Velocity gate in STT appears to be too high

[KenobiOrder]

After the update the gave PD-STT a velocity gate, the F-14 was very good at tracking targets near the beam. Perhaps too good, since 
I could not actually notch it. I reported this issue and @Naquaiisaid he would look at it. 

However it now seems that the velocity gate of the PD-STT mode is excessively large. When tracking targets, It always breaks lock at 100 knots. This is an extremely wide velocity gate for an STT lock. I do not have a source for the tomcats actual STT velocity gate, but this does seem implausible given both the available literature on tradar tracking gates and other jets in the game. The Mirage 2000 for example, which is also HPRF only, has a gate that is +-25 knots or so. Based on the values given in "Radar Guidance for Tactical Missiles", and Skolnik Radar Handbooik, frequency lock loops range between 500hz and 2000hz. Or a potental range of 14-57knots of doppler. On other words, plausible velocity gates are in the department of +-7 or +-28 knots. 

Is is possible the MLC filter was accidently reimposed on the STT lock instead of the vel gate? Because 100 knots or so is around the MLC filter IIRC. 

[draconus]

http://www.heatblur.se/F-14Manual/general.html#pulse-doppler-mode

Quote

The second filter, and second blind spot, of the radar is the zero doppler filter. This blind area is centered around a closure rate of negative own groundspeed, meaning a target moving away from own aircraft at the same speed as own aircraft. This blind area is a hardware limitation as it is a doppler radar mode it cannot detect targets without a doppler shift. The resulting blind area is 200 knots wide, meaning that a chased target moving at a speed of within 100 knots (+/-) of own groundspeed will be invisible to the radar. This means that when chasing a fleeing target it may very well be necessary to use the pulse modes instead.

http://www.heatblur.se/F-14Manual/general.html#pulse-doppler-single-target-track-pd-stt

Quote

The pulse doppler STT works and looks much like the pulse STT mode. It does however have the same advantages and disadvantages compared to pulse STT as the other pulse doppler modes compared to the pulse modes. This means that while much better at tracking a target close to the ground it is however vulnerable to notching.

[KenobiOrder]

28 minutes ago, draconus said:

http://www.heatblur.se/F-14Manual/general.html#pulse-doppler-mode

http://www.heatblur.se/F-14Manual/general.html#pulse-doppler-single-target-track-pd-stt

I understand the difference between the modes and how they generally work. This issue is regarding the STT velocity gate, not the MLC filter. There is no MLC filter in PD-STT. 

[draconus]

But there is still doppler filter:

Quote

This blind area is a hardware limitation as it is a doppler radar mode it cannot detect targets without a doppler shift.

 

[DSplayer]

Just for clarity sake, you talking about the Zero Doppler Filter right? I don't know of any other velocity gate/filter that effects tracking in PD-STT.

Edited February 10, 2023 by DSplayer

[KenobiOrder]

40 minutes ago, DSplayer said:

Just for clarity sake, you talking about the Zero Doppler Filter right? I don't know of any other velocity gate/filter that effects tracking in PD-STT.

 

No this refers to the tracking gate. When a STT lock is formed, there is no broad MLC filter. The radar forms a doppler gate on either side of the target, only a few knots wide. With phase lock loops, it can be even narrower. The only way the lock will be lost is if the target actually flies into the MLC clutter itself, or into a notch or nuller that is placed over the MLC Clutter. The MLC Clutter width will depend on the azimuth look angle and ground speed of the radar. The closer to the boresight, the narrower the clutter. At close to boresight azimuths, the MLC clutter will be extremly narrow, potentially single digit knots wide. At large look angles with typical beam widths of around 2-3 degrees, your talking 30-60 knots or so. If the target approaches the MLC clutter, it can be avoided by using either some kind of dynamic gain control or a notch placed over the clutter to blank it out entirely. In either case, if the target flies into the MLC, the radar will either hang onto it with track memory so that it finds it again on the other side of the MLC, or potentially even track it through the MLC if the signal is strong enough. 

What I was seeing earlier to day was that all locks would break around 100 knots of closure or so, even when the target is off the nose. 

[Naquaii]

I think you're confusing some concepts here.

The rate gate used to track targets in PD-STT follows the tracked target in rate and is not a fixed filter like the MLC or zero doppler filter.

You're right in that the MLC is not used for PD-STT but the zero doppler filter is. The zero doppler filter can't be turned on or off and is always a factor in a PD mode as doppler is what's measured.

The zero doppler filter will always be present and will always be a factor when the closure rate is between -100 to 100 knots, i.e. the filter is centered around 0 closure rate and is 200 knots wide as returns here have no doppler and is treated as either transmitter to receiver leakage or altitude returns.

This means that when you're chasing targets that match your speed more or less and are cold you can't use PD-STT.

 

[KenobiOrder]

3 minutes ago, Naquaii said:

The zero doppler filter will always be present and will always be a factor when the closure rate is between -100 to 100 knots, i.e. the filter is centered around 0 closure rate and is 200 knots wide as returns here have no doppler and is treated as either transmitter to receiver leakage or altitude returns.

It probably should not be present in PD-STT. There is no need for the zero doppler filter in the same manner in which a radar uses this in search. In search this has to be wider because you dont have a target yet and want to avoid random clutter. In STT the tracking gate does this for you. Simply by tracking the doppler of the target, you automatically avoid the ground clutter  centered on the radars ground speed. There is no need for a artificial +-100 knot filter. There would still be either gain control to increase the threshold of detection if the target got close to the MLC, or a null that would by dynamically adjusted based on an estimate of the MLC size. 

Like I said I dont have an explicit source for the tomcat spefically, but I have never heard of a tracking date that would work like that. It would basically defeat the purpose of one of the big advantages of single target tracking. That being that instead of having to place a large filter over a range of velocities centered on the zero doppler, you can now track the target righ up to the actual clutter without losing it. 

 

 

[Naquaii]

3 minutes ago, KenobiOrder said:

It probably should not be present in PD-STT. There is no need for the zero doppler filter in the same manner in which a radar uses this in search. In search this has to be wider because you dont have a target yet and want to avoid random clutter. In STT the tracking gate does this for you. Simply by tracking the doppler of the target, you automatically avoid the ground clutter  centered on the radars ground speed. There is no need for a artificial +-100 knot filter. There would still be either gain control to increase the threshold of detection if the target got close to the MLC, or a null that would by dynamically adjusted based on an estimate of the MLC size. 

Like I said I dont have an explicit source for the tomcat spefically, but I have never heard of a tracking date that would work like that. It would basically defeat the purpose of one of the big advantages of single target tracking. That being that instead of having to place a large filter over a range of velocities centered on the zero doppler, you can now track the target righ up to the actual clutter without losing it. 

 

 

You misunderstand, it's not artificial, it's just how the hardware works. There is absolutely a need for this as otherwise you wouldn't be able to use it at all at lower altitudes and it wouldn't work well higher up either. And a doppler radar can't function when there's no doppler, it's kinda in the name.

The altitude return is from the sidelobes of the antenna and that means that there will nearly always be a strong signal present from the ground directly below you. This is an always present issue with airborne doppler radars.

Edited February 10, 2023 by Naquaii

[KenobiOrder]

15 minutes ago, Naquaii said:

You misunderstand, it's not artificial, it's just how the hardware works. There is absolutely a need for this as otherwise you wouldn't be able to use it at all at lower altitudes and it wouldn't work well higher up either. And a doppler radar can't function when there's no doppler, it's kinda in the name.

The altitude return is from the sidelobes of the antenna and that means that there will nearly always be a strong signal present from the ground directly below you. This is an always present issue with airborne doppler radars.

 

That is not what I am saying. A am not saying the radar does not need doppler to track. 

So you have a target with a closer of 700 knots. The ground speed of the radar is 400 knots. By place a tracking gate centered on the targets doppler, you automatically filter out any unwanted returns from unwanted doppler frequencies. Including main lobe clutter/zero doppler etc. 

I am not arguing that the radar stops tracking doppler. 

[cheezit]

37 minutes ago, Naquaii said:

And a doppler radar can't function when there's no doppler, it's kinda in the name.

 

 

 

Electronically and from a signals-and-systems perspective, this makes zero sense; I yearn for the schematic (which of course will never materialize) to figure out what's actually going on here.

 

In brief: there is no reason for the component of the received signal that lives at the frequency that corresponds to zero closure to have an amplitude of zero; in fact, given that this is by definition the same frequency as you were just transmitting at, the gain is probably highest at that frequency (although of course the Bode gain plot will be pretty flat on either side of that frequency for a good ways).  As evidence of this for those without any signal processing background, there are tons of doppler radars (and other doppler sensors, eg. ultrasound where I have some experience) where a doppler shift of zero is not special and certainly does not represent a blind spot for the sensor.  A doppler radar is not inherently blind at a doppler shift value that indicates a closure rate of zero without a band-reject filter centered at (or otherwise inclusive of) that value.

 

If this is truly a hardware limitation that cannot be turned off, it's because Hughes designed the thing with a band-reject filter that filters out the frequencies that correspond to +/- 100kts closure (the ZDF) and did not design in a way to bypass it some modes or in any mode, not because of some intrinsic or inherent limitation of doppler radars.

 

I can draw diagrams if necessary; hell, the MIT "coffee can radar" can pretty readily be adapted to work as a doppler radar and show objects with zero closure on your laptop for a live demonstration.

[Naquaii]

21 minutes ago, KenobiOrder said:

That is not what I am saying. A am not saying the radar does not need doppler to track. 

So you have a target with a closer of 700 knots. The ground speed of the radar is 400 knots. By place a tracking gate centered on the targets doppler, you automatically filter out any unwanted returns from unwanted doppler frequencies. Including main lobe clutter/zero doppler etc. 

I am not arguing that the radar stops tracking doppler. 

No, you're actually not automatically filtering out all the unwanted returns by doing that, you're just filtering out everything but what's withing the rate gates. That's the key difference.

The MLC region doesn't matter as much as there are multiple factors that have to coincide for the MLC to affect the STT but the zero doppler returns from the ground and the leakage from the transmitter will always be there.

If I can't convince you all I can say is that I can promise you we are not guessing in any way in regards to that the zero doppler filter is not optional, it's a set filter part of the radar hardware and we know that is a fact.

Pulse doppler radars have always had to contend with these factors, it's not until recently that techniques that can bypass these have become common.

The term to google for is altitude return line or sidelobe returns apart from pulse doppler radar if you want to learn more about this.

[Naquaii]

26 minutes ago, cheezit said:

 

Electronically and from a signals-and-systems perspective, this makes zero sense; I yearn for the schematic (which of course will never materialize) to figure out what's actually going on here.

 

In brief: there is no reason for the component of the received signal that lives at the frequency that corresponds to zero closure to have an amplitude of zero; in fact, given that this is by definition the same frequency as you were just transmitting at, the gain is probably highest at that frequency (although of course the Bode gain plot will be pretty flat on either side of that frequency for a good ways).  As evidence of this for those without any signal processing background, there are tons of doppler radars (and other doppler sensors, eg. ultrasound where I have some experience) where a doppler shift of zero is not special and certainly does not represent a blind spot for the sensor.  A doppler radar is not inherently blind at a doppler shift value that indicates a closure rate of zero without a band-reject filter centered at (or otherwise inclusive of) that value.

 

If this is truly a hardware limitation that cannot be turned off, it's because Hughes designed the thing with a band-reject filter that filters out the frequencies that correspond to +/- 100kts closure (the ZDF) and did not design in a way to bypass it some modes or in any mode, not because of some intrinsic or inherent limitation of doppler radars.

 

I can draw diagrams if necessary; hell, the MIT "coffee can radar" can pretty readily be adapted to work as a doppler radar and show objects with zero closure on your laptop for a live demonstration.

You're not wrong in that there are application where you can use low doppler returns or non-doppler returns as data but for pulse-doppler radar applications you really can't. Especially not older non-contemporary systems. I hadn't thought I needed to specify that I meant in pulse doppler radars for target tracking purposes as this is what we're discussing here.

The issue is that the very reasons for the existance of pulse doppler radars are also the reasons why you can't use it when there are none or little doppler on the received signal. There is a reason for a radar like the AWG-9 having both pulse and pulse-doppler modes, they complement each other but the radar operator has to choose which mode to use.

For more moderns systems the controlling computer can do this automatically for the operator or use other newer modes and techniques to bypass these issues.

Edited February 10, 2023 by Naquaii

[GGTharos]

58 minutes ago, KenobiOrder said:

That is not what I am saying. A am not saying the radar does not need doppler to track. 

So you have a target with a closer of 700 knots. The ground speed of the radar is 400 knots. By place a tracking gate centered on the targets doppler, you automatically filter out any unwanted returns from unwanted doppler frequencies. Including main lobe clutter/zero doppler etc. 

I am not arguing that the radar stops tracking doppler. 

You can also stop arguing that the 100kts is too wide of a notch.  Guess what the main doppler width of the eagle is.

[Spurts]

In a PD radar a radar return is processed into two factors, Range Gate and Doppler Gate.  A Range Gate return that has no Doppler Gate component is automatically filtered out as "ground clutter".  This is what allows look-down shoot-down to work at all, because otherwise the Range gate with the strongest return will almost always be the ground.  Pure Doppler modes need no Range Gate and ONLY measure closing velocity and cannot be used for targeting purposes.  Pulse modes only use Range Gate and cannot be used against aircraft near the ground (the ratio of slant range to aircraft and slant range to ground needs to be enough that the notionally smaller return from the aircraft becomes the strongest signal received and by enough margin to pass the SNR Clutter filters)

[KenobiOrder]

21 minutes ago, Naquaii said:

No, you're actually not automatically filtering out all the unwanted returns by doing that, you're just filtering out everything but what's withing the rate gates. That's the key difference.

The MLC region doesn't matter as much as there are multiple factors that have to coincide for the MLC to affect the STT but the zero doppler returns from the ground and the leakage from the transmitter will always be there.

If I can't convince you all I can say is that I can promise you we are not guessing in any way in regards to that the zero doppler filter is not optional, it's a set filter part of the radar hardware and we know that is a fact.

Pulse doppler radars have always had to contend with these factors, it's not until recently that techniques that can bypass these have become common.

The term to google for is altitude return line or sidelobe returns apart from pulse doppler radar if you want to learn more about this.

I understand what zero Doppler and the altitude return are. But there is no reason for a blanket 200 knot wide filter over it in STT. 

The altitude return around zero Doppler is significantly weaker than say the mlc, which is unlikely to be overcome at any practical range. However the altitude return is outside the main lobe, and it's strength is further attenuated by it's distance from the aircraft due to altitude as well as the relationship to target range.

With the tracking gate centered on the extremely narrow band of Doppler the target is in, only clutter inside the zero Doppler range that precisely corresponds to the tracked target can compete with it. You gain nothing by placing a automatic filter over this region because in many cases the radar will be able to complete with this clutter. 

The situation in search is entirely different due to the wide variance of ranges, signal strengths, and lack of knowledge where in the Doppler space the target will be.

 

 

2 minutes ago, GGTharos said:

You can also stop arguing that the 100kts is too wide of a notch.  Guess what the main doppler width of the eagle is.

Far narrower than that. And I'm not guessing.

[GGTharos]

2 minutes ago, KenobiOrder said:

Far narrower than that. And I'm not guessing.

I'm looking in the -34 (I am not guessing) right now.  +/-99kts.

[KenobiOrder]

1 minute ago, GGTharos said:

I'm looking in the -34 (I am not guessing) right now.  +/-99kts.

Nope. 

[GGTharos]

Just now, KenobiOrder said:

Nope. 

Are you saying that -34 has incorrect information?   Doppler gate to 'AUTO' ... +/- 99 kts    You can drop it but you will be paying the piper

[KenobiOrder]

Just now, GGTharos said:

Are you saying that -34 has incorrect information?   Doppler gate to 'AUTO' ... +/- 99 kts    You can drop it but you will be paying the piper

Nah I'm saying I have the same data and your selectively picking which setting you want to go with. And in any case, those values are not used in track. In track it's overridden and dynamically adjusted. Tracking loops do not use gargantuan mlc filters.

[Naquaii]

33 minutes ago, Spurts said:

In a PD radar a radar return is processed into two factors, Range Gate and Doppler Gate.  A Range Gate return that has no Doppler Gate component is automatically filtered out as "ground clutter".  This is what allows look-down shoot-down to work at all, because otherwise the Range gate with the strongest return will almost always be the ground.  Pure Doppler modes need no Range Gate and ONLY measure closing velocity and cannot be used for targeting purposes.  Pulse modes only use Range Gate and cannot be used against aircraft near the ground (the ratio of slant range to aircraft and slant range to ground needs to be enough that the notionally smaller return from the aircraft becomes the strongest signal received and by enough margin to pass the SNR Clutter filters)

Except that in the AWG-9 PD-STT only tracks the target in angle and rate in regards to keeping locked onto the target. Range is measured intermittently and added to the track file but not used to filter out returns in the radar itself.

32 minutes ago, KenobiOrder said:

I understand what zero Doppler and the altitude return are. But there is no reason for a blanket 200 knot wide filter over it in STT. 

The altitude return around zero Doppler is significantly weaker than say the mlc, which is unlikely to be overcome at any practical range. However the altitude return is outside the main lobe, and it's strength is further attenuated by it's distance from the aircraft due to altitude as well as the relationship to target range.

With the tracking gate centered on the extremely narrow band of Doppler the target is in, only clutter inside the zero Doppler range that precisely corresponds to the tracked target can compete with it. You gain nothing by placing a automatic filter over this region because in many cases the radar will be able to complete with this clutter. 

The situation in search is entirely different due to the wide variance of ranges, signal strengths, and lack of knowledge where in the Doppler space the target will be.

 

 

Far narrower than that. And I'm not guessing.

It's funny as the engineers that designed the AWG-9 saw a reason for it. The sidelobes on the AWG-9 aren't really comparable to newer systems and are not as easy to just do away with and you also have to remember that the comparable RCS of the ground that's causing the zero doppler ground return is comparably massive to a target in the air.

Edited February 10, 2023 by Naquaii

[H7142]

57 minutes ago, Naquaii said:

Except that in the AWG-9 PD-STT only tracks the target in angle and rate in regards to keeping locked onto the target. Range is measured intermittently and added to the track file but not used to filter out returns in the radar itself.

It's funny as the engineers that designed the AWG-9 saw a reason for it. The sidelobes on the AWG-9 aren't really comparable to newer systems and are not as easy to just do away with and you also have to remember that the comparable RCS of the ground that's causing the zero doppler ground return is comparably massive to a target in the air.

 

It makes no sense that it doesn't track in doppler with a velocity gate.  The AIM-7 can do this there is no reason the AWG-9 doesn't as well its a PD radar it is getting a good doppler return.  Search and track are not the same thing the MLC filter makes sense for search not for track where your Vgate is going to reject anything outside of it.  The way you had it before made a lot more sense then it does now.

[Naquaii]

10 minutes ago, H7142 said:

It makes no sense that it doesn't track in doppler with a velocity gate.  The AIM-7 can do this there is no reason the AWG-9 doesn't as well its a PD radar it is getting a good doppler return.  Search and track are not the same thing the MLC filter makes sense for search not for track where your Vgate is going to reject anything outside of it.  The way you had it before made a lot more sense then it does now.

I specifically said it tracks in rate. I.e. doppler.

[Spurts]

8 hours ago, Naquaii said:

Except that in the AWG-9 PD-STT only tracks the target in angle and rate in regards to keeping locked onto the target. Range is measured intermittently and added to the track file but not used to filter out returns in the radar itself.

 

 

Gotcha.  Thanks.

[KenobiOrder]

On 2/10/2023 at 9:46 AM, Naquaii said:

Except that in the AWG-9 PD-STT only tracks the target in angle and rate in regards to keeping locked onto the target. Range is measured intermittently and added to the track file but not used to filter out returns in the radar itself.

It's funny as the engineers that designed the AWG-9 saw a reason for it. The sidelobes on the AWG-9 aren't really comparable to newer systems and are not as easy to just do away with and you also have to remember that the comparable RCS of the ground that's causing the zero doppler ground return is comparably massive to a target in the air.

 

What makes you certain of that? Do you have a document stating explicitly that the PD STT has a 200 knot zero doppler filter and the amount of gain specified? These filters generally work by samling the altidue return and associated side lobe clutter and then placing a threshold on it. However these do not have infinite gain. There would almost certainly be a threshold where the target would be visable. The altitude return itself would be the strongest return. SLC on either side of it would be MUCH weaker and at some point, the AWG-9 should be able to see targets in that region as the target in the main beam and the SLC is not. All HPRF waveforms are not as good as MPRF of course, but its not impossible for them to track low closing rate targets, especially in STT where the Signal to Clutter should be far better than in general search. The point being, at some range the AWG9 should be able to track a target with a low closing rate. Maybe not in the altitude return itself, but that would be far narrower than the 200 knot range currently modeled. 

No the AWG-9 is not a newer system, but part of the reason I am not convinced by this model is that even older literature  or literature writtin about older systems suggests this sort of functionality. For example, I spent some additional time testing the PD STT mode and it does not even have track memory. Even the APG-59 on the Phantom J had track memory in STT (this is according to its manual). So as it stand right now, the AWG-9 is useless even agaist at target that does a 180 turn, even if I slow or increase closure dramatically to try to keep track because the radar cannot hold the target for the few seconds it resides in either the MLC or ZCF. 

I am just asking you to take a look at your documention again because it very hard to believe the designers of this radar completely blanked the clutter so that the PD STT mode could not even attempt to compete with it. 

Edited February 13, 2023 by KenobiOrder