I recently was in the position of upgrading from a Tigertronics Signalink USB to an upgraded digital interface. I developed a feature list to guide my shopping. The Signalink has known noise problems which can be fixed by simple mods. One relates to mixing the analog and digital grounding methods. Another involves bypass capacitors on the voltage regulator, which derives its power from the computer USB connection. The mods are out there on the net. There is also a mod which uses the audio to drive a circuit to convert it to an open collector true CW keying line. I did not want to pursue those solutions. I was interested in true FSK RTTY; the Tigertronics cannot do RTTY true FSK, it only does AFSK. This can generate spurious signals from audio harmonics and distortion within the SSB radio passband. True FSK directly shifts the rig's VFO frequency and avoids those problems.

MFJ seems to only offer a low end digital interface, competing with the Signalink. Signalink now offers a second lower cost interface with fewer features. I was looking for a better interface and willing to pay more for higher performance.

Also, the Tigertronics Signalink USB has known frequency response tilt in the inexpensive isolation transformers employed. You can observe this on a quiet band on the spectrum display, which tilts noticeably up to the right (high frequency) side. I wanted a new interface to have better grade isolation transformers with flat frequency response for uniform performance all across the range in the FT-8 window. I am not bashing the Signalink at all here. It delivers a lot of fun for its low price.

The Signalink is a great starter digital interface. The configuration for radio interface is managed by jumpers on a header that plugs into a standard DIP IC socket. It also has the advantage of offering line level output and connection to the rear line input on the radio. The Signalink uses an RJ-45 connector for its output/input connections. I am not sure if the wire used is shielded, a problem that is possible when employing RJ-45 style connectors. I have had no problems with this, but I am still not enthusiastic about it. In fact, I refuse to own a radio which uses an RJ-45 style connector for the main mike connector, as a non negotiable point. I prefer standard 8 pin or 4 pin round connectors, which can be soldered for assembly. Some of the newer mid and high range digital interfaces use a DB-25 style connector instead of an RJ-45; the DB style connectors can be soldered, and are highly reliable.

Some points to consider:

1. Many of the mid to high end digital interfaces use software to set the connections and parameters rather than a hardware jumper system. The hardware system, once it is set, is reliable and simple. The software method can be cumbersome and require drivers and extensive tinkering; it also can be upset by changes to the settings. I prefer something simple and straightforward, easy to set up.
2. In general, interfaces employing the front panel mike connector risk noise and RF pickup, due to the lower signal levels compared to the rear line level input. Also, there is the hassle of swapping connections when switching from SSB to digital modes. If the interface uses the mike jack and the station mike connects to the digital interface for a pass-thru connection, the digital circuitry can introduce noise on the SSB signal via the mike switching circuitry. Many of the popular interfaces employ the front panel mike connector for some of their connections to the rig. This is another of my non negotiable criteria for a digital interface. This eliminated most of the popular digital interfaces from my list when I went shopping for an upgrade from my Signalink.
3. One other feature I was shopping for was a true keyed interface for CW transmit, not an audio method of generating the RF. This meant the keying waveform was determined by the internal components in the rig, not the shape of the audio waveform. It also meant any distortion products or harmonics in the audio would not appear along with the desired CW signal, if it was true keyed CW. Signalink can with modification of the connector header use the PTT relay for CW keying; I wanted a transistor open collector style connection. This was a third non negotiable item on my shopping list. Most all of the mid to high cost interfaces offered this feature.
4. For CW true keying from a computer, some jitter can occur on the waveform edges due to the timing problems associated with a Windows computer. To solve that problem, many digital interfaces and logging programs run the keying through a K1EL WinKeyer. The timing is corrected by the K1EL device. Many of the mid to high end digital interfaces offer a K1EL WinKeyer right in their box. I favored this approach when I went shopping, so it was on my feature list as well.
5. Finally, I wanted an interface that provided true FSK, not AFSK.
6. Most all digital interfaces use a 16 bit audio codec. The microHAM Keyer III uses a 24 bit codec, for more dynamic range, less overload, and lower noise. It also was at the top end of the cost scale. Maybe that will become the standard for the next generation of digital interfaces.

The RigExpert TI-5000 is only about twice the price of the Tigertronics Signalink USB which I compared it to. The RigExpert receive decode is worlds ahead of the Tigertronics, and I would recommend it to anyone for FT8 or RTTY because of the careful attention to grounding and noise issues. I am frequently getting good decodes well below -20 dB on FT-8. The transformers used in the RigExpert are flat uniform response across the audio band; the Tigertronics are not flat, but favor the higher frequencies.

The RigExpert is much lower priced than the other more expensive interfaces you have. The RigExpert correctly chooses to employ the LINE input rather than the MIKE input which the higher priced interfaces use. This contributes to superior noise performance as well. The RigExpert uses a DB-25 connector for its radio cables rather than an RJ-45.

The audio codec did not require any fussing to get it running. Its a standard audio codec. The rig CAT control for my Yaesu FT-950 to the N1MM logger installed perfectly, using the RigExpert TI-5000 accessory cable. It read out frequency and controlled all functions flawlessly. The rig interface cable was high quality and connected easily, with no jumpers to solder or fussy software to figure out. It met my criteria for no RJ-45 or 8 pin mike connectors anywhere in the interface. FT-8 operation was plug-n-play and easily decoded signals in the -20 to -25 dB range. Receive audio response via the isolation transformers is flat on the spectral display. Also transmit power was stable when shifting around in the FT-8 window, and did not require constant readjustment. While this is not as definitive as a real frequency sweep of the audio transformers, it is definitely a convincing end to end performance check.

If you are looking for a digital interface for FT-8 or direct keyed (not audio) RTTY, you should consider the RigExpert TI-5000 if you don't have a lot to spend. In fact, if you are just beginning, I would skip the Signalink and go directly for the RigExpert TI-5000, for not much more cash outlay. The RigExpert TI-5000 is an exceptional bargain, if you do not need a K1EL Winkeyer.

Here is why I am not giving the RigExpert TI-5000 5 stars. I am a rookie contester and wallpaper chaser. I recently got WAS using my Signalink on FT8. I needed something better to up my game for DXCC. I also wanted to do CW contesting, using a digital interface for CW decode and sending pre composed N1MM exchange messages. The standard N1MM message syntax did not work properly with the RigExpert TI-5000 "winkey", since it is not a genuine K1EL Winkeyer. In particular, the ! syntax to send the other operator's call sign caused havoc that required me to reboot the program, or sometime tapping the keyer paddles would exit from the loop. I finally was able to get it running sometimes by inserting an E and a space before the ! syntax in the script for the N1MM function key message. At various speeds, it would send the E or not, and sometimes it just sent random garbage. Changing the speed in N1MM by the up/down arrows or using the control pot on the RigExpert TI-5000 caused problems, so I was unable to configure it so I could send messages at contest speed, or do a slower speed if the other operator had bad copy. Setting the "First Character Extension in msec" in the Configurer/Winkey tab to zero rather than the default 50 helped things, but did not fix it properly. The RigExpert TI-5000 internal proprietary "winkey" sometimes got confused about weight, speed, and farnsworth, even though N1MM does not have those settings. It affected even the keying paddle input to the extent I could not send manually. I limped through the recent CQ WW CW contest this way. This was a serious flaw in my judgement.

It might be possible to bypass the RigExpert TI-5000 "winkey" com port, and access a com port associated ONLY with the keying output. If you do so, uncheck the box in N1MM logger to "use Winkeyer" and select the correct keying output com port. I am not sure how it manages the stored messages activated by the N1MM function keys in that case. If you get it to work, it still is subject to the timing edge jitter caused by Windows 10. A true K1EL Winkeyer corrects those timing problems.

I wanted to verify my initial findings. I downloaded software utilites from the K1EL web site. I attempted to view the internal settings of the RigExpert TI-5000 "winkey". The K1EL utilities would not recognize the RigExpert TI-5000 "winkey". While the RigExpert TI-5000 "winkey" could function with keyer paddles, it could not do anything else I had intended.

My unique goals would have been met better if RigExpert had included a genuine K1EL chip rather than their proprietary implementation. It would have added less than $10 to the price to buy the chip from K1EL and use it on their board, as other manufacturers do. I was unable to get the stored messages in N1MM logger to send properly when using the keyer in the RigExpert TI-5000. It is the most popular freeware logger in the USA. The RigExpert TI-5000 claims to operate with paid software loggers popular in Europe. W8JI reviews one of these loggers and found the logger unsatisfactory, in eham reviews. I have been using N1MM for years, and did not want to migrate to a new logger which cost money and had questionable reviews, and did not support popular USA contests. Dealing with incompatibilies with my existing log files was also not good.

I have been able to disable the internal TI-5000 "winkey" in software. In Windows, go to devices and select the USB COM port associated with the RigExpert TI-5000 internal proprietary "winkey". "Disable" it. Do not "Remove" it or Windows 10 will try to reinstall it. I now use a separate genuine K1EL WinKeyer for CW, and it is compatible with all the software I threw at it, even keyboard sending programs other than N1MM. I now have access to all the rich K1EL Winkeyer features which allow control of weight, farnsworth, high speed shifts in sending boiler plate like 5NN, etc. It was more expensive to have to buy a separate K1EL keyer to complement a digital interface, but now I do have the performance I need. I have a step learning curve to take advantage of all the features K1EL offers, there are so many offered. As a bonus, the K1El WinKeyer will key some of my vintage tube rigs directly when in standalone mode. NICE!

DX Engineering, the US vendor for the RigExpert TI-5000, was in the loop on these findings. They offered to take the RigExpert TI-5000 back for a full refund, no questions asked. Excellent customer service. I elected not to take the offer, because the RigExpert TI-5000 performed so well in all other ways, and there was no other choice in my price range that met my criteria for no RJ-45 or 8 pin microphone connections. I had found a work around that I could live with.

The microHAM KEYER III now looks like the top of the line interface box, with a 24 bit audio codec. This would provide the best overload and noise floor. This is the ONLY digital interface to currently publish actual specs on its transformer audio passband response. It includes a genuine K1EL WinKeyer. It does violate two of my non negotiable conditions: it uses a RJ-45 mike jack on the back; it also uses the 8 pin mike connector on the radio for its digital mode audio. However, the microHAM and the Timewave Navigator seem to be the top choice of experienced DXers for "no cost considered" performance.