a "mystery" blue screen. The reason? No thermal core. Painfully obvious problems are better than the subtle ones, I guess. It did state on the housing it was a body kit, so maybe it never had a core to begin with? Never could find evidence of just a FLIR Apollo Pro kit for sale, but it's a discontinued model anyway. Found out it took a FLIR Tau2 thermal core by reviewing the specs so it was back online. Turns out you can get strictly "as-is" thermal cores for cheap sometimes when broken imaging devices are parted out. You can't be too picky about resolution, refresh rate, etc., as many of these cores were customized for specific applications. The core I picked had a shutter, but that just makes it better for observation and less suitable for use on a weapon so great.

Had to re-machine the front cover of the thermal core to integrate with the 50 MM objective lens of the clip-on, but aluminum is easy to work with. Needed a flat front rather than the lens adapter ring to achieve proper focus. Seated and reconnected and screwed it all back together and I had a fully-functioning FLIR Apollo Pro. The Tau2 worked perfectly with the clip-on circuitry. Had the full Armasight menu available (didn't need the reticles or boresight, obviously) and was able to cycle through the various palettes, digitally zoom, etc. That's where I should have stopped.

My issue was upgrading the Tau2 firmware. The upgrade went fine and I was excited about the 60 Hz refresh rate. Definitely a bigger drain on the batteries, but I had a rechargeable pack that was compatible so no big deal. Unfortunately, I lost compatibility with the Armasight menu in the process. The power and zoom buttons still work, but the menu button is unresponsive. I still have a working thermal that looks great, but can't access any options. I could likely fix this by updating the Apollo Pro firmware, too, or maybe even by flashing the current BIOS of the clip-on but there's no way I'll ever get those files. Seems easier to attempt a downgrade of the Tau2 core firmware since someone here might have that.

Will it work? I don't know. I'd love to find out. Looks like the FLIR GUI software will just flash whatever compatible firmware files I designate. If not, I can just wait for another "as-is" Tau2 core to come along and drop it in without the upgrade. Wish I'd dumped the old Tau2 firmware before upgrading, but you know what they say about hindsight...

Anyway, I have what I set out to build. A cheap, yet high-quality thermal device I can use for wildlife observation. I can tweak the parameters of the Tau2 through the VPC module, but it would be convenient to have the on-board menu back. I'm lucky white-hot is my preferred palette, too, since that's what I'm stuck on. At least it isn't rainbow. I'm open to any suggestions and hope that I may be able to contribute to some of the other fascinating discussions on the forum. Thanks for having me. I'll circle back to this topic and post any updates or breakthroughs I may have on my own 

[Fraser]

I will check my archives but I do not recall ever having TAU firmware. No harm in checking though.

Fraser

[Fraser]

I just checked my archives for TAU2 firmware and found none. This suggests to me that FLIR never published TAU2 firmware on their support site, else I would have downloaded it.

I can only suggest that you try contacting FLIR support with the serial number of your TAU2 and ask them for the firmware that you require. Sadly if your core is salvaged from a product, such as a CCTV camera, FLIR are unlikely to find its serial number on their sales database and so decline support. FLIR CS have previously stated that they did not, and do not, issue firmware field updates for the TAU series so you may be out of luck.

Fraser

[manic515]

Thank you for checking on the TAU2 firmware. I actually found the files to upgrade from version 2.4 to version 2.8 in THIS THREAD on the site and the process worked via VPC module/serialized connection. Went from a 30 Hz core to 60 Hz core. Definitely a performance boost, but I'm kicking myself over losing the Armasight menu I'm new to thermal, though, so it's a case of live and learn. Still have a working optic, after all.

FLIR states that more recent TAU2 cores are "field upgradeable" (I can confirm that mine was) but firmware is only provided via Customer Support. Guess it's to keep consumers (and me, sadly) from getting into trouble with hardware compatibility issues. There's an advanced menu in the TAU2 GUI software you have to unlock to perform the upgrade, too. I came here hoping someone had the older firmware. Not a chance I'd get legacy firmware from FLIR Customer Service and that's okay. I wouldn't hand it out to me, either, for a weird and completely unwarranted project

I have a couple options I can pursue if I can't locate version 2.4 or earlier of the TAU2 firmware. The easiest would be to get another pieced-out random TAU2 core. It would probably have an earlier firmware that would restore the menu. I'll be sure to dump the files for backup if I go that route and provide them (as-is, of course) for anybody who experiences a similar mistake to mine. The other option is to dump the Armasight Apollo Pro board (it has a VPC-compatible port) in Linux and attempt a flash of that firmware to update the configuration. That probably wouldn't work, but would be fun to try.

Anyway, thanks again for checking. I'll keep looking, too, and update this thread accordingly with whatever happens next     

[Fraser]

Manic515,

Thank you for reminding me about the TAU firmware that was published on this forum. Sadly that firmware was provided without any background information on which TAU cores it supported and whether it was an official FLIR release. I likely did not archive it for that reason (not wanting to potentially harm someone’s TAU core in the future). I note that the firmware update/upgrade resulted in increased column noise and whilst an increase in image noise can result from higher frame rates, I would not expect FLIR to issue such a firmware to a user if it caused such a degradation in performance. That TAU firmware download remains a bit of a mystery. I am pleased that your TAU still works and may be controlled by the GUI.

Fraser

[manic515]

You were right not to archive the "mystery" TAU firmware. It had an accompanying .PDF that looked like an official FLIR release, but we'll never know. Seems like a hack author would have tried to take credit somewhere (like a custom boot logo) but maybe not. I'll supply my observations. Started with a FLIR TAU2 640 x 512 30 Hz core. Installed the firmware without issue with a VPC module and the Advanced menu of the FLIR GUI software. Ended up at 60 Hz. Undoubtedly better overall performance, but at the cost of the overlay (no battery indicator, temperature, range-finder, etc.) and Armasight on-board menu. I was able to restore hot-key functionality today by removing and re-inserting the ribbon cable from the control board to the button board while powered on. Now I can cycle through the various color palettes and make various image adjustments with button combinations. Not a menu, but a step in the right direction and a lot more convenient. Still working on a firmware downgrade to see if that restores the missing menu/overlay. I have some ideas. It would also be interesting to try the latest official firmware version from FLIR to determine if the mystery firmware on the forum is a hack or not. Wonder if I could downgrade to the 60 Hz model and keep the performance enhancement? There I go getting into trouble again!   

*EDIT*

I was able to register the TAU2 core successfully with FLIR (guess no one ever activated the warranty) and have requested an official copy of the latest firmware from Technical Support per site instructions. Also requested the previous firmware files. Would be nice to get the files I need through official channels. Thanksgiving is upon us here in the States so it'll probably be a while before I get a response, but no big deal. We can put the mystery of the TAU2 firmware posted here to rest, perhaps, even if I don't get my menu and overlay back. We'll also settle the question of whether or not firmware reversion is possible. Stay tuned!

[Fraser]

Thank you for the update. I hope FLIR CS provide what you need 

Fraser

[hap2001]

I note that the firmware update/upgrade resulted in increased column noise

As I posted in that "firmware" thread:

I already tweaked with that for the whole day...
I suspect it's the problem with bias.
Even before update, if you set the lens aperture too high, it will have column noise, and if set too low, time constant will be very long. So I already played with that magic parameter for days...
Now after update, I seems cannot get it right, with best value I can guess, the column noise is still way too high...

Maybe a gain re-calibration can solve the noise problem?

[Fraser]

Just thinking about the TAU 2 640 core for a minute. IIRC the TAU 2 640 later models had two frame rate configurations depending upon the use of the frame averager function. The TAU 640 2 50/60fps  model could output a frame rate of 50/60fps provided the frame averager was disabled. If the frame averager were enabled, the image would have lower noise levels and so have nicer appearance. With the frame averager enabled, the output of the core is reduced to 25/30fps. This got me wondering whether the host system could become ‘upset’ if the feed from the core suddenly doubled in frame rate. Could it be that the Graphics Overlay system can no longer synchronise to the frame rate coming out of the core ? This is totally off the top of my head without reference to technical documentation so I could be talking rubbish ! If you can enable the frame averager in your updated core, the output frame rate would drop back to 25/30fps and you may regain sight of your graphics overlay. The reason for my thinking here is that your updated TAU 2 core behaves normally when accessed via the PC based GUI and, AFAIK, TAU 2 cores normally use the same communications across the model range. If the host checks the TAU 2 model number for validity, I would have expected issues there to prevent the booting of the scope. Have a look in the advanced configuration menu on the latest TAU 2 configuration GUI and see if an option exists to enable the frame averager. If enabled, you get a nicer image and may regain your graphics overlay.

Fraser

For those unaware, the TAU 2 had an option to average two frames in order to improve the noise levels in the image and generally improve image appearance by reducing image blur. When enabled, it halved the output frame rate so the microbolometer is operated at twice the desired output frame rate to compensate.

[Fraser]

I just checked a datasheet for the TAU 2 and see that the core has the option for external synchronization. More blue sky thinking…. Could it be that your updated core is now running at 50/60fps and the host electronics are creating a video overlay at 25/30 fps with a synchronization signal being sent to the core ? If so, the core may ignore such an invalid sync signal and so the graphics overlay system fails to synchronise to the core output ? Maybe worth checking. I know nothing about your thermal scope host electronics but an incompatibility between your cores output frame rate and the host systems Graphics overlay electronics frame rate does seem to be a possibility.

Fraser

[manic515]

Have a look in the advanced configuration menu on the latest TAU 2 configuration GUI and see if an option exists to enable the frame averager. If enabled, you get a nicer image and may regain your graphics overlay.

Fraser

I just checked a datasheet for the TAU 2 and see that the core has the option for external synchronization. More blue sky thinking…. Could it be that your updated core is now running at 50/60fps and the host electronics are creating a video overlay at 25/30 fps with a synchronization signal being sent to the core ? If so, the core may ignore such an invalid sync signal and so the graphics overlay system fails to synchronise to the core output ? Maybe worth checking. I know nothing about your thermal scope host electronics but an incompatibility between your cores output frame rate and the host systems Graphics overlay electronics frame ratpre does seem to be a possibility.

Fraser

Precisely why I joined this forum -- brilliant suggestions and I'm grateful. And I'll try anything -- part of the fun of these projects. I suspected some sort of incompatibility between the clip-on hardware (of which I know nothing) and the newly-updated FLIR TAU2 (which I'm learning). Had no idea where to begin troubleshooting and this gives me a place to start. A synchronization discrepancy in frame-rate between scope and TAU2 core COULD explain the issue. I'll play with the Advanced menu settings (already have the latest GUI software downloaded) and see what happens. Perhaps I can simply drop back to the previous output that way and avoid a downgrade. Very little risk in messing up the firmware because I can simply overwrite it and start over if need be. Great feature of the TAU2.

Being a bonafide registered/two-year-warranted FLIR customer now , I also opened a query with Customer Service/Tech Support (per FLIR instructions) to request copies of the latest and next-to-latest (for downgrade purposes) FLIR TAU2 core firmware. The initial response was disappointing -- I was told that TAU2 cores could not be field-upgraded and asked why I needed a copy of the firmware files. In response, I cited PG 19 of the TAU2 Product Specification Guide (which specifically addresses field upgradeability and, yes, I read the whole document) and stated that I wished to enjoy the improvements and enhancements of the latest firmware but with the option to revert to a previous version if necessary. Not to mention I've successfully performed a field-upgrade on the TAU2, myself. I stated I was familiar with the procedure and had the necessary equipment. I'll keep the FLIR query open until it gets escalated past the power-it-on/power-it-off gatekeepers. Right now, I can state with complete authority that the forum knows more than FLIR's "expert" representative and the advice here has been far more helpful. Why FLIR would make a fuss over a simple request for firmware on a field-upgradeable core (a feature THEY chose to implement) is beyond me.

Anyway, I'll play around with the Advanced menu settings and see what happens. We'll learn some things about the TAU2 core whether the settings work or not. Thanks again, Fraser!

*EDIT*

No frame averager function I could find, unfortunately, but I believe frame-rate discrepancy explains my menu/overlay disappearance. The FLIR Apollo PRO came in two versions: a 30Hz and a 60Hz model each powered with a corresponding TAU2 640 core. The controller board for my version of the Apollo PRO must have been the 30Hz version (had no documentation to tell me) which is why it worked perfectly with my TAU2 640 30Hz core. The firmware files from the forum were compatible with my TAU2 core, but changed it to 60Hz model. Not faulting the firmware -- it did fine. All the features of the TAU2 core still work in the clip-on with hotkey buttons, but no menu/overlay. If I'd had the 60Hz version of the Apollo PRO, I probably wouldn't have lost anything. Kind of wish FLIR had implemented a direct feature in the GUI to switch between 30Hz and 60Hz. It's an option for some of their other models. So I'm back to waiting on FLIR Technical Support for the TAU2 640 30Hz firmware. I believe that's the solution. Revert to the previous firmware spec and restore compatibility with the Apollo PRO. Wish I'd just dumped the original firmware, myself, but that's hindsight for you.

I'm fortunate (despite what FLIR Technical Support has tried to tell me) that swapping between firmware versions for late-model TAU2 640s is easy. The GUI simply overwrites the existing firmware with the files you designate. It would be a great feature if obtaining firmware weren't so difficult  It'll probably be Monday or Tuesday before I get my next response from FLIR. If I get ghosted, I will simply open a new query and cut/paste. FLIR's instructions are to reach out to Technical Support for core firmware upgrades (why they simply don't provide a link in the firmware section, I don't know) and as a duly registered/warranted FLIR customer I see no reason why they can't provide the requisite firmware files to me for archival purposes.

Anyway, we'll see what happens next.

[CRC]

I flashed that firmware hosted here to a tau 640 without issue. I didn't notice any increase in vertical line noise.

[manic515]

I flashed that firmware hosted here to a tau 640 without issue. I didn't notice any increase in vertical line noise.

Did the same thing and didn't notice any performance issues. Everything still works on my TAU2 640. Wonder if the vertical line noise increase was an issue on the hardware/body side rather than the camera side. What that firmware DOES do is upgrade 30Hz cores to 60Hz and that was my issue, I think. Trying to drop back down to 30Hz to restore hardware compatibility. Easier than getting the 60Hz version of the FLIR Apollo PRO. Wouldn't recommend the files hosted here for 30Hz versions unless you know your hardware is compatible with 60Hz signals as well.

I will say that FLIR endorses v2.4 (released in 2015) as the best firmware candidate for field upgrades. I'm going to revert to that release and stay there if I'm able to obtain the files. Customer Support is helpful once you get past the first tier so patience is key.

**UPDATE**

I successfully reverted back to v2.4 firmware via the field upgrade feature of the TAU2 core and restored full functionality to the FLIR Apollo PRO. Overlay is back and menu is operational again. I'll do a more in-depth post with some pictures and analysis of the signal (curious if I returned to 30 Hz -- assuming so).

This has been an exciting intro to thermal imaging. Definitely some tense moments, but I made some new friends and learned some important lessons:

Don't tamper with the firmware on your thermal core without a plan, folks! 

[Fraser]

Manic515,

Excellent news 

Did FLIR customer support help you in the end ?

Fraser

[manic515]

I had an interesting exchange with FLIR. Opened a question and asked for the latest firmware update for my TAU2 core and the older v2.4 firmware I had previously. Can't be downloaded by registering with the site -- you have to request core firmware through Customer Support. The first tier wasn't very helpful: the representative told me my core didn't support field upgrading and they didn't typically provide core firmware to customers, anyway. I cut and pasted the relevant section of the TAU2 product spec document pertaining to the field upgrade feature and just sat back and waited for the question to be escalated.

Got a much more knowledgeable representative in tier 2 support who apologized for the confusion and acknowledged the field upgrade capability of the TAU2. He advised against field upgrading to the latest version of the firmware (v2. citing instability and a higher upgrade failure rate than the v2.4 firmware. So I asked for just the v2.4 firmware and was able to download it. The v2.4 firmware flashed successfully over the v2.8 firmware (a downgrade but there's no version check performed by the GUI software) and restored full functionality to the FLIR Apollo PRO clip-on. Got the overlay and menu back and everything works as it should. Completely disassembled it to clean the LED display and ALL optical surfaces. Reassembled it and I'm about to begin field testing.

The signal output is still 60 Hz. Thought I had a 30 Hz model, but maybe not. The only question the FLIR representative asked was whether I had the slow or fast video version of the TAU2 so I guess it's just those two versions of firmware. As most here probably know, attempting to upgrade a slow video unit to a fast video unit locks it and requires a factory reset. Probably an ITAR feature. My background is in night vision, but it's illegal to export certain image intensifiers from the United States due to ITAR.

So, in conclusion, I would advise EXTREME caution before attempting any modification of firmware on a TAU2 core. The only reason I was able to correct my mistake was the field upgrade capability which not all TAU2 cores have and my own VPC module. FLIR endorses v2.4 of the firmware and will practically beg you not to update beyond that. Good advice. I still believe the v2.8 firmware posted on the forum here is legitimate. All the file nomenclature matched the v2.4 firmware I downloaded direct from FLIR. The v2.8 firmware just carries a significant risk of instability and/or failure compared to earlier releases. I would not advise upgrading unless you're certain your TAU2 can downgrade if necessary AND you have a hard copy of v2.4 firmware as a backup.

FLIR came through for me, but there was a lot of luck involved. All I asked them for was a copy of the TAU2 core firmware as a backup for troubleshooting. I figured out the rest thanks to the great advice (and no harsh judgement) I received here in the forum. I didn't dare ask FLIR for advice about my actual predicament. They would have probably told me I got what I deserved for slapping a bunch of random components together from eBay and then installing an unsolicited upgrade from a web forum. I would have agreed. 

Anyway, a couple pics. Output from the FLIR Apollo PRO MR 640 (60Hz) to a field monitor and the clip-on itself.

[Fraser]

Thank you for the excellent insight into the firmware versions and potential issues with Version 2.8 

Fraser

[hap2001]

So the FW in the forum is actually 2.8? And 2.7 is not the last version?
I saw no version info after upgrade, and the official document only record up to 2.7.

[manic515]

That's part of the mystery. The official Tau 2 v2.4 firmware from FLIR was a zip file called "Tau 2.4 fast Field upgrade.zip" -- probably safe to assume the firmware version number is included in the files they release IF what's on the forum is a FLIR release. My understanding was also that v2.7 was the latest firmware for the Tau 2, as well. I referenced v2.7 as the latest firmware in my chat with the FLIR reps and they didn't correct me. They were very helpful once I got past the script-reading gatekeeper tier. The rep advised me to stick with v2.4 which is definitely what I should have done

Thankfully, I had a newer TAU 2 and was able to revert back to v2.4 via VPC module/software.

So...when I installed the v2.8 firmware, it worked with my TAU 2 but was not compatible with the hardware of my FLIR clip-on. I had an image with no issues, but lost all the onboard menus and overlay. I also failed to back up the original firmware (v2.4) on my TAU 2 and thus was unable to restore full functionality without seeking help from FLIR. I told them I just needed an archival copy of my original firmware to take advantage of the field upgrade feature of my TAU 2 to troubleshoot any um...hypothetical...firmware issues I didn't really deserve help for the big mistakes I made, after all. What's cool is that I was able to register my TAU 2 for a two year warranty.

In retrospect, I wonder if the v2.8 firmware might have been a leaked beta of some sort. I don't believe it's a hack (Fraser may feel otherwise) unless the hacker wanted the firmware release to look as close to an official FLIR release as possible. Usually they take credit somewhere for a "custom" firmware or software release. The documentation with it seemed legit and the instructions worked. The v2.8 firmware installed without incident and didn't affect the behavior of the Tau 2 core except when I installed it back in the clip-on body. My issues with it could have been hardware compatibility between the clip-on and the TAU 2. Had no way to upgrade the firmware on the clip-on and perhaps there was a version conflict as a result. Don't know. Suspiciously, I couldn't pull the official version number from the GUI software info section. The GUI software would display v2.4 as the firmware version I was running when I had v2.4 firmware installed. Again, beta or hack? I don't know. 

In conclusion (and I think Fraser and I agree here), the v2.8 firmware can potentially cause trouble. There's np way to verify its legitimacy and FLIR told me don't upgrade a Tau 2 past v2.4 anyway. I'm glad it's posted on the forum as a resource for people seeking to test the limits of their hardware and perhaps further our knowledge of TAU 2 cores, but exercise EXTREME caution with it. Have a contingency plan. Check to make sure whatever device you may be considering "upgrading" runs at 60 Hz. Back up ALL current firmware and be sure you have a means to restore it.

Short answer...I'd leave it alone

[Fraser]

I looked at firmware 2.8 and, if memory serves, it looked to be an firmware update for a Maritime FLIR camera !

The TAU and TAU 2 are used in many FLIR products, including their Maritime Pan Tilt heads for boats. A TAU core that is part of a system may use slightly tweaked firmware and so not be compatible with other systems that also use the TAU core. I suspect that this is the case here, but have no proof. Whilst Retail TAU and TAU 2 are standard cores, the versions used in other products can differ in resolution and functionality. A “TAU2” version 2.8 firmware update for a Maritime use camera is not something I would want to install on a standard TAU 2 as the differences are not documented. I believe the Version 2.8 firmware is likely genuine, just not intended to be used on anything but the intended platform. The TAU core in the FLIR HS scopes had a firmware patch issued by FLIR to prevent it crashing….that is the same sort of situation and I would not install that patch on a standard stand alone TAU core.

Fraser

[Fraser]

As a side note on core version differences in FLIR cameras…….

I recently purchased a FLIR CCTV camera that I suspected contained a nice FLIR BOSON core with 60fps frame rate and and 18mm lens. A really nice combination 

Upon opening the CCTV camera, I found the expected FLIR BOSON core and it was marked up as 60fps. Whilst I was very pleased with this news, I also noted that the BOSON has a non standard part number. Whilst the part number appears to be the standard number and letter format found in the cores integration documentation, it has one important difference….. the first digit. My BOSON is labelled as a 30320A version. The standard retail BOSONs are the 20320A and 21320A (PLUS version). Now whilst this may just be an indicator to show that the BOSON was part of a “system” rather than a stand alone core, it may also indicate a different firmware package used with the host IP camera electronics package. I have not carried out any tests on the BOSON to discover any differences to the standard 20320A

When FLIR decide to integrate a core into a larger system, they can modify the cores firmware to better suit that systems needs. The same is true when they provide custom configured cores for 3rd party product manufacturers. Hence why there is an area at the end of the cores part number for custom configuration codes.

Fraser

[manic515]

Excellent insight -- I hadn't considered that FLIR could subtly modify firmware to suit the needs of specific models/products but that, of course, makes perfect sense. In fact, the FLIR reps I dealt with were curious about what specific firmware features I was looking for in an upgrade. There are great cores out there hiding in random devices.