Return Of The Tagarno
HTML-код
- Опубликовано: 27 авг 2024
- The Tagarno microscope is back from Denmark with a new control PCB, will it fix the flicker issue?
Support the EEVblog on:
Locals: locals.com/mem...
Patreon: / eevblog
Odysee: odysee.com/@ee...
EEVblog Web Site: www.eevblog.com
Main Channel: / eevblog
EEVdiscover: / eevdiscover
AliExpress Affiliate: s.click.aliexpr...
Buy anything through that link and Dave gets a commission at no cost to you.
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.co...
T-Shirts: teespring.com/s...
#ElectronicsCreators
UPDATE: I can actually get this to happen with just the shiny reflected PCB soldermask, no need for the die. Although the die works a lot better. And no it's not the auto white balance.
So an exposure issue?
I have no clue, how this camera works, but if there is a brightness control that limits the values the sensor can spit out, this could make sense for me. Maybe it uses auto brightness for whole sensor area and compensates for the average, not taking max values into account. Then the sensor might be able to generate larger numbers and firmware later on might corrupt the signal in a way that breaks it for capture card, but the screen just works somehow. What if you have as dark background as possible and a small super bright object? Again leaning on somehow capturing the raw HDMI data and some expert to look at it, if there is one that can actually understand it.
I think the problem is the mismatch between the lens and the image sensor. Angled light reaches the silicon surface which normally should not be illuminated and causes a malfunction. It is not without reason that the packages of integrated circuits are opaque and the windows of old EPROMs are covered with stickers.
@@piotr86 that would be applicable if overexposure wasn’t only in the center of the frame. Also, if there was scatter from the original projection bouncing off of the CCD/CMOS causing this, then you’d have some representation in the image we’re viewing of that secondary/tertiary reflection.
Given that we see the same problem with a regular PC screen display on our ATEM Mini Pro it seems unlikely to be anything to do with the oddities of the video signal. We only get it on the 4th HDMI input and the PC is displaying white text on a blue background, switching the PC to input 3 fixed it for us but I have not had much time to experiment.
Its odd to me that tagarno was like "the biggest EE influencer is having problems with our product. Should we get him squared away? Nah. Fuck em. Have em ship it to us and then ship it back with a new hdmi cable."
At his expense.
Lack of an IO shield around the HDMI/microUSB3 sums it up. "Doctor, it hurts when I do this". Doctor/Tagarno says. "Then don't do that"
a new meme has born: Dr Tagarno says. "Then don't do that"
At 3:04 in the video. Came here to comment on that. A shield would also mechanically support them somewhat.
Got a laser pointer, Dave? The reflections off of the silicon are somewhat polarized and possibly coherent. If a laser light causes a similar reaction, then I'm wagering on a lower level firmware issue with the image sensor and its firmware due to getting "overloaded" by coherent light.
Tried it, no problem.
@@EEVblog2, the plot thickens! Fun!
@@EEVblog2 What happens when you have the menu up with the problem scene before the lens. The menu covers up part of it. That would determine whether it's the pixels themselves, or something else that triggers. (Also if it doesn't trigger it, move the chip to the edge where the picture shows around the menu).
Tagarno should've put their best engineers on a plane, with a board and another brand new unit straight to Dave's lab. The positive marketing from that would've offset any costs and then some. Bean counters man, bean counters!
At least it shows a human readable magnification info, not some semi-random hex value
HOLY CRAP!!
I didn’t know that Tagarno (named after Tage Arnø, who was the original inventor of their cinema projector optics, their initial/primary type of products) was a mere 6 minute drive from my childhood home?? (I was born in 1978.)
Be interesting for Dave to have a visit to black magic team with the Tagarno and the board that demonstrates the problem. Then hopefully Black magic will have a test rig to see the data flow.
Tagarno may not know a lot with probably bought a chip off the shelf to output the hdmi and don't want to admit it. Also silence is always not good as have worked for corporate and got sent updated board's to fix issue only not. Usually to get more time as when I visited HQ they knew it would not fix my problem just buying time.
What about a strip down of Tagarno so you can find the chip set which you can pass onto black magic. Problem is if the use FPGA then you can only assume they use the standard HDMI code or popular open source hdmi code.
What about other HDMI sources not from camera like Raspberry pi playing retro games as some emulators push the limits of the SOC.
Waiting a full minute for the upside down joke and then it comes out 🤣🤣
It has the magic pattern that disallows you from viewing it :-)
Similar to the pattern used in money to make scanners and copiers reject it :-)
Did you try the different white balance settings?
Yes. There is only one settign which is Power Up and I tried that, problem remains.
@@EEVblog2 You know what you have to do now... scope the HDMI output to check the waveforms... how deep is the rabbit hole.
Blank section 40 to 52 seconds.
Yep, edit goof. Fixed in RUclips editor shortly.
#blackmagic 😉
Thank you Dave, I now consider it conclusively proven :)
If this was composite analog, I'd think the output of the overexposed sensor would cause the waveform to exceed specs and resulting in a loss of sync which hoses the A/D conversion for HDMI.
Every time I go to the Tagarno website to price their products and see if it's something I'd buy I lose the will to live. The annoying never ending scrolling and needing to request a price for each and every item so I can't simply compare and contrast products based on price makes me absolutely not ever want to deal with them.
To me it feels like an overflow issue, when some data do not fit in their memory location and overflow maybe in sync data or something like that. Different hdmi capture and display solutions might explain it. Only if there was a way to capture the raw hdmi data and send it to an expert.
I agree with data overflow due to a missing sync which more commercial decoders calculate out. Dave needs a day out with the tegarno to black magic as they should have the tools. Tagarno may just buy a chip off the shelf whereas black magic roll their own.
Quite strange that they didnt want to send a PCB board as replacement eLAB equipment for EE people should have the required knowlenge to swap a board.-.
If that PCB had some infos that they dont wanted to disclouse after all therse years that would be the only reason I can think of. (They didnt scrubbed the parts # off the chips?)
But sending a pre production unit whith "trade secrets" stuff on it to EEVBlog with his motto... is a dumb move...
Nah it looked like a more anti repair move to me...
But he can simply open it and look at the board after they replace it if he wanted. It's not hiding anything.
Why couldn't Tagarno do the same flicker test before sending it back ?
They probably tried but could not recreate the conditions, Dave did say it only happened with that one chip when it was in close up so it's probably a unique issue or a one off.
@@ianc4901 That makes sense... It must be a one off...
It seems to me that a high intensity light cause some weird effects in the sensor / processing board. I would take a look at the current absorbed by the microscope over time when the problem occurs. You'll probably see that the issue is caused by an overload of the sensor amplifier causing the rail's voltage of the main processor/video converter to drop. Best wishes, Pietro
Few seconds at the start of the video are blank?
Nope. Must have been a RUclips release thing.
yeah I have that too
No issues for me
Is the problem still there if you leave the OSD on? If so, I would bet my money on something before the OSD menu is added.
Not the first rule of troubleshooting ("always check your voltage rails"), but high up on the list is "don't trust anything automatic" - my guess is the auto white balance. When you tilt the board and get the light reflected of the die at an angle, it may upset the algorithm that does the white balance calculation. Try turning the white balance to manual and see if it still happens.
It will be one of the two devices lying about which spec of the HDMI standard they support and yolo'ing it hoping nothing bad happens. This is very common on cheaper HDMI capture chipsets and I assume it is on cheaper HDMI encoders. My money is on the Tagarno as they seem like a bit rubbish as a company. I run my work laptop through one of these cheap capture devices and it does this exact thing for some excel spreadsheets and is totally fine at other times.
Its either that or some kind of magenta paradox.
To me it looks like it works great, maybe to well, as the flicker seems to be initiated by the reflectivity / refraction of the ic, which is very bright and rapidly changes due to movement. Turn down the illumination when putting a mirror in front of the light.
I agree, it seems to be misinterpreting the reflections/refraction and struggling to keep the image from inverting to negative. I have no idea how any of that is possible but it only happens with that particular chip under that light with that much magnification, it reminds me of a welders mask struggling to react to fast changing lighting.
Could it be too much light for the sensor ?
@@ianc4901 Maybe filters would make it better..
I get the feeling that the HDMI stream hasn't enough bandwidth for the changes between frames and overflows and causes the flickering.
Isn't HDMI a constant data rate, though? It's not delta.
A quick test would be to see if you could slowly blank out the video with a static overexposed image, or is it only the transitions that does it.
Display stream compression is only a very recent thing needed for 8K video, anything below is just a raw datastream with a constant rate.
I am betting you found the same issue as in the wallpaper that kills phone, you can RUclips it
ruclips.net/video/iXKvwPjCGnY/видео.html
I'd be interested to know if you recorded the targarno video and then played it back into the black magic, would it do it?
If it does it bypassing the blackmagic, then i'd assume it would - as it'd be the Tagarno....
I'm not sure if he's tried it without the blackmagic, just directly into a display or something yet though...
this would have been one of the first things I tried, if it doesn't do it, then perhaps doing what you said may be an option, or passing it through a 3rd party 'capture' device before the black magic may be a work around etc...
Would have to try and test each component individually some how for sure, which I don't think he's done.
The recorded video will be compressed and will not be the same encoding that going over the HDMI cable.
@@backgammonbacon yeah, im just wondering if its some sort of intelligent AI gain control inside the blackmagic which is buggy and getting all messed up with patterns like that
In the previous video he mentioned that he did try this and it was fine. Something in the HDMI data stream is triggering it. Whether it's non-compliance, or compliant with the decoder not being able to handle it, is difficult to tell.
You should do a teardown video of the Tagarno microscope, maybe there is a problem with the hdmi chip.
they shouldn't send you"a board", you should send them "The Board"!!!
Play with the Menu options some more.
Try turning the auto white balance off. Maybe the fast changing colors coming off of the IC are more than the camera can handle.
Dave , seems like only on exstream close up , maybe some electrical or mechanical issue at that point ?
Maybe something inside the mechanism of the lense is pushing against something ?
Try using a HDMI splitter as a buffer, which might just provide enough extra buffering to not trigger this bug.
Dodgy Tagarno time again! Pity that it didn't solve the problem. Maybe you'll have their technician on site... and make an episode of trying to fix the issue? Would be nice.
Please do an HDMI bus capture! My bet is the video signal is sent continuously without interruption (good frame structure), but that the encoded video data violates CEA-861 (bad frame content) during extreme transitions (due to reflections in this case).
the fact that the magnification was in hex, and the firmware was non upgradeable indicates these people are incompetent in software design.
We bought one at my old job, and it was supposed to be USB Class Compliant video, and when it turned out it didn't work on Mac as promised they just shrugged and said they couldn't fix it.
If I'm Tagarno I'll send a new one for Dave, a good review better than 100 grand on advertisement.
Aussie power is upside down
Maybe send the Atem folks photos of the main Pcb and info of the Tagarno Hdmi chipset.......might help them do their own investigation short of talking to Tagarno.
........or send them your Tagarno for a while!
Does this happen zoomed in with the OSD screen up? I know you had it up but you were not zoomed in enough for it to happen.
Dave - when are you gonna upgrade your 7 1/2 digit meter to an 8 1/2, so you can be a true Volt Nut?
He needs to catch up with ZenWizardStudio, AND get a GPS-Rubidium time standard.
I thought the flickering didn't happen when you connected the Tagarno to a monitor so are you sure it's the Tagarno that's at fault?
If it's a non-compliant HDMI stream that doesn't necessarily prevent everything from displaying it.
Maybe something with the firmware or algorithm thinking the image from the camera is glitching or something out of sync and auto adjust itself and messed its output.
Just like my old TV that has a faulty RCA cable from DVD thinks I changed the source input and triggers the OSD.
Ah sorry Dave . Rabbit holes goes on .
Richies and their problems.... 🙃 I'm happy with my cheapo Andonstar.
If you take the black magic capture out of the loop does it still do it? Change the monitor does it still do it?
It's weird the image can cause an error.
Yeah, but then again it causes suddenly alternating electrical, analog signal from sensor to AD/driver chip. That change in level would normally be balanced to stay within certain range but because it's so sudden change (I guess for only part of the image makes difference too) the system fails to sense it properly and compensate.
Problem could also be higher up in processing, but for me it looks like clipping / overdrive on early stages.
box is upside down. all electrons will fall out :)
photons in this case!
Did you try toggling the 'white balance'? maybe it does more that just white balance
The old FFFF to 0000 flop over problem.
congrats, its an easter egg
Could it be a problem with the power supply?
Why not try turning off auto white balance too?
I still think its a video compression issue. Slowmo guys did a video where they deliberately tried to break the compression using glitter.
Its going over a HDMI cable so its only using the HDMI spec compression. It will be one of the two devices lying about which spec of the HDMI standard they support and yolo'ing it hoping nothing bad happens. This is very common on cheaper HDMI capture chipsets and I assume it is on cheaper HDMI encoders.
@@backgammonbacon Internal software compression / buffering etc has nothing to do with the hdmi spec
Does it only happen with the Tagarno and ATEM? I assume you've tried the USB and going direct to a monitor?
Yes, it's not the ATEM. Some monitors handle it better though.
@@EEVblog2 Interesting, could be interesting to hook a signal analyzer up to the HDMI signal and see what it looks like when the weird stuff happens
Hmm would be Nicer if your Head was not in the way of the bloody Menu Screen... And Glad your Micro Camera Still doing it's Frame Twitching..
The zoom display used to be on the left side.
@@EEVblog2 yes, but you unpacked it upside down, so of course it will be on the other side!
It's just a Gremlin, no worries...
whazzup with the audio/video missing 0:39 to 0:51 ?
it is fine for me
@0:37 - 52 it's black screen david? cheers
Shipped the microscope to the other side of the world to replace the board?
I'm thinking of a microscope to help me with board repairs but I can't justify the 6k price tag this has. Can you suggest something reasonable around maybe 1k?
Get this one ruclips.net/video/MBFyGzrClsI/видео.html but get the one with the extending arm base.
White balance issue?
It's the black magic switcher not the TagaRno magnifyer
Nu må i sgu lige ta jer sammen Tagarno og hjælpe manden.
first :-)