LPE-200 modem or ProVoice issue?

[Cognomen]

I am using an LPE-200 with the Digital Voice Conversion Method. At random times, the LPE stops passing clear voice. Switching the LPE off and back on gets it working again.

When the problem occurs, the LPE remains powered up and the radio display does not change. Occasionally some data noise is heard through the speaker, but it is mostly silent.

The mean time before failure can be anywhere from a few minutes to several hours.

The LPE does not appear to be getting warm. I've tried adjusting the RF signal generator's output from full to maximum attentuation. I've tried different power supplies.

Finally, I have substituted the LPE with another one, and the same problem occurs. Both LPEs worked flawlessly on their own (outside of the DVCM), happily trunking for hours at a time.

Is anyone else experiencing this issue, or does anyone have suggestions as to what might be causing this?

 

[Radioman96p71]

I think you are the first person ive actually heard of doing this. My speculation on it would be that maybe the radio is getting too much erroneous data and causing the dsp to hang, crashing the radio. The LPE was a decent radio but it, along with the first-gen 700p's with the C56 DSP were prone to weird problems like crashing and just doing goofy stuff. Do you have the signal constantly being pumped in to the radio, input carrier or not? As the static, being perceived as a modulated carrier, would cause the DSP/CPU to constantly try to make order out of chaos of the random noise. Eventually the right combination or bits or simply a buffer overflow would cause the stack to barf and the radio would halt on the last thing it was doing.

As usual, all speculation, but from what ive seen with buggy computers, they tend to behave the same way when exposed to large amounts of junk data.

 

[Cognomen]

Thanks for the reply. I'm not actually sure what the LPE is hearing. The RF signal generator is set to constant output, the computer-controlled radio has its squelch set to open only when receiving a transmission. I have programmed a receive tone into the LPE to eliminate the speaker pops caused by the radio's modem switching from analog to digital. So I "believe" the only time the LPE is actually receiving is during a digital transmission, once the modem switches to digital. But if I'm wrong, what you're saying makes sense. Perhaps I need to try a newer radio.

 

[davidbond21]

I think you are the first person ive actually heard of doing this. My speculation on it would be that maybe the radio is getting too much erroneous data and causing the dsp to hang, crashing the radio. The LPE was a decent radio but it, along with the first-gen 700p's with the C56 DSP were prone to weird problems like crashing and just doing goofy stuff. Do you have the signal constantly being pumped in to the radio, input carrier or not? As the static, being perceived as a modulated carrier, would cause the DSP/CPU to constantly try to make order out of chaos of the random noise. Eventually the right combination or bits or simply a buffer overflow would cause the stack to barf and the radio would halt on the last thing it was doing.

As usual, all speculation, but from what ive seen with buggy computers, they tend to behave the same way when exposed to large amounts of junk data.

That's a good theory but I'm not really sure if this is the case. The way the digital voice conversion method works is that you are combining the voice channel receiver's 10.7 MHz IF with a constant carrier(I use 862.7) in a frequency mixer that outputs the difference mixing product of 852.000 MHz, which you have the radio tuned to. The thing is, until the receiver is active on a voice frequency, the rf carrier is sent to the radio as 862.7 MHz so it shouldn't be unsquelching and trying to decode anything at all.

Since the problem seems to be not with the LPE-200, but with any radio in that position in the DVCM, let's work back up the line. I'd like to ask what type of voice receiver the OP is using, and then a follow up if you are using a DC block on the 10.7 MHz IF output. Also, has this problem always existed for you since you assembled the setup, or did this start recently?

 

[Cognomen]

Sorry for the delay; I'm actually replying but you won't see the posts until they're approved as I'm a new member under probation.

The voice receiver is a Yaesu VR-5000. The inner DC block is located between the Yaesu's IF output and the mixer. This problem has existed since the system was assembled last weekend.

I am using the same frequencies as you; the LPE is programmed for 852.0000 MHz and the RF generator (NovaSource M2) is programmed for 862.7000 MHz. The M2's attenuation level is currently set to 30 out of 31 steps, although I've tested it at several levels with no difference in performance.

The DC block is a Mini-Circuits BLK-222+, and the mixer is a Mini-Circuits ZAD-2+. Coax is RG-58/U with BNC connectors, except for the cable from the M2 which has an SMA on the M2 end. Cable assemblies were purchased, not home-built. Cable assembly length is 12" per cable. There is an SMA-to-BNC adapter at the LPE's antenna connection.

 

[Radioman96p71]

I guess that makes sense. I was under the impression that the RF generator was making the constant carrier, which will cause the radio to unsquelch. How else does the radio know that there is a signal present? Does the 10.7 MHz IF change the strength of the mixed signal enough to open it? What i was saying was if the RF generator is set too high that the LPE can't squelch, the 10.7 IF will be the only kind of modulation, and if there is no actual modulation on the 10.7 IF then you will essentially be mixing a strong carrier with white noise. Similar to putting a preamp on an antenna, if there is no signal, it will still boost the noise.

 

[davidbond21]

The DC block is a Mini-Circuits BLK-222+, and the mixer is a Mini-Circuits ZAD-2+. Coax is RG-58/U with BNC connectors, except for the cable from the M2 which has an SMA on the M2 end. Cable assemblies were purchased, not home-built. Cable assembly length is 12" per cable. There is an SMA-to-BNC adapter at the LPE's antenna connection.

Those are the same components that I'm using(and signal generator as well; using an Icom for the voice receiver though). I have a ProVoice LPE-200 but I've never used it in the setup, only my Jaguar 725M. The reason I've never tried it is I don't have the proper adapter for the antenna connection to bring it to a BNC. If you are using a SMA-to-BNC adapter instead of the special GE-to-BNC adapter I've seen sold on ebay, then that could be the cause of your problems. Have you gotten any successful voice out of the set up before the radio is failing and you are needing to cycle the power?

 

[mitaux8030]

RF mixing follows the basic principle: signal A and signal B will give two mixing products A+B and A-B (where A > B) as well as the indivual original signals A and B. Without B, the only signal present is A.

However, where signal B is a 10.7 MHz (or similar) IF output, there will be a fair amount of noise present in the form of a high noise floor when there is no received signal on the tuned frequency, especially if the IF output is wideband for connection to a spectrum monitor, for example. That wideband noise floor could mix with your mixing carrier 'signal A' and produce a cachopony of detectable noise to the end receiver, in this case your LPE200. It would be better if you can restrict the 10.7 MHz IF to a narrowband output of 30 to 50 kHz bandwidth, if your host receiver can support that. That'll present less noise to the mixer overall, and may ease - but probably not eliminate - your problem. If you've tuned the LPE to the mixed product of A-B, then it should be quite capable of filtering out A, B and A+B... but the noise component of A-(noise from IF of B when not present) will certainly be there.

 

[davidbond21]

I guess that makes sense. I was under the impression that the RF generator was making the constant carrier, which will cause the radio to unsquelch. How else does the radio know that there is a signal present? Does the 10.7 MHz IF change the strength of the mixed signal enough to open it? What i was saying was if the RF generator is set too high that the LPE can't squelch, the 10.7 IF will be the only kind of modulation, and if there is no actual modulation on the 10.7 IF then you will essentially be mixing a strong carrier with white noise. Similar to putting a preamp on an antenna, if there is no signal, it will still boost the noise.

Well the signal generator is making a constant carrier, and in this case it's creating a carrier signal on 862.7 MHz, which is constantly being fed into a frequency mixer, whose output is connected to the radio. The radio though is set to a single channel, with a frequency of 852.0 MHz. Until the voice receiver is active (and therefore the 10.7 MHz IF output is active), there is no mixing product of 852.0 MHz being sent to the radio.

I would also say that for a ProVoice setup using the DVCM, you inevitably have to use an EDACS radio, and as a trunking radio it's receiver is active pretty much the entire time the radio is locked onto a control channel, so I would think that these radios would be able to stand a constant signal placed on it(given it's not in excessive strength).

 

[Cognomen]

Have you gotten any successful voice out of the set up before the radio is failing and you are needing to cycle the power?

Yes, sorry if I wasn't clear in the original post, the system works just fine while it works, which can be for only a few minutes to up to around 6 hours. Nice clear voice. Regarding the adapter, I believe it is the unit sold on eBay and at any rate it is designed for the LPE; I just thought it was SMA. Somebody supplied the adapter to me; I didn't purchase it myself.

Also, the RF signal generator is running at maximum attenuation, 30 out of 31 attenuation steps. I have tried it at various steps from full strength to 31 out of 31 attenuation steps; the system performs the same at all levels, and changing the attenuation does not affect the MTBF.

Mitaux, I'm not sure how I can set the Yaesu to do that; I don't remember seeing it in the User Manual but I wasn't looking for it. Will check.

 

[mitaux8030]

Some receivers can be set to pass the entire buffered IF baseband output before it reaches the IF filter, or be set to post-IF filter buffered output. Your Yaesu may not give you the choice... if so then post-IF filtering, so long as its not too narrow, would be the best choice.

I was thinking about your problem last night. When does the lockup actually occur? Only when in the middle of receiving something, or could it also happen when not actually receiving? That might offer a few clues for us to think about what is happening.

Given that an IF output has plenty of amplification (and a high noise floor, RF level wise, as well) and you've got a mixer thats wound back in attenuation as much as it will go, I'd add in even more attenuation at the output of the mixer, just before it enters the front end of your LPE200. Start with 10dB, and go up in 10dB steps and see if that helps. You should aim for about a -60dBm input signal as the butter zone.

Also, are there any local signals on 852.000 MHz that could be competing with or interferring with your converted signal? You never know what sort of wideband / pulsed signal could be there which may be upsetting your system. I've found all sorts of rubbish RF from unapproved / unauthorised devices in parts of the spectrum that they shouldn't be in. Just because they're not supposed to be there, doesn't mean they're not.

 

[Cognomen]

The IF Out jack is mentioned exactly twice in the Yaesu's manual. It states:

"This jack provides a low-level 10.7 MHz IF output."

and

"A portion of the 10.7 MHz Intermediate Frequency (IF) stage signal is available via this jack. This may be used for observing signal characteristics, or a separate receiver may be connected here to monitor FM broadcast sub-carrier signals, etc."

What do you recommend for post-filtering?

I am not sure when "exactly" the problem occurs, just that after a while I realize that I am no longer hearing voice. I do not believe the problem occurs in the middle of a transmisson as I don't remember any transmissions being rudely interrupted, but I will have to be paying attention to determine if the problem occurs pre or post transmission, during idle, or whenever. The problem is that I could be waiting a while. A third LPE-200 installed today lasted 5 hours for the first period, 5.5 hours for the second period, and about 3 minutes for the third period.

I will see about installing a step attenuator at the output of the mixer.

I did listen briefly to 852.0000 MHz before I switched on the RF signal generator; I didn't hear anything but that doesn't mean there isn't an intermittent source that coincides with the MTBF.

 

[mitaux8030]

Unfortunately the Yaesu IF output could still be either way; it could be post-IF filter, or it could be tapped and buffered just before. The only way you'll know for sure is to either study it's schematics or block diagram, or attach a spectrum scope or spec-an to the IF output. If you do use a spec-an, take note of how high the IF noise floor is, compared to RF levels - it might suprise you!

If you do find that the IF output is tapped pre-filter, then you can add your own post-IF filter tap & buffer amp (ie within the receiver), or build your own outboard buffer, filter and another buffer. A simple single transistor amp with IF coil to broadly preselect the signal as it goes into a crystal filter (I'd recommend a filter width of 15 to 20kHz between 6dB points) and the same arrangement as the signal leaves the crystal filter. There could even be designs on the net ready made for this application, or you could 'lift' a portion of an IF stage from an existing designs schematic... even 'lift' a section from a surplus / old two way radio, so long as it had a crystal filter of about the right bandwidth & frequency.

The only complication is that you'll be looking to minimise group delay (or phase distortion, whatever term you're more comfortable using) with whatever arrangement you decide to go with... though at 10.7 MHz it won't be quite as critical as if you were working at a lower IF freq like 450kHz.

Definitely be worth monitoring 852.000 MHz and see if a burst of noise etc coincides with your system lock up. The simple solution then is to just reassign your frequencies to avoid any clashes.