New Radios, higher duty cycle, more reliability!

With the rework underway, and additional time on my hands, I looked into dumping the conversion of commercial radios for a home brew replacement. Most repeaters I have worked on, use either a commercially purchased repeater, and you mold it into what you want, or some convert older Motorola Micor, or  Maxtrac gear. While these work, and you can certainly turn the power down for better duty cycle on the commercial gear, there are a few problems.

 Obtaining the gear. Usually you have to search out Ebay, or a reseller, matching serial numbers so you get the right band, then hunting down someone to do the conversion, or if you are lucky enough and have the test equipment, doing it yourself. I have neither readily available to me, so I decided on another route. Even my local Motorola shop has closed.......Im in the middle of a rural area, Im out of options going this route. I do have some test equipment I can borrow, but this is even limited.

 The solution:

I did some research, and while its not totally perfect, it will work, and indeed it does! It will take a little extra filtering, and a look at some preselectors later, but its been fun learning and designing the replacement.   Meet the DRA818V/U

 This module is a fully functioning UHF or VHF radio commonly found in the Baofeng and other Chinese ham radios. Yes, I know what your thinking, Chinese knock off, it sucks. I happen to own 2 Baofeng UV5RA handhelds. While these dont have all the bells and whistles of the " big 3 radio companies" radios, the general opinion is they do work, for what they are used for, short, mobile handheld communications.

The modules though are available commercially to folks that want to work with them. What others have found out is that while the T/R switch in them isnt the worlds fastest, and there are no output filters at all. They can be used if you are willing to take the extra time and work with them. I decided that one could be used strictly as  a receiver, and the other as a transmitter for the repeater. They were going to need massive amounts of filtering, and even more amplification, but at least they were cheap and easy to obtain, fully programmable with Linux, and if they blow up, who cares, for 15 bucks, get another one.

I decided to use these at full rated power. That is simply 1W RF......driven at full output has a cleaner RF output, and pulls a whopping 350ma of current. Ok I sent in an order. By the way there are vendors on amazon, and while I have no stake in Amazon, you can get them fast and if there is a problem they have a return policy....perfect.

I did more research, and a fellow ham W0ANM has already made some leaps and bound using these modules, as well as KP4TR and WA3DSP. Many have designed nano nodes for use with Allstar, and these are great starting point. Look these guys up, they have plenty to read and have done some of the work already.

I set out to design my own radio module. With a basic working knowledge of the module I was able to come up with a small easy to construct design I liked. Initial test were good, and I am still playing with the setup.

The top is the first generation board. I have since revised this to allow for even better filtering. The module is easy to use and the schematic is really straightforward. There is really not a whole lot to discuss here other than the COR( squelch) circuit.   Q1 and T1 make up the squelch circuit. When a signal is received, you will get two outputs, why? easy of use. I made it so that you can have either a HI( 5V0) output for COR, or a LO ( 0V0) output depending on your needs. This eliminates the need for further conversion if you needed one or the other type of signal. All of the signals, PTT, CORs, Audio out and MIC in are all available at the small header on the right. The other header is just for power. Power is 5V0, a diode knocks this down to 4V4 volts since the module uses 3.3-4.7Vdc. The header at the top is for connecting a 3.3V FDTI module to. This is the programming connection, and is a simple 3.3V serial connection. For programming the module all that is needed is a computer USB port, the FDTI module ( 3.3V) and the connection to TX/RX and a ground. For programming the module for your frequencies, W0ANM had a easy to use python program. Its available on his website, just goggle W0ANM.

Filtering:

The modules alone when powered up have terrible harmonics. The second harmonic usually shows up about -30dBc and the third and fourth arent far below that in power. This is totally unacceptable. At worst case, something on the order of -60dBc would be a starting point, and -70dBc would be better, heck no second or third harmonics would be great but this isnt the case.

On the web there are many great free Chebyshev low pass filter calculators. I opted for using a 5 pole Chebyshev for the first generation boards. This consisted of 2 5pf and a 11pf capacitor sandwiched with 2, 22nH coils. I chose in the first generation board to use chip (SMD) filter parts. I  used 805 SMD chip parts, the next generation will use 1206 smd chips, and Im considering just chip caps and hand wound inductors, that has yet to be determined. In any case, the low pass filter worked really well. I measured the second harmonic at -67dBc, and the third and fourth harmonics were really not even hardly noticeable on the scope. Im pretty certain that this setup will not tolerate a huge amount of usage, but the first generation board is a prototype, not a final production model.  Also note, the prototype board is designed for a 7 pole low pass filter. I added this in case I wanted more filtering or the 5 pole wasnt enough. In the end I bridge soldered the last coil pads together, and omitted the last smd cap.

There is a 1206 LED used for both power, and SQL, for visual indications only. These can also be omitted if you want to further reduce power consumption, its your choice. I like to see when things are working, and Im a big believer in smd LEDs for this purpose. You can troubleshoot a pieve of gear pretty fast simply by knowing the gear, and looking at the indicators.....its always the first place I start, that and meter readings.

Thats the transmitter for now, it will be improved. I have the receiver done, and more test on sensitivity and a preselector design are in the works. More on those, as well as the 30W amplifier design will be posted later.

In the meantime,

Happy Coding!

Behind the scenes, work continues

So life gets in the way a little, thats normal. So goes it when you are working on projects. For me it seems I get started, then have to put it away for a while, and come back, such is the life story of the Repeater Controller.

So I had the prototype working, and it was ok, the boards I designed worked, but there is always room for improvement. I petitioned the University I work for to start a Maker/Radio Club, and asked if the repeater could be located on campus. Initial indications were that it was going to move forward. They asked me questions about the installation, my network needs, and initially rejected my request for space on the old club tower. I replied with a new location on the same building, and told them I would reduce power, and monitor for interference. If any occurred, I was already prepared to deal with it. All indications were they committee was happy, and then the President retired, a new one was sought out, hired, and everything fell through. Its been 5 months now and I havent heard anything from the committee. Im thinking, my request has been scraped. Its ok, Ill move forward, I still have a couple of options, back in the town I live in, that havent been explored. The University location would have covered both cities with minimal power, and ease, but I will now have to look into other avenues to cover the area I want.

Old Board, New Clothes:

The old design hinged on the use of the main computers parallel port. While this worked I did discover that parallel ports are tricky, especially if you accidentally send it a wrong signal, or in my case cross a wire. On the old computer I was using, I did this while wiring up a header connector, hence I lost that input on the parallel port. Not really a huge thing, but got me to thinking. I was using alot, I mean a lot of header connectors. I decided since the committee, was dragging their feet, or just plan quit on me, it was time I addressed that problem. Hence the next generation of boards. What started as fixing just a connector problem, blossomed into a full blown rework.

New Main Board

Old Main Board

The old board had everything external to it. There was the radio interface board, and the computer interface and tons of just boards, and more boards. What I did was reduce the clutter, and take those boards and incorporate them into just one board, complete with its own controller, and computer interface. As you can see by the top  picture, at the top of the board,  there is a radio interface port. This takes the old radio card, and puts it on the main board, its just plug and play. I also dropped all the extra channels. There is a interface for the repeater itseld, and just one other radio, whether it be a control link, for split site operation, or just another remote base, but the two additional radio ports were dropped. If I am having trouble getting a site for just the repeater, Lord knows I am not going to get a site thats going to allow 2,3 or even 4 additional receiver/transmitters.

All of the radio headers are now reworked and use the DB9 connectors at the bottom. All of the signals are routed through here ( COS, PTT, even a 5Vdc and a Ground, should an external key signal be needed. This removed 4 multiple pin header connector from the original design. Power supplies were removed and added externally to this new design. I reused the old supplies from my car project, just ordering more boards, and parts thus saving a ton of space.

Also I switched computers I am using. The new design uses the BeagleBoneBlack Version C as the brains of the controller as opposed to a full blown PC board. This saved space, and allowed the use of USB for signalling which Ill discuss in another blog. By doing this I saved power, space and time. Also by using the BBB VerC, I was able to get rid of some of the signaling ICs, and allow the computer more control, instead of having the computer tell the Arduino what to do. I still use the arduino to control some functions, which Ill get to later, but the use of the BBB really streamlined things! With all of this, cost was also reduced. Original boards for the old system were over 30 dollars for 3, making them a little over 10 bucks a piece, the new boards were just 15 dollars for 3 and cost 5 dollars per board, a 50% savings.....not big bucks, but hey I can buy more parts now.

Thats about it for the new main board design. Its working like a charm, and has added features as well which Ill outline in another blog as well. So work is continuing, and coming along nicely.

Happy Coding!