Tweet
As mentioned in in the IDX-PRO+VDI thread I worked on a direct sampling based VDI.
Straight from the sinus signals of the transmitter and receiver instead of the x/y.
Just for fun and to find out if it brings us possibilities for new designs in the cheap.
The goal for me is to get the analog part as simple as possible and do all the logic and smart things with the processor.
Modifying the functionality and updating the software shall be easy on the end. And I like fancy LCD screens
Discrimination, beeps, VDI maps, selectivity and slow-fast motion compensation will be a breeze.
I was asked to open a new thread for this.
What I did:
Normally I work with a 16f877a processor but for this purpose it's to slow and small.
So I bought a 18f4520 pic for a few euros and a lcd display for 6 dollar (!) on ebay and started experimenting.
The results for now are on youtube.
What you see is the pic connected to the graphic LCD and a pc based signal generator to emulate the TX and RX detector signals.
Those 15kHz signals are phase shifted by hand. While I shift the phase you also see the new VDI calculated and displayed on the LCD.
Digitizing is on every wave, 15000 times a second, and the lcd is updated around 25 times a second.
Most fun is the graphic screen showing the spectrum in a bargraph. Like all those expensive md's do.
Excuse me for the bad quality but filming with a mobile phone is harder than you think...
One thing in the digital world is that the signals have to be 0/1.
So what I did was connecting the signalgenerator on the pic, pumped up the volume to around TTL level. And there was my digital signal.
Then I programmed the pic to detect the downslopes of the sinuswaves and calculate the timeshift between TX and RX downslopes.
Calculate the VDI from there and displayed the results. Simple as that.
The pic is pretty fast and keeps a steady sampling up to 17,5 kHz detecting a phase shift from 1 to 179. Enough for my homemade 14.2 Khz TGSL-EDU.
Got that done the time has come to connect this experiment to my TGSL-EDU.
But I'm a digital guy and my knowledge of analog electronic is below basic. Building my TGSL first time right was more luck than wisdom.
And so the hard part for me begins. When poking around with a oscilloscope it was easy to find the TX signal. Strong enough to feed the 4024, thus also strong enough for the pic. The problem is the RX signal. 18mv from the coil is way to weak and after the opamps the sinus wave seems to be gone or transformed. I couldn't find a descent signal but that doesn't mean it isn't there, I just could not find it.
And thus I reprogrammed the pic using the onboard two analog comparators for detecting the slopes.
These comparators can react on lower voltages and are programmable on 15 different tresholds. Nice to adjust the sensitivity.
I managed to let them react on 220mv but they couldn't handle the frequency. 5kHz was the max to keep it steady. Does not work.
And now I'm stucked.
To get this going I need 3 signals. A TX wave (found), a RX wave and a simple signal to calculate the signal strengt for the a/d converter.
The RX signal have to be amplified to a logic TTL level for processing. I dont know how to do that. Experimenting with a opamp didn't work for me.
The signal can also be used for the strength signal.
Does anyone have some hints for me to do that? Any help is appreciated.
The sourcecode of my project is attached for anyone who likes it.
Regards,
Willem
Straight from the sinus signals of the transmitter and receiver instead of the x/y.
Just for fun and to find out if it brings us possibilities for new designs in the cheap.
The goal for me is to get the analog part as simple as possible and do all the logic and smart things with the processor.
Modifying the functionality and updating the software shall be easy on the end. And I like fancy LCD screens
Discrimination, beeps, VDI maps, selectivity and slow-fast motion compensation will be a breeze.
I was asked to open a new thread for this.
What I did:
Normally I work with a 16f877a processor but for this purpose it's to slow and small.
So I bought a 18f4520 pic for a few euros and a lcd display for 6 dollar (!) on ebay and started experimenting.
The results for now are on youtube.
What you see is the pic connected to the graphic LCD and a pc based signal generator to emulate the TX and RX detector signals.
Those 15kHz signals are phase shifted by hand. While I shift the phase you also see the new VDI calculated and displayed on the LCD.
Digitizing is on every wave, 15000 times a second, and the lcd is updated around 25 times a second.
Most fun is the graphic screen showing the spectrum in a bargraph. Like all those expensive md's do.
Excuse me for the bad quality but filming with a mobile phone is harder than you think...
One thing in the digital world is that the signals have to be 0/1.
So what I did was connecting the signalgenerator on the pic, pumped up the volume to around TTL level. And there was my digital signal.
Then I programmed the pic to detect the downslopes of the sinuswaves and calculate the timeshift between TX and RX downslopes.
Calculate the VDI from there and displayed the results. Simple as that.
The pic is pretty fast and keeps a steady sampling up to 17,5 kHz detecting a phase shift from 1 to 179. Enough for my homemade 14.2 Khz TGSL-EDU.
Got that done the time has come to connect this experiment to my TGSL-EDU.
But I'm a digital guy and my knowledge of analog electronic is below basic. Building my TGSL first time right was more luck than wisdom.
And so the hard part for me begins. When poking around with a oscilloscope it was easy to find the TX signal. Strong enough to feed the 4024, thus also strong enough for the pic. The problem is the RX signal. 18mv from the coil is way to weak and after the opamps the sinus wave seems to be gone or transformed. I couldn't find a descent signal but that doesn't mean it isn't there, I just could not find it.
And thus I reprogrammed the pic using the onboard two analog comparators for detecting the slopes.
These comparators can react on lower voltages and are programmable on 15 different tresholds. Nice to adjust the sensitivity.
I managed to let them react on 220mv but they couldn't handle the frequency. 5kHz was the max to keep it steady. Does not work.
And now I'm stucked.
To get this going I need 3 signals. A TX wave (found), a RX wave and a simple signal to calculate the signal strengt for the a/d converter.
The RX signal have to be amplified to a logic TTL level for processing. I dont know how to do that. Experimenting with a opamp didn't work for me.
The signal can also be used for the strength signal.
Does anyone have some hints for me to do that? Any help is appreciated.
The sourcecode of my project is attached for anyone who likes it.
Regards,
Willem
Comment