Contacter l'auteur / Contact the author

Recherche dans ce site / Search in this site



The Kenwood TS-570D(G) transceiver

DSP functions (II)

DSP features are the heart of the TS-570D transceiver - as of all recent transceivers - and have greatly anticipated its reputation. DSP gather complex digital functions we usually find in dedicated and external device that may cost 5 times the price of the most expensive transceiver (the price of a good car). But for some years, due to the higher integration of electronics components, radio amateurs manufacturers were able to include this technology in most of their digital receivers.

TS-570D explores this world in depth while being limited to its category, and the marketing has been so well organized that many amateurs bought first this rig because of its DSP features. Therefore these functions need a deep coverage to evaluate their performances. As you already understand, there is a gap between what the publicity claims and the reality in the field.

One day, in the summer, it happened that in the afternoon I switch off all digital filters as the QRN disappeared in the 20-meter band. For the first time in three months there was no white noise, no interference, nothing except the pure signal. That was a rare occasion to feel exciting sensations; imagine speaking with or hearing directly your correspondent without the least interference ! A great moment I remember for long. Usually I encounter such conditions a couple of time each year.

Unfortunately our HF environment is rarely so quiet. Listen on any frequency on weekends and you will understand what I mean... (see these waterfall spectrum) But don't blame amateurs against this brouhaha as this loud activity is rather a sign of good health for our hobby. However it is sometimes accompanied by QRN, atmospheric, solar and geomagnetic disturbances that generate many noise on all frequencies. So it is very important that numeric systems be able to reduce or even suppress all these noises, natural and artificial in order to extract and amplify the Q signal.

At left the reponse of the 7 MHz bandpass RF filter that improves intermodulation rejection. A right the result of noise reduction DSP filters N.R1 (SSB) and N.R2 (CW) Documents Kenwood.

In short the DSP is used when you need to tailor received signals, e.g. rejecting interferences, or when you transmit audio and want to reshape your signal. Complementary, DSP is also able to perform spectrum analysis, extracting signals lost in the noise in a wide range of frequencies or timescale. The famous SETI project for example from Berkeley University is one of the applications of this technique.

But it is not necessary to go so far, hams fan of EME activities, signal processing or satellites tracking work sometimes the same way. And near to you, the TS-570D provides similar functionalities too ! However the TS-570D is not equipped for spectrum analysis but software like CommCat can perform this task when connected to the serial interface of the transceiver.

A 16-bit DSP circuit on the AF-stage

How works the DSP features of TS-570D ? This transceiver uses a DSP that works with a well-known 16-bit circuit. In a high-end transceiver, the DSP is considered as the last IF-stage in the receiver chain (e.g. the 4th IF), but it is purely software. In the TS-570D serie the DSP is placed on the AF-stage : this is equivalent of placing a DSP filter on the output jack of your speaker.

Although globally the DSP provides extensive control over signals whatever the mode use, SSB, AM, FM, CW or FSK, its performances are limited due its position in the chain. However the result on the audio is not attainable by analog devices, and it is thus a serious advantage on models that don't use this technology. 

In fact the functions of audio/HF filters are of the uttermost importance. Without them this is like if you listening frequencies is a very noisy environment. You hear all sort of parasits excepting Q signals from stations : static, QRN, RFI, etc. When those filters are enabled, suddently the silent arises and the signal appears... Yes, it is fantastic ! It is fantastic... for an audio level DSP. It is even better than many external devices (e.g. MFJ serie).

But there is one major problem with this kind of solution. A performing receiver must intercept the interference before it reaches the detection and mixer to get the highest selectivity as possible. With TS-570D this solution is impossible to achieve; serie signals have to pass through the all chain of reception before being processed by the DSP.

Take a simple example to well understand the problem : imagine a strong station from southern Europe developping a 20 dB signal at 0.2 kHz up from the most wanted DX emitting a weak signal at S3. The automatic gain control, AGC, working behind the IF-stage will reduce the overall gain because of the strong station. Result, you will no more hear the DX because this time it is lost in the background hash. In other words the DSP is useless...

So we can wonder why Kenwood and M.Torri in particular, the designer, choosed an AF filter in place of an IF one ? For the same reason that you have choose a TS-570D rather than a TS-2000 : because is is cheaper !

Hopefully contrarily to other so-called performing receivers, like Elecraft K2, the first IF-stage of TS-570D is at 73 MHz, far outside amateur bands (in K2 for example there are a lot of images right in the 15 and 17m bands...).

So what to do to improve the DSP ? Logically select a DSP suited to your activity (Kenwood states that it is, but we know now that there are limitations). There are several ways to improve the signal that you receive : install either optional filters in your transceiver (e.g. the 1.8 kHz SSB filter in the TS-570D) or, as a second choice,  install external Collins filters or even an external DSP.

The other solution is to select another RTX which receive module is equipped with a DSP on the IF-stage completed with several filter controls (ATTN, AGC, IF-shift, variable BFO, Notch, RIT, etc) on the front panel and optionally a DSP speaker. But at what price ? Usually you enter into the high-end class transceivers... This is like selecting an external speaker or a microphone : do you select an Electret mike in the Kenwood line of products, a Heil Sound or a high-end using a polyethylene teraphthalate film dynamic element ? Personally I should select the third, but for financial reason I selected the first...

Filtering : DSP or Audio stage ?

If you are only interested in CW, due to the short bandwidth of the tune, a filter placed on the audio stage is more than enough to get good results. Autek QF-1A is for example an excellent product.

But a DSP doesn't work the same way, and in SSB it removes automatically carriers that fall in the current passband, and this without using a notch circuit... a great improvement.

The passband of a DSP is also programmable to get abrupt wings and short skirts contrary to an audio filter that displays a bell shape. The first is thus "better".

At last, the elimination of noise is ruled by signal processing protocols (maths), whereas an audio filter depends of its bandwidth... which affects the audio and therefore, it is not always suited to the mode used.

In selecting a filter, keep in mind these reference parameters when comparing performances of your filters.

Generally speaking we can however say that a DSP is more versatile as it extend to most amateur traffics.

Cut-off frequency control

Beside the standard IF Shift function that rejects interferences in shifting the center frequency of the filter passband, and CW functionalities that we will see later, TS-570D comes with a true DSP slope for SSB, FM or AM modes.

This button provides two options, High and Low to modify (raising or lowering) the level of cut-off frequency of received passband. The High control for example, that you will probably set as soon as you will switch on the rig, shifts the cut-off frequency by step of 0.1 kHz from 1.0 to 5.0 kHz. I personally shifted it quite high near 4 kHz to get a clearer SSB signal. I reduce its passband to about 2.5-3 kHz when there are too much QRM in order to only isolate the stations signal.

In spite of limitation of this system, that helps sometimes. But when the low bands are crowded, this passband is still too wide. The Low control works by step of 50 Hz from 10 to 1000 Hz and is set by default at 500 Hz. This is here that notch should be very useful in addition.

In CW or FSK mode the bandwidth is changed using two other buttons, FILTER and MULTI CH. Turning this latter clockwise you wide the bandwidth and vice versa. As we will see later, by default the CW DSP filter bandwidth is quite wide, 600 Hz. Optionally you can install a narrowest filter (ref. YK-88CN-1) providing a bandwidth as small as of 50 Hz. A very narrow CW filter is very interesting to work weak tone signals or to extract them from the background noise. Even listening to beacons, a bandwidth of 100 Hz is very appreciate.

Above the DSP slope reduces interferences in changing the receive bandwidth. Below the Beat Cancel attenuates cyclic interferences.

Noise reduction control

The DSP slope is completed with several noise reduction capabilities named N.R1 for SSB/AM and N.R2 for CW. Basically these filters reduce random noise that interfere with the signal you try to capture.

Unfortunately, as we told, there is no manual notch on this transceiver and you cannot select your level of filtering. This option being requested by many amateurs, TS-570D was upgraded to model "DG" or "SG" (price $125 at that time) with a new built of the software providing a more accurate control in gradual steps. Today this upgrade in included from factory.

Using Menu #51 you can adjust the noise level to get a better sound, opening up the RX audio to a full 0-5 kHz bandwidth when conditions allow. The fact to be able to tailor the N.R1 feature in 9 steps is very exciting ! The same option is available for CW. However, results are far to be perfect.

Beat Cancel

Last but not least, the Beat Cancel filter (B.C.) offers you a very appreciated help in attenuating or suppressing any repetitive interference, or beat, hence its name. Technically speaking an adaptive filter is activated within the receive passband, changing its characteristics according the nature of the signal at a given moment. When switched on, in a fraction of a second the jam is reduced and the signal suddently raises. Fantastic ! Like the N.R filter, the B.C. should be ON permanently when there is some noise. These two filters are probably one of the more powerful functions of this transceiver.

Spectrum analysis (FFT) of signals before/after activating the N.R (left) and B.C. (right) filters. Clic on images to enlarge. Documents JA6QXW

In conditions where QRN is heavy, the sun well above the horizon or when the thunder strikes near your location sending static everywhere, the RX continues to extract weak signals, but of course with much more difficulties. When signals become too weak or too noisy, there is no miracle, the DSP reaches its limits and you miss your contact.

But I can say that I recently hear the amateur radio club of Shang Hai (BY4BZB) for example while other stations could not hear the operator using another model of RTX, proof of the excellent performance of the DSP. But I didn't say that TS-570D was able to work stations that were not readable by other receivers. All depends on your working conditions, especially your antenna properties, performances of your receiver and propagation conditions.

DSP filters

TS-570D accepts optionally other filters. However you have only one slot available at a time and you cannot for example install a narrow CW and a narrow SSB in the transceiver at the same time. But thanks to the built-in DSP, the use of these narrow Inrad CW (500 or 270 Hz) or SSB (2.1 or 1.8 kHz) filter is not mandatory, although it could improve the selectivity.

In fact the DSP CW filter for example can be set all the way down to 50 Hz and is nothing but useless ! At the narrowest bandwidth the sound is still good, with a slight tin-can sound but without echo. Amazing ! A similar filter is available for SSB. If you want to install it I warmly recommend you to select the narrowest one too, 1.8 kHz.

Documents JA6QXW

CW mode

As we told before, TS-570D comes with two input jacks at the rear to digitally processing signals from a straight keyer or a paddle. More complete than some other rigs, this one is able to manage itself dots. In the upgrade to model "G" dots/dash weighting is also customizable via Menu #26, making your keying easier for other operators to copy.

The changing bandwidth of the CW/FSK filter.

The DSP provides also a very appreciate feature, the automatic CW TUNE that scans the current band from up to down in search of tones and it adjusts the frequency according your preset pitch. However, at several occasions, in presence of QRM, it missed the frequency within a few Hz or made a jump of 5 kHz to select the noise ! In fact it works fine when the CW tune is strong but miss detection as soon as there is QRM or noise.

As we told earlier the DSP provides also a customizable CW bandwidth down to 50 Hz and a bandwidth control via the FILTER + MULTI CH buttons that help in rejecting interfering signals. 

This filter working also in FSK mode makes crystal filters in the IF almost unnecessary, and the CW receiver is as good as any I have seen for tens of  years.

The N.R2 filter is dedicated to CW and performs very well. It uses the SPARC method to extract CW signals buried deep in the noise, and that works ! In fact it helps in clearing the static and makes the CW tone very clear with very little background noise. All CWers fans say that they appreciate much this function and still better when upgrading to model "G" in which both N.R. filters correlation time can be adjusted to reduce the level of noise according the band quality. It can be processed manually via the Menu or set automatically. However it is recommended to set a short correlation time in SSB to hear signal with more clarity and a longer correlation time in CW to allow a more reliable reception.

Minus side, the LCD displays also the key speed but seems optimistic, 30 WPM when you work at only 22 WPM... Some users observed erratic keying, with spurious characters being sent. This event being not often noticed it could be due to dirt on the paddle or a lack of screening on the paddle lead causing RF pickup. If that occurs again, try another paddle.  

At last the CW memories are short, storing a maximum of 50 characters in three memories, nothing more than a CQ call. This problem is however balanced with a quick access. Indeed all three CW memories can be activated at the press of a single button and be played in sequence.

The effect of the attenuation (ATT) on  signals. Its activation allows a better selectivity when signals are close to the QRM level.

Completed with the preamplification switch (which of course amplifies both the signal and noises as it is not placed at the antenna output), these digitals filters allow you to hear with more or less ease weak stations that transmit close to the noise level.

However, in condition of heavy QRM like in contest or during pile-ups, I regret that Kenwood's engineers haven't equipped this mid-range transceiver with a custom NOTCH filter, a more sophisticated AGC/TONE control that could be switched off instead of pumping all signals from adjacent powerful stations, and more filterings, so much features that are very useful to discriminate weak signals or signals flood in QRM, and that we find to its competitors.

For the loudest signals or RFI interfering with an intelligible signal, you have the possibility to use the Attenuator (ATT) as well as to reduce the RF audio gain to reduce the level of QRM while keeping the signal over the hash. 

As many settings, this information is stored automatically in the current band. But as this attenuation practically reduces all received signals and is not associated to a DSP on the IF stage, I presume that this option is rarely used, excepting for ragchewing in local QSOs between stations arriving 59++. I suggest to Kenwood to replace this ATT with another feature, more useful, like a custom DSP NOTCH filter.

At last spurious emissions are reduced up to -50 dB, an excellent value, but carrier and sideband suppressions are reduced up to 40 dB below peak output what is up to 37% lower that its competitors, some of them reducing modulation up to 55 dB or more.

The frequency stability and accuracy of TS-570D is far to be excellent, and is 20 times lower than its competitors (10 ppm vs. 0.5 ppm at room temperature). Its sensitivity reaches only 4 mV instead of 0.16 mV in SSB to its competitors, knowing that 0 dB = 1 mV. However in practice these limitations do not much influence ragchewing but will have some effects working DX stations arriving with a weak signal.

Last chapter

Page 1 - 2 - 3 -

Back to:


Copyright & FAQ