General > General Technical Chat

Review: Hantek DDS 3X25. Anyone own one?

<< < (88/108) > >>

marmad:

--- Quote from: zibadun on December 17, 2012, 10:10:57 pm ---Thanks, I think I understand the formulas and arithmetic behind them. But  I don't see how I can avoid downloading a new waveform every time I need to change the frequency.  Even if I'm just changing the number of points with the same wave buffer don't I still need to call DDSDownload?   It seems transfering of the buffer takes up most time.   May be I will try with the minimum number of points and periods that will give the desired frequency and save on the transfer time a bit.
--- End quote ---

I don't think you're getting the formulas. Once you've downloaded a waveform, you can change the frequency by using the DDSSetFrequency command, as noted in the code at the bottom of the post from the second link. But without using the DDSDownload command, you're going to be limited to what frequencies you'll be able to get (as noted in those posts). And because the Hantek uses a clock divider - the number of frequencies you can hit by just changing the clock rate (without downloading) gets smaller as you get closer to the maximum DAC clock of 200MHz - it's just a function of math.

As far as speed goes, the speed at which you can change the frequency will be limited by your interface/software/OS/computer. If you want to do 600 steps in a minute, that means 10ms 100ms per step - which would probably be close to the maximum. I think I was able to get that speed with tests I ran. The Hantek software accomplishes sweeps with faster speeds - but they're definitely using commands that they haven't documented to the general public (plus they don't allow sweeps above 100kHz). Boy, I need food :D It's easy to get that speed!

torch:
Using Goltek, I set the start at 1Mhz, the end for 30Mhz, step at 50Khz, interval at 1ms and single-cycle. I hit "run" and the stopwatch at the same time. 15 seconds to complete one cycle -- on the laptop with USB2. On the older slower laptop with USB1.1, the identical setup took 1 minute, 55 seconds.

marmad:

--- Quote from: torch on December 17, 2012, 10:44:59 pm ---Using Goltek, I set the start at 1Mhz, the end for 30Mhz, step at 50Khz, interval at 1ms and single-cycle. I hit "run" and the stopwatch at the same time. 15 seconds to complete one cycle -- on the laptop with USB2.
--- End quote ---

That's ~25.8ms per step.

Edit: Clearly my brain needs food (I've delayed my dinner) - my post above was wrong by a factor of 10 :o He needs 100ms per step, duh - so well within reach by Goltek or any software.

marmad:
Here's an image showing a big problem with sweeping with the Hantek.

The image shows the frequency change point from 200Hz to 300Hz - but because the frequency doesn't change at the zero-crossing of the waveform, you get a spurious frequency injected - in this case, it happens to be ~287Hz - as shown in the cursor measurement window.

Edit: BTW, I was incorrect in what I posted in a previous message. I did some more testing and it turns out that the Hantek software isn't capable of doing smooth sweeps either. Honestly, with the various problems inherent in the device (no zero-crossing of frequency change, no sync signal for sweep, etc, etc) I really don't think it's a good thing to use for sweeping - especially above 5MHz - but perfectly fine for basic signals and simple arbitrary patterns.

zibadun:

--- Quote from: torch on December 17, 2012, 10:44:59 pm ---Using Goltek, I set the start at 1Mhz, the end for 30Mhz, step at 50Khz, interval at 1ms and single-cycle. I hit "run" and the stopwatch at the same time. 15 seconds to complete one cycle -- on the laptop with USB2. On the older slower laptop with USB1.1, the identical setup took 1 minute, 55 seconds.

--- End quote ---

thanks for the stopwatch test torch. wow something is really wrong with my setup. The same scan took over five minutes here.  I'd be happy with 15 secs. 

update:  looks like a bad port/hub on my pc.  i switched the cable to another port and got 50ms instead of 500 to switch frequencies.  32 seconds on goltek.  Not as good as 15 but much better than before!  hmm I thought usb port is a usb port ;)


--- Quote from: marmad on December 17, 2012, 11:41:22 pm ---The image shows the frequency change point from 200Hz to 300Hz - but because the frequency doesn't change at the zero-crossing of the waveform, you get a spurious frequency injected - in this case, it happens to be ~287Hz - as shown in the cursor measurement window.

Edit: BTW, I was incorrect in what I posted in a previous message. I did some more testing and it turns out that the Hantek software isn't capable of doing smooth sweeps either. Honestly, with the various problems inherent in the device (no zero-crossing of frequency change, no sync signal for sweep, etc, etc) I really don't think it's a good thing to use for sweeping - especially above 5MHz - but perfectly fine for basic signals and simple arbitrary patterns.

--- End quote ---

isn't this a transition problem? like a DDS settle time which equals some N number of samples.  I've seen this in a digital mixer in another device. every time you retune the mixer there was all kinds of garbage injected to the signal,but it lastsed only a few clock cycles.  It may even be specified in the datasheet..  In some cases you may be able to account for that transition in whatever is on the "receiving" end of the sweep..  At least that is what I'm hoping to do in my project...


Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod