Page 1 of 1

Recording & Playback of NBTV, DC-coupled.

PostPosted: Wed Jun 19, 2019 5:47 pm
by Steve Anderson
PC NBTV Interface 1.pdf
(23.71 KiB) Downloaded 618 times
I needed a recording method for a client that went down to DC, yet also encompassed frequencies up to the limits of what could be called audio.

We all use PCs here and unless they're 20 years or more old they'll have a USB port. Those that have even glanced at the USB specs and documentation wilt at its depth of comprehension required...that includes me...have a look here, "USB in a nutshell"...must be one heck of a nut!

https://www.beyondlogic.org/usbnutshell/usb1.shtml

One way to circumnavigate this is to use a USB-to-UART converter. Load the driver (simple), plug it in and the USB external device appears as a COM port. That I can handle.

Your interface is now a UART port, you can record and play back files easily.

The two devices I have used are USB-UART 'cables from FTDI in Glasgow UK, (type TTL-232RG-VSW5V-WE) and a PCB version available from a local company here about 3km from me. Pics below. They both use the same FTDI chip so you can swap between the two no problems, just one driver needed for both as they both use the same chip. The specs go up to 3Mb/s, yes megabits per second...if your terminal software can handle it.

First off most might think of Hyperterminal which used to come bundled with Windoze many years ago, you can still find it on the web, but it does have some problems. It refuses to handle certain bytes which in ASCII it treats as control codes and promptly ignores them and will not record them. I use RealTerm (free) and it works fine. It takes a while to get used to, but it does all that's required. I'm sure there are others.

So armed with the software, the USB-UART hardware interface you're potentially in business.

The block diagram attached shows the external hardware. Sorry it's a pdf, my version of Autocad has all of a sudden refuses to out put raster-files which I usually turn into a gif file. No idea why.

The A-Ds and D-As are 12-bit devices, fine for the video, but some may cringe at 12-bits being used for audio. But actually it's not that bad, when you're using 'multimedia' speakers and the like you (or at least I) would say it's fine. However, it's designed for 16-bit devices when I can find some from a local supplier at sane prices.

This is still at a conceptual stage for NBTV, I need to sort out my workshop/office before I make a start on the hardware.

Those interested - I would like to know what else you may wish to add/subtract. So all those DC-restorers can go in the bin...

Some may say at around 15 quid for the FTDI 'cable is expensive, but being the wire-ended version I put a 9-pin D-type on the end so I can use it with potentially many different devices. The small PCB is about 7 quid locally. The FTDI cable you can buy on-line via their website or the usual large distributors, Farnell, RS, etc...

Note: In both cases mentioned above, the voltages are NOT RS232, but 5 or 3.3V, as if they had a MAX232 (or similar) built in. Direct connection to a 3.3 or 5V micro is the designed usage. DO NOT connect a 'real' RS232 port to them, you'll fry it! And maybe the port too...

Steve A.

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Thu Jun 20, 2019 4:41 pm
by Steve Anderson
OK, I've found a work-around for my printing to GIF problem...I hope...

Steve A.

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Sat Jun 22, 2019 10:06 pm
by Steve Anderson
Update on previous, a bit more detail being defined...the pdf version should be better quality than the gif...more still to do, note the notes in red...

Steve A.

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Tue Jun 25, 2019 5:23 pm
by Steve Anderson
Here's more detail on the filter arrangement though I still need to sort out the I/O levels. On the output that's just a simple resistive divider. You can build in gain to the input filters but as you change the gain the filer components need changing too, not easy. More later...

I've also sorted out my printing-to-gif problem...

Steve A.

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Wed Jun 26, 2019 8:19 pm
by Harry Dalek
Looks interesting Steve I am just watching seeing what you come up with.

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Thu Jun 27, 2019 2:45 pm
by Steve Anderson
I'm considering using FIR (software) filters on the input side, this would reduce the hardware requirement a bit.

I've been thinking of using them for some time but have never tried them out - I guess there's a first time for everything...

I still need to clear my workbench!

Steve A.

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Tue Jul 09, 2019 8:50 pm
by Steve Anderson
Well, a few design considerations later...I've dropped the logic supply volts to 3v3 with a change of micro that is more than capable and half the price, and changed the analogue supplies to +/-9v which is OK here. It might be possible to make this single-supply and usable on batteries with a change of op-amps. Still dwelling on this...

Steve A.

Update, nothing suitable in terms of op-amps for single-supply operation here, sadly...

Re: Recording & Playback of NBTV, DC-coupled.

PostPosted: Wed Jul 10, 2019 3:17 pm
by Steve Anderson
I'm still considering using FIR (software) filters on the input side to reduce the hardware component count.

I've come up with a plan which has the frequency response shown below. It starts to cut in at 9khz, is -10db at 10kHz and all above 11kHz is at least at -40db. The ripple in the passband is +/-1db.

To achieve this level of performance with an analogue filter would be complex (i.e. expensive), so my plan is to initially build this device with the analogue filters shown above, then remove them, add in the FIR filters to the code, and see what the difference is. The processor has enough memory and speed to do this.

Steve A.

For those interested the FIR filter coefficients can be generated here...

http://t-filter.engineerjs.com/