Jump to content
secretspy711

Oscilloscope recommendation for video system troubleshooting

Recommended Posts

Hello,

I've been having a problem with my FPV video where, when flying at range (4 or 5 miles), the picture in my goggles looks like it's being smacked by a gremlin on the right side of the screen, and "bounces" left.  When I return and fly closer, the bouncing stops.  This only happens when using my Immersion Uno 2400 receiver, wired to a 5.8GHz relay to my Dominator v2 goggles.  If instead I use a different receiver, or hardwire my goggles, or use my old v1 goggles, the problem seems to go away but I do not consider any of those to be "solutions."  Using the other receiver causes the audio to be terrible (probably due to sideband modulation differences), and I do not like being tethered to my groundstation, and I prefer the image quality of the DomV2 goggles.

I can replicate the problem at home using attenuators on the transmitter and unscrewing or placing my hand on the receiver antenna. The UHF radio (Rangelink) is OFF for these tests at home so it's not external interference.

I am considering buying an oscilloscope to help me diagnose the problem, as I suspect the root of the problem is due to improper video levels or impedance mismatch somewhere.  According to Mr. RC-Cam's page here, I need one that can measure at least 10MHz:  http://www.rc-cam.com/lawmate.htm

Is the SainSmart DSO203 an acceptable tool for this job?  I have some concerns with it based on this video:

 

Edited by secretspy711

Share this post


Link to post
Share on other sites

A basic 10MHz scope is great if all you want to do is check & adjust video levels on a **working** FPV system. For troubleshooting bad video I recommend purchasing something with TV-SYNC features (not provided on basic scopes) and higher bandwidth (>50MHz analog BW). Don't forget you'll need good scope probes too; their bandwidth should be better than the scope. FWIW, I use an analog Tektronix 2245A (100MHz bandwidth, has TV-Sync).

In your installation you have a wireless relay on top of a FPV link. This requires well behaving vTx & vRx choices. That is because wireless video links introduce some video signal distortion and the amount of distortion will vary. Satisfactory performance depends on how much distortion is present and the tolerance to it by the video display.

Just curious:
Is the UNO2400 being used with a Immersion 2.4GHz vTx?
Is the 5.8GHz relay vTx and the vRx module in the Dominators both FatShark brand?

 

Share this post


Link to post
Share on other sites

The man himself!  Thank you for the reply Thomas.  I will see if I can find one with TV sync.  But to answer your questions...

Quote

Is the UNO2400 being used with a Immersion 2.4GHz vTx?

Yes, the vtx is Immersion's 700mW 2.4 unit and has been tested recently with my Immersion PowerMeter, and is outputting very close to 700mW.  I typically make my own antennas for the aircraft side, and I have been using a RHCP airblade that I made which is measuring 1.2 SWR at the frequency I am using.  Looks like this, though that's not the exact one.  I'm using a VAS pepperbox on the ground.

Quote

Is the 5.8GHz relay vTx and the vRx module in the Dominators both FatShark brand?

The relay vtx is an Immersion 25mW 5.8 transmitter with a VAS airscrew on it.  I own 2 different modules and have tried both in the goggles:  an original Fatshark 5.8 module as well as an Immersion raceband module.  Both behave the same.  I'm using a VAS mad mushroom on the goggles.

So as you can see, I have tried to stay with Immersion and Fatshark as much as possible, since they both use NexWave modules, which should theoretically present the best chance for compatibility.  The only components in the link that are not, is the PZ0420H camera and the Vector OSD.  So I would be most interested in hooking up the O-scope to the camera first to see if that's the culprit.  I suspect that some configurations at the groundstation just tend to mask the problem better.

Sander recently suggested adding a PowerBox to the groundstation which has an adjustment pot and apparently a sync restore circuit to clean up distorted pulses, but as they say, "garbage in, garbage out" so I would rather find the source and fix it there.

-Brian

Edited by secretspy711

Share this post


Link to post
Share on other sites

It's great to hear you are not mixing vRx / vTx brands. I agree you need to put a scope on the video and investigate.

I expect that you will find that each part of the video chain has partially contributed to the issue. Most likely you have breached "the last straw" threshold and the display is no longer happy with the reduced video quality. So anything you do to help improve the video will probably eliminate the video tearing issue.

If you haven't removed the Vector OSD and tested without it (directly connect camera to 2.4GHz vTx) then that should be next on the list. If things work great without it then you may be able to reduce/eliminate the tearing by adjusting the Vector's video level settings (it has several) in its OSD menus.

You also need to test for video tearing under a wide range of camera lighting conditions. That is because image content can exaggerate some problems in the raw video signal. For example, some users may have unexplained video smearing/tearing/sync problems when flying in certain directions and/or orientations. This is not always RF related, sometimes it is camera lighting. It is possible to bench test this with a bright flashlight; Besides overly bright or dark images, rapid changes in image brightness should not affect image stability. Likewise, OSD data shown on the screen should be altered, enabled, & disabled, to see that all possible conditions do not affect image stability.

If the problem is a simple video level issue (and not sync distortion) then ordinarily you would just calibrate the video levels. But I can confirm that the 5.8GHz link cannot be adjusted, so this fix is not an option on it. I don't know if your 2.4GHz is adjustable (but check inside the vTx and vRx to see what's in there). It is possible that an external video amp between the 2.4GHz vRx and relay vTx would help, so maybe something like the PowerBox might save you. But indeed, amplified garbage in is just bigger garbage out, so there is a limit to what things can be solved with a video buffer amp.

 

Share this post


Link to post
Share on other sites

My workplace has a hackerspace of sorts and allows employees to check out and take equipment home.  They have an Agilent MSO9104A...  I gotta admit, this thing is slightly intimidating, but if I can figure out a VNA, I think I can figure out how to work this thing.

Share this post


Link to post
Share on other sites

If the Agilent's menus do not have a defined TV-Sync profile then hopefully that is something you can create without too much difficulty. But before hauling the $25K scope to your house I recommend you start your troubleshooting with the OSD. If the problem goes away without it then there is hope that revised OSD settings may help solve it.

FWIW, on my first Vector installation I had unreliable FPV video. The problem was caused by Vector's default video settings. Through experimentation I found some values that made it work much better with my FPV monitor. So playing with its video settings menu might be the big money maker for you too.

I'm not saying that the Vector OSD is to blame. Video troubles like yours are usually the result of  more than one contributor. So I think a good plan would be to find one or two easy things you can cleanup that will eliminate the tearing.

 

Share this post


Link to post
Share on other sites

Well I hauled it to my house anyway...

I should make it clear that the bouncing I am seeing is distinct from what I would call "tearing."  The problem I originally set out to solve is the entire image shifting left by about 1/10 of the image width and going back, over and over again.  Tearing, to me, is when part of the image distorts and produces some waviness at the edges.

Vp-p directly from the camera is a little high, about 1.1 to 1.2V.  Levels out of the Uno were around 950mV, a little low.  When I put a PowerBox in the system, I had to turn up the pot all the way to even get close, at about 900mV, and saw significant tearing.  Turning it down made things worse until the image went black and white, and as expected, Vp-p went down with it.    Levels out of the RMRC receiver were 990mV to 1.1V (this is my go-to receiver right now, it seems to produce the best image when flying at range) but the waveform was not nearly as clean looking.  I have some pictures here.  I played with the settings in the camera a bit, but it seemed to want to stay a little high no matter the settings.  Something else interesting is that when I had the powerbox in the system, I cranked up the brightness in the camera which caused the tearing to stop, but then the bouncing started and levels were at about 870mV.  Mind you, this is without any attenuators, as before... this is full power of the vtx about 20 feet from the receiver.

I didn't try physically bypassing the Vector since it's sorta buried in the plane and a pain to do, but I did toggle the switch to turn off the OSD text.  That didn't seem to make a noticeable difference in the levels, and didn't make a difference to the bouncing when using attenuators to simulate flying at distance with the Uno.  I played with the vector settings a bit, and the only setting that seemed to change things noticeably was setting the black level to 0 which produced some horizontal white bands in the image.  Everything else didn't seem to do much, but it's getting late and I need to play with all these settings more deliberately in order to say that conclusively.

So, lesson learned tonight is that the camera is a little high, the Uno is a little low, the PowerBox makes things worse, and the RMRC receiver has good levels but the waveform is dirty (go figure).

Share this post


Link to post
Share on other sites

Your description of "tearing" is the same as mine. Let's just call your "bouncing" issue poor video stability.

In a perfect world the output of the vRx would be identical in every way to the output of the camera. The intention of the wireless link is to simulate a hard wired copper connection and not alter the video waveform in any way. In practice our FPV video links cause some video signal impairments (varies, depending on quality of the vTx/vRx).

My comments:

1. The video levels from the camera are very good. Full whites (ignoring color info) are at about 1.05Vpp and sync is 285Vpp. Color burst a bit high at 365mV, but that is not a problem in this case.
2. The video out of the RMRC vRx is a wreck; The price you pay when mixing brands (video de-emphasis not compatible).
3. The video out of the UNO has very low sync amplitude (60% of requirement) with poor quality. And the entire waveform is poor/disappointing with reduced bandwidth.
4. The power box is not helping in this case. Garbage in garbage out.

Your measurements are incomplete because you only show one TV-line. You also need to look at the TV-Field. This requires using the TV sync features on the scope. There is a lot of useful info in the TV Field that could tell us more, so get some screenshots of it at the camera and UNO. BTW, if you have two scope probes then you can monitor both camera and vRx at the same time for quick visual comparisons.

Most importantly, you need to bypass the OSD and get it out of the video chain. If removing it does not have a substantial effect then the issues are in the vTx, vRx, or their wiring.
 

Edited by Mr.RC-Cam
Use two channels.
  • Like 1

Share this post


Link to post
Share on other sites

Very helpful!  A high color burst makes some sense, I have the color gain in the camera turned all the way up. 

I'll work on bypassing the Vector tonight and see if that changes anything.

I only know enough to be dangerous...which is a shame because I do realize what a powerful piece of equipment I've got here.   Regarding setting up the TV sync profile, I am not quite sure what to look for on the scope.  Is this a triggering thing?  What should it do differently when it's set up correctly?  From the little bit of reading I've done, it sounds like the higher-end scopes rarely have a dedicated TV sync feature because of the more versatile triggering options available.

I will pick up a 2nd probe from the hackerspace today.

Edited by secretspy711
acknowledging 2 probe edit.

Share this post


Link to post
Share on other sites

Attached is the TV trigger info for your scope (excerpt from manual).
Agilent9000Oscilloscope.pdf

Set it up for TV Field. Instead of the horz sync pulse every 63uS you see in your TV-Line screenshots, the NTSC TV-Field vert sync will be at 16.6mS intervals. The screen shot should show one full field so we can check for problems such as video SAG and whatever else might be harming it.

 

 

  • Like 1

Share this post


Link to post
Share on other sites

Thomas, I've added many more screenshots over at FPVlab, posts 24-26 in various configurations.  Please let me know what you see, I'm not quite sure what I'm supposed to be looking for.

http://fpvlab.com/forums/showthread.php?47983-Troubleshooting-bouncing-image-from-Uno-2400-receiver&p=797225&viewfull=1#post797225

Share this post


Link to post
Share on other sites

EDITED:

Your field waveforms are odd looking to me. I suspect it is because of the digital scope's behavior. The waveform out of your PZ0420 camera was captured on your digital scope as this:

ntsc_waveform2.jpg

 

BTW, I read that you removed the Vector OSD. Did that make the sync stability problem better, worse, or no-change?

 

 

Edited by Mr.RC-Cam
My fault, looks ok.

Share this post


Link to post
Share on other sites

When I was attempting to get the field to show, inside the "Trigger" menu, I found a submenu called "Shortcuts", and inside that was an item called "...a TV or video signal."  In the dialog box that popped up, it had several picks for different video standards.  One of them was "NTSC 525/60."  I picked it but the waveform would constantly scroll to the left.  Another option was "User defined" so I picked it, at which point the waveform stayed put and this is what I used to generate the images, which I see now is probably not correct.  I selected "all lines" in a drop down menu, and I zoomed out to see if the vert sync pulses were at 16.6mS as you said they should be, and they were, and then I took a screengrab.  Notice the shot of mine shown above has a scale of 2 mS/div.  I see that yours remains at 10 uS/div, so apparently I need to stay zoomed in to 10uS/division to get the shots you are looking for?  I think I was almost there, I just need to put it back on NTSC 525/60, figure out how to get the waveform to stop scrolling, and zoom back in to 10uS/div.

I used a servo Y harness to go directly from the camera to the vtx and bypass the vector, while providing power to both using a 3S battery.  I was focused on making oscilloscope measurements and I didn't actually try to replicate the bouncing in this condition by simulating flying at range (installing attenuators, placing my groundstation across the room, and putting my hand on the receiver antenna).  Of course the real test would be flying at 4 miles or so, but there are only so many hours in the day.

Edited by secretspy711

Share this post


Link to post
Share on other sites

Must have been half asleep last night because I posted a line waveform. And I hadn't done your test with a flashlight in years (color bar generator is used instead) so your waveforms took a bit to get used to. Sorry for the detour.

Your camera's field waveform is fine. The UNO's field waveform suffers from reduced bandwidth (but the field images you posted look better than I expected). I was hoping we would have found something that would have given us hope of an easy solution.

Do you use the 5V output on the 2.4Ghz vTx? If so, what is on it?

The OSD removal sounds hopeful. Be sure to check for the bounce problem now and see if it has gone away.

EDIT: Can you swap the 2.4GHz vTx with your 5.8GHz vTx (easy/simple connector access) and check the waveforms? You should find that the 5.8GHz performance (TV-Line waveform) is much better. If it looks just as bad then perhaps there is something wrong in your installation or test setup.

Edited by Mr.RC-Cam
5.8GHz test

Share this post


Link to post
Share on other sites
Quote

Do you use the 5V output on the 2.4Ghz vTx? If so, what is on it?

No.  I've used it in the past on other aircraft but am not currently using it here.

Quote

 Can you swap the 2.4GHz vTx with your 5.8GHz vTx (easy/simple connector access) and check the waveforms? 

I can...  but I'm not quite sure I completely understand what you would like me to do.  If I put a 5.8 vtx on the plane, I can measure the waveform between the camera and the vtx.  Measuring at the groundstation as before would require an Uno5800 which I don't own.  I can attempt to measure it directly on the goggles through the AV-in/out jack, terminated with a 75 ohm resistor, or I could probably rig up a way to connect the goggle module with jumpers instead of directly plugging it in, in order to access vid and gnd between the module and the goggles (the goggles would be the termination load).  Something of note, several weeks ago I measured the goggle's module sockets with a multimeter and found the goggle video input to be 43.6 ohms.  A far cry from the 75 I expected.  But I'm not smart enough to know if that's a valid measurement or not, or if anything can be done about it if that is the problem.

 

Share this post


Link to post
Share on other sites

I don't trust using the goggle's shared Vid-Out for the scope measurements. So if you don't have access to a UNO5800 then the 5.8GHz vTx test is a no-go. So rather than throw more busy work at you, maybe this is the place to stop and review the problem & propose solutions.

I would summarized your situation as this: The video signal presented to the goggles has been impaired by the two wireless video links and possibly the OSD. Each has contributed to the reduced image stability (by lowering the sync's operating margins). The last straw scenario occurs when your RF link signal strength is too low (more noise on top of marginalized syncs causes unstable image).

I expect that nearly anything you do to improve the RF link or video signal will help out. It sounds like you have optimized the 2.4GHz antennas so the most bang for the buck will probably be with a different 2.4GHz vTx/vRx (the biggest offenders seen from your waveforms) and/or some OSD tweaks.

 

Share this post


Link to post
Share on other sites

I bypassed the Vector once more, using the original setup with the Uno and Immersion 25mW repeater.  With attenuators and putting my hand on the antenna I was still able to produce some bouncing.  For sanity's sake I wired my goggles to the Uno, did the same thing, and the picture was much more stable.  Then I tried using a boscam 200mW transmitter as the repeater, and the bouncing was back.  Routing it through the powerbox again did not help.

I found an old 2.4 700mW RMRC transmitter I had laying around and put it on the plane.  Receiving on the Uno (mixing brands), the waveform was much stronger looking (actually quite high at about 1.5 vpp, and the difference between the sync "bucket" and the color burst was much more well-defined.).  Saw the same thing when receiving on the RMRC receiver.

Routing these through the powerbox and adjusting the pot to approximately the middle of it's range, the vpp went from 1.5 to ~1v.  But now the difference between the sync bucket and the color burst was lower, but maybe good enough?

pictures here

The only problem with this one is that it gets quite hot without the bulky blue anodized case, though some are having success with it.  One of my flying buddies uses one of these and drilled 4 holes in the interior (silver) case for airflow.  Not sure if that's a good idea or not... I tried removing the lid of the interior case on a 1.2GHz transmitter recently, and it cause the video to be very distorted.  I think some heatsink fins would be better.

Edited by secretspy711

Share this post


Link to post
Share on other sites

It's a relief that the Vector does not contribute to the bounce. I agree you need to swap out the 2.4GHz.

FWIW, I recommend that you revise the way you measure the video levels on the scope. Do not use its auto measuring function, it must be done manually.

Do not include the color information in the active video region and use the average noise at the sync tips. Do it like my marked up image below, where your original measurement was stated as 1.49Vpp, but is actually closer to 1.25Vpp:

ntsc_waveform4.jpg

 

BTW, The final adjusted level out of the PowerBox appears low (see images in your link). It is about 0.88-0.90Vpp.

Also, rather than use the PowerBox to reduce the video levels I suggest you leave its adjustment at unity gain (1X). Instead reduce the RMC's vTx's level using the video adj pot inside of it (assuming it has one). If that is not possible then calibrate video levels using the PowerBox adj.

Finally, it would be wise to capture some TV field waveforms with the brand mixed vTx/vRx configuration. If there are any significant video emphasis compatibility problems then they can often be identified in the field waveforms.

Edit: I forgot to mention that your sync levels are still quite low. I'm not sure if this brand mixed setup will be the holy grail for you. But worth trying out anyways.

 

Edited by Mr.RC-Cam
sync levels

Share this post


Link to post
Share on other sites

For some reason it won't let me upload pictures to this thread.  Even though the 2 images I want to post are about 170kb each, it says "you are only allowed to upload 501.76kb."  In that case I have posted 2 shots of the TV field from both the Uno and RMRC receivers (both without using the powerbox) using the RMRC transmitter.  I apologize for all the linking away from your forum.

When you say the sync level is still low, you are referring to the difference between: the middle of the lowest flat part (where your horizontal blue line is drawn) to the middle of the color burst, correct?

 

Edited by secretspy711

Share this post


Link to post
Share on other sites
Quote

For some reason it won't let me upload pictures to this thread.  Even though the 2 images I want to post are about 170kb each, it says "you are only allowed to upload 501.76kb."  

Sorry about that. I don't see anything wrong in your member account. As a sanity test, try to upload a small image. The test forum is a good place for that.

 

Quote

When you say the sync level is still low, you are referring to the difference between: the middle of the lowest flat part (where your horizontal blue line is drawn) to the middle of the color burst, correct?

With the video adjusted to 1Vpp (full white level without chroma, to bottom of sync tip) the ideal sync amplitude is approx 0.28Vpp. It is measured from sync tip to the blanking level. See drawing below.

FWIW, the official composite specs are based on units of IRE, where one IRE is 7.2mVp-p. For example, full white video is 140p-p IRE. So 140 x 0.0072 = 1.0Vp-p. Here's how the specs show the NTSC TV line waveform:

ntsc_waveform7.jpg

 

Here's a practical view using a scope.

ntsc_waveform6.jpg

Here the blue to red lines show the overall full-white video measurement (1.0Vpp). The blue to yellow lines are the h-sync amplitude (a shy 240mV in this example). The orange lines are the measured color-burst region.

Edit: More comments.

What is ironic is that the RMRC 2.4GHz vTx seems to have better compatibility with the UNO2400 than the RMRC vRx. The reason I say that is because there is evidence of poorly matched video emphasis between them. The tell-tale sign is this SAG-like hump in the field waveform you posted (per your links):

ntsc_waveform5.jpg

The waveform's bottom edge should be flat, not wavy or humped.

SAG in a traditional hard-wired composite video signal is a sign the video output coupling capacitor is too small. In a wireless link it could be that problem too, but is often due to incompatible modulation methods (video pre-emphasis in vTx does not match de-emphasis in the vRx). This issue occurs often when vTx/vRx brands are mixed, which is why I advise to avoid doing that. But in this case it seems the brand mixing might be helping (or hurting less :) )

 

 

Edited by Mr.RC-Cam
Added another image
  • Like 1

Share this post


Link to post
Share on other sites

Thank you Thomas.  You have been a huge help and I've learned more than I could have hoped for.

It does seem as though the RMRC transmitter with the Uno receiver produces the best sync level (though still short of the 280mV ideal) out of any combination I've tried.

Alternatively, I wonder if I could replace the pot in the powerbox with a different one, to make it "go to 11" ? (spinal tap reference).  The goal would be to make the entire signal, including the sync, taller, and may get me above the threshold needed to maintain stable video.  The Vpp out of my Uno, when measured correctly as you pointed out, seems to be about 840mV.  My friend's Uno is about 920mV, which means the sync portion on his is also slightly taller.  Just an unfortunate luck of the draw I guess, but as you said, not the entire cause of my problems, most of which seems to be the transmitter.

Share this post


Link to post
Share on other sites
Quote

Thank you Thomas.  You have been a huge help and I've learned more than I could have hoped for.

You're welcome, my pleasure to help out.

Quote

Alternatively, I wonder if I could replace the pot in the powerbox with a different one, to make it "go to 11" ? (spinal tap reference).

What video amplitude can you achieve with the stock PowerBox's video gain set higher?

But before you add the PowerBox, any chance you can pull the cover off the vTx and check to see if there is a level pot in it? I have a hunch your Chinese made 2.4GHz vTx has one of those. Might even be accessible through a hole after you remove the pretty anodize case/heatsink. Just be careful, there might be an audio subcarrier or antenna trimmer in there too. Choose wisely!

BTW, it is bad practice to hap-haphazardly crank up the video level beyond 1Vpp in an effort to increase the sync amplitude. The waveform bits are ratio-metric and each piece of the pie is expected to adhere to the standards. Otherwise reliable compatibility with all display devices will be affected. But having said that, sometimes any port in a storm will do; So if you have to use a non-standard video signal to solve your problem then go with it.

 

Share this post


Link to post
Share on other sites

What video amplitude can you achieve with the stock PowerBox's video gain set higher?

Using the Immersion vtx, not high enough.  Setting the power box on its highest setting results in just slightly lower levels than what I get straight out of the Uno, so it's like the power box is only meant for bringing down signals that are too high, not too low.  I do have a Pyle amplifier/splitter I could try, from an old ground station build.  Just seems odd to have to go to all this trouble to get in-brand components to work together, when the reason I went with Immersion is because I thought it would just work.

The Chinese vtx has a pot yes, and it's overall level is a bit high while the sync is low. The Immersion does not have any adjustment.  Ideally I would like to come up with a solution that uses he Immersion transmitter due to owning 4 of them already (with a 5th sunk in a lake) and because the Chinese one gets quite hot if uncased (I've heard they burn up easily) and is quite heavy and bulky with the case on.  I'd also like to think the Immersion is more reliable, but who knows.  I'm getting frustrated with them now, after all this scope testing with the levels still being low.

Share this post


Link to post
Share on other sites

I understand the frustration. To be fair, the IRC FPV gear is working up to five miles even with the added burden of the 5.8GHz goggle relay. So you have to give it some brownie points. And on the plus side, you learned some useful video measuring skills this past week. Unfortunately now you know that a lot of popular analog FPV gear has room for improvement.

If the syncs are low with 1Vpp video levels, and all your IRC 2.4GHz vTx's experience this, then maybe your UNO2400 vRx is defective. Regardless, a buffer amp wouldn't normally be the go-to solution in your situation. But maybe something like the Pyle amp will provide a band-aid fix.

Share this post


Link to post
Share on other sites

The signal out of my Uno is measuring about 840mV.  My friend's Uno is showing about 920mV.  Both syncs are very low, and make up the same percentage of the total signal, but since the entire signal is taller from his Uno, therefore so is his sync.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×