VCO issues in P7100 UHF low to ham conversion

[willbartlett]

Hi All.

So, I picked up a couple uhf-lo P25 7100 system radios at Dayton last year. I' ve been able to move them to to 440-450 with minor tweaks to the VCO coils accessable under the stickers on the top of both the tx and rx vco. In RX, the radio gets progressively more deaf as you go higher in frequency, but not more than 3db down from nominal expected sensitivity vs. my EFJ5100. I'll pull the RX BPF assembly and revork it later.

The RX vco seem unconditionally stable up to 450, but I'm having an issue with the TX vco.

first off the radio seems to be a bit low in frequency in TX, 600hz with one radio and 800hz with the second.

The transmitter will come on upon PTT, and will stay on for approximately 20-30 seconds, at which time the transmit LED will go out and RF carrier dissappears. If you press the PTT immediately, the radio will do nothing. No Synth unlock, no error, simply nothing. Waiting 3-5 seconds will result in the radio transmitting again, but for no more than a couple seconds before dropping TX again.

The lack of error message is very odd. I would expect a synth unlock.

I'm tempted to depop the VCO ( I can handle that) and test it off the board to see if it's an issue, but is there more going on in the antenna switch/Low pass assembly?

I know that there is some TX power feedback going on there, but it doesn't look as though it goes any farther than PA power control. I'd expect to see a carrier at significantly reduced power bleeding through the PA even if it went into some sort of shutdown mode, but it looks like the TX is shutting down completely. is the power sense handed back to the controller, or is it a small closed loop independent of the rest of the radio?

I know this is fairly advanced stuff, but I'm real curious about this kind of stiff.

Will

 

[mitaux8030]

I wouldn't worry about the radios being off by a few hundred Hz here or there. You'd be able to alter the DAC word that pulls the reference oscillator at a later stage.

I don't have any experience with the M7100, but if it follows more conventional radio design that Shinwa does (I think the M7100 has its origins in Shinwa?) then you can safely isolate the TX PA strip after the first buffer (might require a SMD component to be removed) That way you'll be able to check the synthesis & VCOs alone without any influence of the PA. If you get a service monitor proble close to the PCB, you'll be able to pick up enough signal from the VCOs themselves to see / hear what is going on.
Then start looking at things like VCO output levels at design vs your modified freqs; perhaps enough VCO output for the buffer to pass on to the next stage (for RX mixing or TX amp), but not enough for the prescaler to reliably divide down to a workable freq to do the dividing and phase comparison.

Just some general thoughts... having done similar things to other radios.

 

[willbartlett]

So, Seems it was a PTT membrane button issue. Put some packing behind the hard plastic button and now it works FB. Still need to look into the low pass response, and will need to look at the RX filters. I'll post info as I learn more.

I do think that I'll look at the VCO output once I get to a point where I can measure with the board out of the back casing. Would be interesting to see how it performs. For the record, between 2 radios, The tx vco locked up to 448mhz and 449mhz with no tweaks, didn't take much to push it up a couple more mhz.

 

[ElroyJetson]

Would you please document your conversion efforts? It's absolutely no problem to get UHF low range radios. Converting them to work in the ham band would actually make them worth something.

I have two sitting here, myself. They're paperweights. Brand new condition paperweights. I'll take most any reasonable offer for them.

But if they can be converted over, that makes them a bit more useful.

Be aware, getting that LPF module out is a real bear. And a few SMT components are soldered right
between its pins on the board, so they have to come off, too. It's not a task for those with weak
soldering skills.

Reworking the LPF module will be even more fun but I'd like to see the results when you're done with it.

Be aware, there's also an issue with the oscillator warp adjustment in the maintenance tools software. Put simply, that adjustment doesn't work at all. It's all that's keeping one of my otherwise great 800 MHz radios from being very useful. To get anything out of it I had to determine the frequency error and then adjust all programmed frequencies by that much. A real pain. It's NOT the ref oscillator.
I've tried three different ones.

Elroy

 

[smackdaddy]

Hey Elroy,

Just for what it is worth.. I have found that some C5416-based radios will not pay any attention to the reference oscillator warp values if you have an incorrect ESN and/or FE combination... Just throwing it out there.

Cheers,
SD.

 

[willbartlett]

Hey All.

So, for what it's worth, there are only 2 components affecting the critical response for this modification.

In the low pass portion of the ANT-SW/LPF assemply, spreading out turns on one series inductor ( it believe in the upper left corner) a fair bit seems to be sufficient to bring the attenuation from 440-450 back into line with the attenuation you might find within the rated bandsplit. I did need to bias the PIN diodes with approximately 1.5v to get the module to move into TX state. I think Pin 8 is the critical pin for applying this control voltage.

Once this had been done, I was seeing just .25db more attenuation up to 446mhz than I saw at 430. most of our repeaters near Boston are low in high out.

In the bandpass assembly, again, I found one series inductor that , when spread, would move the passband 3db point from 430 up past 450. it's in the upper left corner. The radio went from being 10db deaf in the upper portion of 440-450 to only ~2db deaf. It is very close, but still noticeable when compared to the EFJ-5100, both on a service monitior and in real world operation. RX sensitivity is a relative measurement taken against the 5100, not against specifications. YMMV

Also, When looking at the RX filters, component side facing up, pins at the bottom:

The left side is a low pass which will require the modification.

The right side appears to be a notch or bandstop, centered around 295mhz, presumably to mitigate 1st LO energy from re-radiating. It continues above the notch with no real appreciable attenuation untill you're well above the operational range of the radio, like 800mhz or so. Again, a picture is worth a thousand words. More to follow.

Both VCO's were able to be "tweaked" to allow operation within the ham band, the stickers on the top peel back to reveal a coil which, in both cases, had been 'spread' at the factory to tune the vco range.

All said, this isn't a perfect solution, just a stopgap. I feel that component change is more appropriate, certainly in the rx filtering. A change of the PA brick might be advisable, as power obviously falls off as you progress up towards 450mhz, maxing out at approximately 3.25 watts at 449mhz, while it has no problem generating 4+ watts in rated bandsllit.


I took pictures of both modules with the covers removed and will put something together to put up on repeaterbuilder when time permits. If anyone wants to see raw pictures, I'll put them up at picassaweb this weekend.

Elroy, see where you get with one. I'm going to take a closer look at the second one I have, and will try to get network analyzer plots for all to see. certainly I can get screen captures with the service monitor and tracking generator.

Will

 

[ElroyJetson]

The radio that won't warp does have a valid ESN and feature string set. A very good set, including P25 trunking, CAI, and practically everything else you'd want...but it's not a "40 option": feature set.

The TAC says that the warp feature wasn't even implemented prior to ProGrammer 20! But I can't get it to warp with ProGrammer, any version from 7 to 20, or even RPM up to version 3. (The newest version that I have.)

It'd be a perfectly good radio if it would just warp to frequency like it should. It's something like 1.2 KHz off channel. Not very good.

Elroy

 

[willbartlett]

Images of the filters are here:

Picasa Web Albums - will - filters