0
\$\begingroup\$

Yes, another question about this DDS. However, I think mine is sufficiently different from the others I've read.

I purchased an AD9850 board on Amazon for $17 from "NooElec" (see here). I've been testing it out tonight, and it operates very strangely. I've read where these AD9850 PCBs (often made in China) sometimes use fake chips or have bad designs. Mine seems to have shipped from the US, but its country of origin is questionable.

I'm given to understand that the DDS output attenuates at higher frequencies. However, I've been playing with my board hooked to my oscilloscope, and there is gross attenuation output for frequencies as low as 5 MHz. (The chip itself is designed for 60+ MHz). In fact, the output voltage increases and decreases repeatedly with increasing frequency. Some sample points:

  • 100 kHz: 1.13 Vpp
  • 500 kHz: 1.08 Vpp (same)
  • 1 MHz: 1.1.08 Vpp (same)
  • 2 Mhz: 1.15 Vpp (slight increase)
  • 3 MHz: 1.38 Vpp (mega increase)
  • 4 Mhz: 1.32 Vpp (slight decrease)
  • 5 MHz: 0.975 Vpp (mega decrease)
  • 6 MHz: 0.660 Vpp (mega decrease)
  • 7 MHz: 0.220 Vpp (mega decrease)
  • 8 Mhz: 0.400 Vpp (mega increase)
  • 9 Mhz: 0.400 Vpp (same)
  • 10 Mhz: 0.150 Vpp (mega decrease)
  • 20 MHz: 0.340 Vpp (mega increase)
  • 30 Mhz: 0.150 Vpp (mega decrease)
  • 60 Mhz: 0.150 Vpp (same)

This is the only frequency synthesizer I've used (except for some lab-quality function generators). Should I be expecting this kind of weird behavior from a AD9850 board? I understand the output can attenuate, but can it really attenuate so much so quickly and then increase again? Or do I just have a bad board?

Note that I'm not using filters on the output. If you're curious, I'm operating the board with a Raspberry Pi 2 and code from Tom Herbison.

\$\endgroup\$
7
  • \$\begingroup\$ What kind of wave is it? Sine? And how are you connecting it to the scope? \$\endgroup\$
    – uint128_t
    Jan 26, 2016 at 6:19
  • \$\begingroup\$ It's a sine wave. I hooked it to the scope several ways: direct connection to out pin, in series with a capacitor on the out pin (of the size recommended by Tom), and on the capacitor with a resistive load on parallel with it forming roughly a 70 MHz LPF as recommended on one web page I'd read. \$\endgroup\$
    – wisner
    Jan 26, 2016 at 12:43
  • \$\begingroup\$ Looks like a rather long and un-terminated cable connection to the scope, with several reflections. \$\endgroup\$
    – user16324
    Jan 26, 2016 at 12:55
  • \$\begingroup\$ Unterminated? The RC filter was terminated. I mean, I could try putting a resistive load I'm series with a DC-blocking capacitor. (And terminate it, of course.) \$\endgroup\$
    – wisner
    Jan 26, 2016 at 13:02
  • \$\begingroup\$ How would you suggest testing it? I know the concepts, but frequency synthesis is new to me. \$\endgroup\$
    – wisner
    Jan 26, 2016 at 13:05

2 Answers 2

2
\$\begingroup\$

I use the AD9850 board from NooElec, purchased on Amazon. However, the unit I bought was described for up to 40 MHz. I used it to build Antenna Analyzer

enter image description here

I measure voltage at the junction of C1 and R2. There is a Low Pass filter on the 9850 board. So I am measuring after the filter. On the 9850 board, there is a 200 ohm resistor to ground before the LP Filter, and also another 200 ohm resistor to ground at output of filter. This contains schematic of what I believe is on my 9850 board.

I measure + 0.300 volts peak, and - 0.300 volts peak (.6v peak to peak) very flat from 1 MHz to 30 MHz. Then voltage drop off begins at 30 MHz up to 40 MHz. I don't see the wobbling output similar to yours from 2 to 20 MHz.

Not sure if this qualifies as an answer, but too large for a comment. But at the least it might give you a comparison to your results.

EDIT 1 : The output of the AD9850 IC (not the board) is a current, so output voltage will be dependent upon the resistor you choose to use as a load on the AD9850 output. "Pin 21 IOUT Analog Current Output of the DAC"

\$\endgroup\$
10
  • \$\begingroup\$ 1) A homebrew antenna analyzer? AWESOME! \$\endgroup\$
    – wisner
    Jan 26, 2016 at 16:15
  • \$\begingroup\$ 2) This actually does help me quite a bit. Even at 10 MHz, I can't get a steady sine wave output but rather a funky-shaped wave with other waves visibly inside of it and an FFT showing harmonics out the wazoo. I tried this morning to measure the wave across various capacitors and a 200 Ω resistive load in series. I'm aware the output is current and not voltage, but even adding this load doesn't help me. \$\endgroup\$
    – wisner
    Jan 26, 2016 at 16:20
  • \$\begingroup\$ 3) Based on your seemingly easy use of the board and my difficulty getting it to even put out a pure-ish sinusoid, I'm concluding it's defective. \$\endgroup\$
    – wisner
    Jan 26, 2016 at 16:21
  • \$\begingroup\$ 4) It's worth noting the following new piece of information in the description of the NooElec Amazon listing: "After the 20-30MHz frequency harmonics increases, the waveform will be less and less clean." \$\endgroup\$
    – wisner
    Jan 26, 2016 at 16:25
  • \$\begingroup\$ Not sure what you mean by "resistive load in series". If you are measuring at pin 21 of the AD9850, the resistor would be from pin 21 to ground. \$\endgroup\$
    – Marla
    Jan 26, 2016 at 16:26
1
\$\begingroup\$

I've done a similar test with mine and I get a 1 V peak to peak voltage up to 10 MHz. I want to use it as a VFO for ham radio HF bands, so far so good, but its output voltage drops by half around 24 MHz. The DDS module crystal is supposed to be 125 MHz. The doc says that output amplitude is affected by a function of both output and clock frequency: \begin{equation} y = sin(x)/x \end{equation} where: \begin{equation} x = \pi (f_{out}/f_{osc}) \end{equation} However, the calculated attenuation factor is 0.94 at 24 MHz, not 0.5 on both filtered and unfiltered outputs! Something is wrong with these bargain DDS modules for sure...

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.