Premium Rcvrs, FlexRadio, HPSDR, SoftRock, SDR, DSP, PCs, OSs, etc.!

[Mike/W8BAC]

I picked up the other Cubic (R-3038) receiver this guy was selling. It's not an SDR but it is a dual receiver with filters included.
Anybody have a suggestion where to find a manual and schematic?

Mike

[VE7 Kilohertz]

Well Frank is the Cubic wiz here, but that 3038 sure looks like a 3030 or 3080. I know the 3030 manual is online, but can't quite remember where. If you get stuck, let me know and I'll email you the 3030 one.

cheers
Paul

[Mike/W8BAC]

it occupies the position of being the single most expensive piece of radio gear that I've ever purchased! Tongue

Hay Mack,

You could have spent that kind of money easily on a Howard Mills restomod 75A4 or some other highly prized bit of fluff. Not knocking Howard's work or the 75A4 at all.

That 3280 is going to be a lot of fun Mack. Your going to love it. I know you'll keep all of the SDR fans here (I include myself) up on your progress.

It sounds like you might be able to figure out how to communicate with it and I might need help with the 3038. I'll be asking a few questions from time to time as I learn about it.

The manuals for the 3030/3080 will probably give me enough information to keystroke though the menus to light up and run the two receivers. I think I found enough information to hook up my GPS disciplined 10 MHz oscillator as well but the rigs probably have a number of differences as well. Thanks for the post. I suspect the manufacturer doesn't want the manuals and schematics for the more esoteric equipment passed to the public or competition. We will have to figure out the radios ourselves I guess.

Tell you what. In your searches for manuals for the 3280, ask for information on the 3038 as well. I'll do the same. Have fun. Way to go!

Mike

[WA1GFZ]

My guess the 3038 is just like a 3030 only 1 front panel control to the CPU driving 2 rf sections in parallel. I have a bunch of cubic stuff so let me know if you have a problem. Like Racal the weak link is the caps on the supply rails. My guess the left side is missing a panel interface, CPU and remote module and the control signals are hard wired across motherboards. There are not many control  signals required. 4 for preselector, 4 for the filter module and 4 or so to detector. Then serial lines to BFO and synthesizer modules. Heck you could eliminate half of them too.
This would make a cool diversity module. BTW I have a friend in FLA who just had key pad overlays made. That is a must have spare. I try to use a knuckle when I press them and not a finger nail to prevent damage. I think he is getting $10 or $15 each

[KA1ZGC]

Mack,

As far as there being no SDRs for unix:

http://www.gnu.org/software/gnuradio/

Looks pretty complete to me.

Personally, I'd use this (or any other SDR) off the IF of an existing transceiver to prevent front-end overload. There's plenty of code out there to control modern transceivers, all one would need to do is set the software for the IF and have it command the rig to set the target frequency.

That would be minimal coding effort, and would get around what Frank (very correctly) calls the "TRF"-ness of current SDR methodology.

If you don't care about front-end overload, the hardware is out there to run this straight off the antenna. Haven't looked at the prices, but it looks pretty straightforward.

[WA1GFZ]

There is a nice lady at Cubic but it is very hard to get stuff out of them. This is new stuff so it must be PDF files. Cubic also has a 455 KHz output. They use a single tighter roofing filter usually 10 KHz wide. Racal usually uses two 20 or 25 KHz wide.....unless you are very lucky to get a pair of 8 KHz units that are very tight. I was lucky to find a pair for my hot rod.

[KA1ZGC]

Jack, I'm not at home now and can't look at my files on the subject but it seems that I did find that website a while back and investigated it. Just from memory here, it was a fairly basic affair on the hardware & software and didn't seem to offer much.

I went searching through the thread looking for someone named "Jack" and didn't find one. Were you referring to the link I posted?

If so, then you should look at it again. It does pretty much everything under the sun, even GPS, PCS, and HDTV. Works on Unix, Mac, and windows. There are pre-built executables for all the OSes it runs on, or you can download the source code and build/modify it yourself.

The front-ends are all there, too. All you need is the hardware.

The primary interface is a little over $700, which is a little higher than I'd expect, but not too bad. There are daughterboards that cover HF, VHF, UHF, and SHF. Couldn't find the prices on them.

Looks like a pretty freakin' slick package to me. I've been going through the source code for the last few hours and I like what I see.

Certainly worth looking into.

[Steve - WB3HUZ]

Good thread. Probably should have been in the tech section all along.

Some background on digital radio/digitizing RF at the link below. It's a little dated by the basics are good reading.

http://www.amwindow.org/misc/pdf/DigitalRXConcepts.pdf


And a short, but fairly good intro to SDR (with their own approach touted) from Vanu. Having had some involvement with the company in the past, they do some pretty good work.

http://www.vanu.com/wp-content/resources/intro/SWRprimer.pdf

[WA1GFZ]

HUZ,
Very good article. I need to read it a dozen or so times.

Cubic is weird about letting too much data out. I have a pile of 35XX series VHF/UHF stuff but try and find a manual for it. My part of the cubic deal was to build a 3512 without any information. The only thing I had to go on was the stripe painted across the top of the modules. It took a pile of work to get it going. JN is very close to it while on vacation. It was sold to a retired friend in P.I. What a beautiful design. A friend just bought a similar unit off ebay and tried many calls to Cubic with only a simple users manual. Getting info on the innards of the modules requires inside contacts I have been unable to make. The closest I got was a guy who works there and he does repairs for big bucks and wasn't about to let anything out.
It seems Cubic got rid of most of the hand drawn data files. I'm sure the 3280 is in a format that could be copied as PDF but someone will have to spend the bucks to get it. Good luck if it is USG property.
Interesting thing about Cubic looking at the 3280 block diagram. I bet many circuits in the 3280 came right out of the 3030.
Makes me wonder what the new stuff looks like if this is hitting the surplus market.

[KA1ZGC]

The primary interface is a little over $700, which is a little higher than I'd expect, but not too bad. There are daughterboards that cover HF, VHF, UHF, and SHF. Couldn't find the prices on them.

Thats a pretty high price in the SDR world. The Perseus is only $1000 (jam up hardware & software), SDR-IQ & SDR-14 even less, SoftRock kits are dirt cheap but the pits for an old geezer like me to build.

Sorry about the name change Thom, senility is really cool, I think?

Only $1000?

So, $1000 bucks for hardware & software is less than $700 for hardware & software?

You confused me less when you called me "Jack"!   

All kidding aside, for two seperate (mostly-)full-duplex transceive chains that can be run on a 1 or 2 GHz machine with a half-gig of RAM and any OS you like, that doesn't seem that bad a deal to me. The high-speed number-crunching is done in an FPGA which allows a fairly modest host machine to run it. It's a USB 2.0 attachment, which is actually pretty sensible (I've never been a huge USB fan, but in this application it makes sense).

I've been looking for a development platform that will allow for a rig with a software IF that has auxiliary digital ins/outs for things like external AGC, filter switching, and so on, and this looks like exactly what I've been looking for.

The main reason I brought it up was because you said (quite emphatically, in fact) that there were no SDR packages for unix systems. Once I got looking into it a little further I realized that this would give me the means to build an SDR system that can address the shortcomings of entirely-software SDR setups.

Not to mention, it runs on unix. 

Hey Frank, a couple of questions for you, while I'm on this topic:

When we last discussed this on the air, I mentioned that I've got piles of SBL-1s and SRA-1s from back in my microwave days. You said those would be poor candidates for mixers in this kind of application. I'll take your word for it, but I'm curious as to what their shortcomings are, and what would be better mixers for this kind of application.

Also, what's the word on phase noise with DDS chips? Good/bad/what? Should I look elsewhere for frequency conversion?

Now, I'll be good and sign my name so you old buzzards don't start calling me "Craig", "Allan", or "Johnny". Just try not to call me "Dave"*, okay? 

--Thom
*Anyone who's heard my "Dave" incidents on 75 knows who I'm talking about.

[W2INR]

Okay Dave

[WD8BIL]

Dave

[WA1GFZ]

SRA1 and SBL mixers work fine but not the greatest for dynamic range. They could be used as a I/Q demodulator. A good DDS has spurs even the newest ones have them. A good clock will limit phase noise to a very good value. All synthesizers have phase noise.
The direct A/D with a clean clock is the future.
BTW, not that I'm bashing Flex but look at the power connector on the 5000. A little molex connector for 25 amps WUWT.
HPSDR guys are about to release the 1/2 watt exciter soon. I've had a number of emails with the designer over the few days. He figures the BB spurs down around 75 dB. I asked him to post a picture. There is talk of a spectrum analyzer software package with a tracking generator. Covington's RX sounds cool but his price is right up there with Perseus.   

[KA1ZGC]

SRA1 and SBL mixers work fine but not the greatest for dynamic range. They could be used as a I/Q demodulator.

No, I'm talking about just using them as straight mixers.

We were discussing building real RF sections and mixing down to a filtered IF and digitizing from there (and the other way around for transmit). SDR without the "TRF" factor. This platform I'm looking at allows for just that. It's got two DACs, two ADCs, all coupled to an FPGA. Gives you two RF ins, two RF outs, and spare analog and digital lines for other applications (like AGC, DDS, other control lines).

The hardware: http://www.ettus.com/.
The software: http://www.gnu.org/software/gnuradio/

This is the perfect launching platform to build exactly the kind of system you and I were discussing. It allows us to do what software does best while still doing in hardware what really needs to be in hardware. Just adding an AGC can keep us out of the saturated-ADC scenario (which we agree is the biggest weakness).

Best of all, absolutely nothing about the hardware or software is proprietary, and there's no requirement that one be used with the other.

A good DDS has spurs even the newest ones have them. A good clock will limit phase noise to a very good value. All synthesizers have phase noise.

So if I wasn't sloppy about clocking it, I'd probably be okay?

I'm tending to stick with the most used of the SDR line, where you have large user groups that provide a lot of info and help if needed and no programming skills required. The higher priced units take a lot of the processing load off your PC and the cheaper ones put all the load on your PC.

No programming skills required. I don't know where you got the idea there was. The Ettus unit costs $700 but takes a heavy load off the PC by doing most of the work in an FPGA.

Some of the SDR stuff seems slanted towards consumer grade electronics users, the GNU Radio website is showing a lot of screen shots of HDTV captures, I can plug a coax into this PC and use the factory provided Hauppauge WinTV card for that purpose.

...but can your WinTV card generate and transmit HDTV? How about radar, GPS, PCS, WiFi, or any other service you can name?

I don't want or need a 'consumer grade' SDR rcvr.

What is that supposed to mean?

The GNU Radio seems to be a new website since their "timeline" starts on 12/25/07 and shows first release as 10/22/07, it's origin seems to come from some academic papers dated Jan. '06, that pretty much explains why I was unaware of it

HUH?!? Where are you getting this from?

Mack, go back and look at it again. The project started in 1998! Where did you get this "first release" information? You've got your facts completely wrong, here.

and everything there reads like "for i in `find $(exampledir) -type f ! -perm 755`; do \" and that's a big 'no thanks' from me.

So download the pre-compiled Windows executables and you're on the air instantly, just like any other SDR! You don't even need to use that evil, nasty command line for anything! Just because something allows you to modify or improve it doesn't mean you have to!

I get the idea you're not reading as much as skimming to find the first excuse to rule it out. Your loss.

I can't see having an A/D sample the whole of the HF spectrum at once, as the cheaper models do, that's why I ordered my SoftRock rcvrs in the 455khz & 500khz IF version, to use a standard rcvr as the front end. It just doesn't make sense to try to sample that much at once in your PC hardware & software or in the SDR box, unless your using it for test gear, the spectrum analyzer & panoramic display functions.

Then don't! Nobody's got a gun to your head!

Appearently your interest in SDR is limited even more than you think some of the options limit you (and no SDR option limits you to anything, by the very nature of SDR).

Buy whatever you want, it's your money. But after making broad, sweeping (and often insistent) statements like "the only SDR software out there is written for Windows only", "there is no SDR software for unix", "there is no SDR software for the Mac", and just about everything you've said about GNU radio being entirely wrong, I would recommend you do a lot more studying before sinking much more money into this stuff, as a lot of the conclusions you're basing your decisions on appear to be the product of someone else's mis-information.

Not trying to give you crap, Mack... just think about it, okay?

--Thom

[WA1GFZ]

Don't waste your money on anything over a year old.
Stick with softrock if you don't want to invest in the newest hardware. Things are changing fast as new parts appear.

[KA1ZGC]

Thom, my ONLY interest in SDR is in rcving HF. So all the stuff that a lot of the SDR guys are doing up high doesn't interest me at all, I could care less about rcving and displaying the signals from stars. No sir, radio astronomy, VHF, UHF & SHF are of no interest to me.

Then don't use those functions. That's all in software, not hardware. It's the same with any SDR.

I just couldn't understand why you were so quick to cast stones at it given the amount of money you were eager to throw at a receiver you can't get any support for.

Statements like "I don't want or need a consumer-grade SDR" or that this package was "a fairly basic affair" that "didn't have much to offer" are out of place, given that you only want an SDR to receive HF. Any pile of rubbish can accomplish that. All SDRs are capable of doing everything you say you don't want one to do, and they all have those functions built in.

From reading on the ettus website, $700 gets you a motherboard, enclosure and PS (wall wart), you need daughterboards to actually make it function.

Not for what you want to do. The HF daughterboards are really just a straght shot from the motherboard to the RF connection. One uses transformer coupling, the other uses differential amps. No other work is being done at that level, so you can accomplish the same thing with a hunk of wire. You don't even need the Ettus motherboard, it just takes a lot of the load off the PC itself. You can plug the software into anything you want to plug it into, even the SoftRock units you're already purchasing.

There is no iron clad 'best way' to accomplish any of this SDR stuff.

Then it seems a bit disingenuous to immediately dismiss any approach as somehow doing it wrong.

There is also a "catch 22" in implementing FIR filters. To get the best shaping, you have to turn down the sample rate, a counter productive move in my opinion. My limited knowledge leads me to believe that lowering the sample rate is a bad move, my concepts lead me to believe that the maximum possible sample rate is always to be desired.   

That's the catch with SDR, no matter who made the hardware or software. You can't change the laws of mathematics. An FIR is an FIR is an FIR, no matter how it's implemented or who wrote the code for it.

Once you get to a certain threshold with the sample rate, going beyond that doesn't gain you anything. Since you're talking about a large bandwidth being whittled down to a small one, the sampling rate only needs to satisfy the Nyquist theorem. Since your FIR is setting a max frequency, you have no need for any real resolution beyond that. Going much higher than that falls into the realm of audiophoolery, it's just doing something that makes no practical difference for sake of perceived gain that's immeasurable in reality.

--Thom

[WA1GFZ]

I suggest reading the link HUZ provided. There is some good information in it. I learned something today but will need to read again by morning.
I'm only interested in raw performance and don't care how you get it analog or digital.

[KA1ZGC]

Thom, if you can straighten out all the dummies at FlexRadio, HPSDR, RFSpace, Perseus, SoftRock, etc. on how to properly execute FIR filters, then why the hell don't you do just that, telling me about it won't accomplish a thing!? Here I am repeating the laments of long experienced SDR users from the email reflectors about the FIR filter problem and you've been hiding the cure all along!

I didn't say I had the solution, Mack. Calm down, go back, and read it again. What I said was: the problem is mathematics itself. Ask anyone on those reflectors and they'll tell you the same thing.

What I also said was that once you've got a pre-determined bandwidth (therefore, pre-determined max frequency), the sampling rate only needs to be so high, and going higher yields rapidly diminishing gains. Ask anyone on those reflectors that, and they'll tell you the same thing. It's only your opinion that the sampling rate needs to be any higher than that which yields a measurable benefit.

For that matter, ask any math professor that, and he'll tell you the same thing.

If you believe that exposing an ADC to DC-SHF all at once results in the same I&Q stream that only sampling a small portion of spectrum does, then either you or I need to look over some basic hardware info!?

Go back and read what I wrote again. I said the exact opposite.

Thom, you need to quit hiding your light under that basket, step out and shine, make millions of $$$$$$$$!!!

You need to calm down, go back, and read what was written instead of just skimming over it and coming up with something completely different. I'm not the one who made all kinds of false claims about whose SDR package does what, and what it doesn't do. When I show you examples that counter those claims, you blow your top and shoot the messenger.

Since I'm clearly wasting my time sharing anything factual with you, this conversation is over.

[KA1ZGC]

Thom, I'm all  !!!!!!!!!!!

I was just pointing out the obvious, if you've got the superior knowledge on all these SDR problems, make your move, cash in on it.

Thars gold in dem SDR hills!

Again, I never claimed I had any solutions to anything. I only stated what the actual problem was, and why the solution is so elusive. You're putting words in my mouth.

You can throw all the smiley faces on it you want, personal attacks are personal attacks, and I won't carry on a conversation with someone who is going to stoop to that level when presented with facts that don't agree with their opinions.

Good night.

[WA1GFZ]

read the huz article and learn about oversampling.
Aim then shoot
Or as the OM says make sure brain is engaged before putting mouth in gear.

[WA1GFZ]

The problem with up conversion then down conversion is the close in phase noise of the synthesizer. Go to a DDS source and you live with spurs down only 80 or so DB.
The A/D of the antenna needs to be very high dynamic range or have a very good preselector ahead of it.
Making very fast A/D parts with many bits is easy to say but quite hard to do. As noted in the article HUZ provided. It is a bit dated and L.T. has caught up a bit. Even interleaving is hard to do because the parts need to track perfectly.
I have followed this stuff for years when parts could barely track a missile head as a plane flew over a target. So now that we are decoding RF is a very big deal a couple decades later.

[WA1GFZ]

Check the next issue of QEX and Drentea's RX article if you think hardware guys been resting. I've been following synthesizer design since about 1975 and have designed enough of them to know the pain. Kind of hard to duplicate the best HP synthesized generator in less than 100 cubic inches.

[Steve - WB3HUZ]

The 8640 is VERY clean.

[w3jn]

The 8640 is one of the cleanest sig gens ever, but it ain't synthesized.  It uses a very high-Q tunable cavity as the element in a free-running oscillator.  That, and plenty of attention to filtering the power, makes the difference.

[WA1GFZ]

Show me a synthesized source cleaner that a ham can afford and I'll buy it. Till then A half hour of drift is a minor issue.
John, You still in P.I. I would love to Visit Ken Gaskin out there. He has quite a collection including the R3512 I built..