New NRG prototype

Everything technical about radio can be discussed here, whether it's transmitting or receiving. Guides, charts, diagrams, etc. are all welcome.
nrgkits.nz
Neckmin
Neckmin
Posts: 337
Joined: Fri Oct 17, 2014 10:35 am

New NRG prototype

Post by nrgkits.nz » Mon Oct 22, 2018 8:52 am

This will eventually become the second revision of the NRG pro5, the main difference is the output device is now RD06HVF1 instead of a 2SC1971. The oscillator is still the same using a dual gate mosfet BF981 going into a BFR96, and I've also added a second BFR96 for some extra gain to provide enough drive to the RD06. I quickly knocked it together and without much tweaking it's getting 5.5w out, it's also relatively flat across the band.

The pll chip I'm also going to change and instead of the TSA5511 it will use the MC145170, there's also going to be reverse and forward power detection with VSWR readout. The whole board will be SMD as much as possible and I'm planning on getting a bulk lot done by pick and place/reflow.

I'm also not sure if anyone here knows if or uses SimSmith but I've found it very useful for calculating input and output matching networks.
image.jpeg
You do not have the required permissions to view the files attached to this post.

Albert H
proppa neck!
proppa neck!
Posts: 2737
Joined: Tue Apr 05, 2016 1:23 am

Re: New NRG prototype

Post by Albert H » Wed Oct 24, 2018 2:31 am

The choice of the MC145170 as your PLL IC may come back to bite you - it's been obsolete for several years, and it will become unavailable just when you want to order 500 of the little blighters! I've recently been using the LMX1601 from Texas Instruments with a little AT Micro controlling it. It's a very cost-effective IC set-up.

There's even a newer version of the PLL IC which includes a VCO - the LMX2954. This can make for a very compact exciter board. I have a prototype etched board experimental exciter that's just 5cm by 2.25cm and delivers 5 Watts! The PLL / VCO IC and the amplification stages are on one side of the board and the AT Micro and the heatsink are on the other.

One of the prototypes had its corners trimmed off, so that it would fit - very precisely - into the smallest Eddystone diecast enclosure, using the box as the heatsink! This is now a very small, waterproofed rig that runs for several hours from a gel-cell, and drives a Slim-Jim aerial that's made from 300Ω twin feeder. This pocket-sized 5 Watt FM broadcast rig will be put to good use at music festivals next year. There's also a miniature (also pocket-sized) "studio" to go with it, using a couple of big capacity SSDs to contain the music, a microphone input with compression, a stereo limiter and a stereo coder, with Bluetooth to use a mobile phone as the display. The whole system can be assembled and on the air in minutes. Setting the output frequency can be done using a mobile phone application as this version also has a Bluetooth module talking to the AT Micro.

We were thinking that a little radio station like that could also be used for sporting events, carnivals and parades.
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

nrgkits.nz
Neckmin
Neckmin
Posts: 337
Joined: Fri Oct 17, 2014 10:35 am

Re: New NRG prototype

Post by nrgkits.nz » Wed Oct 24, 2018 7:02 am

Thanks Albert, I've got several lots of the MC145170 (about 200) so I might use these up first - they're easy to get out of China, the only problem is the vendor supplied them loose in bags not on a reel but all the ones i've used so far (manually soldered) have worked without any issue. I had a quick look at the datasheet for the LMX2954, I see the VCO freq range only goes as low as 1990MHz for part number LMX2541SQ2060E - are you using the built in VCO or your own VCO? I also see that it uses Microwire which is a some kind of SPI, I may have to rewrite my SPI routines in my PIC code, i'm going to investigate this more as I may also have to use a different PIC. Also where do you usually purchase these from? All the usual places e.g. Digikey don't seem to have them. I'll be looking to get around 500 to 1000 of them on a reel.

Edit: Found them on Digikey https://www.digikey.co.nz/products/en/i ... ?k=LMX2541

They're prohibitively expensive though.

radium98
proppa neck!
proppa neck!
Posts: 907
Joined: Fri Aug 26, 2016 7:01 pm

Re: New NRG prototype

Post by radium98 » Wed Oct 24, 2018 5:05 pm

What k8nd of bfr96 is this .i never seen it before

Banus_radio
no manz can test innit
no manz can test innit
Posts: 124
Joined: Wed Aug 24, 2016 9:58 pm

Re: New NRG prototype

Post by Banus_radio » Fri Oct 26, 2018 4:07 pm

radio-berlin wrote: Tue Jul 28, 2015 5:03 pm Heres some radio porn.... A sneak preview..... Not built yet as im abroad and dont have all the stuff with me. As soon as I build first one ill post some pics. Ignore the website on it, its not up n running yet, just have my rds stuff on there for time being...

The large tin plated area is where the heatsinks sits with 3 self tppers from underneath the board (2 eitherside of the rf transistor for good grounding). There is pre-empth jumpers, power jumpers and power pot, usual rf dc and lock leds. Fingers crossed that I didnt make any mistakes as i didnt etch one up before manufcture :mrgreen: ... to much hassle lol .. The pic has uploaded shit quality, but writting on the left says "RIP Stephen 'NRG' Moss - A True Radio Legend".. Thought id pay some respect as used his oscillator ;)

propll5.jpg
You should contact Berlin as he got a run of PLL PRO 5 manufactured. I know he no longer doing stuff so he might send you the boards for free considering where you are based in NZ.
Last edited by Banus_radio on Fri Oct 26, 2018 4:10 pm, edited 1 time in total.

Banus_radio
no manz can test innit
no manz can test innit
Posts: 124
Joined: Wed Aug 24, 2016 9:58 pm

Re: New NRG prototype

Post by Banus_radio » Fri Oct 26, 2018 4:09 pm

not sure why the picture didnt show, its here, copied from the old posts on this forum
You do not have the required permissions to view the files attached to this post.

Albert H
proppa neck!
proppa neck!
Posts: 2737
Joined: Tue Apr 05, 2016 1:23 am

Re: New NRG prototype

Post by Albert H » Sat Oct 27, 2018 2:25 am

nrgkits.nz wrote: Wed Oct 24, 2018 7:02 am Thanks Albert, I've got several lots of the MC145170 (about 200) so I might use these up first - they're easy to get out of China, the only problem is the vendor supplied them loose in bags not on a reel but all the ones i've used so far (manually soldered) have worked without any issue. I had a quick look at the datasheet for the LMX2954, I see the VCO freq range only goes as low as 1990MHz for part number LMX2541SQ2060E - are you using the built in VCO or your own VCO? I also see that it uses Microwire which is a some kind of SPI, I may have to rewrite my SPI routines in my PIC code, I'm going to investigate this more as I may also have to use a different PIC. Also where do you usually purchase these from? All the usual places e.g. Digikey don't seem to have them. I'll be looking to get around 500 to 1000 of them on a reel.

Edit: Found them on Digikey https://www.digikey.co.nz/products/en/i ... ?k=LMX2541

They're prohibitively expensive though.
The 2541 is the one we're using at Band I, II and IV. The 2954 is being used for the microwave links. They're not all that expensive when you consider how many separate components the IC replaces, and the reduction in construction time. They're also just about the cleanest PLL / VCO combination I've ever seen. Phase noise is below the analyser grass, and there are no spurs or harmonics to worry about at all. We buy them through a couple of component brokers that deal with Texas parts.

Two of our lads have been messing with these chips, and developed PIC Code that will set them up. Microchip have published some Microwire driver code, so it really is just a question of chopping it around a bit to make it work as you want. The board they showed me earlier today uses the Texas chip driving an MMIC for about 120mW, into a Mitsubishi FET for 5 Watts at 12V. It makes for a ridiculously tiny PCB, and this prototype even tried SM inductors, but they managed to melt a couple, so gave up that idea! The spectral purity and low phase noise makes for an exciter that matches up to the very best I've seen. There was a Bext beast that we used to use as a reference, but we're several times better than their best now in every parameter.

The most expensive part for us these days is the hardware,the mains transformer and smoothing capacitors. The components for the boards are relatively cheap - the Philips and Mitsubishi FETs are really cheap per Watt these days, and their high gain has reduced the number of stages required. A fully "no-tune" transmitter is easily possible up to a few kilowatts these days - without driving stages into saturation to try to get the same out of them right across the band (I'm looking at you, Paul).

Good quality 19" format cases are still really expensive, and everyone seems to want their gear in 19" cases since it's the "Industry Standard". I was considering doing a range of gear in 66cm (26") cases, to set a new standard. I found some fan assemblies and suitable heatsinks for 66cm format, and the extra width inside the cases would be very useful to help prevent cramming things in! Unfortunately we'd also have to get custom-made cabinets and other hardware, so the prices would be insane. However - if anyone wants to move to 66cm panel width, I'll chip in for the tooling costs, and we can set a new standard!
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

nrgkits.nz
Neckmin
Neckmin
Posts: 337
Joined: Fri Oct 17, 2014 10:35 am

Re: New NRG prototype

Post by nrgkits.nz » Sat Oct 27, 2018 7:52 pm

I’m going to order some samples of the 2541 and give it a try. Albert how are you modulating the built in VCO, I’m assuming you’re probably using the vtune pin?

Albert H
proppa neck!
proppa neck!
Posts: 2737
Joined: Tue Apr 05, 2016 1:23 am

Re: New NRG prototype

Post by Albert H » Mon Oct 29, 2018 1:27 am

The board mixes the output of the loop filter with the modulation. LB discovered that the time constant of the loop filter needs to be about 20 times the lowest modulating frequency, or the PLL freaks out, trying to correct the mod as short term drift. There's also an FSK input, and we've used that for a high speed digital data stream - we're trying fully digital links, with the analogue stereo coding happening in the studio, then digitised and squirted down the link in digital form. We have a micropower link set up that works over about 18 km, giving a noise-free recovered stereo composite multiplex with RDS from a signal that's in the grass at the receiver!
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

nrgkits.nz
Neckmin
Neckmin
Posts: 337
Joined: Fri Oct 17, 2014 10:35 am

Re: New NRG prototype

Post by nrgkits.nz » Mon Oct 29, 2018 5:14 am

Albert H wrote: Mon Oct 29, 2018 1:27 am The board mixes the output of the loop filter with the modulation. LB discovered that the time constant of the loop filter needs to be about 20 times the lowest modulating frequency, or the PLL freaks out, trying to correct the mod as short term drift. There's also an FSK input, and we've used that for a high speed digital data stream - we're trying fully digital links, with the analogue stereo coding happening in the studio, then digitised and squirted down the link in digital form. We have a micropower link set up that works over about 18 km, giving a noise-free recovered stereo composite multiplex with RDS from a signal that's in the grass at the receiver!
I thought that's how you might be doing the modulation which means it will probably have to be AC coupled. The design i'm working on now using the MC145170 (before I looked into this new pll ic) has the mod DC coupled onto one of the varicap anode's - 0v side, with a large value 1mH inductor and series resistor going to ground to correct the mod freq response, i'm getting a very flat responce from 30Hz to 57KHz, and there's absolutely no tilt when modulating a 30Hz square wave. I'm also switching in some high value resistors in the loop filter after it locks, so initially its got a very fast lock time when the resistors are bypassed - the PIC reads the phase pulses to determine lock status and then switches the resistors in.

Looking more at the LMX2541, the Microwire timing requirements look to be the same as the MC145170 - shifted in on low to high transition of the clock.

LMX2541
timing1.JPG
MC145170
timing2.JPG
You do not have the required permissions to view the files attached to this post.

Albert H
proppa neck!
proppa neck!
Posts: 2737
Joined: Tue Apr 05, 2016 1:23 am

Re: New NRG prototype

Post by Albert H » Mon Oct 29, 2018 8:52 pm

I need to look at the code the lads were using. Some years ago, I published a paper on recovering (digital) signals from severe noise (it was part of a PhD thesis), and did a huge amount of work on digital correlators, error correction schemes and data stream recovery filters. Some of my work ended up being used to get the images back from some of the deep space missions, and is still in use in difficult environments to recover broadcast quality signals - one instance is the system used on the Formula One circuit, to get the pictures back from the cars.....

My earliest work on very low power links was done in Band IV and V, using hacked PMR gear and TV Yagis. My preferred receive IF strip used a PLL for demodulation, because this increased the effective receive sensitivity by nearly 14dB. We used to link insane distances with just 100mW transmit, over some very noisy paths!
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

User avatar
3metrejim
no manz can test innit
no manz can test innit
Posts: 147
Joined: Wed Feb 22, 2017 3:00 pm

Re: New NRG prototype

Post by 3metrejim » Wed Oct 31, 2018 11:23 pm

Nice stories Albert. Now have you any pictures, references (online) , etc. To back up your claims?

RF-Head
tower block dreamin
tower block dreamin
Posts: 271
Joined: Wed Oct 15, 2014 9:25 pm

Re: New NRG prototype

Post by RF-Head » Thu Nov 01, 2018 7:44 am

Lol i was thinking the same thing about albert
many stories but i never seen any picture of his high Q transmitters
maybe he is only making them in his head.......
I believe that he knows a lot but also like to see some things he made

nrgkits.nz
Neckmin
Neckmin
Posts: 337
Joined: Fri Oct 17, 2014 10:35 am

Re: New NRG prototype

Post by nrgkits.nz » Thu Nov 01, 2018 8:10 am

I've also got a new PA on the way, it will use the MRFE6VP5300NR1 and deliver 350w across the band.

Albert H
proppa neck!
proppa neck!
Posts: 2737
Joined: Tue Apr 05, 2016 1:23 am

Re: New NRG prototype

Post by Albert H » Fri Nov 02, 2018 4:17 am

3metrejim wrote: Wed Oct 31, 2018 11:23 pm Nice stories Albert. Now have you any pictures, references (online) , etc. To back up your claims?
Look back in the history of this site - I put a few rig photos up here. Pictures of 19" rack cases with ducted coolers on top are pretty dull, though, and the PCBs just look like any other surface-mounted, double-sided board. There are some nifty thermal schemes in the bigger PAs though - two of which are subject to patents. You're welcome to read the papers on data recovery from noisy paths - the mathematics is somewhat abstruse, but if you can understand it, the way the correlated recovery is done is quite beautiful. I might just send you the links privately though (they are rather revealing..!).

The original UHF links from the late 70s used Pye PF1 "pocket phone" Police radios that could be bought from a company called "Garex" for about £5 a set. The PF 1 was a two part effort, with the transmitter (with its silly pop-up aerial) in one box, and the receiver in another. This was convenient for our purposes! Judicious application of a hacksaw separated the useful from the useless parts of each board, and we used to add a 10.7 MHz WBFM IF strip to the receiver - the "Ambit" IF PCB was favourite for this, though we did get a load of car radio IF boards from a company called "Hart" which we used for a time. Each unit was assembled into a small folded aluminium box, and 75Ω Belling-Lee (TV) aerial connectors were fitted for the aerials. The modulation in and out were through phono (RCA) sockets, and a 3.5mm power socket handled the power. There was also a relay output on some receivers that could be used to switch the main rig. I think that I might still have a TX / RX pair in a box in the cellar - I'll look them out and snap a few pix. They really are museum pieces, though.

My earliest Band II rigs were valved, of course. High power transistors were way beyond anything that I could afford, and the QQE06/40 could be obtained for a few quid and deliver the better part of 100 Watts if over-run a bit! The only tricky parts to get were the "butterfly" trimmers. The power supply parts came (mostly) from knackered TVs, as did most of the canned coils that I used. The hardest part of building those old rigs was doing the metalwork! I found some old circuit sketches from the '70s in a notebook last week - I'll put them up here if anyone's interested. The valves we used are still available, and a bottle like the 06/40 only needs a few hundred milliwatts of drive to get it going - a bit like the modern FETs.

PS: I used to be known as "ZIP van Tilburg" (Zeer Illegaal Persoon)
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

User avatar
3metrejim
no manz can test innit
no manz can test innit
Posts: 147
Joined: Wed Feb 22, 2017 3:00 pm

Re: New NRG prototype

Post by 3metrejim » Fri Nov 02, 2018 8:41 am

OK Albert, it would be helpful for people to actually see PCB's as from that you can learn how a suitable layout should be done - you can even cover up sensitive components if you really need to to stop a design being lifted. It just sometimes pains me (and probably others) to hear of great things, but then not even getting a look - your tiny 5W sounds really nice, not like it is hard to make if you can program a microcontroller - but the layout is important for RF... How about some layout guidelines with examples (pictures are nice) keeping in the spirit of this site?

I remember you almost teasing people with the CMOS diode switched PLL (very old tech now), so much so, that I had to go and reproduce it myself (with modifications) to make it accessible to others here. The picture of the three steps is below - there is a problem with the site only allowing a single picture attachment to work...
3pics.jpg
You do not have the required permissions to view the files attached to this post.

MC Spanner
big in da game.. trust
big in da game.. trust
Posts: 93
Joined: Thu Mar 23, 2017 12:01 pm

Re: New NRG prototype

Post by MC Spanner » Fri Nov 02, 2018 11:17 am

3metrejim wrote: Fri Nov 02, 2018 8:41 amThe picture of the three steps is below
Reminds me of some of Stephen Moss's prototypes from the 80s! Nice. :)

Albert H
proppa neck!
proppa neck!
Posts: 2737
Joined: Tue Apr 05, 2016 1:23 am

Re: New NRG prototype

Post by Albert H » Fri Nov 02, 2018 12:45 pm

3mJim - the CMOS PLL is actually about the cheapest solution for frequencies up to about 70MHz. There is a 74F90 that can be used as a prescaler that gets you to well over 100 MHz.... There's a neat enough PCB layout somewhere here for a general purpose exciter - can be built for Band I II or III. It's designed to be cheap - as in "throwaway"! The original version used all through-hole components, and plans for a surface-mount version never really got beyond the basic planning stage. There's diodes to set the PLL division ratios, no switches and no trimmers - it's fairly broadband. The output filter and diode-based doubler needs differing capacitors and coils according to frequency, but almost everything else remains the same. There was also a higher power version that used a pair of transistors in the doubler to get more power to drive the final (about 400mW at Band II to drive an MRF237 to around 5 Watts)

I used to use Stephen's two-transistor, self-doubling oscillator, but later on moved to a dual-gate FET with a gain control loop to flatten the power output across the band of operation. Back when I used to supply rigs to the pirate market, this PLL board was used in hundreds (possibly thousands) of rigs. The board was usually gooped with black or deep blue epoxy resin, both to stabilise the coils mechanically and to prevent the solder-jockeys trying to copy them. I have all the artwork (and Gerber plots) for the boards on a drive here somewhere, and got lots of them made by "PCB Train" at the time, usually in batches of 50.

The 1 Watt exciters for the pirate rigs were designed to cost around £12 including the PCB. The expensive bits were always the output devices, the output high voltage capacitors, the power supplies and the heatsinks.

I'll dig out the original PCB layouts, and if anyone wants to replicate them, they're welcome. The only real caveats are that they'll need an understanding of the PLL logic and they'll need to be able to solve some fairly abstruse equations for the component values in the oscillator, doubler and output stages - it's not just a case of throwing in the right parts and getting it going like one of Stephen's kits. I'll throw in the design equations too.

These days it's seldom cost effective to build broadcast rigs from the ground up - we supply Bext, Harris , Siel and a couple of others. We make our money from the design of the antennas for shaped coverage, installation and configuration services, studio design and construction, and complete "turnkey" broadcast solutions. The only RF stuff we make ourselves these days are the more specialist link gear and some audio processors / stereo coders. We don't supply pirates any more - our current customers (with a few more notable Far Eastern exceptions) are all licenced. We're one of the few teams that will do gear from Long Wave to UHF. One current contract (as an example) is to provide a moderately powerful, shaped coverage AM broadcast facility in the Far East, that's going to be used (as I understand it) as a "Border Blaster" operating into another country. That's an interesting set of problems!

As I said - photos of 19" rack boxes and ducted cooling systems are pretty dull and pointless. The internals of the gear is just like the stuff you've seen from all the big manufacturers.

However - just for fun, I'm building a fairly powerful AM rig for use over Christmas. It will be a pirate, and if you're in South-East England, it should be easily audible in the daytime. Frequency is still to be decided, but one of us will post details here nearer the time.
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

nrgkits.nz
Neckmin
Neckmin
Posts: 337
Joined: Fri Oct 17, 2014 10:35 am

Re: New NRG prototype

Post by nrgkits.nz » Fri Nov 02, 2018 7:48 pm

Very interesting read Albert, will the border blaster be used in South Korea by any chance? I know they got all sorts of transmitters there sending stuff north with some extremely high erp’s

User avatar
3metrejim
no manz can test innit
no manz can test innit
Posts: 147
Joined: Wed Feb 22, 2017 3:00 pm

Re: New NRG prototype

Post by 3metrejim » Sun Nov 04, 2018 7:12 am

MCSpanner - It was a device made in the 90's that I had lying around in amongst other bits and pieces of home-made prototypes. I added the CMOS pll just to prove that it would work, when I became fed up with Albert raving on about it, but never providing sufficient detail to actually construct one; well at least if you can't reverse engineer a photo - the top pic is from one of the 'old forum' posts (a band I link) and the test jig is from 2014. The diagram is my own work - even though it has been posted with the text removed elsewhere.

Albert H - Please post away. Any information, even if out of date, is a valuable learning resource. Even if FM transmission will soon become a thing of the past (in the UK anyway), any technical information also makes a great piece of 'pirate radio' history, that should be preserved. If someone like that Mr. H guy decides to use it, you can always claim some sort of copyright infringement. Make it free for personal use but not for sale without express permission (and royalties) to the author - hard for a business to get out of - just like for patents, I believe.

Post Reply