Paul,

An undersampling system needs some kind of sample & hold circuit. IIRC from
digital audio, the aperture time should be less than T/10 (T=sampling
period) in order to keep the high frequency rolloff below 1 dB.

Yes, the figures sounds reasonable.

So if we assume a hypothetical 300 kHz sampling rate (for 0 .. 140 kHz BW),
the aquisition time should be less than 300 ns, which should not be too hard to
implement:-). Since the hold time is about 170 us (6 kHz), we are talking
about 1:500 aquisition/hold time ratios, which can be a problem due to hold
time leakage.

Yes, the 1:500 ratio could turn the S/H "droop rate" into a problem.
I was thinking about using a cheap 74HC4066 or similar analog switch as sampling
gate but after giving it a second thought I realised that it is probably out of 
the question
due to the high "charge injection" from the control pin to the switch i/o pins.
The hold cap' has to be quite small in order to meet the acquisition time 
requirement
so the control pulse leakage will probably upset things... Assuming S/H driver 
source
resistance + switch R-on = 50 ohms, C has to be about 10nF for -3dB at 300kHz 
but...
At an aquisition/hold time ratio of 1:500, C must be <= 20pF. Among other 
things, this
calls for a buffer amplifier having almost zero bias current :~)

Do you have any suggestion for a hombrew S/H circuit? Diode ring gate? Ready 
made
S/H chip? I have no experience of building fast S/H circuits.

By adding dithering noise, low and medium frequency tones could be reproduced
well below -100 dB, when the teoretical limit for a 16 bit system would be -96 
dB.

Yes, isn't it funny...The system performance can sometimes be improved by ADDING
noise! The dithering noise spectrum has to be outside the wanted passband of 
course.

If some super-selective front end is used (say 250 Hz bandwidth) in the LF
receiver, the digital signal may contain signal related quantisation noise,
thus some form of dither noise may be required. However, if the analog

bandwidth is larger, I guess that the band noise may be used as dither noise.

Yes.

Steve VK2ZTO did some extreme narrowband experiments on longwave using a 1-bit
ADC. I think it was a simple comparator hooked on the RX audio output. Dithered 
by natural
band noise, the signal appeared to have many bits of resolution after the 
decimation.

Fs=6kHz seems like a good choice. The selected Fs harmonic should be just 
outside
the band of interest (lowest acceptable beat note) and the preselector BW 
(-many dB!)
must be less than Fs/2. Pse correct me if I'm wrong.

From this description, it appears as if this device is actually a 13 bit

flash ADC and the three extra bits are generated by the Fs/8 downsampling
:-).

Right. It's not a true 16-bit ADC. It is some kind of combination of a 
sigma-delta
with 5-bit flash & 5-bit feedback and a 12-bit pipelined ADC.

Some data:
- Fs <= 20MHz
- 1 MHz signal BW
- SNR 89 dB
- THD -98dB
- SFDR 100dB
- Equiv. input noise 0.6 LSB
- Stopband attenuation 85 dB

Unfortunately, my preliminary data sheet doesn't tell the ENOB (effective 
number of bits).

I actually have two of these chips and a HSP50016 DDC too. I hope that I 
someday will
find some time to play with them...

Anyway, the harmonic sampling idea is *very* appealing.

73
Johan SM6LKM