Avidyne TAS605 Installation in a Socata TB20GT
This article describes the installation of the Avidyne TAS605 traffic warning system in a Socata TB20GT.
There are several fairly obvious reasons for wishing to "see" traffic:
1) To reduce the chance of a mid-air collision
The probability of a mid-air is extremely small, but most of them are 100% fatal. The UK has around 2 per year. The number of hours flown by GA pilots in the UK is not known, but taking the known figure of about 20,000 pilots with valid medicals, and the generally believed annual average flying time of about 30 hours, gives us an order of magnitude for this, of about 1 mid-air per 300k hours. Not a good case for spending money on it!
What suprises most people is that, as far as is known, all of the UK mid-airs since WW2 have been in VMC. This is probably because of the far lower traffic density in IMC. Enroute traffic is limited to pilots with instrument qualifications and circuit flying is "officially" not allowed in IMC.
The case for spending money on it gets even worse when one looks at the circumstances of the mid-air collisions that have happened. Most of them are at very low levels, usually below 1500ft. There are some notable exceptions such as the RAF Grob training mid-air which happened at 2900ft, but it appears that most of those are between flights that are not exactly unrelated. There seem to be almost no collisions between unrelated aircraft flying enroute at 3000ft or more.
No transponder-based traffic detection device will work unless the target has a transponder, it is switched on, and for a useful warning (relative altitude included) the transponder has to be Mode C or Mode S. A proportion of GA has no transponder, or the pilot turns it off (due to stupid training, or a misplaced concern for "civil liberties" i.e. he doesn't want to be seen on radar). So a percentage of traffic will never be detected anyway. However, from years of flying under a radar service several things are quite obvious to me, for the UK:
- most reported contacts are never spotted (this is normal
- everybody has that problem)
- of those that are spotted, nearly all are well below me (are estimated to be at 2000ft or below)
- of those that are spotted and reported "altitude unknown" (non-transponding i.e. primary rardar return only, or transponding Mode A only) nearly all are estimated below 1000ft
It would thus appear that the biggest benefit of a traffic warning system is at low altitudes, with the benefit reducing at very low altitudes. An exception to this may be helicopters which routinely fly very low, but most will have a transponder.
2) To improve awareness of traffic in the circuit which one is joining or departing
A lot of mid-airs occur near airfields where one cannot avoid being at a low altitude if arriving or departing.
It's often hard to spot other traffic in or near the circuit. Aircraft are inherently hard to see, and a target on a genuine collision course will be a stationary point in the field of view (for both pilots involved) so doubly hard to spot. Also some pilots do not report their location with any accuracy; sometimes accidentally, and sometimes deliberately to get a joining or landing clearance before somebody else!
A fairly common situation is where e.g. someone reports "downwind" but you (joining crosswind) can't see him. He could be anywhere from several miles upwind (especially if doing a downwind join) or he could be late-downwind. Or he could be right where you are joining... Some people fly massive circuits, extending miles away from the airfield...
Therefore, a traffic detection device should improve one's awareness of circuit traffic - even if it fails to show non-transponding aircraft.
3) To improve awareness of traffic, while one is on the ground
With Avidyne or Garmin systems, you can see transponding traffic (both on the ground and airborne) so e.g. you can see traffic on the final approach before you enter the runway. This gives an extra layer of safety over ATC keeping track of who is where, etc.
4) To offer some protection when flying an instrument procedure in Class G
Many instrument procedures are outside controlled airspace and one often finds traffic flying merrily straight through the procedure track. The platform altitudes are often 1500, 2000 or 2500ft which a lot of people fly at anyway. I've had several near-misses when flying an instrument approach in VMC, but in UK Class G there could be non-radio traffic there in IMC too.
It's clearly much better to first spend money on a ground proximity warning device but I already have that.
Other Traffic Detection Systems
In light GA, no traffic detection system has achieved a "critical mass".
FLARM has raised a lot of hopes, is cheap, but market penetration has been negligible outside the gliding community, and both aircraft must be equipped with it. Functionally it is a good system because it uses GPS and, with each aircraft continuously radiating its 3D GPS position, it is possible to compute mutual trajectories very accurately. Unfortunately there is no internationally approved frequency on which to radiate the stuff, which has prevented certified equipment from emerging, which in turn has prevented anything resembling wide adoption. Also the trajectory algorithm is not in the public domain; it is somebody's "intellectual property". If somebody developed a certified product which combines FLARM with the detection of Mode C targets which displayed the azimuth for Mode C, they would have a winner.
ADS-B does not yet exist in Europe (for light GA) but is being introduced and legally mandated in the USA. Functionally it is similar to FLARM but works on internationally agreed frequencies (on which there are two incompatible systems, just to make life interesting) and is certified. It will be many years before it is mandated in Europe, and in my view it is virtually certain that VFR/OCAS traffic will be exempted. ADS-B data is already being radiated by large aircraft, and the data returned under an Enhanced Mode S interrogation amounts to the same thing, but stupid European (EASA) regulatory practices have prevented European-registered light GA participating in it.
The only system which has a a relatively wide adoption is the old-style TCAS like the one described in this article, which interrogates nearby transponders. These are expensive (usually over €10,000 installed) but they benefit from Mode C or Mode S transponders being legally required in many airspaces. Today, it is difficult to do much "going places" flying in Europe without Mode C or Mode S. There is also a benefit to GA in getting equipped with a transponder: all aircraft used for public transport, over a certain number of seats, are legally required to have TCAS, so if you accidentally end up in the way of something big, they will be able to avoid you. So, for someone contemplating a TCAS system, the "other half" of the system is already in place to a large degree.
TAS or TCAS?
The terminology is confusing.
TAS systems advise you of nearby traffic, visually and, with certified systems, verbally via the aircraft intercom. You don't get advice on what to do to avoid it. The system described in this article is "TAS".
TCAS systems are mandatory on large aircraft and they offer advice on avoidance. The advice (called an "RA" - Resolution Advisory") is mandatory and over-rides ATC instructions and clearances. In fact it over-rides everything except a GPWS (ground proximity) warning With TCAS, the two systems communicate with each other and agree an avoidance strategy, so one pilot is told to climb and the other is told to descend. Standard practice is that all avoidance is done by climb/descent only, with no directional changes.
There are several main contenders: Avidyne TAS600 series, L3 Skywatch, Honeywell KT870, Garmin GTX800 series. The following comparison (produced by Avidyne ) lists them. No UK installer seemed to know anything about the L3 product. Honeywell was widely installed about a decade ago but the company has during that time virtually withdrawn from the GA avionics business so even though the KT870 works well it is difficult to recommend it. The KT870 installations I know about were also billed at around €30k which is 2x higher than the Avidyne option. So it really comes down to Garmin or Avidyne.
A page from the above comparison is here
I made extensive enquiries, in the UK and the USA, regarding Avidyne v. Garmin. There was no obvious performance difference between the two systems, and it was apparent that in both cases a careless installation could result in a poorly working system. In the UK, the Garmin was much less well known among installers, and the box is a lot bigger and heavier than Avidyne's.
The Avidyne unit was chosen.
I chose the TAS605 over the cheaper TAS600 because I was told by the installer that the TAS600 has no heading input and thus its data cannot be overlaid on the moving map on the KMD550 MFD, and that it works only in a dedicated MFD page which has to be north-up always. This turned out to be substantially false. The difference is that the TAS600 accepts "heading" via the GPS data i.e. it treats the GPS track as the heading. In the context of azimuth accuracy in these systems, and fixed-wing aircraft, the difference between the GPS track and the actual aircraft heading is rarely relevant. In a helicopter (which can yaw around without the GPS track changing) the TAS600 would not work well.
Also, one TAS600 owner (TB20 aircraft) reported that its "7nm" range is rather optimistic...
The chosen TAS605 has a reasonable range of "13nm", a 5500-foot vertical separation maximum and a 55,000-foot certified ceiling. It accepts heading data via ARINC 429 which could be obtained from the SN3500 EHSI.
Installer Choice and Costs
Initially I considered doing a DIY design and installation, as I did with the Sandel SN3500 EHSI. Purchasing the equipment in the USA, from discount avionics outlets, and using a freelance installer to install it, would have saved me about 1/3 of the total cost. And certification would not have been an issue because the AML STC covers the basic job and extras like antenna mounting are done via AC43-13-2B (local copy). This was not done for the following reasons:
1) The job would require a hangar, and would require it for about 2 weeks. For me, this was very difficult. I am hangared but I am not allowed to do any maintenance work in the hangar. Now, I know of a hangar I can rent for £20/day where they permit work to be done but getting there is 1hrs' driving each way.
2) The finished installation needs specific test equipment to test the range and azimuth. It turned out that the installer chosen did not have this equipment anyway (despite earlier assurances), and using the ground mode plus a flight test was the only way to see if it actually works!
3) There is a certain amount of experience involved in achieving an installation which works well, and I didn't want to risk it. With a DIY installation there is no warranty. The installation clearly matters; even a small number of enquiries digs out a good percentage of owners who say their system basically doesn't work.
4) Equipment bought from discount avionics outlets in the USA has little useful warranty due to the high cost of sending it back - of the order of $200 each way. Even small items are expensive because virtually all US firms are incapable of using e.g. registered airmail and always send stuff via the most expensive courier option. Buying from the USA is great for equipment for which there is a backup on the shelf at home.
I narrowed it down to two well-known UK installers. Both hold every imaginable FAA 145 and EASA 145 approval and both quoted almost exactly the same price. Both quoted 2 weeks, which was important because I have seen TCAS installs take 6 weeks. One of them was chosen but after a month (and before I took the plane there) they came back saying they wanted another £1447+VAT for "structural substantiation reports and drawings (8110-3s) required for the 2 antenna locations and the unit location" which was outrageous since this extremely basic structural work is done in accordance with (IAW) AC43-13. So they were dropped instantly and the second company was selected. The first company then came back saying they can do it IAW AC43-13 after all, but it was too late
The costs, excluding 20% VAT, were:
TAS605 system (controller, 2 antennae, splitter) £6175
Comant CI-2480-200 GPS/VHF antenna (non WAAS) £525
Installation and sundries £3420
Additional items, which I free-issued to the installer, were:
1) KAC504 ARINC interface card
This (Honeywell P/N 071-00166-0411) plugs into the KMD550 MFD and provides it with the ARINC 429 input capability required by the TAS605 system. This was purchased from South East Aerospace for $1150 (overhauled, with an 8130-3). A new one would have been $2500. There is a fair number of these cards coming onto the used market because a lot of people are removing the 1990s Honeywell/King avionics and installing "glass" cockpits.
2) RG400 cable
This double-shielded solid-PTFE coax is used for the four antenna cables and is the best cable one can reasonably use. Most UK installers don't like to use it because it is expensive, heavy and stiff. It was £6.64/metre and a total of 25m were used. I bought a 50m reel and it's easy to sell the remainder to freelance installers. There are several "RG400" fakes on the market which have a very similar name but use a lighter and cheaper insulation and their attenuation is worse.
A factor in the choice of a very well shielded cable is the apparent higher incidence of burnouts of KFC225 autopilot servos in aircraft which have a TCAS system installed. This sounds bizzare but I now believe that the burnouts are caused by RF interference affecting the KC225 autopilot computer and causing it to drive the servo(s) with a rapid waveform which the servo cannot mechanically follow and which drives it into current limit, which (due to a defective design of the servo) causes the servo to self-destruct.
The operation of a TAS system is simple.
The system uses two antennae, as shown below, to get direction.
The upper one has two antennae within the single "blade" and the relative power received by these two resolves azimuth front/back. The lower one has two antennae which using the same principle resolve left/right. The system does not use propagation delays to determine azimuth. The altitude resolution is done by comparing your pressure altitude (obtained via a connection to your altitude encoder) with the other aircraft's Mode C transponder return. The distance is done by measuring the time delay between the interrogation and the target's response.
To avoid picking up transmissions of one's own transponder, a small "T-box" coupler (shown above) is inserted into the transponder RF cable which picks off the transponder transmission, and this is used to inhibit the TAS system transmitting. A similar inhibiting function is done with the DME but modern DME units have an output signal which is wired directly to the TAS system.
The information can be displayed on a variety of devices. Avidyne offer various options, from dedicated units such as the ATD150
or the MHD300
all the way through their glass-cockpit products such as this
A number of 3rd party displays are also supported such as the Garmin GNS series, and my KMD550 (if it has the KAC504 ARINC card fitted)
This is not stated in the manuals, but there are subtle variations in traffic indication functionality between displays. The TAS605 unit transmits the traffic data to all display units and each of those has its own configuration and can choose to display data according to its own criteria e.g. one might show traffic only 2500ft above and 5000ft below, while another might show traffic 5000ft above and 5000ft below. Of course, it is also possible to disable traffic display on a particular device!
Also, the TAS605 can output the traffic data in RS232 or ARINC429. These differ in both the electrical interface and the protocol. The RS232 data includes the target aircraft registration (as configured in the Mode S transponder) but the ARINC data doesn't. From what I saw on a flight with this system in this Cessna 400, this feature shows the registrations only for N-reg targets which suggests that the system generates the tail number by decoding the 24-bit Mode S ID. Such decoding is algorithmically possible only for N-regs; other countries have gone for a "random" mapping relationship. It would be nice to see the tail number, even if only for N-regs, but the RS232 format works only with Avidyne MFDs...
On a US-registered aircraft, unpressurised, the antennae are installed IAW AC43-13 which is a generic aircraft repair/modification manual from the FAA. The single-blade one goes on top and the dual-blade one goes on the bottom.
On the composite roof of a TB20GT, there are prescribed antenna locations which have to be used. These are identified by removing the ceiling trim; the locations are obvious.
I did a lot of "due diligence" on the antenna location, prior to selecting the installer. The problem with TAS systems is that almost every well equipped aircraft has so many antennae that it is impossible to comply with the TAS manufacturer's requirements for minimum spacing to other antennae! The wording is "avoid mounting either antenna within two feet of other antennas, or physical obstructions (three feet for DME or transponder antennas)". Obviously most systems are installed in breach of these requirements. Most work, but some work better than others...
When my aircraft was new, in 2002, it looked like this
The single GPS antenna feeds the KLN94.
There were rather more antennae on the underside... transponder, DME, ADF.
Shortly after delivery, I had a 2nd GPS antenna fitted, which is at the very front of the roof as shown here, in front of the COM1 antenna
That antenna comes out on a BNC connector in the panel and is currently used for a bluetooth GPS receiver.
That, I was told by the Socata dealer at the time, exhausted all available antenna positions on the composite roof. The roof is about 20mm thick and antenna locations are "scalloped out" i.e. the material thins out by about 10mm. The following pic shows two of the locations provided; the one on the right is a combined GPS+COM antenna
It is believed that fitting antennae elsewhere, into the full thickness of the composite roof, would present additional difficulties because of the need to mill away a lot of material from the top surface to reach the ground plane. It has also been argued that AC43-13 would no longer cover it.
Later, once the ceiling trim was removed, it was discovered that the dealer (who went bust several years ago) was wrong and another antenna location exists
which is in front of the COM2 antenna... but it was discovered too late to make use of it. Moving the rear GPS antenna to it would have saved the cost of the combined GPS+VHF antenna (see below). For future projects, it may be useful for a GPS antenna, and/or a satellite phone antenna as described here.
In many TAS installations, these issues are solved by relocating one or more VHF antennae to the underside of the aircraft, but that is more work because the antenna cables should not have any connectors within them so would need to be completely replaced, all the way to the instrument panel.
In this case, prior to the discovery of the extra antenna spot, it was decided to remove the rear GPS antenna and replace the COM1 antenna with a combined GPS+VHF antenna.
These combined antennae are a relatively recent introduction. They use an internal resonant cavity notch filter to eliminate interference between the 11th or 13th VHF harmonic and the 1575MHz GPS frequency which is a major problem - see notes here.
They are not cheap. With the KLN94, I did not need the WAAS version of this antenna, which would cost even more money.
These pre-installation pics show the proposed locations
It can be seen that the proposed location lies behind a slight "curve" in the roof which may affect performance in the forward direction. I sent these pics to several US installers, Avidyne and their distributor DAC to check this is OK, and to the limited extent that clear information could be extracted by an "end user" the consensus was that it would be fine.
There was a similar issue on the underside where the bottom antenna would also be behind a slight curve, being directly below the top one, and this was "OK" also.
The antennae need a ground plane as per the TAS605 Installation Manual. On the underside this is a non-issue because the belly panels are metal. However, the TB20GT roof is a non-conductive composite which contains a ground plane embedded within the composite material. This ground plane is exposed at the prescribed antenna locations, so that the bases of antennae mounted at those locations make a direct connection to it.
My earlier attempts to get a drawing from Socata showing the extent of their ground plane were unsuccessful. I managed to get only these drawings 1 2. Their ground plane is believed to run the whole length of the roof and its width is believed to be the left-to-right distance between the door hinges, but the person who supplied that information (on a US forum) works for Socata USA who are prohibited from supporting European owners
In retrospect it would have been easy to determine the extents of the Socata ground plane, with a cheap (£100 or less) capacitance meter, with one lead connected to the airframe and the other with a piece of metal foil on it (say 50mm x 50mm) and by just moving the foil over the roof and seeing where the capacitance suddenly changes. Or one could have used a cheap mains cable detector from a DIY shop...
In what appears to be a misguided attempt to treat the composite roof as totally-composite (no ground plane, as one might find on some aircraft types not certified for IFR) the TCAS installer added another ground plane, about 20" wide, to the underside of the composite roof, visible here as the silver colour
An extra ground plane is pointless because the existing embedded ground plane will prevent the antenna "seeing" the new one...
The correct way of doing it means either (a) milling away the composite layer, on the upper surface, all the way down to the metallic layer and mounting the antenna directly on the metallic layer, or (b) counterboring the composite all the way to the metallic layer and mounting the antenna on metal spacers whose length is just right for the antenna to sit flush. I have not been able to establish what exactly the installer did (subsequent correspondence was inconclusive) but it appears no effort was made to connect to the internal layer. The end result appears to be working, however, and the two ground planes will be only a few mm apart, over an area of roughly 0.3 sq. metre, which will couple them well enough through capacitance alone. I calculate the capacitance to be roughly 3000pF, based on a relative permittivity of fibreglass being about 5, which is plenty enough at the GHz frequencies involved. Both ground planes are connected to the airframe, with a low resistance, but that is via wired connections which won't do anything for RF.
For the best job, the antenna bases should be sealed with PR1422 which is a much higher grade sealant than the white silicone normally used.
Other Installation Issues
The TAS605 box itself normally goes into the luggage compartment. In this case it ended up next to the WX500 stormscope box (the WX500 is the smaller unit with the chrome handle)
The black metal cover was fabricated by the installer and is merely an enlarged version of the one which originally covered the stormscope box.
One could mount the TAS605 in the void behind the luggage compartment but that would require some metalwork because there is no mounting surface there, and it would be subjected to a much wider temperature and humidity range. The lowest specified temperature is -20C and this could be easily exceeded in high altitude flight.
Due to the KFC225 factor (mentioned earlier) I made sure the two belly panels were properly grounded to the airframe. They are aluminium and are held down by many screws, but they are painted and there is no guarantee that there will be a connection. The TAS605 installer fitted one grounding wire (on which they stripped the thread on the stud so badly we had to grind off the nut at the subsequent Annual ) and a second ground strap was fitted later during the Annual. The lower TCAS antenna is mounted on the belly panel (which acts as a ground plane for it) and if the belly panel is not thoroughly grounded to the airframe, it will be radiating heavily into the airframe and the RF will be picked up by all the wiring inside the aircraft.
The TAS605 installation has some weird antenna cabling requirements:
All four antenna cables must be at least 5m long. The actual wording is "the attenuation of each antenna cable should be three 3±0.5 dB at 1GHz" and together with RG400 attenuation being 0.6db/m this gives the 5m minimum length. With more exotic cables such as this, the minimum length will be even greater. Moreover, the two top antenna coax cables must be the same length, and the two bottom antenna cables must be the same length. The tolerance on "same" is 5cm. Where there would be a lot of spare cable, attenuators (example) can be used but they are fairly pricey, and very pricey if bought from "aviation" sources. In most light GA installations there will be a lot of spare cable (but not enough to justify attenuators plus the extra connectors) and the simplest solution is to coil it up somewhere. In the TB20, the best place is to locate the coils in the void behind the luggage compartment
The antenna cable is the brown cable in the above pic, shown before it was relocated on my request to not protrude into the access hatch (see later notes on Installer Performance).
Connections to Other Avionics
There are some connections between the TAS605 and the rest of the aircraft:
Pressure altitude input. There are at least 2 ways to do this: ARINC 429 from the Garmin GTX330 ARINC output, or connecting up to all 10 of the gray code wires which go from the altitude encoder (a KEA130A encoding altimeter in my case) to the GTX330 (GTX330 IM). The former method is much neater, but there is a mysterious incompatibility which neither Avidyne nor Garmin wish to sort out. So the gray code connection had to be wired up...
Heading input. This can be done in several ways and in my case I was able to use the SN3500 EHSI's ARINC output, which provides this (Label 320) and unfortunately very little else. There is a missed opportunity for Sandel to enhance the SN3500 as an ARINC data concentrator and re-output all the data it has connected to it.
Traffic data output. In this aircraft this goes to two display devices: the KMD550 MFD, and the SN3500 EHSI.
The KMD550 needs the KAC504 ARINC card which is installed using the instructions in the KMD550 Installation Manual. All configuration is via the front panel; no laptop etc is needed.
The KMD550 needs firmware version 01/03 which is actually very old (2001 or 2003?). The KMD550 database cartridge, which can be purchased as a one-off update from Honeywell, also updates the KMD550 firmware. A one-off update is $285 for an exchange flash card (2012 pricing) which did not stop one pilot in Germany being quoted €1000 for it by a very well known avionics shop Because the KMD550 database is non-critical (basically land/sea outlines, airports, navaids and airspace) and all flight-specific data comes from the GPS, many pilots update the KMD550 database relatively infrequently. At present (1/2013) Honeywell still support the KMD550 with database cards, even though software development on it ended some 10 years ago. If Honeywell stopped supporting the KMD550 with database cards, and your KMD550's firmware was too old for the TCAS requirements, somebody else's database card could be copied to the older one using the CSM OmniDrive USB2.0 Linear Flash writer from CSM. Unlike the database copy protection schemes used in most IFR GPSs, the KMD550 database is not locked to the particular KMD550, or to the FLASH card.
The SN3500 does not need any hardware changes but does need a "Traffic Enable" software key, which Sandel sell for nearly $1000... This key is obtained from Sandel, by supplying them with your EHSI serial number. I decided this feature was worth having because it provides a second means of displaying the traffic data.
Ground mode sensing. The TAS605 has a "ground mode" in which traffic is displayed (which is very nice e.g. you can see any transponding traffic on the final approach before you enter the runway) but audio warnings are disabled. There are several ways to do the ground mode sensing. One can wire to the landing gear squat switches, or one can install a differential pressure switch which activates when the pitot pressure exceeds the static pressure by a certain amount. I had the latter; the switch is a type widely used on Cirrus aircraft, P/N 14128-001 which is one of these, P/N visible here:
The same pressure switch was used to switch both the TAS605 and the GTX330. The use of a single switch to do both requires the use of two diodes
to guard against current flow from one instrument to the other, under fault conditions.
Audio output. This needs to go to an unswitched intercom input, so it cannot be pilot-disabled. In my case there were no unswitched inputs left on the PMA7000 intercom, so the TAS605 was connected to an input pin (Unswitched Input #2) which was shared with the KFC225 autopilot.
To ensure that both sources drove the intercom input properly, a 470R resistor was added in series with the TAS605 output, to match the 500 ohm output resistance of the KC225. Because the KFC225 annunciations are very loud, I fly with all headset volume controls set to MIN, and use the various radio controls to adjust the volume as required. The TAS605 output level thus had to be adjusted upwards, to match the current KFC225 volume, which is done with a laptop connection.
Socata documentation practices mean there is a question mark about the availability of the two unswitched inputs. This diagram shows what is supposed to be in the aircraft, but there are at least two very similar diagrams covering the late-GT version (i.e. with the KFC225 autopilot referenced). Despite selling the TB20GTs for about £200k, Socata never produced an airframe-specific set of wiring diagrams, so an installer has to examine the existing wiring to see which diagram is a "best match". So, it may well be that Unswitched Input #1 was available after all.
Mute/Test switch: a small momentary switch is installed which enables the TAS605 to be momentarily muted. If pressed twice, it generates a "ground mode" annunciation.
The TAS605 installation was planned to be carried out without extracting the TB20GT centre avionics stack, shown in the centre here
Removing the centre stack takes about 1 hour if you know exactly how to do it, and if no previous installer has run any cables directly in/out of it. Various connectors and unused connector locations are provided for such cabling
but since the backs of these are inaccessible until the whole stack is removed and stripped down on a bench, most installers just run cables straight through the various holes and leave a problem for the next one... It is OK to run cables like that provided the installer puts in an inline connector!
One can access the intercom without removing the centre stack, because it is at the very top and just removing the grey trim gives access to the connectors on the back of it. One can also access the KMD550 connections, by pulling out the KMD550 and loosening its connector backplate; the KMD550 hole is big enough to get a hand through.
Overall, I would rate the wiring difficulty in this installation at the "easy" level and the wiring is about 2 days' work. Installing the two antennae, with all the interior trim removal and refitting etc, could take about a week, and attention must be paid to the required ground plane areas. So 2 weeks for the whole job is a reasonable estimate, for 1 person doing it.
The final antenna locations are here; top
The lower antenna is roughly 200mm more forward than the upper antenna, which does not matter at all.
I do a lot of long trips in the aircraft and thus I am very fussy about who works on it, and I personally inspect all work done. This installer was carefully selected, based on good reports from customers I knew. The price was not haggled over, and the price paid was a "top price" for a supposedly high quality job.
The installer did it on time - just over 2 weeks.
Unfortunately their workmanship was poor. The list is long and every subsequent opportunity to work on the aircraft - an Annual service was done shortly afterwards - turned up more bodges. I quickly lost any confidence in the work and decided to do a very thorough inspection, which included paying a freelance avionics man to check stuff while he was doing other (unrelated) work on the aircraft.
A lot of the wiring was untidy, with one wire cable-tied to the emergency gear release cable sheath, coax cables running across the hatch that leads into the rear of the aircraft, etc. But that's just cosmetic...
Minor damage was done to trims, and a number of screws turned out to be stripped. A heater hose attachment was broken right off and together with the hose no longer secured with the original cable ties, the hose was close to fouling control linkages. Another heater hose was damaged and will need to have a section replaced.
On the initial delivery, I found that the autopilot audio annunciations no longer worked, due to incorrect wiring to the intercom (the 470R resistor was required in series with the TAS605 audio output). Also the GTX330 GND/AIR auto switching didn't work, due to the wire to the GTX330 being missing altogether.
I requested a "radio check" to make sure the basic navaids and the autopilot still worked, but this was not possible because the company had no LOC+GS test set (the Sandel SN3500 EHSI will not display a glideslope if there is no localiser signal, so one needs a test set which generates the full ILS signal, such as an IFR4000) and their VOR tester worked only intermittently. I arranged for a radio check to be done later with a freelance avionics man, and all navaid reception (and autopilot tracking of them) was later verified in-flight in VMC conditions.
On taxi out, the instrument backlighting was found to be not working. This turned out to be a short between one of the instrument light wires and "somewhere" so they just re-ran that wire and cut the original off at both ends and left it in the wiring harness! This is obviously totally unacceptable (anything like that is a fire hazard) so at the Annual a few weeks later the wire was traced and pulled out (took only about 10 minutes) and found to be undamaged, so the short must have been in one of the pieces they cut off.
After I discovered the first batch of bodges I flew the aircraft back to them and left it there for a number of days to fix the various issues and even on the final collection I immediately found yet another bodge: the circuit breaker panel was not screwed back in place and 3 screws were missing from it.
However, potentially life-threatening stuff was discovered during the subsequent Annual. This video shows one of the air pressure switch hoses was barely attached and came straight off. If this happened in flight, it would have connected both the pitot and static lines to the cockpit pressure which would have been "interesting" in the TB20. It would have been much more "interesting" in a pressurised aircraft which this company often works on, with the resulting loss of airspeed, altitude, and many aircraft systems such as the autopilot and the capability to fly any instrument approach that uses baro assisted VNAV.
These videos 1 2
show the wiring harness (where the gray code wires were taken off the back of
the GTX330) is fouling on the control linkage. This was not obvious on my early
checks and is clearly visible only by inspecting with a mirror, or when the
RH instrument cluster is tilted forward.
This video shows another aspect of the same thing where a sharp edge of the control linkage is rubbing on the underside of the harness.
This pic shows a sharp edge of a jubilee clip cutting into wire insulation. There is already a dent there and eventually it would have cut through.
Later, it was noticed that the centre avionics stack was showing severe vibration. It turned out that the trim which goes over the top and down the sides and on top of which the vertical card compass is fitted, was forced back on. It was jammed between the stack metalwork and the windscreen, with the compass touching the winscreen material! The windscreen gets a lot of vibration from the propeller slipstream and this was being transmitted down via the compass. It took about 30 mins to fix it, and it turned out that the bodger who refitted the trim left some inline connectors trapped underneath it, which should have been tucked away into some cavities.
Obviously the whole job was done in a mad rush.
The company is one of UK's biggest, has a good reputation from the past, has EASA Part 145, EASA Part 21, FAA Repair Station status, but obviously I cannot recommend them on the basis of my experience. Their attitude in attending to the issues has been positive but even after their attention was drawn forcefully to the lousy work they seemed to be unable to deliver either reasonable workmanship in fixing it or exercise any form of quality control. Eventually the various known defects were fixed and the aircraft was thoroughly inspected by myself and others, but I am not going anywhere near them again.
The system works well and is accurate in relative altitude, distance and azimuth.
Distance appears accurate. There is an easy way to verify it: as you approach an airport, you can see aircraft on the ground, at the expected altitude, if they have their transponders on.
Relative altitude is hard to verify but traffic acquired visually appears accurately depicted, and one can use the same trick as above with targets on the ground whose altitude is known.
The azimuth is easy to verify with visually acquired traffic which appears generally accurate to within about 10-20 degrees. One can also check it by flying near a big airport (outside controlled airspace) and seeing the traffic lined up on its ILS.
Within this 10-20 degree azimuth band, targets do sometimes wonder about, and an occassional sudden change in the azimuth of say 10 degrees makes them look like they have suddenly turned into an SR71
One limitation of the system is that it sometimes cannot see traffic which lies in a cone above the aircraft, or in a cone below the aircraft. These are potential blind spots. It's hard to say how big the cone is but a 30 degree included angle is about right.
This appears to be caused by such targets not being seen by both antennae (due to airframe masking of one antenna or the other) and the TAS605 suppresses such returns. This should not be a problem in practice unless somebody descends into you while remaining directly above you the whole time, or climbs up into you similarly. It's most likely to be noticed when joining a circuit from above, when targets vanish as they pass below you.
Targets approaching the above cone are more likely to suffer from azimuth errors, before/after their disappearance/reappearance.
The system works impressively well on the ground too, picking up both ground based and airborne targets, but the azimuth is likely to be way off if there are any metal building nearby - due to reflections from the buildings.
The claimed 13nm range is difficult to test because so much depends on the transponder power of the target, but in reality anything over about 5nm is beyond visual acquisition (for light GA aircraft) except under the most favourable circumstances e.g. at night, landing lights on, head-on, and with nothing illuminated on the ground. And at 5nm the TAS605 appears solid.
For its intended purpose (avoidance of conflicting traffic) the TAS605 system appears very good. For example, traffic within a +/- 1000ft vertical range and a few miles away, and approaching, is depicted in a stable way which facilitates visual contact acquisition. The fact that a visual contact is rarely acquired in practice is just something which makes one feel the money was well spent
As mentioned above, the TAS605 transmits (via ARINC 429) all data it has to all display devices. Each device can then select its own scaling or its own vertical limits on the displayed traffic. I found that the only way to get the KMD550 to display the same as the SN3500 is to select "normal" on the KMD550. I found the best setting for the SN3500 is "ON"; additional information is in the Traffic section of its POH. Sending "everything everywhere" is also how the WX500 stormscope data distribution works, and is very good because if one display device failed, the other remains fully available.
Here is a video (25MB) which shows the end result, in flight, and on the ground. The comment about "non transponding" traffic actually refers to non-Mode-C traffic on which no altitude is indicated. Totally-non-transponding traffic is obviously invisible to the system.
Apart from selecting the options on the display devices, there is no "user interface" to speak of. The system has no on/off switch, it comes on with the avionics master switch, and the default settings of the display devices are fine.
The Pilot's Operating Handbook for the TAS605 is a very generic document which talks a lot about Avidyne's own traffic display devices and contains little other useful information, so it is just as well that the operation of the system is very simple. It's more important to read the Traffic sections in manuals of the display devices.
Paperwork - US Registered Aircraft
As stated earlier, the TAS605 is covered by the AML STC and the extras (antenna mounting etc) are done IAW AC43-13.
The STC also comes with an FAA approved AFMS (Flight Manual Supplement) which is a legal requirement to have in the aircraft flight manual (AFM). You print this off to the required size, punch the holes down the edge, and insert it into your AFM. The AFMS is so generic as to be almost useless but it is a legal requirement.
Any Major Alteration needs a 337. If the 337 is supported by an STC, it is sent directly to the FAA in Oklahoma for filing. This is not a Field Approval so the FAA does not need to pre-approve the job. After the work is completed, the installation is checked for conformity to the 337 by either an A&P/IA or by an FAA Part 145 Repair Station, the 337 is signed and sent off to the FAA.
A W&B sheet was also drawn up by the installer. The complete system, including cabling, weighs only about 6kg.
Paperwork - EASA Registered Aircraft
Many TAS600-series systems have been installed on G-reg TB20s.
According to some UK installers, approval on an EASA-reg aircraft is via a Minor Mod. The AFMS is approved under an EASA Form 36. The minor change cost will depend on whether the avionics shop already has some prior similar approvals, but if generated from scratch, should cost £900-£1200. The Form 36 approval for the AFMS costs about €260.
The Avidyne TAS system already has an approval on all Part 23 single-engined aircraft via a German LBA validation of an FAA STC, so no additional design and certification fees are applicable.
Paradoxically, therefore, EASA treats the system as a Minor mod whereas the FAA treats it as a Major mod... but in practice the FAA route is much simplified by the AML STC.
A quick google digs out numerous approvals for various aircraft types. For example it suggests that the UK CAA appears to have approved a TAS600 system at some stage in the past, under # LA101077, as found about halfway down here. EASA is required to accept any such prior approvals.
Update 8/2013: I found a note from Avidyne from 2007 stating the TAS600 range is covered by the same UK CAA Aircraft Equipment Approval" as the old Ryan 900BX. The approval number is stated as "LA101077" which oddly enough is the same as above...
The TAS605 is capable of driving most avionics display products on the market, so this is not an issue. Regarding ADS-B, Avidyne claim that this will be supported, for a few-k $ extra. Update 7/2015: still no news on this "upgrade". For certain (probably Garmin) display products, this foot-dragging by Avidyne can be sidestepped because there is a Garmin box which can combine the ex-TAS605 ARINC429 traffic data stream with the traffic data stream from a Garmin ADS-B receiver and feed the combined data to the display device. However, I can't see anybody radiating ADS-B who will not be Mode C equipped; ADS-B penetration in European GA is miniscule and is likely to be so for many years.
Would I do it again?
Because of the hassle I got with a supposedly top-class company certified for every imaginable type of aviation maintenance, I am tempted to say NO. But it does work well, does not clutter the aircraft (is virtually invisible) and has a negligible W&B cost.
However, if I was doing it again, I would rent a hangar for a couple of weeks and pay a good freelancer to install it (closely monitored, obviously) saving a few k in the process. However most owners don't have the expertise to design and manage such a project, and actually neither do most freelance installers whose bread and butter is avionics repair or screwing in the occassional GNS430 or similar. It may also be difficult or impossible to do such a closely managed job on an EASA-reg (Part M) aircraft; owners of those have no choice but to trust the company approvals, which can evidently be a trust totally misplaced...
Update 8/2013: the system has been working very well, and really drives home how much traffic is never seen. It doesn't matter how many people you have on board, all knowing roughly where to look, but they cannot see the target aircraft. At least 90% of traffic is never spotted.
But equally one is often in a situation where one has spotted something and then lost visual contact with it. The TCAS system is likely (if the target has Mode C/S) to show it in a way which enables one to not worry about it. This is also a great benefit. However, obviously, there could be more than one plane out there and the one you briefly spotted was the non-transponding one
Flying around Europe IFR (example) one sees virtually nothing on the TCAS, if the range is set for 3000ft above and below. The separation between light GA and jets in controlled airspace is normally so great that nothing shows up even when overflying major airports like Frankfurt or Schiphol. The system really earns its keep in Class G, especially in the UK.
2+ years later (July 2015)
The system has been working well and both myself and any passengers think the money has been very well spent - when flying around in Class G airspace especially in the UK at weekends. One still finds a good number of Mode A targets (no altitude) which I think is very socially irresponsible because it denies visibility of altitude to ATC too.
The only issue is that head-on traffic at same or similar altitude is reported quite late. There is a partial blind spot on the straight-ahead bearing, and often such traffic doesn't appear till it is just 1 mile or so away. However, often it does appear at say 5 miles before than and then disappears! This isn't consistent so is probably related to the quality and power of the target's transponder installation. 1 mile is enough at GA speeds and much better than nothing. I wonder if this is due to propeller shielding?
With maximum distance, there is a strong correlation between light GA traffic and jets. Obviously, the quality of transponder installations varies a lot. The 15nm figure for the TAS605 is a software limit; the actual distance is much less and varies between 5nm and 10nm for light GA.
Update January 2019
I had an opportunity to revisit the compromised range issue, and based on e.g. this the location of the upper antenna was suspected.
Because the original installer had to make all four coax cables the same length, there was plenty of spare length going to the upper antenna to enable it to be moved forward on the roof - about 60cm in this case. The combined GPS+VHF antenna was then moved to the old TAS antenna location; this was also not difficult because one of the two cables was long enough and the other was extended with a short BNC extension cable made up with the proper RG400 cable.
This is the end result
Flight tests immediately showed a dramatic improvement in the range all around, especially at 12 o'clock where I now get airliners up to the full 15nm range and GA targets to around 10nm.
The mechanicals of the antenna installations on the TB GT composite roof are not trivial. The antennae are all mounted in the scalloped-out locations, and these are OK for normal antennae, but the TAS antennae are slightly too big (in length) for all four screws to fit inside the scalloped-out circles. And one cannot just use longer screws, because tightening the nuts crushes the honeycomb composite. So one has to use spacers, around 13mm long, and these prevent the composite being crushed. The length of the spacers needs to be just right so tightening the screws squeezes the composite just a little. The doubler plate (supplied in the TAS kit) works fine with this. The previous installer did some bodges there but on this occassion proper brass spacers, 8mm OD and ~5.5mm ID for the screws, were machined. The TB GT roof is a hard composite of about 3mm and then about 11mm of honeycomb, except in the scalloped-out areas where you have just the 3mm hard material.
The next issue was the ground plane which every antenna (of the types used here) requires. The original TAS installer stuck some conductive tape to the underside of the roof; this probably works well enough. However, when that TAS antenna was moved to its more forward location, the grounding had to be checked. A standard electrician's metal detector immediately established that the whole roof has a metal layer in it, but it was not obvious how to connect to it because - looking in the various holes - it was obviously very thin. Fortunately the new forward location contained an exposed metal surface there which connected to the airframe. It is a mystery why Socata provided that in the forward location and nowhere else, despite providing several antenna locations.
Note also that Avidyne revised their installation manual, with this new requirement which represents a useful concession over the old one where all four had to be the same length.
All avionics manuals here were found freely on the internet.
Last edited 29th January 2019
Any feedback, reports of dead links, corrections or suggestions much appreciated