frustration on TMA
 

I have some frustration on TMA
"Redbook" says the dual sensor tracking should represent the truth, but looks like it doesn't work in DW
https://i.imgur.com/GsrRkB3.jpg
I am tracking an oiler ship, I got it on sphere passive and TA. I try to find the place where the two line connect but it doesn't work. After 20 min 's frustration on where to put the ruler, I turned the truth on. Mark the accurate location, and this is a TMA solution to reflect the ship's accurate location.

I don't understand why there is a huge difference from the two dots on the top. Did I do something wrong?

Thank you

You probably turned or changed speed. :up:
That will throw off the towed array data.

Don't think of TMA like its a radar, TMA works over time and is, at its best, an estimate.

In addition, the two lines from each sensor must come in at the same time, otherwise it will not depict an accurate intercept. Also, the towed is known to produce data with a noticeable error at times. However, I have yet to figure out the conditions for this to happen. With the slim angle between the two sensors' data, any small error will have large implications for where the intersection lies. In my experience, sensors like the spherical/cylindrical or conformal don't get this noise in their data. Traditional TMA techniques are by far the more reliable method than the intersection technique. Intersection can be good for a ballpark range, but really just for contacts that are rather close where the large angle between sensors' data will far outweigh the towed bearing error.


For clarification, I was not talking about the bad towed data you get when the towed is moving around from a depth or course change. That is also something that invalidates the intersection range method. Sometimes even when the towed is in steady state, I'll get noisy data. Typically for faint, distant contacts.

Hi ET2SN and FPSchazly. Thank you for your answer
I reload my save. I can confirm there was no course change. I wish I can upload the save file here.
the save was recorded on 12:10, there was a depth change on 12:07. I am not sure if that causes the error.

I can image that in RL and in game, the TA's data will have error. But for this case, it is the sphere passive generated the most error. If you can take a look at that TMA again, you will see there is gap between the head of the ruler and the first white line. And the dot leanning to the right side is the data from sphere pass.
I can take a guess on what's going on. But I am a newbie, so I will need vetern divers' opinion, my guess the root of cause is : I assigned the cotact via broadband on sphere pass.


My understanding is the sensor in DW will provide an update to TMA once a minute, so the updates come in at different time should not be an issue?
From my save , I assigned contact through sphere pass' BB at 12:04:55. assigned contact through TA's NB at 12:06:05


Is there any good read or viedo to learn this Traditional TMA Techniques?



Thank you

Another a little OT question:
Does LA class have DTA indicator in RA mod?

Yes, watch Chazly's vids on YouTube. :yeah: (blatent FPSChazly plug)
Honestly, Charlie does a really good job of explaining TMA and target tracking. Check his YouTube page for his "basics" vids and mission play-thrus.

You need to think about how that towed array works, its a long cable that streams out of the back of the sub. Since its a cable, it bends when you turn.
DW also tries to model how sound travels underwater and that discussion would take about a week. :doh:

:D


https://www.youtube.com/watch?v=h5ZU582-bP4


I've wanted to come up with a more advanced tutorial but it's tricky. I'm not sure it would be anything more than just going over several tracking scenarios. So much of TMA is just developing a feel for it, once you get some basics down.

Looking at that plot up there again, the towed data is diverging towards the end there (by "end", I mean the arrow of the ruler and the top of the data in the dot stack, as both of those places are where the most recent data show up). I also notice that each sensor has 6 pieces of data but the data don't appear to be for the same track - it doesn't look like the latest towed data and spherical data would intercept anywhere based on the lines of bearing I can see on that plot. Were there any other contacts? You know about mirror contacts, yes? Merging two contacts that aren't the same target can be the hardest error to diagnose when doing TMA. I don't always do this myself, but it would be a good idea to try and develop solutions for contacts independently to see if they indeed line up before merging.

Also, I could be completely wrong, but I think redbook was written for Sub Command, yes? From my experience, that game models no error in sonar data and it doesn't even model the towed giving bad data when it's turning. Both phenomena happen in Dangerous Waters, as mentioned.

Hope this helps!

I've already watched your tutorial video.

I'm still confused how to deal with bad bearing datas given by TA after ownship changes course, which is not mentioned in this video.
I mean I have to know which LOBs are incorrect so that I can ignore them,but how can I know that?

It takes a few more LOBs after your towed has straightened out to see if some data is bad. One assumption you'll make is that the target has not changed solution during your turn. If so, the bad towed data will bump out and look like a little hill. You'll need to try and line it up with when the towed is turning in your sonar, too, to verify it's just from the turn.

One of the things you look for is a contact turning or changing speed, also known in TMA as a Zig. A contact may zig because they reached a way point or a point on their charts where they need to slow down. Or, they may have detected a submarine tracking them. :o

So, zigs are important. The big thing to realize is that your current solution is now garbage and will need to be changed or refined.

One last point to ponder, TMA is always relative to your hull. Try not to put too much faith into a solution, the speed you can assume is only accurate to within one knot.

:Kaleun_Cheers:

Hi all
Thank you for your answers.

After four months of play I finally got some feeling about TMA. Well, TMA still drives me mad. Sometime looking at all the dots in a curved sharp makes me feel very bad. Fortunately I am not a perfectionist, I understand not all dots can be straight and I can live with that. As you guys mentioned, many times the problem is not about TMA ruler, but is about data accuracy.

About the TMA ruler, I still have a hard time to understand how it works. But it doesn't matter, my brain doesn't know but my hand knows how. Just like how my middle school geometry teacher trained us, "You don't understand how I solved the problem? Fine, hand copy my formula 100 times then I will explain why." Asian education.....:03:

I guess after another 4 month I will consider myself an experienced student in TMA :)


Here I have another question about TMA
After entering a firing solution, LOB contacts will change into a solution. Can I change the solution back to LOB contacts? For example, I found out this solution is not correct, or I do not have time to keep tracking a zig-zag fishing boat, so I would rather to have LOB contacts on the map to increase situation awareness?




I usually extend the TA to 1/3, running at 5 kt. I noticed that usually it will take 2 min for TA to be straight after you hear "steady on course xxx” , of course the TA is already in a curved shape before that . So at least the one dot before this and two dots after that are very likely bad data.

Like most other skills, I can tell you all you want about it but until you start actually practicing, it won't really stick. So, I'm glad to hear you're getting "seat time" with it! The best way to improve is to keep doing it.

The ruler is sized to a certain speed. When the dots line up in the ruler, that means the solution works at the speed given. Does that help at all?

Unfortunately, I don't think there is a way. What I end up doing is dropping the track and re-marking it.

This is good advice. The towed array doesn't gain any sensitivity past a certain point (about 1/4 to 1/3 for the TB-23) so keeping it only partially deployed will allow it to become straight more quickly after a turn. The only reason to extend the towed further is if you wish to dip it beneath a layer or through the shadow zone.

Yes, that makes sense. Now I realized it is a simple tools based on vertical angle pairs.

I spent some time study the TMA, Ekelund ranging , I watched your very helpful tutorial again, and a video called “TMA An Accurate Solution”, read “target motion analysis” from SubGuru, “Submarine approach and attack tactics” that is in RA manual folder, and a paper called “Passive ranging and TMA” found online. Now I see the light. I know where my frustration came from:

1, In most of the case a solution will need 3 legs zig-zag. I failed to follow this requirement before. I usually track the target without any turns or just execute a 2 legs turns. You probably won’t need a 3 legs zig-zag if the target’s speed is known and there is a large bearing rate changes. But if the bearing rate change is low, you will need a 3 legs zig-zag.

2, I didn’t know the tricks to straight the dots until recently. For example, I don’t know where I should move the ruler if the dots line are curve to the right, or in case the first three dots looks like in a straight line, but the others are not. A lot of my frustration came from this. Now I know how to deal with it. First makes all dots form a line (in any direction), after that position the ruler to make the first dot fall into the middle line, then turn the ruler until all the dots line up on the middle line. It’s just that simple.
:Kaleun_Applaud:

Yesterday in a game I apply my knowledge, and got a very satisfied result


https://i.imgur.com/kZQnnUu.jpg

I was in a LA Flt I. After captured a "fishing boat" signal, I execute the first Ekelund ranging calculation. I found out the target was at 20nm away. Then I began a "sprint and drift" under the layer to get closer. In the process I was ambushed by a Foxtrot. After the threat was eliminated, I continue the trip. At 10nm away, I did the second Ekelund ranging calculation, it proved that my pervious range estimation was correct. I also ID the target as a Victor I SSN, finalized the solution with an updated course. Then I did a maneuver to get into the Victor's baffles zone. Get above the layer and double checked my solution about 7nm away. Fired two torpedoes at 4.5nm.
Mission accomplished :)

Having more distance between two sensors should reduce range ambiguity I think. In RA the sensor of the TB-29 makes up 2/3 of the full array, so (in theory) it should be extended to at least that length. Like you say though this might not have an effect in DW.

The TB-16's sensor is much shorter in comparison to its cable, which together with its higher resistance to breaking and to own-speed noise make it more useful for red book style high speed TMA once a target is in range as far as I can tell. I have seen it referred to as a tactical array or something similar somewhere.

KungPao, since you have presumably seen FPSChazly's videos (which I also recommend), you could have a look at some of JiveTurkey's videos too, he has a series on TMA. Edit: Here are a couple from a quick search. Apologies if I'm cluttering up the thread I can't find how to remove the youtube embedding:
https://www.youtube.com/watch?v=dsOg0tE7el4 https://www.youtube.com/watch?v=V8d8CY9_6mM