Basic SDR Questions

[Electro Fan]

This post is an attempt to foster discussion about the basics of SDR.  I have a bunch of questions but for starters:

1. The RTL-SDR products seem to be based on RTL2832 ICs that demodulate digital signals and then send the demodulated signals to computer via USB.  The "RTL" dongles apparently include a tuner chip with the Elonics E4000 having been an early favorite - but Elonics went out of business - so what are the prevailing and preferred tuner chips that are used in place of the E4000?

I saw that the NooElec units are based on the R820 (R820T2) tuner IC made by Rafael Micro. Apparently these use the RTL2832 (RTL2832U) IC as the demodulator and USB interface - so maybe the R820 and R820T2 are the new tuners of choice (and they are effectively a demodulator integrated with a tuner on a small board)?  (Also, what is the difference between a RTL2832 and a RTL2832U?)

2. Can anyone summarize broadly the most typical choices of antennas that might be used with a RTL USB dongle?  Apparently they need to connect to the dongle using a MCX cable for starters but then could be extended to various types of antennas - perhaps some shorter, some longer, some indoors, some outdoors.  Just looking for a "101" discussion of the typical antenna choices with respect to performance differences at the various frequencies supported by the RTL USB dongle devices. 

3. In the image below it looks like a PCB with a USB connector with two main chips:  The RTL2832U and the E4000.  This is what one could expect to find inside a typical USB SDR dongle?  In a dongle that uses the R820 are the demodulator and tuner still two separate chips or are the two functions integrated on one chip?

Update:  found the answer, I think - in the second image.  Looks like the RTL2832U is still there by itself and the E4000 has been replaced with the R820T.

Next question:  how does the functionality and performance of the R820T compare to the E4000?  (I know, somewhat academic since the E4000 isn't made any more but apparently you can still purchase an E4000 based unit.)

Many more questions to come..... but this would be a good starting point.  Thanks

- saw this (looks informative):  http://k4fmh.com/wp-content/uploads/2014/10/FMH_SDR_Cheap.pdf

[tggzzz]

2. Can anyone summarize broadly the most typical choices of antennas that might be used with a RTL USB dongle?  Apparently they need to connect to the dongle using a MCX cable for starters but then could be extended to various types of antennas - perhaps some shorter, some longer, some indoors, some outdoors.  Just looking for a "101" discussion of the typical antenna choices with respect to performance differences at the various frequencies supported by the RTL USB dongle devices. 

That entirely depends on what signals you want to examine and which you wish to the antenna to reject.

Be aware the R820 devices are 75ohm input impedance, not 50ohm.

[Electro Fan]

2. Can anyone summarize broadly the most typical choices of antennas that might be used with a RTL USB dongle?  Apparently they need to connect to the dongle using a MCX cable for starters but then could be extended to various types of antennas - perhaps some shorter, some longer, some indoors, some outdoors.  Just looking for a "101" discussion of the typical antenna choices with respect to performance differences at the various frequencies supported by the RTL USB dongle devices. 

That entirely depends on what signals you want to examine and which you wish to the antenna to reject.

That's a fair response, so maybe the question can start with: 

"In addition to AM and FM, what are likely to be the most interesting and useful frequencies to receive?"

Or, "Are there a few antenna designs that would in the aggregate give good reception for all or most of the available SDR spectrum so that a new user could experience the various bands?"

Or, "How many antennas would you need to reasonably cover the SDR spectrum?", ie, "What would be the fewest number of antennas recommended for an initial attempt at full SDR coverage, and what would they be?"

In other words, it's kind of a chicken and egg deal.  Hard to know what frequencies are worth listening to/gaining experience with if you haven't been listening to them, and it's hard to pick an antenna if you don't know what frequencies are most interesting/useful to listen to.

(I love how sometimes EEVers seem to like to make you beg ask for help very precisely). 

[tggzzz]

Just look and see what's available on ebay Or consider vivaldi and discone antennas.

But if you only want a quick look without worrying about efficiency and directivity, then just use a piece of damp string that is roughly the right length

As for which frequencies, look on the various SDR forums to see what people are doing with them.

[cdev]

For $10-12 you cant go wrong. They are a spectacular deal.

A good antenna is something like this: http://www.wa5vjb.com/references/PlanarDiskAntennas.pdf

Note that the feedline is bonded to the disk down to the bottom. the coax shield needs to be welded or soldered or covered with aluminum or copper (conductive) tape.. Thats called an infinite balun. Commercial antennas tend to be ludicrously overpriced. You can design and build filters to your hearts content for a few pennies in parts. Money spent on HF reception is better spent on something like a softrock than on upconverters and so on, because you're always going to run into the lack of dynamic range and bandsplatter, images. OTOH, an LNA is a good idea because they arent that sensitive, compared to what you get with an LNA. Also, if you can use an IF bandpass filter in your software, use it, it makes a big difference.

In my experience, there is no difference between the cheapest ones and the expensive ones. Actually the cheapest one I bought ($10) was the best. Don't fall for all the hype.

[tggzzz]

A good antenna is something like this: http://www.wa5vjb.com/references/PlanarDiskAntennas.pdf

Elegant; bookmarked.

...because you're always going to run into the lack of dynamic range and bandsplatter, images. ... Also, if you can use an IF bandpass filter in your software, use it, it makes a big difference.

How can a software IF bandpass filter help if the front end is overloaded by "uninteresting" signals?

[cdev]

Well, then don't get one

[Kalvin]

Creating a set of bandpass filters for the ham bands should not be too hard. Manual control with rotary switch or relays is easy to do. And the filters do not have to be very precise as long as they pass the band to be used. Or, if there are one or two transmitters nearby, building a bandstop-filter or notch filter is not too hard either. With a little filtering tricks one should be able to increase the sensitivity quite a lot as one can use higher gain in order to listen distant stations as the input stage is not overdriven any more.

[tggzzz]

Creating a set of bandpass filters for the ham bands should not be too hard. Manual control with rotary switch or relays is easy to do. And the filters do not have to be very precise as long as they pass the band to be used. Or, if there are one or two transmitters nearby, building a bandstop-filter or notch filter is not too hard either. With a little filtering tricks one should be able to increase the sensitivity quite a lot as one can use higher gain in order to listen distant stations as the input stage is not overdriven any more.

Just so.

But to avoid overload they have to be between the antenna and front-end, hence cannot be in software.

[Kalvin]

Creating a set of bandpass filters for the ham bands should not be too hard. Manual control with rotary switch or relays is easy to do. And the filters do not have to be very precise as long as they pass the band to be used. Or, if there are one or two transmitters nearby, building a bandstop-filter or notch filter is not too hard either. With a little filtering tricks one should be able to increase the sensitivity quite a lot as one can use higher gain in order to listen distant stations as the input stage is not overdriven any more.

Just so.

But to avoid overload they have to be between the antenna and front-end, hence cannot be in software.

Yes, indeed. For example, as the popular SDR-dongles come with two SMA-connectors (low band and high band) it is an easy task to attach an external filter-box between the antenna and the SDR-dongle.

[Electro Fan]

Did some research on my questions and found the chart below.  A newer version of the chart is at this link:
http://icomamerica.com/en/amateur/amateurtools/2015_BandPlan.pdf

For the time being I've decided to focus on receiving in the 160, 40, and 20 Meter bands and to also give some of the VHF/UHF bands a try.  This has led me to trying 2-3 simple antenna types.

While FM probably isn't very interesting to a lot of people here I've used FM as my entry learning point for SDR.  My first impression is that the $20 NooElec 820T dongle does a pretty good job on FM with both SDR# and HDSDR.  (I haven't tired a 820T2 but I think it is a newer version and it's probably as good or better than the 820T; my guess is better.)

Using the NoolElec 820T dongle with a regular cheap FM dipole antenna, a regular Windows 10 PC, a nothing special sound card, and PC speakers FM sounds "ok".  Not bad, not great - but I think the limitation might be the speakers.  With real hifi gear (especially the speakers) on the back end of the audio system the front end SDR system might prove to perform pretty well sonically but that's just a guess.

So with a dipole antenna and the Nooelec dongle as the radio hardware and SDR# and HDSR as the radio software and FM as a frame of reference I have established a "baseline" regarding how easy/difficult it is to find radio frequencies and how easy/difficult it is to tune for a good signal and ultimately how good or not so good the sound is - ie, is it well defined with low noise or is it distorted and noisy?  My initial impression is that if this is as good as it gets with local FM stations then listening to signals in more esoteric bands from who knows what transmitters is probably going to be an exercise in trying to pick out not much of a signal from among lots noise - but that's just a guess.  (Or it might be that SDR is just a modest introduction to real radio.)  We'll see, I mean, hear.   

So, my next move has been to acquire a SDRPlay.  My first impression is that it offers more functionality and more control than the NooElec dongle.  It also seems to provide more sensitivity - to the point that with FM the challenge is to avoid overloading the system.  The SDRPlay ExtIO software seems to offer more control and insight to what is happening with the signal chain vs. the more basic NooElec 820T (I'm using HDSDR with both NooElec and SDRPlay).  I'm really just at the very beginning of the learning process but what I think I'm seeing is the IF section; this aspect of SDR, the ability to use software to see and control radio graphically while experiencing the impact of various settings seems to be a major benefit of SDR (along with the relatively lower entry cost.)  On the downside, although admittedly I'm just getting started, the quality of FM signals doesn't seem to better on the SDRPlay than the NooEelec and it's possible the NooElec is better with respect to sound - but it might be that I haven't figured out to tune and tweak the SDRPlay.  It's my understanding that NooElec uses an 8 bit ADC and that SDRPlay uses a 12 bit converter so I was hoping for SDRPlay to be a step up in sound quality.

I'm going to keep trying to squeeze better sound out of FM as my baseline on sonic quality and start to experiment with other bands and antennas.

All of this is just a bunch of entry evel learning and experimenting.  I'm happy to answer questions and I'm up for any info and guidance.... YMMV, EF

[cdev]

Have you ever used any of the more professional SDRs? (with the exception of the HackRF which is also 8 bits)

the reason I ask is because the difference is dramatic, in terms of weak signal reception. I like the dongles, I just don't think its worth spending a lot of money on fancy ones or add ons for them that would be better spent on simply getting a better SDR.

As the cheapest ones in my experience work as well or sometimes even better than the more expensive ones.

The dongles are great. For $10-15 they offer fabulous value.  They also make a good IF for single purpose custom hardware- making it very easy to make all sorts of dedicated radio devices cheaply.

I'll reserve judgment on "SDR-play" until I've tried one. I have heard mixed reports.

There is a lot of benefit to be gotten with almost all SDRs by using non-digital, physical band pass filtration.

[cdev]

Creating a set of bandpass filters for the ham bands should not be too hard. Manual control with rotary switch or relays is easy to do. And the filters do not have to be very precise as long as they pass the band to be used. Or, if there are one or two transmitters nearby, building a bandstop-filter or notch filter is not too hard either. With a little filtering tricks one should be able to increase the sensitivity quite a lot as one can use higher gain in order to listen distant stations as the input stage is not overdriven any more.

Just so.

But to avoid overload they have to be between the antenna and front-end, hence cannot be in software.

This design (below) by YU1LM works well. Its not quite a sharp as I would like (I could probably improve it by moving things around a bit) but its sharp enough to make a big difference.

[Electro Fan]

Have you ever used any of the more professional SDRs? (with the exception of the HackRF which is also 8 bits)

the reason I ask is because the difference is dramatic, in terms of weak signal reception. I like the dongles, I just don't think its worth spending a lot of money on fancy ones or add ons for them that would be better spent on simply getting a better SDR.

As the cheapest ones in my experience work as well or sometimes even better than the more expensive ones.

The dongles are great. For $10-15 they offer fabulous value.  They also make a good IF for single purpose custom hardware- making it very easy to make all sorts of dedicated radio devices cheaply.

I'll reserve judgment on "SDR-play" until I've tried one. I have heard mixed reports.

There is a lot of benefit to be gotten with almost all SDRs by using non-digital, physical band pass filtration.

My only experience with SDR has been NooElec and SDRPlay.  I saw that you could add the NooElec HamItUp converter to the NooElec dongle but after reading lots of reviews I decided to go for the SDRPlay.

What SDR is doing is providing a learning test bed.  I don't know if I will go further up the food chain; kind of depends on what I learn.  I like good quality signals (used to be into hifi).  My experience with the best FM (analog) tuners were that they sounded better than the digital tuners - that was several years go so maybe things have changed.  People here will probably say phooey but that was what my ears told me.  In any event, I'm very sure that some systems - from signal creation to signal recreation - sound better than others.  It depends on everything from how the source was captured to the amplifying electronics to especially speaker design, speaker locations in a room, and room acoustics.  On the other hand, my interest in EEVblog type electronics and SDR is not to achieve super hifi quality music it's to gain some hands on experience with RF, baseband, and A to D / D to A - so the only reason I'm mentioning audio quality is that I think it makes sense to establish a frame of reference on what is (admittedly subjectively) discernible on a pretty strong and well know signal before listening for ET on the airwaves.  (Actually awhile back I discovered HAMs have been doing EME "Moon Bounce" - so I realize there is a point at which just sending and receiving a signal can be an extremely impressive feat even when the end result is truly barely discernible.)

Anyway, back to SDR....

One of the things I'm enjoying with the SDRPlay is the ability to look at 8MHz of bandwidth and then to narrow in and tune in once a signal of interest is identified.  The relationship between the IF bandwidth and the sampling rate along with the IF gain are all very interesting; the ability to graphically as well as acoustically experience the results makes a very good learning platform.  And it's raising TONS of questions.

Attached is a screen shot of a FM station signal.  I think it's being sampled at rate of 192 kHz.  The ExtIO graphic says the bandwidth is 600kHz which I presume must be the entire left to right blue section of the signal shown in the large panel (minus the rolloff on either side).  Or perhaps the IF bandwidth is what is shown in green rather than blue?

As the second image shows the light blue vertical highlight is narrowed when the sampling rate is lowered from 192 kHz to 107795 Hz.  In some respects the sound improves when the sampling rate is reduced - perhaps the lower sampling rate is reproducing less noise and effectively improving the signal to noise ratio, or maybe it's just putting more music into the signal and raising the gain?  (Sometimes louder sounds better even if it isn't.)  I noticed that changing the FM-BW sampling rate (reducing it from the max setting via the vertical slider) not only changes the signal audibly but the characteristics can be seen in the lower right panel that I think is the 15 KHz FM signal.  To me, the ability to see the signal from all these perspectives is fascinating.   

Net, net:  what is the thinking on optimizing the tuning when using the IF bandwidth and sampling controls?  And the IF gain control?  (I've decided to uncheck the LNA EN in check box ExtIO - it provides too much gain, I think.)  I'm betting there is some preferred sequence and technique for tuning into signals - I just haven't figured it out yet.

On a related question, I'm guessing that the two humps next to the main signal are the upper and lower side bands.  I'm wondering why they have more of square top and less of a pointy top?  Also, all the small vertical white lines, are those what are referred to as "birdies"?  What causes them?   Back to amplitude, it seems that where the blue turns to green might be about where the blue sampling highlight intersects?  Any significance to this being where the sound quality and gain sounded good, or just a coincidence?

Thanks for any insights.  EF