Sangean HDT-1X

The Sangean HDT-1X is an update of the HDT-1. This writeup covers the differences between the tuners. See the HDT-1 writeup for common properties. This writeup covers HDT-1X firmware version 0.2. At the end I mention changes for version 0.3. Version 0.4 exists, but I know nothing about it.

Outside

The HDT-1X enclosure and rear panel are black metal. I noticed that the finish texture was rougher than that of the HDT-1, and it seemed to have a slight odor. I didn't think any more about it. However, after testing the unit for several hours I noticed white marks and smudges on the finish. Some were located where equipment placed on top had mounting feet; others appeared where I had handled the unit. I wiped the stuff off and continued, but it kept reappearing as I moved the unit around in the lab. After a day of this I finally removed the cover and washed it inside and out with detergent and a soft brush. The photo above shows what happened when I ran my fingers across the finish after I had washed it. What is this persistent stuff?

The rear panel of a second HDT-1X was rough and exuded the mysterious white stuff. The enclosure finish was much smoother and the small amount of white smudging that developed with normal handling wasn't an issue. Since I haven't heard anyone else mention this problem, I suspect its incidence is low.

Inside

The only difference visible inside is on the power supply board. The white cable is the SPDIF digital output. In the prototype HDT-1X I evaluated, the AM RF amplifier on the right was enclosed by a metal shield with removable cover, probably intended to isolate it from the high-bandwidth SPDIF signal. The shield mounting holes and outline are visible on the PCB, but the shield was not installed in this production unit.

I noticed a weak heterodyne on a couple of AM stations. When I put my finger near the hot conductor of the internal SPDIF cable, it got much louder, as did the background noise on several other stations. With my hands away from the cable the interference was minimal though audible. I believe this very sparse, low-level interference will not bother ordinary users, but it may annoy AM DXers.

Features

The SPDIF optical output is new. It delivers 44,100 16-bit samples per second. I saw no sudden onset of distortion for an FM test signal deviated beyond 150%, which suggests plenty of digital headroom. Connecting both the optical and analog outputs to a Mitsubishi M-PF5200 preamp, the output level was about 10 dB lower for optical than for analog. The deemphasis error was considerably lower for the optical output (see the curves below).

With the HDT-1X you can force analog reception of an HD Radio signal. Analog often sounds better than digital on AM, and it may sound better for some multicast FM signals. Forcing analog also will keep the tuner from switching back and forth between modes when the digital signal is marginal.

You can now force monophonic reception of a stereo signal. While the tuner progressively blends the stereo channels as a signal gets weaker, a strong signal with co-channel or multipath interference may be unlistenable. Forcing mono can clean things up. Forced mono reverts to stereo whenever you change frequency (except with the manual tuning button) or cycle power.

Forced analog and forced mono are not saved independently for each memory preset.

The LCD backlight operates differently in the HDT-1X. When you switch to standby, the backlight goes off. This lets you sleep in the same room as the tuner. Pressing any key turns on the backlight so you can check the clock. The backlight goes off 15 seconds later. During normal operation an option automatically dims the backlight 15 seconds after the last keypress. Pressing any key brings back full brightness, after which it eventually dims again. This scheme solves all of the HDT-1 backlight problems.

The HDT-1X has a special split-audio mode where the analog signal appears in the right channel and the digital signal in the left. This feature allows a broadcast engineer to measure amplitude and delay differences between the analog and digital signals. It lets a consumer instantly compare analog and digital sound quality with an external switch (normal mode switching takes several seconds). Split-audio mode uses the left source channel and the audio output level is 8 dB higher than normal.

A cute dual-speaker icon indicates stereo reception. It works for both FM and C-QUAM AM stereo, which is very helpful.

The carrier to noise ratio display is even more useful in the HDT-1X because it displays values in tenths of a dB.

The HD Radio station ID display shows just two of the four numerical components the HDT-1 displays.

The bit error rate display still wanders near 0.5 on AM but now says INVALID on FM. This function is useful only to a broadcaster.

Frequency scanning is much slower in the HDT-1X. It takes about 50 seconds to scan the FM band and 28 seconds for AM.

The HDT-1X keyboard scan rate has been lowered so much that rapid keypresses are lost. This is very annoying when tuning manually, entering frequency digits, or even setting the clock. According to Sangean, the problem is due to an updated iBiquity software development kit.

A service manual for the HDT-1 can be downloaded here.

Analog FM

The HDT-1X uses the same stereo channel blending and lowpass noise filtering algorithms as the HDT-1, although the thresholds may differ. These automatic functions cannot be disabled. There is no stepped muting.

S/N at 44.2 dBf just before the onset of channel blending is 56 dB. Assuming the noise level changes linearly in this region, estimated stereo 50-dB quieting sensitivity is 38.2 dBf. I estimated mono 50-dB quieting sensitivity as the 56.3-dB quieting point, allowing a calculated 6.3 dB noise reduction for the lowpass noise filter that progressively engages below 29 dBf.

Adjacent-channel signals can affect stereo separation. For a 45-dBf signal (1 dB above the onset of channel blending for an isolated signal), a 100%-modulated 74-dBf signal at either adjacent channel dropped separation to 20 dB. With signals at both adjacents, just 52 dBf was needed. For a 65-dBf signal, a single 73-dBf adjacent reduced stereo separation 10 dB. A 92-dBf adjacent dropped it to 20 dB. I found a couple of stereo signals on the air well above the normal blend threshold that had blended to mono due to a strong adjacent. However, on one occasion I noticed that an unusual number of signals had blended to mono. Resetting the tuner restored normal operation, confirming a firmware bug.

A strong adjacent-channel signal can also muffle the sound. For a 50-dBf monophonic desired signal, the high end started to roll off as I increased a 100%-modulated adjacent-channel signal past 66 dBf. By 73 dBf, the response at 15 kHz was down 16 dB and the frequency response curve sloped downward above 1 kHz (it looked like a single-pole LPF with a 3-kHz corner). For left-only stereo, the whole curve dropped 6 dB.

Tests on weak signals scattered over hills with adjacent-channel interference revealed a remarkable advantage for the HDT-1X over conventional tuners using two 110- or 150-kHz ceramic IF filters. While all of the tuners were able to reject the adjacent-channel signals, the HDT-1X always yielded cleaner audio, often strikingly better with no audible distortion. I believe the high group-delay variation of narrow ceramic filters aggravates residual multipath distortion. The HDT-1X uses a digital IF filter with uniform group delay.

On other rare occasions, the HDT-1X audio was somewhat distorted when receiving a weak signal next to a strong one. A conventional tuner was undistorted. The cause is unknown.

For the following measurements I used IEEE 185-1975, modified as noted here. I used the test equipment listed here. See the HDT-1 review for additional measurements that should apply to the HDT-1X.

50-dB quieting sensitivity, mono    15.9 dBf (estimate)
50-dB quieting sensitivity, stereo  38.2 dBf (estimate)
THD, 1 kHz, mono                    0.02%
THD, 1 kHz, stereo L+R              0.04%
THD, 1 kHz, stereo L-R              0.2%
THD, 1 kHz, stereo L or R           0.25%
Stereo separation, 1 kHz            52 dB
AM suppression ratio                81 dB
RF intermod                         84 dBf  (97.7 + 98.5 -> 96.9)
RF spur                             87 dBf  (92.2 -> 96.9)
RF image                            74 dBf  (118.3 -> 96.6)
Modulation acceptance               150%
Channel blend onset                 44 dBf
Noise filter onset                  29 dBf
Treble response, analog output      +0.4/-3.6 dB
Treble response, optical output     +0.5/-1.3 dB  (uncharacterized D/A)
Bass response -1 dB, analog output  44 Hz
Output level                        0.7 V
Output impedance                    2.2kΩ

See the HDT-1 review for an explanation of the sensitivity measurements.

RF intermod, RF spur, and RF image are the 50-dB quieting levels for a third-order intermodulation product, an untuned signal, and a mixer image. I made these measurements in a way that sidesteps the phase-noise sidebands of the tuner and signal generators. For these measurements I used a 50-dB quieting level of 15.2 dBf, which disregards the lowpass noise filter response. It is not much lower than the 56.3-dB quieting level because of the large and frequent impulses that begin near the FM threshold. I averaged the noise level with the 5-second time constant of a true RMS meter.

The RF intermod figure is similar to what I've measured for other synthesized tuners with good front-ends. I had no trouble with RF overload at my location. But others with several FM transmitters nearby have reported severe overload with the HDT-1 and HDT-1X. The problem seems to be due to local-oscillator spurs or IF A/D overload, not ordinary front-end intermod. At signal levels above 80 dBf, the tuners respond to untuned single signals. Zero to several spurs occur and their pattern completely changes when retuning 0.1 MHz. RF spur is 9 to 33 dB worse for the HDT-1X than for other tuners I've measured.

Modulation acceptance is the modulation level of a 1-kHz tone for 1% THD.

Front-end alignment at 96.9 MHz dropped the signal level where channel blending begins by 1.6 dB, but 50-dB quieting sensitivity remained about the same. The factory FM alignment was much more accurate than for the three previous Sangeans I had checked. Compared with 96.9 MHz, sensitivity was 1.7 dB worse at 90 MHz and 0.7 dB worse at 106 MHz, tracking less well than previous tuners.

The sensitivity of a second HDT-1X improved 2.0 dB at 90.1 MHz, 1.3 dB at 96.9 MHz, and 1.5 dB at 106.1 MHz with realignment. 50-dB quieting sensitivity was 16.2 dBf at 96.9 MHz. It was 0.5 dB worse at 90.1 MHz and 0.3 dB worse at 106.1 MHz, tracking better than the first HDT-1X.

This compares the HDT-1X deemphasis response with that of HDT-1 version 1.2F (1.4F is the same as HDT-1X analog). The HDT-1X optical output fed a Mitsubishi M-PF5200 preamp (response unknown).

This is the distortion spectrum after alignment for a 1-kHz left-channel stereo signal deviated 75 kHz with 9% pilot. Note the absence of even-order harmonics even though I had adjusted the mixer transformer for maximum signal, not minimum distortion.

Analog AM

Sangean extended the −3-dB bass rolloff point for analog AM from above 200 Hz in the HDT-1 to below 50 Hz in the HDT-1X. This is a major improvement. However, while they were at it they also extended the high end.

The HDT-1X and HDT-1 sound very different on AM. Despite its wider frequency response, the HDT-1X sounds less natural to me. It has way too much treble and can sound shrill. The curves show that Sangean extended the high end flat without paying any attention to the standard NRSC-1-B AM deemphasis curve. This is a gross design error in a tuner intended for the U.S. market. (HDT-1 frequency response is for version 1.2F. 1.4F has the same response as the HDT-1X.)

The broad blue bulge peaking at 4 kHz shows the extent of the problem.

Here is one solution. This is the measured response with the analog outputs loaded by 0.027 F. The curve provides a close match to NRSC-1-B deemphasis to 4 kHz, with an error of −3 dB at 5 kHz. Switch out the loads when receiving FM or HD.

Here is an even better solution. This shows the BSR EQ-3000 octave equalizer settings needed to flatten the response to within 0.5 dB to 8 kHz. I modulated a signal generator with 40 simultaneous preemphasized tones from 250 Hz to 10 kHz and adjusted the equalizer for the flattest overall response.

This is the spectrum of the test signal to 8 kHz through the equalizer. The vertical scale is 1 dB/div. Set up this way the HDT-1X sounds remarkably good on AM.

The HDT-1X AM RF-amp varactor drive voltage is monotonic with frequency. It does not have the anomalous backsteps I found in HDT-1 version 1.2F. This was a welcome fix of a minor bug. Less welcome was what I discovered next.

The AM front-end alignment was way off. This graph shows how far down from the resonant peak the response was at the tuned frequency. I obtained this data by loosely coupling a tracking generator to the AM loop and connecting a spectrum analyzer with an active probe to the RF-amp output. The narrow front-end resonance was easy to locate and measure.

This shows how far the front-end resonance was offset from the receive frequency. In contrast, two HDT-1s each had a worst-case tracking error of 30 kHz. The factory alignment of one HDT-1 was spot on, while the other needed only a slight tweak.

Before I generated these curves, I had compared the HDT-1X and HDT-1 side by side on AM during the daytime and at night. I had not noticed any difference in sensitivity. After I discovered the tracking error, I connected the HDT-1X to an octave equalizer and matched its sound to that of the HDT-1 using a signal generator modulated by white noise. Then I compared the sensitivities of the two tuners by judging their relative levels of background noise, orienting each loop for best reception. The HDT-1 was noticeably more sensitive in the top part of the band, but only slightly so in the lower part. These observations are not entirely consistent with the tracking measurements reported above.

I measured the inductance of the HDT-1X loop as 12.9 H with a Q of 65 at 1 MHz. The loop supplied with the HDT-1 measures 12.2 H with a Q of 70. Using the HDT-1 loop with the HDT-1X did not alter the tracking.

Adjusting T301 to zero the tracking error at 600 kHz yielded the blue curve shown above. Midband and high-end tracking losses are 4 dB greater than those of the HDT-1.

A second HDT-1X had a worst-case tracking error of 55 kHz with 8 dB of tracking loss. The factory alignment was better than that of the first HDT-1X, but still improved with realignment.

When listening to AM with the HDT-1, occasionally I'd hear a distinctive click. The clicks were infrequent and not loud, but still somewhat annoying. I had never been able to locate the cause. Then I noticed the HDT-1X do the same thing on a fading skywave signal. Suspecting an AGC problem, I connected the HDT-1X to a signal generator with continuous control of output level and observed the waveform shown above while slowly increasing the signal level past several specific points. Usually I heard a thump rather than a click, but it clicked clearly if I caught it just right. I think I'm hearing the AGC make step changes in the gain of an amplifier with DC on its input. The horizontal scale is 100 s/div.

HD Radio

The HDT-1X and HDT-1 sound the same and behave identically on HD Radio signals.

I measured the peak audio amplitude of several HD Radio signals as 1.5V. Then using a 20015,000 Hz bandpass filter per IEEE 185-1975, I measured the residual noise level while the tuner tried to lock to an HD-2 channel with no antenna connected. The noise was 85 dB below the RMS level of a 1-kHz sine wave whose peak amplitude was 1.5 V.

Version 0.3

 1. Pressing a single digit recalls a memory
 2. Pressing a digit for two seconds instead of five overwrites a memory

HDT-1X 0.3 vs HDT-1 1.4F

 1. SPDIF optical output
 2. Forced analog
 3. Forced mono
 4. Split-audio mode
 5. Backlight off in standby
 6. Backlight can be dimmed
 7. Stereo indicator
 8. Minor internally generated AM interference
 9. Stereo channel blending begins at 44 dBf instead of 50 dBf
10. No stepped muting

Issues

 1. Severe FM intermod may occur when transmitters are nearby.
 2. AM response does not follow standard NRSC deemphasis and can sound shrill.
 3. Loss of signal for more than a few seconds retunes HD-2 to HD-1.
 4. The FM analog-output passband droops and is 3.6 dB down at 15 kHz.
 5. The clock resets if power is removed.
 6. The AM synchronous detector is not always in lock when the tuner unmutes,
    occasionally loses lock during a selective fade, and will not lock to a weak
    signal when the signal on the adjacent channel is strong.
 7. The extended-info selection is lost when tuning to a new frequency. The
    display reverts to regular info.
 8. Extended-info parameters are not in most-used order.
 9. It should not be necessary to press PRESET to recall a frequency. Any press
    of a digit shorter than 2 seconds should call up the station. [0.2]
10. The distinction between a memory being empty (press 2 sec to store) or full
    (5 sec to overwrite) is unnecessary. The memories will eventually all fill.
    Instead, store the memory after 2 sec, overwriting whatever's there. [0.2]
11. Occasionally the tuner crashes, emitting only chugging noises.
12. 0.1-MHz steps on FM slow automatic tuning and make manual tuning awkward. 
13. When tuning manually, the tuner steps through multicast channels even when
    not attempting to lock to HD. This slows tuning and upsets the two-push
    cadence necessitated by 0.1-MHz tuning steps.
14. SSI indicates maximum except for weak signals and does not reflect RF level.
15. The clock is not synchronized with the line frequency. Eventually the time
    becomes inaccurate.
16. AM tracking was off as much as 85 kHz with 11 dB loss. After alignment
    tracking loss was as much as 4 dB greater than for the HDT-1.
17. The tuner does not retain the last station tuned when power is removed.
18. When power is restored, the tuner comes up in standby. This makes it
    impossible to power up an entire stereo system with a single switch.
19. The polarized power plug does not fit the unpolarized AC outlets on older
    stereo equipment.
20. Reset may be required to prevent unnecessary blend to mono.
21. The frequency scan rate is slow.
22. Forced analog and forced mono are not saved for individual memory presets.
23. Although likely inaudible, FM stereo distortion is rather high.
24. Split-audio mode uses the left source channel instead of left plus right.
25. The output level in split-audio mode is 8 dB high.
26. A weak heterodyne due to SPDIF is audible on some AM frequencies.
27. The LCD font is rather primitive (e.g., raised descenders).
28. AM intermod can occur for nearby transmitters, mostly noticeable between stations.
29. Occasional clicks are audible on AM. They are infrequent and not loud, but
    annoying when they occur. They coincide with discrete AGC level changes.
30. Rapid keypresses are lost.
31. SSI drops when there is audio near 15 kHz on FM and varies wildly with any
    modulation on AM. The display is ambiguous at best on FM and useless on AM.
32. When in split-audio mode, the tuner receives no signals after changing from
    AM to FM unless you cycle power.
33. A large internal analog/digital delay offset may occur if the tuner loses HD
    lock after tuning to a new frequency in split-audio mode. Cycle power to cure.
34. On AM the display sometimes does not indicate HD when receiving it.
35. The -1-dB bass response frequency of 44 Hz is rather high.
36. There is no way to disable the automatic stereo channel blend or the
    high-frequency noise filter.

August 28, 201388108 MHz