Feedback for version 0.9a here

Information and discussions on Spectrum Analysis
SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Fri Aug 17, 2018 1:08 am

DaveR wrote:Some good news and some bad from me: the second AD9851 module I ordered arrived from China this morning, and it looks really nice - 3 times the size of the usual boards, with four SMA connectors, which would make it a very good board to build into a box as a small signal generator. Unfortunately, the board itself was obviously originally designed for the AD9850 chip, and has a 125MHz oscillator fitted instead of the required 30MHz one!! Consequently, due to the mismatch, the module will not work as either an AD9851 or AD9850.
That is a pity. The board looks very nice, project-case ready as you say. It looks as it would be easy enough to create your own schematic from the board layout. We can only hope that the wrong crystal was fitted in error and that they can send you a board that has the correct oscillator fitted.
DaveR wrote:Have you actually managed to get a Nano board fully working with Trackgen?
I'd be very interested to know that as well.
DaveR wrote:In theory, the Nano should work the same as an Uno, but I can't get mine to be recognised by Trackgen no matter how I wire it up. I can set any default frequency in Steve's code before the sketch is uploaded, and the output is correct - so I know that the Nano board works using Uno code - but I can't get the red square in RSP-SAS to turn green, so can't control the Nano afterwards.Is there something obvious I'm missing, Steve, or will another mod to the code be necessary to allow the Nano to be used?
A wise man once said, "In theory, theory and practice are the same thing, in practice they are not" ;)

I can't think of anything obvious, certainly nothing in the code should need to be changed. Having said that, I've noticed that sometimes when the analyser is first launched, after my PC has just booted, the Trackgen comms port will not select (turn green). The Trackgen comms port needs to be selected, then de-selected by selecting another comms port, then re-selected, then I get a green light. I've noticed this behavior before with another project, and it seems to be related to the Microsoft comms driver. It appears that once my PC is booted, the driver ignores the first block of data received - from anything. This 'feature' is mentioned in the Microsoft comms port documenation. I've addressed this problem for the next release.

The only code differences are the clock frequency that has to be set when the code is compiled for an AD9850 or an AD9851. Perhaps there are differences in the libraries used between Arduino modules. Unfortunately I don't have a Nano here to test code with. I'll recompile the code for both the AD9850 and AD9851 for the Nano board and post a zip file here later. I have to go out for the rest of the day, so it will be around 8 hours before I can get that done.
K9AQ wrote:My AD985-'s appear to be same design. I attempted to attach a photo but the file was too big. I will have to play around with the file to make it smaller.
Dave posted pics of two type of module, perhaps you can just make a copy of the pic from the webpage, or indicate which of his pics match the module you have. I'm just curious to see if your module is different to the modules Dave and I are using. BTW - I'm still waiting for my AD9850 module to arrive - might be time to chase up the supplier I think.
K9AQ wrote:Like Dave, I would love to have the trackgen software working with a Nano. For this application you don't need any more IO.
I agree, anything else is wasted. I only used the Mega because it was all I had available at the time. I'll get a Nano on order so I have something I can test future code with.

Reason: No reason

K9AQ
Posts: 37
Joined: Sat Jan 14, 2017 1:20 am

Re: Feedback for version 0.9a here

Post by K9AQ » Fri Aug 17, 2018 1:17 pm

Steve,

My AD9850 module looks exactly like the top photo that Dave posted.

Don
K9AQ

Reason: No reason

DaveR
Posts: 25
Joined: Fri Jul 27, 2018 1:07 am

Re: Feedback for version 0.9a here

Post by DaveR » Fri Aug 17, 2018 7:43 pm

Hi Steve,

I'm still waiting for the seller to respond to my message, but my plan for the second AD9851 board is to replace the oscillator with the 30MHz one from the first board to see if I can conjour up a working one, assuming that the innards of the 9851 chip haven't been mangled by trying to run at 750MHz! As a final fallback, I've also ordered yet another AD9851 - this time the same board as yours - but delivery will depend on the slow boat ....

Looking forward to your next files to see if we can get a Nano working!

Regards,
Dave

Reason: No reason

SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Sat Aug 18, 2018 8:32 am

K9AQ wrote:My AD9850 module looks exactly like the top photo that Dave posted.
Thanks for that Don. As Dave and I have used the same modules without experiencing problems (AD9851) I can only assume (see my previous comment regarding assumptions ;) ) that your original module was suspect in some way. I still can't work out why it worked with the code snippet you posted some time ago though.
DaveR wrote:I'm still waiting for the seller to respond to my message, but my plan for the second AD9851 board is to replace the oscillator with the 30MHz one from the first one to see if I can conjour up a working one, assuming that the innards of the 9851 chip haven't been mangled by trying to run at 750MHz
It has to be worth a try while your waiting for another module to arrive - I'm still waiting for my AD9850, time to chase up the suppler I think.
DaveR wrote:Looking forward to your next files to see if we can get a Nano working!
I finally got the files sorted out. Sorry about the delay, I had to re-install Visual Micro and it took me for ever to find where it was putting the final HEX output files.

First, DDS_Arduino_Test.zip contains source code for a simple test I made up for Don some time ago. I've made a few changes so it includes support for the Arduino Nano. Define flags allow it to be compiled for Nano, Uno or Mega2560, with either AD9850 or AD9851 support. It can be compiled using the Arduino IDE. All it does is sets the target DDS to an output frequency of 1MHZ. This allows users to at least check that their Arduino/DDS combination is working without needing to use the analyser. It includes an option to make the Arduino continuously send the frequency settings to the DDS, this might be of help if fault-finding.

Arduino_HEX_Collection.zip contains HEX files for the Nano, Uno and Mega2560. Each board has two hex files, compiled to support either the AD9850, or the AD9851, and should work with the analyser. I still don't have an AD9850, or a Nano, so any feedback regarding those files would be most welcome. From what I can see, the Nano uses either an ATmega328 processor or an ATmega168. The files have been compiled to support the 328 variant of the board.
Arduino_HEX_Collection.zip
(64.61 KiB) Downloaded 1264 times
DDS_Arduino_Test.zip
(3.76 KiB) Downloaded 1257 times
Let me know how it goes.

Fingers firmly crossed - Steve

Reason: No reason

DaveR
Posts: 25
Joined: Fri Jul 27, 2018 1:07 am

Re: Feedback for version 0.9a here

Post by DaveR » Sat Aug 18, 2018 7:29 pm

Success, Steve!!

I've tested a Nano with both the AD9850 and AD9851 hex files, and they both work as intended (making due allowance for my half-working AD9851, of course). I also tested an Uno with the AD9850 and it worked as well, so I guess the other three hex files will also be ok.

Have you any timescale in mind for when the next release RSP-SAS will be available for testing?

As regards your long awaited AD9850 board, both of the AD9851s I received recently came from AliExpress in 10 - 14 days, but I've recently had stuff from ebay suppliers in China and HK which took eight weeks to arrive. You tend to pay a liitle more for things from AliExpress, but I'm impressed with their overall service and backup, and the extra you pay does appear to cover a faster journey in the post. (Which is great, especially if it works when you get your hands on it ;) )

Regards,
Dave

Reason: No reason

SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Sat Aug 18, 2018 11:18 pm

DaveR wrote:I've tested a Nano with both the AD9850 and AD9851 hex files, and they both work as intended (making due allowance for my half-working AD9851, of course). I also tested an Uno with the AD9850 and it worked as well, so I guess the other three hex files will also be ok.
Great news Dave - Thanks for your efforts. I know the other hex files work, I've tested them on a Uno and a Mega2560, both with an AD9851. As the AD9850 works for you, I think we can say with a high level of confidence that all 6 files should be good to use with the analyser.

As regards the analyser not always detecting the Trackgen module: As mentioned, I think it's due to an 'undocumented feature' of the Microsoft comms driver as I've experienced similar problems in the past. That problem is now fixed, but required changes to both the analyser and the Trackgen code. The only workaround at the moment is to select the Trackgen comms port, then select another comms port, then select the Trackgen port again.

The revised code is getting closer to being ready to release. I'm reluctant to quote a time frame as the unexpected could (and usually does) arise. I can only say that I'd hope to have something ready in around 4-6 weeks, but please don't hold me to that.

I've just had a quick look at AliExpress. Prices seem pretty reasonable given the wide variations in price seen on Ebay between similar products. I think I'll try them when I next need to order something. I'm interested in the AD9910. Up to 420MHz output, plus a whole range of in-built functions would make this an ideal tracking generator. I'd like to get it now, but it would only be an unwanted distraction at this point in time.

https://www.aliexpress.com/item/1-pcs-A ... autifyAB=0

As you say, faster shipping has got to be worth paying the slightly higher cost.

Regards

Steve

Reason: No reason

DaveR
Posts: 25
Joined: Fri Jul 27, 2018 1:07 am

Re: Feedback for version 0.9a here

Post by DaveR » Sun Aug 19, 2018 1:13 am

No problem , Steve - I'm doing things that I enjoy!

As for Trackgen not always starting, my workaround is simply to start the analyser then click on the Trackgen com port again, which always seems to get it going, but it's good that you've now got it sorted for the next release. Time is passing by so quickly now, that 4 - 6 weeks (or whatever) is but the blink of an eye. I'll have a chance to get some other stuff done while I'm waiting for your announcement!

Happy coding!

Regards,
Dave

Reason: No reason

K9AQ
Posts: 37
Joined: Sat Jan 14, 2017 1:20 am

Re: Feedback for version 0.9a here

Post by K9AQ » Sun Aug 19, 2018 2:52 pm

I would gladly pay the extra cost to get an AA9910. The upper frequency limitation of the AD9850/51 will limit the usefulness of the tracking generator. I will also be willing to wait for this functionality. Getting the next release of the spectrum analyzer software out is much more important. For my use, the AD9850 works.

Don
K9AQ

Reason: No reason

DaveR
Posts: 25
Joined: Fri Jul 27, 2018 1:07 am

Re: Feedback for version 0.9a here

Post by DaveR » Sun Aug 19, 2018 3:13 pm

The AD9850 fits the bill fine for me as well, Don, but getting a working AD9851 has now become something of a quest for me - just for the sake of having one, I suppose. I'm staying clear of the AD9910 and its cousins for now because I know it would end up as just another trinket waiting for a use to be found for it. When Steve eventually gets one integrated into Trackgen I suppose I'll succumb to curiosity and temptation, though :mrgreen: .

Regards,
Dave

Reason: No reason

DaveR
Posts: 25
Joined: Fri Jul 27, 2018 1:07 am

Re: Feedback for version 0.9a here

Post by DaveR » Thu Aug 23, 2018 2:43 am

I got my second AD9851 module (the green one) working today! I fitted a 30MHz 25ppm clock chip in place of the 125MHz one it came with (probably 50ppm), and did the D0+D1 and D2 mods to force it into serial mode, and it now works far better than I ever hoped. It's probably down to the board being larger and well laid out, as opposed to cramming everything into the smallest possible area, but the signals are strong and stable, usable up to about 65MHz, and remarkably free of spurii compared to the outputs from the other two boards I've tried. I ran it with a Nano and did a few sweeps with it, all without any problems and good results. Definitely a good buy in the end!

Regards,
Dave

Reason: No reason

Post Reply