POCSAG paging not working

[helmutforren]

(((Hopefully I found an acceptable forum for posting this question. There are so many...)))

BACKGROUND:

I have an existing commercial product that uses a TEKK brand transmitter, in combination with custom POCSAG encoding hardware and software, to send a page to a paper receiver. This has worked great for over a decade.

Now the TEKK brand transmitter is being discontinued. I am having difficulty getting an alternate transmitter to work properly. Since the TEKK transmitter was costing us under $100 each, I'm trying to find something that's at least under $150, as opposed to one I found that works but costs $700.

Note that I have over 40 years electronics experience, but relatively little radio experience.

PROBLEM:

I've had partial success with a Friendcom (Chinese, of course) FC-302 transmitter. But I am not getting any customer support to help me get all the way to full success, even with the promise of volume purchases.

When I try to send a page through the FC-302, the pager receiver most often receives the page, but after 10 or 20 out of 50 chars, the page gets corrupted and truncated. Sometimes, the pager receiver doesn't get the page at all, presumably due to corruption within the CapCode portion of the page. Sometimes, the pager simply says "tone only", presumably due to some other kind of corruption.

This is definitely a problem with the transmitter, and most likely my method of driving it. But it's NOT a problem upstream, with my software or hardware specifically. That is, that I can get my software and hardware running, connect the old TEKK transmitter, and get great perfectly formed pages on the receiver. Then I can unplug the TEKK and plug in the FC-302, and start getting corrupt pages. The 12V power supply level, the PTT signal, and the data signal all look good on the scope.

So I need help figuring out why the FC-302 pages are partially OK but not totally OK. Chinese support hasn't helped, so perhaps one of you can help.

Note that I put a spectrum analyzer nearby. The trace on it looks very similar between the two. I can't see any difference. It's a circa 1996 spectrum analyzer (BK Precision 2630) that seems to have at best a 20kHz bandwidth and 100kHz/division display. There might be a difference in the two radio signals that's just too small for me to see with this tool.

Note also that I can put the spectrum analyzer in "scan mode". I can only guess what that does. I guess that it shows me a signal across time, for demodulating at the knob-tuned frequency. When I do this, I see a square wave that's a facsimile of my input data signal. So I believe this is in essence doing the signal demodulation to give me a real time signal, just like the pager receiver does. Well, these look very similar between the TEKK transmitting and the FC-302 transmitting. However, the vertical position on the screen depends a lot on the center frequency chosen by knob (and drifting). The peak-to-peak magnitude of the signal is never more than half a vertical division (no scale provided by tool, that I can see), for both transmitters. So, again, I can't see a difference with this tool.

What else might I try? What else might be my problem? Surely, the FC-302 signal is close enough that the pager receiver *almost* gets it. But not quite. There ought to be some other thing I can do to bring this home. Any ideas?

Note some more things. I don't know the spec for frequency shift that I *should* have for POCSAG. I don't know the shift created by the working TEKK. I don't know the shift created by the not-quite-working FC-302. I don't think I can see this with my spectrum analyzer. I don't know if I need to rent another tool to check this, or if I can get the problem solved without a new tool, just some expert advice instead. The FC-302 should, I think, be causing a shift of 2.5kHz to 3.5kHz.

I have Friendcom's "FSK" modem inside the FC-302, but I think I'm bypassing it anyway. They also have a "4FSK" and "GMSK" modem. However, since I'm encoding the POCSAG myself and simply driving a data line, I don't think I'm using the modem at all. I don't know if I can twiddle my voltage input levels to get more frequency shift. I'm currently providing contact closure to ground again the radio's 5V pull-up.

MEANWHILE, I've tried a much more expensive transmitter from Apollo, their TX125EN. It works flawlessly, but costs too much. On my spectrum analyzer, however, rather than seeing just a peak, I get a rather wide flat top. So this tells me the TX125EN is outputting a much, much wider band signal than the TEKK or FC-302. The TEKK works with the smaller bandwidth. Maybe the FC-302 is much smaller, but I can't see it with my tools. Am I grasping at straws now???

Any and all advice is greatly appreciated.

Oh, one more thing. As described, I'm encoding the POCSAG myself, and proven successful by the working TEKK transmitter for the last decade. I could instead actually let the Friendcom FSK modem do that POCSAG encoding instead. But when I try to talk directly to the FSK modem over RS232, I get no response whatsoever. Friendcom provides zero doc about the RS232 protocol. I tried "AT<CR>", but get nothing. I also have no clue what baudrate. I've tried 1200 and 9600 and 19200, perhaps even 115,200. I tried a message as used with the TX125EN, that begins with <STX> and ends with <EOT>, but I get no response or radio frequency out of the FC-302. If only they documented the protocol on their RS232 input. Any guesses?

Thanks VERY much,
Helmut

 

[freddaniel]

No specific information provided such as band, power, deviation, etc, but I will offer the following:

POCSAG is a digital modulation signal and when input into an analog transmitter, it will occasionally work. This is because some digital transitions from high to low level occur over a longer period of time. These long transitions on an analog radio allow the frequency modulation to drift back towards the center frequency and not remain high or low, as would be the case with a transmitter with a digital input. Your transmitter MUST use a digital input.

If you are using a transmitter with a digital input, then your modulation level is probably incorrect.

I presume your required transmitter is low power, due to it's low price. I therefore suggest you consider a Ritron Inc. DTX series telemetry transceiver. They are American made and they do provide customer support.
You will need a frequency and deviation meter to properly setup your transmitter.

 

[helmutforren]

Fred,

Thanks very much for your reply. I'll look into the DTX for sure.

BEFORE ANYTHING ELSE, would you consider consulting to me on this? That is, allow me to pay you to get it working?

Regarding your other comments, the POCSAG signal includes NRZ encoding, but my decades old recollection says it also includes run length limitation to minimize the drift you mention.

You do imply a significant point, however. Maybe the FC-302 drifts faster, which means the decoded [somewhat] digital signal will droop faster.

Next, I'm trying to parse your meaning. Please tell me if I understand you correctly. You're saying that because the transmitter must be using a digital input, that my digital input voltage levels shouldn't matter. And because of this, it therefore must be my modulation level that's incorrect. Is this what you meant?

This may indeed be the case, but the reason I brought it up is because what little doc I have says that the input should be "250--300mVp-p POCSAG (if it is possible, you can change input signal level, use the instrument to calibrate the frequency offset to 2.5 -- 3.5 kHz)." This implies that the input might be analog. I can provide the doc, but I don't see an attachment button on this forum.

When you say the "modulation level is probably incorrect", I take it to mean that the frequency deviation is incorrect. Maybe it's too low, and I can't see that on my spectrum analyzer. (This goes in to your mention of a frequency and deviation meter.)

Note this is 469.7250 MHz, 1W or 5W transmission. Deviation should be defined by POCSAG. However, I read of different values used over time, so I don't know exactly what any of these transmitters use.

Finally, I've never used a "frequency and deviation meter". A $100 price is mentioned at https://www.youtube.com/watch?v=8IBOYoIV5m8 , as well as then a way to do it without that expense. (That expense is NOT an issue for me, at all.)

Regarding the video, might I be able to use that method and simply look on my spectrum analyzer for the carrier to seem to fall to zero? I'm not sure, however, that I could see discrete side bands on my spectrum analyzer. For now, I'm always only getting a few seconds of transmit every 30 seconds, from my software and hardware. I'd have to rig up separate drive, to turn on PTT indefinitely and drive the data line indefinitely. I could do that of course. But first, perhaps, I'd like to know if you can simply take over the job for me.

Thanks very much,
Helmut

 

[freddaniel]

You really need to learn more about digital transmission. If the info is limited for paging, then look at data transmission. From what I recall from 20 years ago, POCSAG requires plus and minus 4.5 KHz deviation. On many paging receivers, they will not decode a 3.5 KHz signal, or a 5.5 KHz signal, so measuring your transmitter deviation is rather critical. Being on frequency is also critical. If you are off frequency more than 1.5 KHz, it probably won't work. It is like watching a movie screen through a window. If you are not centered, you will lose some of the picture or information.

Most FM transmitters, unless specifically designed to modulate low-frequency [about 20 Hz] non-uniform square waves will not work for POCSAG. This normally means being a digital data transmitter. However, if you just need one or two transmitters, most can be modified to modulate the frequency reference as well as the normal modulation input. If this is a commercial project for many transmitters, you really need professional help.

I have no time to take on another project, but another fellow I know has a business selling paging equipment and has vast experience in design, and building paging transmitters. [Yes, there are still people using paging]. He is Mike Glaser at Glaser Electronics 909-669-5050 east of Los Angeles.

You might also find some useful data on a web book called "Pager_Handbook_for_the_Radio_Amateur" at www.qsl.net/n9zia/wireless/pager/Pager_Handbook_for_the_Radio_Amateur.pdf

In the end you will need a real frequency and deviation meter, generally referred to as a Service Monitor. The equipment examples you referenced were fine in 1972 but not today. The service monitors from the era of 1985 through 2010 will work just fine for you. Some of the more reliable units were Cushman CE-5, IFR COM-120 or the 1200, HP 8920, and many more. Look on Ebay for some Cushman units for $100 to $300, and the more popular IFR and HP units selling for $1,000 to $2,500.

Again - on frequency - 4.5 KHz deviation - digital input is a must.

 

[ff2sherman]

I have an existing commercial product that uses a TEKK brand transmitter, in combination with custom POCSAG encoding hardware and software, to send a page to a paper receiver. This has worked great for over a decade.

When you use the word Custom that makes for a challenge in helping.,
did you get this working or do you still need help

 

[helmutforren]

The company Ritron provides domestic customer support. I got that reference from this forum. They spoke to me and sent me a transmitter. It worked great right out of the box, fully configured correctly for my drive levels and other things. So I'm past this problem. Thanks for your concern.

 

[ff2sherman]

The company Ritron provides domestic customer support. I got that reference from this forum. They spoke to me and sent me a transmitter. It worked great right out of the box, fully configured correctly for my drive levels and other things. So I'm past this problem. Thanks for your concern.

Glad everything worked out,
Our Company uses Visiplex and Waveware Paging systems as well as Scope

 

[nzmike]

Just spotted your posts re the FC302 transmitter for POCSAG.

I have designed a sample POCSAG encoder for the FC-302 that replaces the internal modem and have no issues with the FC-302 transmitting messages up to 255 characters.

I do have all the RF test equipment to correctly set up the DC coupled encoding. It required careful setting to ensure the transmitted NRZ signal didn't have huge overshoot or "ringing" causing corruption.

Glad you got the result required, if you need to revisit the FC-302 let me know.