SDR UPDATE

[WA1GFZ]

I am building the boards for the RA6830 SDR interface. A number of guys have asked about it but here is a cool option. Check out "soft Rock 40" site. Soft rock 6 is coming out later this month. The Kit price will be $15. Say you want to hang it off a 455 KHz if...just change the crystal. This is far cheaper than anything I could produce. The RA6830 interface module will be over $50 when all is said and done. The stock kit is set up for 40 cw but the frequency can easily be changed. This will work with flex software and a number of other free packages. This is by far the cheapest way to get into SDR. gfz

[WA1GFZ]

Hi All,
2 functional SDR boards up and running! now I can test the real performance of the Flex front end in the gfz RX torture chamber. I did my bias on the mixer a little different that flex but should not effect the dynamic range tests. Running I2PHD software on 3.885.
The only problem is I need to increase the attenuation at the input because the 6830 IF output is saturating the sound card. I used the same op amps as the flex so
here is a big clue on real performance. I had to set the sound card levels as low as possible and put some 40 dB of attenuation ahead of the converter to handle the minus 15 dBM of 455 KHz signal from the RA6830. This is using the low gain setting flex used in their op amps. This tells me the flex is saturated at about minus 60 dBM
input.  (-67 dBM is 100 uv.) No wonder why W3DUQ has to add an attenuator to the front end of his flex. It does sound very nice and the spectrum display is very cool. So if the flex will hear about -135 dbM and saturates at -65 dBM that is 70 dB of input range in my book who knows what the ip3 will be.
Soft rock 6 will be out soon. BTW I copied their oscillator and a color burst crystal works great. I added the required series cap for the crystal in my design. This means a soft rock 6 with only a crystal change will provide a sdr interface to 455 KHz. for $15
kit charge plus a buck for a crystal. An input pad will also be required to drop the signal down. It will not have the dynamic range of the flex but will be fine connected to an IF. you can't beat that price to get into sdr. They plan 1000 Kits.  fc

[WA1GFZ]

Performance with a stock Dell sound card
board only with 455 KHz input.

MDS = -135 dBM

Dynamic Range
2.5 KHz.   89 dB
5 KHz.      90 dB
10 KHz.    94 dB


my input is a 6 dB resistive pad to set the input z.
The only gain in the circuit copies the flex in low gain mode.

The Flex circuit has no AGC ahead of the mixer so I wonder how it can handle signals above S9 or so without generating mixing spurs. It could be my sound card but need to check further. The circuit should handle a volt RMS of signal.

The circuit hanging off my receiver IF out at -15 dBM needs a bunch of attenuation before the module functions properly. more to come when I'm less stupid.

[kz0e]

Frank,

What do you mean by 'flex circuit', the sound card + software? Also, is there a prefered source for the color burst xtals?

Thanks,

Larry

[WA1GFZ]

flex circuit is the Flex Radio mixer and post amplifiers.
I think my crystals came from mouser. The Guy who did the drafting bought the parts. I think he paid about $1 each. 18 pf series cap.
more data to come.

[WA1GFZ]

Well last night I interfaced the SDR ahead of the 455 KHz IiF filters just after the second mixer. The UA733 had no problem driving the normal path and the SDR
with a resistive splitter. I was listening to the guys on 3825 last night and the software filters worked fine  while I could monitor 25 KHz of spectrum.
I still have too much gain in my interface but should be able to adjust it lower.
Looks like I need to shift the operating point up about 30 db with my post amps are running at 40 dB of gain so should be easy to do.
It looks like the Flex guys should have put some agc ahead of their mixer rather than adding a preamplifier.  I easily hear  -135 dBM with a 6 dB pad and resistive splitter ahead of the Tayloe mixer. gfz

[WA1GFZ]

Tonight I  wanted to see what the Flex front end would do with power.
The high gain setting it goes into hard saturation at just under -10 dBM (.08V) input
The mixer was not saturating but the post amps were clipping at the peaks.
at 25 volts peak to peak..
So if I go to the low gain setting from 40 dB to 20 dB I should get to about plus 10 for an input. It seems the tayloe mixer is not the problem just the post amplifiers.
I don't see and third order products below the saturation input.
I was able to shift the operating range from -135 dBM / -40 dBM up 30 dB by setting the sound card inputs to minimum. This also hurt the sensitivity but that is not a problem since the Racal will have gain ahead of the converter. Flex gets the sensitivity with the high gain audio amps but pays a bit of a price with dynamic range.
This could be due to the sound card noise floor but I still can't understand why they added more gain with an rf stage.

[WA1GFZ]

well hit the sweet spot tonight. dropped the gain of the post amps from 40 dB to 10 dB and the level is right on the money. Listened to a number of stations on 75 and 160 with plenty of noise then tuned to wtic 1080. I could listen and watch wtic mudulation and watch the two stations either side up and down 10 KHz on the spectrum display. Line level gain slider was at 1/4 up from minimum.
This interface will also work with the RA6790 since the circuit is the same as the 6830.
The $15 softrock 6 kit will be out in a few weeks and all it will need is a gain change and maybe an input pad after the crystal change. I can't match that price and will be banging nails soon. This would be an easy way too try out SDR without spending a lot of time or money.  I'm sure this could be connected to even an R390A IF output.

[kz0e]

Got this in the mail from Tony this morning:


SoftRock v6.0 orders now being accepted

My initial testing with the new SoftRock v6.0 is far enough along to
order production boards and kit parts. 

The v6.0 SoftRock is a single board receiver kit that provides for
up to two band coverage without changing oscillator frequency or BPF
components.  The board size is 1.5 inches by 1.9 inches and a built
unit has a height of about 0.6 inches.  Band changing is by moving a
jumper plug on a three-pin header to give a divide by four or divide
by eight clock to the QSD circuit.  A two-pin header on the board
provides for QSD mute if the two pins are open circuited.  This
should help make the SoftRock more useable with a separate
transmitter.

Here are the two kit options that are being offered at this time.

40m/80m kit option
40m/80m two-band kit at $14 US/Canada and $15 for DX will tune the
following ranges:

40m with a soundcard that samples at 48 kHz - 7.032 to 7.08 MHz
40m with a soundcard that samples at 96 kHz - 7.008 to 7.106 MHz

80m with a soundcard that samples at 48 kHz - 3.504 to 3.552 MHz
80m with a soundcard that samples at 96 kHz - 3.48 to 3.576 MHz

160m kit option with selective 160m BPF
160m single band kit at $14 US/Canada and $15 for DX will tune the
following ranges:

160m with a soundcard that samples at 48 kHz - 1.819 to 1.867 MHz
160m with a soundcard that samples at 96 kHz - 1.795 to 1.891 MHz

Other band coverage will be offered after additional testing.

The audio cable has become a sourcing problem since the supply of
cables I was using with the v5.0 kits is no longer available in
distribution.  Kit builders can get suitable cables from local
consumer electronic stores at a reasonable price. 

Orders will be accepted starting today by PayPal to user name
raparks@ctcisp.com or by mailing a check or money order to me at:

Tony Parks
1344 E 750 N
Springport, IN 47386

I will get a picture posted in the files area in a couple days. 

Thanks for your interest the SoftRock projects.

73,
Tony KB9YIG 

[KA8WTK]

I wonder which one will be most easily modified for use with a 455 IF?

[WA1GFZ]

Bill,  I copied the soft rock 40 oscillator and all it took was a crystal swap. I will know better when Tony posts the schematic. I have the same jumper option on my board but don't need to use the additional divide by 2. I think he has a low pass filter on the input so anything you get may need a modification or elimination.
I have my board working great now. I can interface it to the output of the first IF or the output of the 455 KHz iF strip with only a change to the sound card line level slider. I'm not sure which way I want to go yet.

Sure can't beat Tony's  price. fc 

[KA8WTK]

  Nope, you can't beat the price. I ordered 3 of the kits, 2 of the 40/80 and 1 of the 160 kits.
  I figure to build one of each and use the extra kit for playing with IF frequencies. Tony told me via e-mail that the only difference between the kits is the crystal supplied and the bandpass filter section. It may even be possible to build a 40/80/160 unit with one board and switch bandpass filters and crystals. That will leave 2 kits for playing with.........only time will tell.

 

[WA1GFZ]

Yup, I plan to buy a couple this week myself. My SDR board is working great
but much bigger. It eliminates the function of 3 modules in the 6830.
I'm looking at another design to eliminate the second mixer but don't think the tayloe will work at 40.455 MHz. I think it will take a couple double balanced mixers.
fc

[WA1GFZ]

Soft rock 6 pictures were posted last night very nice board for $14
First thing monday morning,  I ordered a pair of them, one of each type

[WA1GFZ]

Well the winter project is complete and a board is shipped to my friend who did the art work and parts search. I am quite impressed with the SDR close in dynamic range and cool operating features. The software makes it a fun thing to use. I would never consder hardware sync detector now.
The wideband dynamic range is less than impressive. My card would do -135 dBM to   -40 dBM dynamic range for signals spaced 20 KHz. This is due to the noise floor and the maximum signal a sound card can handle. I duplicated the Flex front end with the same gain settings used in the high gain mode but without an RF amplifier.
Reducing the op amp gain so the sound card didn't saturate raised the MDS. This meant shifting the dynamic range was required which can't be classified as performance. -40 dBM into the SDR module with a gain of 100 (40 dB) is just about the saturation point of a stock sound card. A high quality card gets you a bit higher.
The cool thing was the SDR output would make a clean 20 volts peak to peak signal output but the sound card would be hurting with that kind of an input.
All this confirms the best performance is when the SDR is interfaced to an IF with roofing filters to deal with the wide band crud while the SDR does a better job with the close in stuff. The Flex is a cool radio but I don't think it is the hot rod set up.
At least not on my dynamic range test set up.
The soft rock 6 schematic is similar to my design but it is only a single balanced tayloe and op amps run at 5 volts. So what it can't make 20 volts peak to peak since the sound cand can't handle more than a volt. It should be a cool way to play with the nice features of SDR and draw your own conclusion for $14.

[WA1GFZ]

I've been buying Racal RA6830 DF receivers for a while now and converting them to useful receivers. I've provided a number of them to happy customers with a pile of useless DF components stuffed in boxes.
Today it hits me. Why not interface the SDR to a DF radio missing all forms of demodulation. After a few hours of testing this looks like it will produce a very cool receiver having the advantage of an external preselector.  The DF radio has a digital manual AGC and step attenuators on the input.
It looks like my last chassis will end up like the first one I bought in the middle 90s.

The soft rock guys will be shipping the latest V6 module soon and there is a report of V7 coming soon.

[KA8WTK]

Which brings up a thought I have.....Take an otherwise "not premium" (weak audio, poor selectivity, etc.) boatanchor and shunt the IF out to a SoftRock. All you really need the receiver for is pre-selection, RF amplification and down conversion to the IF. After that you have an SDR with all the features! There may be hope for that HR-10 or (insert your own receiver here) sitting on the shelf. Looks like a good way to breath new life into some old rigs.

[WA1GFZ]

Bill, it would work but I find the better the front end the better the whole system works. I find the software filters have about 60 dB of ultimate rejection. I know my old motorola DSP board was 90 dB so who knows what a good sound card will do.
Anyway I think a good analog section into the SDR will provide great performance.
I started building a RA6830 DF RX last night that should do the job. The DF software is set up to control the DF preselector for even more performance.
I look at my TCI 8174 and it is lacking in the analog section for crystal filter performance. Wide band dynamic range suffers when signals blow around th efilter. This can be easily fixed. The Ten tec and other DSP receivers all cheap out in the analog section filtering hoping the DSP makes up for it.
I'm going to take an analog hot rod and connect a SDR on the output and have real performance. I suggested V8 last night to the soft rock guys with  H  mode mixers.
Or I could just buy a Cubic R3150 for about 2 grand.  fc
A hot rodded R390 with the tuned preselector would rock.

[KA8WTK]

  Actually, the R-390A here is the first radio I would stick the afterburner on. After that I would start working my way down through some of the other radios here like the HR-10 (S-38 maybe?)
  It will be interesting to see what the SDR will do for them. But, you are correct. The better the frontend, the better the results following the the old adage of "garbadge in, garbadge out". While I know I can't make a silk purse from a sow's ear, I might be able to make a nice leather wallet. (Man, what a day at work, all these trite sayings popping out of my head! Better quit typing now.)

[WA1GFZ]

I find the limit of SDR performance is the saturation point of the sound card around 1 volt. I can get 20 volts peak to peak out of mine but the computer wants to barf with that kind of signal. The better the stuff in front of th econverter the less chance a signal can sneak through and saturate the sound card. This means gain distribution is key in the SDR performance. You don't want any extra gain.  fc

[WA1GFZ]

My old 3725 KHz.  Novice crystal has beed sitting on a shelf since 1967 when I moved up to a VFO. I think it will fit on my SDR board moving my sound card IF to 10.6 KHz.
It has hung off  6V6GT, 6L6GT and 6AG7s. I wonder if that tiny surface mount 2N3904 can vibrate that FT243 rock. What a fitting place to mount it . An SDR interface.

[WA1GFZ]

My Soft Rock V6 and V7 kits arrived today. Very nice little boards and you just can't beat the price. You would spend more money buying the parts. V6 $14, V7 $20

Bill K8wtk just asked for the 455 kHz modification with a V6 and will post it here.

1. Replace the crystal with something that will land near 455 KHz. Better to be off a little
to reduce noise around zero hz..
    A color burst crystal will give a 7.625 KHz IF to the sound card. A 3720 novice crystal will give you about 10 kHz into the sound card. Your software bandwidth will be just a little less than twice the IF into the sound card. You have the option of divide by 8 or 16
with jumpers so a 40 meter crystai would also work. It can land above or below 455 KHz.
2 Remove the input filter ahead of the transformer.
3. determine your IF output High Z or low Z. A high Z use a resistive divider with 50 ohms across the SDR input and a series resistor to the radio output. You are driving a low reisitance transformer so use a blocking cap in series if there is any dc voltage present.
Use a 50 ohm pad if your 455 KHz. signal is 50 ohms. I did 6 dB 2- 150 to ground and 36 ohms series. Either way the input is forced to 50 ohms.
4. Now connect to the sound card and look at the spectrum display level on I like I2PHD software.
 The no signal level should be below -80 dB. and above -100 dB. If the signal is too high then parallel the op amp feedback resistors to reduce the gain. You don't want to saturate the sound card which will limit the dynamic range of the SDR (I)
5. The goal is to just raise the sound card background noise a few db when the board is installed to allow maximum dynamic range.

My RA6830 I'm running a 6 dB pad and the gain of the two amplifiers is about X10 or 20 dB. My background noise is still a bit high at around 80 dB when connected to the radio.
I'm building a radio around the module so will set the gains then. My board has a resistive splitter so I can also interface the board ahead of the 455 KHz filters for a wider input bandwidth. I did 3- 33 ohm resistors star connected one in one out and the third to the sdr(I). Still have not decided on the final configuration of the interface but the spectrum display alone is well worth the trouble.

[WA1GFZ]

I think 1/2 the kits have been sold in the past two weeks....

[KA8WTK]

Thanks Frank!!

[WA1GFZ]

Started building my soft rock 6 last night high quality stuff.