TB20 KLN94 GPS Approach Approval (FAA)

Before an IFR GPS can be used for flying GPS/RNAV approaches (example) it needs to be approved for this purpose, by the aircraft's certification authority. Merely installing an IFR GPS, with the required separate annunciators, is not enough. This is true for both USA (FAA, N-reg) and European (EASA, G-reg etc) registered aircraft, though the process is different between the two.

This article describes the process of obtaining such an approval. It was done with a new AFMS (approved flight manual supplement) for a KLN94 GPS in a US registered Socata TB20GT aircraft


The KLN94

The Honeywell (formerly Bendix-King) KLN94 is a good basic IFR GPS which does everything operationally relevant to flying IFR around Europe. Many thousands were installed (by both new aircraft manufacturers and for upgrades) during the 1990s and up to around 2003. It was finally discontinued in 2011. It interfaces to many industry standard fuel totalisers and airdata computers, and interfaces elegantly to the KMD550 MFD (the upper unit in the picture below) which delivers excellent IFR+VFR mapping (including European visual reference points). The KLN94 is authorised by the manufacturer for all LNAV-only IFR operations including GPS/RNAV approaches. The KMD550 in turn interfaces to standard TCAS products such as the Avidyne 600 series, upmarket GPWS systems, and its NTSC video input supports weather radar, or whatever can output NTSC video

Today, one would not choose the KLN94 for a new installation.

Suprisingly for an IFR approved GPS, its database does not contain any RNAV SIDs/STARs, which precludes its legal use for such procedures, but bizzarely the database does appear to contain all their individual waypoints which can easily be manually inserted into the flight plan. It is a mystery why Honeywell introduced that limitation, but it is not currently operationally relevant, not least because most if not all airports that publish all-RNAV terminal procedures do not operate them and use radar vectoring. The database does contain non-RNAV SIDs/STARs.

The KLN94 does have analog vertical (VNAV) outputs (from its installation manual: pic1 pic2) and obviously the intention was to deliver a "GPS derived glideslope" capability, but Honeywell's GA division collapsed around 2000 and they stopped product development other than trivial fixes. So, the KLN94 does not support LPV (GPS/RNAV with a synthetic glideslope) approaches or variations thereof e.g. GPS/LPV or GPS/LNAV+V. This is also not operationally relevant in Europe, where almost no LPV approaches exist, and I am not expecting this to materially change for perhaps 10 years. However LNAV+V is a very handy capability which is available at many (not all, for complicated reasons) locations where there is an existing GPS/LNAV approach.

Due to a lack of a specific Letter of Authorisation (LoA) from the manufacturer, the KLN94 cannot ever be approved for PRNAV, even though it is perfectly capable of the required accuracy (RNP 1.0). This legal technicality is another mysterious Honeywell decision. PRNAV is perhaps the biggest cloud on the horizon for European IFR GA, and this aspect alone may force an expensive avionics refit one day. But not yet...

Europe has been gradually introducing ordinary GPS/RNAV approaches - more than 15 years behind the USA Progress in the UK remains crippled by the legal requirement for ATC for any instrument approach, which prevents instrument approaches (of any kind) being introduced at nearly all airfields that would benefit from them most. I also do not recall a single instance (from my flying) of an airport outside the UK which has Customs (required for any flight between the UK and the rest of Europe) and whose only instrument approach is a GPS approach. There are some in Germany; EDMS is one example. But GPS approaches are slowly becoming operationally relevant because navaids (NDBs etc) are sometimes out of action, airfields with funding issues can be slow to repair them, and GPS approaches tend to offer a small improvement on the decision height.

Unfortunately for me, the aircraft was originally delivered (2002) with a KLN94 AFMS from Socata which approved only IFR-Enroute (BRNAV) operations, not SIDs/STARs or GPS approaches. Every TB20GT came with that infamous GPS Supplement. The following placard was present

It is not known why this restriction was imposed, but enquiries via Socata suggest that it was at the insistence of the DGAC. The same "certification authority" also demanded that the WX500 stormscope display does not rotate according to the aircraft heading, because if it did, the pilot might use it to avoid thunderstorms The same "certification authority" also demanded that the Shadin fuel totaliser transducer is mounted on the passenger side of the firewall, in breach of the Shadin STC, where its reading was sufficiently affected by turbulence to result in errors of 20-30%

To be fair to Socata, there were no GPS approaches in Europe (or probably anywhere in the world outside the USA) in 2002, and aircraft which they exported to the USA were modified locally, as mentioned below.

The actual aircraft and the GPS installation have always been fully capable of flying GPS approaches, so this is a legal (paper) restriction only.

It means the aircraft cannot legally fly standard GPS approaches such as this one. However, because there is no law prescribing the method used to actually navigate, the GPS can 100% legally be used to fly the NDB/DME approach to the same runway In the USA, on such approaches, the GPS may similarly be used in lieu of DME, ADF, or VOR, but is not authorised for the final approach segment.

It is very unlikely that one would be picked up for flying a GPS approach using an unapproved GPS but where is one to draw the line? This could be debated for ever but in this case the request to fly the GPS approach is made openly on the radio and someone familiar with the aircraft who happens to be listening on the frequency could report it or otherwise cause trouble. A worse scenario might be where an IFR flight plan is filed to an airport whose only instrument approach is a GPS one, and then the flight is illegal before departure, which may impact the insurance...


Changing the AFMS

Fortunately, the AFMS can be changed, or a completely new one can be created, but it is a nontrivial job because under FAA (and to some degree similar EASA) rules, changing an AFMS is a Major Alteration. In FAA-land this is done with a Form 337 supported by Approved Data.

In this case, the lack of an aircraft-specific STC (the KLN94 STC is for a Beech Baron) meant that this had to be done as a Field Approval i.e. a Form 337, with some supporting documents, which has to be FAA approved (signed). In the case of a "field approval" one submits any data which the FAA finds acceptable and when the FAA inspector signs block 5 on the 337, that makes the data "approved data" for that installation.

The AFMS is custom written for the particular aircraft and is specific to its registration number.

Generally, a specimen text of the supplement is provided in the back of the installation manual for the particular GPS and this is used as the template for the new AFMS. The KLN94 version is here. The installation manuals are not in general circulation and only authorised dealers are supposed to have them, but they are easy enough to find on the internet. I have a huge collection of them myself

Assuming the GPS installation was done correctly, nothing physical on the aircraft changes; all this is purely paperwork! The exception to this may be where the GPS is initially configured for VFR-only (possible with the KLN94) in which case one needs the Installation Manual for the special key sequence required to enter the configuration page where IFR operation can be enabled. And the application for the modification should state something like "Placard stating 'GPS to be used VFR only' removed".

This article (local copy) has some background information. This is a well trodden road...

Over a period of several years I had contracted no less than four UK avionics shops to produce the custom AFMS for the KLN94. All had agreed to do it within a budget of £500. The first three got bogged down in technicalities and after some months each of them gave up. The fourth started on it and stopped when it got busy on other projects...

All tried to sell me avionics replacements... there is little doubt that even a used GNS430 or GNS430W will make approvals easier, and it can be approved for PRNAV. And nowadays a GTN650 comes with an AML STC which makes the whole "Euro IFR GA paper collection exercise" trivial. However my KMD550 MFD is not Garmin compatible (in the OBS mode, in particular) so it would be a big job, changing the whole centre stack and installing a GTN750. £30000, perhaps. Once you have flown with an MFD, you will never want to fly without one; all the screens on the aforementioned GPSs are too small to be used alone. It's also a job involving considerable downtime, which only 1 or 2 UK avionics installers are capable of doing well, so I am leaving it until it becomes absolutely necessary. I would also prefer others to wring the various issues out of the new GTN650/750 At time of writing, The GTN650 is unaware that Europe works in litres, for example. The reality is that the KLN94 does everything needed for European IFR.

Eventually, 10 years into my ownership of the aircraft, and with experience and contacts gained from this EHSI project, I decided to have a go at it myself.

A field approval is a lot easier if one can find a previous field approval for an identical aircraft and GPS. I tried to obtain a 337+AFMS, or even just the AFMS, for another US registered TB20/KLN94 installation, by contacting some US TB owners and some US Socata dealers. It is known that many TB20s/TB21s exported to the USA had this modification done by the Socata dealer in Florida. Despite there being perhaps 100-200 of such aircraft "out there" I never managed to find anybody willing or able to supply the documents. Many pilots had apparently never even opened their flight manual... I did find some US pilots who thought they had a fully-IFR KLN94 installation but it turned out they had the standard Socata AFMS. I did not get the slightest co-operation from Socata USA who should have the records but they are prohibited by the factory from supporting European owners. All very frustrating...

Had I been able to locate even just the N-XXXX tail number of an aircraft with the correct paperwork, there is an easy process whereby for about $10 one can get a CD from the FAA with all 337s filed for that aircraft, but I had no obvious way of locating the tail numbers. There is a database out there of TB aircraft which is current to maybe ~ 2009 (edit the URL for different serial number ranges) but it provides no clue as to which ones are located in the USA and if so which US dealer might have sold them. I did in fact get an FAA approved KLN94 AFMS from the one US Socata dealer who could help, which contained a tail number, but stupidly I lent it to the one of the above mentioned avionics shops (without keeping a copy) and they lost it

So I collected as much supporting documentation as I could.

My first attempt at a new AFMS was using the FSDO which I used for the above EHSI field approval. Unfortunately, the very helpful FAA inspector there had recently retired and I spent several months educating several of his colleagues on avionics trivia. Initially, one inspector claimed that a KLN94 is not IFR approved How many thousands of Cessnas, Pipers, Beech, and other fully IFR approved FAR Part 23 aircraft were shipped with a KLN94? I sorted that one by sending them the installation manual and several FAA approved AFMSs for other aircraft (the Beech Baron for example). Then they found something else, and then something else... Eventually I "almost got there" and airmailed a 337 with all the supporting documents to the FSDO, but when I started chasing it a couple of months later it turned out they lost the package! So I emailed them the documents, which the inspector rejected, mostly because various signatures were missing, which was true (my fault) because the documents I emailed were not the signed originals... At this point it became obvious that this FSDO was going to be very hard work... It's obvious that there is a lot of variation in FSDO expertise but how does a European pilot find a helpful one? The aviation approval business runs on relationships built over time.

The FSDO officially responsible for Europe (the NY IFU) told me in 2010 they no longer do avionics approvals, which is scandalous.

EASA does not make this any better. You get essentially a single point of contact but you never know what they will object to when you send in the application... the process is set up to channel €€€ business to EASA Part 21 companies.

However, out of my numerous aviation contacts I found a very experienced retired US avionics engineer, and with his help I was able to move forward. He visited his local FSDO and quickly sorted out what was required. He then presented the documents on my behalf. Some edits were requested by the FAA, to the Honeywell sample AFMS. The 337 went back and forth several times to get the wording consistent.

Clearly there is a business opportunity for a US-based avionics "agent" who can act in this way, on behalf of European owners. The FAA process is the same regardless of where in the world a US registered aircraft spends its time. It would not suprise me if some European avionics shops already have such a contact.

Due to these hassles, many European avionics installers sidestep the unpredictable FSDO field approval process by paying a DER (example) to generate the approved data. He generates an 8110-3 form which is sent off with the 337 to the FAA in Oklahoma for filing. No FAA approval is then needed. However, while this process produces a 337 supported by "high quality" approved data which nobody is likely to argue with (albeit at a hugely inflated cost to the end customer) very few DERs are authorised by the FAA to generate an AFMS, and this is where Garmin, with their FAA AML STCs which include a pre-approved AFMS, have a huge commercial advantage. Now that Garmin have an EASA AML STC (GTN650/GTN750 products) they are obliterating the competition in Europe... In any Major Alteration scenario, most European avionics shops are simply not interested in installing anything other than AML STC products, and most of those are from Garmin.


Checking Proper GPS Operation

This is the first step. The approach approval requires the GPS installation to be checked for proper operation and immunity from potential interference sources in the aircraft.

The basic document for IFR GPS approval for both enroute and approaches is FAA AC20-138a (local copy). AC20-138a supercedes AC20-138 which (for the curious) is here. Various local CAAs have developed their own procedures from this; for example this (local copy) format is from the New Zealand CAA but is old and seems loosely based on the old AC20-138. This is another flight test form which was used by Socata in the USA in the late 1990s.

The current version is AC20-138C (local copy) which doesn't appear to add anything relevant in this case.

The most notable requirement is a test for interference from VHF transmitting equipment, on specific frequencies e.g.

121.150 MHz 121.175MHz 121.200MHz
131.250MHz 131.275MHz 131.300MHz

and some others as described on page 41 of the AC20-138a PDF. These frequencies are, in general, ones whose 13th or 11th harmonic (respectively, for the two lines of frequencies above) lies on or around the L1 GPS frequency of 1575.42 MHz. Interestingly, AC20-138a defines this requirement in the ground (post-installation) test and not in the flight test.

Also interestingly, the common BRNAV approval involves similar interference tests so any GPS installation which was properly BRNAV approved should not have any problems.

VHF Interference

I quickly found that a transmission (pressing the PTT switch) lasting more than a few seconds from either of the KX155A radios on 121.150, 121.175 and 121.200 affected both the KLN94 IFR GPS and the yoke-mounted Garmin 496. Both COM1 and COM2 radios did it, with COM1 (which is is mounted directly underneath the KLN94 GPS) being slightly worse than COM2. Both GPS receivers recovered within a few seconds, but the gap in reception could have inconvenient effects. A loss of GPS reception is never useful but it could result in a real jump in pilot workload if flying with the autopilot, in NAV mode, tracking a GPS track (which could be enroute or when flying a SID/STAR/approach procedure) when the loss of GPS signal would terminate the autopilot's NAV mode and return it to (probably) a simple wings-level mode.

This discovery was suprising since the GPS installation was BRNAV (enroute only) certified from the French factory (by the French DGAC) under the TB20 Type Certificate. I have no idea what (if any) VHF immunity tests the DGAC had specified; this document (local copy) suggests the requirements were based on AC20-138. Page 6 of the original Socata AFMS also references AC20-138...

More significantly, this particular aircraft was manufactured for the U.S. market where the Socata dealers were routinely doing the full-IFR field approvals. When I bought it, it already had the N-number on it.

So, either there is something odd on my particular aircraft, or Socata France skipped the immunity test in their original BRNAV compliance tests, and when Socata USA did their customised GPS flight manual supplements they either did not do the immunity testing or they developed a fix which never found its way into the Socata maintenance documentation. Enquiries to the usual contacts drew a blank.

A swap-over of the two KX155A radios eliminated the possibility of a faulty radio, and an installation of a brand new KX165A (8.33) radio did not eliminate the interference either.

To the limited extent I was able to enquire via contacts at Honeywell (always a difficult company for support, they practically abandoned GA avionics about 10 years ago) they were unaware of any relevant modifications (SBs etc) on the radios, but there is much anecdotal evidence that there was a major redesign between the older KX155/165 and my KX155A/165A radios; the non-A versions were notorious for GPS interference. Unfortunately an upgrade to the "A" versions is a re-wire because the rear connections are different. There is a general awareness in the avionics business of this issue; for example Collins sell radios with claimed low GPS interference - example.

The KLN94 installation manual also has a section on VHF interference. It lists a TED GPS filter P/N 4-70-54 inline notch filter which is installed in the VHF antenna cable, as close to the radio as possible. It costs from $80-$500 depending on where it is purchased from. An alternative and slightly more compact and cheaper filter is sold by Garmin, as shown on the last page of this PDF. Note that the issue can be caused by an ELT's output filter resonating around the 121.xx frequencies and re-radiating the transmission at a high enough power to overcome the GPS receiver.

The installation instructions suggest installing the filter as close as possible to the radio(s). Unfortunately, on the TB20, there is no easy way to do this, given the way the radios and the GPS are located (the bottom three items in the centre avionics stack in the following pic)

(the above pic doesn't show the Sandel SN3500 EHSI which was installed in 2011)

There is around 50cm of RF cable running from the output of each radio to bulkhead-mounted BNC connectors. This cable is not accessible, short of extracting the entire centre avionics stack. On the TB20GT, the centre stack is actually cleverly done, with all connections passing through large circular mil-spec connectors (visible in the pic below) and can be extracted whole in about 1 hour; this assumes the installer is familiar with the procedure, and that no previous installer has done any bodges. A common bodge is running wires directly out of the stack to the rest of the aircraft, without passing through any connectors, which makes it impossible to remove the stack without cutting those wires. If one extracted the centre stack, one could replace those cables with a higher grade type (the state of the art solution would be a mil-spec semi-rigid coax e.g. RG402) with the GPS filters installed very close to the radio outputs. On a Garmin avionics stack, the filter can be plugged directly into the other side of the floating RF connector which the radio plugs into, but Honeywell stacks use a hard-wired connector. This item might help, but you still have to get your hand in there...

A skilled installer could access the back of the radios without removing the centre stack, by pulling out the KMD550 MFD, removing its connector backplate, and sticking his arm in through that hole. Very very tricky.

A quick and interesting test was to temporarily fit standard 50 ohm terminators at these bulkhead connectors, to see if the elimination of most of the antenna cable runs and the antennae itself reduced the interference. It didn't... This proved that the radiation was coming out of that inaccessible cable run inside the centre stack, and eliminated the possibility of ELT resonance as the cause.

The breakthrough was eventually made by fitting the notch filters to both radio outputs, not just the worse offending one. This reduced the problem to a level where GPS reception was not affected. Clearly there is some cross-coupling involved...

The nearest location to the radio outputs that the notch filters can be fitted into easily is at the bulkhead connectors, where it is a 1-minute job.

The above pic shows the filters in place. The filters are free to rotate around their BNC connectors, and their resonant cavities (the bits that stick out at 90 degrees to the filter body) could end up touching each other. It is not 100% clear from an inspection of the filter whether every part of its casing is actually a continuous ground with the BNC connectors. Therefore, black heatshrink sleeving (better visible in the following pic)

and was placed over each filter's resonant cavity to prevent such contact.

The filters need a logbook signoff - a Minor Alteration.


Generating the new AFMS

The new AFMS is most easily done by taking the above mentioned one from the back of the KLN94 IM and annotating it using the full version of Adobe Acrobat and under Comment/MarkUp Tools use the TextBox tool - or some similar method. The original AFMS is a PDF made from a scanned document and is not editable but that doesn't matter because most of the editing is blanking-out or replacing small amounts of text. Alternatively one could OCR the original and edit that...

The final edited version of the AFMS is here. The edits were done as per the Honeywell instructions at the start of the document in the IM, with several edits which were requested by the FSDO. One of these was a new placard specifying no "precision" (i.e. LPV) GPS approaches, and no RNAV SIDs/STARs.

On advice from the avionics specialist I changed a paragraph on page 6 of the PDF (page 3 of the AFMS) from

Instrument approaches must be accomplished in accordance with approved instrument approach procedures that are retrieved from the KLN 94 data base. The KLN 94 aeronautical data base must incorporate the current update cycle.


Instrument approaches must be accomplished in accordance with approved instrument approach procedures that are retrieved from the KLN 94 data base. GPS instrument approaches using the KLN94 are prohibited, unless the KLN94 unit's approach data is verified by the pilot or crew to be current.

That is consistent with the AFMSs for other GPSs in the USA and permits e.g. the flying of a GPS approach using a non-current database provided the pilot has verified that the latest approach plates for that airport pre-date the GPS database.

The final 337 form is here. I have removed stamps, signatures etc whose publication is inappropriate.

When the 337 has been approved by the FSDO, it is signed by the A&P/IA and returned to the FAA for filing. With a normal avionics installation, the IA has to check that the installation conforms to the text of the FAA-approved 337. In this case, he only had to check the new placard

In accordance with FAA procedures, these two documents (in their final FAA-approved form with all the signatures, which I can supply to someone embarking on a similar project) can be directly used to support a field approval for another TB20 aircraft with a KLN94.


Last edited 4th February 2014

Any feedback, reports of dead links, corrections or suggestions much appreciated:
Contact details

Aviation Home