Page 4 of 4

Re: Negating LF problems of NBTV.

PostPosted: Thu May 20, 2021 4:28 pm
by Steve Anderson
A slight error above...the AD9833 using a 25MHz crystal (as per photo) is tunable in approximately 0.1Hz steps up to 12.5MHz. Actually 0.0932Hz increments..

The photo is of a general purpose oscillator I recently built for bench-work, it generates sine, square and 5V HC-level TTL signals. The AD9833 can also generate triangle (but not sawtooth) waveforms, but I have no need for them.

It's controlled from a PC via an ASCII serial link (USB to UART) either manually or under automation for frequency response checks for example. The 14-pin chip is a PIC16F688 micro that does the maths and handles the serial link.

This with a bit of 'massaging' could easily handle the FM NBTV subcarrier generation...it wouldn't need the HC540 chip, that's just a buffer for the TTL output.

Steve A.
AD9833 DCO 1.jpg
AD9833 DCO 1.jpg (162.97 KiB) Viewed 5555 times

Re: Negating LF problems of NBTV.

PostPosted: Fri May 21, 2021 6:50 pm
by Steve Anderson
Due to some cobwebs within my skull I forgot about Carson's rule. This basically scrubs my FM-NBTV idea. So it's back to my original plan....

Steve A.

Re: Negating LF problems of NBTV.

PostPosted: Fri May 21, 2021 8:27 pm
by Harry Dalek
Some times plans don't go the way we want ,i would say give it a rest and come back to it having a think about it but i think you have least there's always a plan 2 to fall back on !

Re: Negating LF problems of NBTV.

PostPosted: Sat May 22, 2021 12:12 pm
by Steve Anderson
...So it's back to my original plan....hang on...my wife has a desktop PC with an on-mainboard audio I/O. I'll see if it will run at 96kHz. However I think it still uses those horrible 3.5mm jacks.

Steve A.

Re: Negating LF problems of NBTV.

PostPosted: Sat May 22, 2021 4:42 pm
by Harry Dalek
I found a massive lap top 2 to 3 times the size and thickness of todays or even yesterdays laptop s the other day...best of all it was free.
My sons now pinched off me since he noticed it was a Gaming PC i think it has a quadcore processor ,the sound card for the Alienware 18 has a built-in Realtek ALC3661, it's a -Sixchannel High Definition Audio device all i know so far this would be 96khz if not more
This link below gives you a better idea of the size of this thing
https://au.pcmag.com/laptops/33753/alienware-18-review
But alas my son has taken owner ship ! :?
Perhaps NBTVers should look into gaming computers ~!

Re: Negating LF problems of NBTV.

PostPosted: Tue May 25, 2021 5:09 pm
by Steve Anderson
I seem to go though mice (the PC versions) like nobody else. I routinely have to replace mine about once a year. If they can make the scroll-wheel optical why not make the left & right buttons optical instead of microswitches? I wouldn't mind paying twice the price for an all-optical mouse if it lasted five years and not just one. There's no rocket science involved!

I don't want wireless as I know I'll walk off with it in my hand and put it down somewhere, later I'll have to search the house trying to find it. While it's tethered to the PC via its 'tail' it's not going far...plus the batteries don't die just when you need to use it.

Now in connection with Harry's PC above, I have seen one of these. But an answer to my failing mice was suggested by a friend...."Buy a gaming mouse, they're designed for rapid and abusive input." I looked at a few but they've got more buttons on them than I have fingers (I still have all ten).

All I want is a basic Microsoft or Logitech rodent that'll last a while...it's not the cost, it's the hassle of buying a new one. Next time I'll buy five maybe...

Steve A.

Re: Negating LF problems of NBTV.

PostPosted: Tue May 25, 2021 7:42 pm
by Steve Anderson
...continuing on the theme of this thread...I looked at a few external audio I/Os for PC's...I found a few which offered 192kHz sampling, ideal I thought. But although their bandwidth could be up to 96kHz most quoted a frequency response of up to around 50kHz. OK for 'real' audio (in fact way more than enough) so why bother with 192kHz when 96kHz will basically do the job? Specmanship and sales ploys I suspect...damn infuriating...

...also possibly to make the anti-aliasing filters simpler (or omit entirely) and do away with the sinc filters on the output...the burden shifts from them making decent filters (hardware or software) on to you - creating huge unnecessary files in the process..

Sinc hardware filters are required on the output of D-As or done in software prior to the D-A to preserve the frequency response up to almost the Nyquist frequency. Google 'zero-order hold'.

Steve A.

P.S. Sinc rhymes and is spoken the same as 'sink'.

Re: Negating LF problems of NBTV.

PostPosted: Sun May 30, 2021 12:19 pm
by Steve Anderson
Not having used my laptop at 921,600 baud rate before I thought I should make sure it did work at that speed before I continued...
So I grabbed one of my USB-UART 'cables' and tried it out...fine up to 460,800, but no-go at 921,600.

"Try another USB port Steve." I thought...and it worked!! My laptop has three USB ports, two USB 3.0 and one USB 2.0. Sod's law, the first one I tried was the USB 2.0, typical!

USB 2.0 is supposedly 12Mb/s so it should work, but for some reason it doesn't...a word to the wise...don't get caught out by this when playing with USB!

Steve A.

Re: Negating LF problems of NBTV.

PostPosted: Sun May 30, 2021 5:43 pm
by Steve Anderson
Having sorted out the PC-UART baud rate hiccup for 921,600 I need the try a PIC microcontroller UART at that speed. The datasheets only show examples up to 115,200, but the derivation of the programmed value needed it is possible at 921,600 with the formula given in the datasheet.

I cannot see any reason why the silicon should object to such a data rate - there's only one way to find out...given a high enough (but legal) clock rate...

Steve A.