Galaxy DX2517 10.240Mhz osc off frequency

[Tolteca]

Hello all,

It seems lately I get nothing but strange cases and I can't catch a break with a routine alignment.

This time around I have a Galaxy DX2517 base station. Problem with this one is that for some reason the 10.240Mhz oscillator is off frequency by about 1khz. Basically when I go to probe Pin 4 of IC3 I get about 10.2405Mhz. Service manual says to allow only about 200hz of tolerance tops. I tried swapping the crystal only to find that when I do that, the issue gets worse and I'm off by about 1khz or more vs the 500hz I have now. When I probe the crystals directly I find that they are oscillating within tolerance or close to it but the actual output from Pin 4 comes out altered. Furthermore when I put back the previous crystal I find that someone must have previously tried to deal with this issue too as this crystal is about 500hz too low. Basically the MC145106P seems to add an extra 1khz to the signal and the previous person trying to resolve the issue put in a crystal that is 500hz too low to try and get as close to the right frequency as posible and this way we are only high by about 500hz vs the full 1khz we get if we put a within tolerance crystal.

Anyone have any idea what this may be besides the actual IC? Don't have a replacement so I might have to resort to some kind of hacked repair. Unfortunately there also seems to be no voltage chart for this radio or an actual full service manual. The info I'm getting is posted on Cbtricks.

 

[nomadradio]

The real side effect of this error will be with the spacing between channels.

The manual says 200 Hz max errror. But if it's wrong by exactly 500 Hz, remember that this crystal's frequency gets divided by 1024.

The result of this divide-by-1024 inside the PLL chip is that your crystal's frequency error is also divided by the same factor.

The 10,000 Hz result of this divide-by function is what sets the spacing between channels. 500 Hz divided by 1024 is just a bit less than 2 Hz. So, rather than the dead-on 10,000 Hz spacing between channels, it will be 10,002 Hz per channel.

Don't see a lot of folks reading frequencies with 8 digits, but that's how many you'll need to measure that error from the radio. If you set the radio exactly to 27.205.000 on channel 20, the frequency would read 27.215.002 on channel 21. 27.205.004 on channel 22, etc.

But this error is cumulative. If you set the radio on frequency at channel 20, it will be 20 times this error on channel 4, and 25 times that much on channel 1.

But 2 Hz times 25 is still only a 50-Hz error. Takes a counter with 7 digits or more to see that.

This means that if the radio is set exactly on frequency on channel 20, the frequency on channel 1 will be 26.964950. Channel 40 would be about 27.405.040

My experience with frequency counters our customers bring here to get calibrated is that they only start to drift in a big way around 5 years of use. A cobra 2000's internal counter made in 1978 might or might not set correctly at all with the adjustment all the way to the limit. The newer a counter is, the less error it will tend to show, if you don't calibrate it to a proper reference from time to time. If the last time it was calibrated is a few years, don't expect accuracy better than the first 6 digits.

Not trying to scold, but the effort to set the radio for more accuracy than the counter can offer will get you close, but only close.

73

 

[Handy Andy]

Is the 2517 the 6900 series?

One - check all your "Gated" lines, the thing can shift down to 1kHz thru the input pins alone.

2nd - the VCO
(the CB Tricks site, THAT IS THE Repair Manual.)

Have you gone to this section yet...
http://www.cbtricks.com/radios/galaxy/dx2517/dx2517_epf000210z_vco_pcb.htm

Wonder if a power supply line cap is affecting this...Hint TP2...graphic attached...




I wound up getting into a big argument over a radio similar to what you're working on - someone had turned a radio like this into a "Monster" and used a transverter to unleash h*ll on locals and hams alike - they never saw what I wound up undoing and many had to take sides - and all I was doing was telling my side of the story.

I made a vow, a pact - not to go there again but I can share some of the fun I had. I also will not take pics of those types because they are a customer - not my enemy. But they did make enemies - so I too to those they affected, was also Their Enemy.

:+> Andy <+:

 

[Tolteca]

Hey,

This is my first time working on one of these Galaxy radios so I don't know if it's a 6900 series. How can I tell?

Anyway as Is I have to compensate using the clarifier to get it on frequency.

Andy, what do you mean by the gated lines?

Hmm am I missing something here? I was assuming that the output of pin 4 should be exactly the same as what crystal oscillator is doing as I thought internally all this was is the output from a buffer stage that should show no change and stay fixed no matter what the frequency of the radio or other oscillators. Is my 1khz difference a compensation for a fault in one of the other oscillators?

 

[sonoma]

see if I can load this voltage chart here for you. are you sure your freq counter is spot on. I ran into this one time and my counter was off.

 

[Tolteca]

I'm sure the counter is accurate. I'll admit it's been a while since it was calibrated but considering every other radio I align with it comes out dead center I can't imagine it's off by more than a few hertz. Thanks a whole lot for the voltage chart. Another thing that tells me the freq counter is alright is the fact that when I read at the actual crystal, the frequency is within tolerance.

 

[sonoma]

if the freq shows good at the crystal it could be one of the caps causing this like Andy stated. what I have done is like in the C84 position it shows a 27pf cap there. I have some variable caps that i have had to put in that location and adjust it until the freq shows right. I had a 29xlr that was way off a few years ago and put one of the variables in and the radio is still just fine. it is one of my old radios. also a 959 for a friend of mine to get it to align properly. just food for thought.

 

[Tolteca]

It seems the issue was C84 27pf ceramic. Not everyday I see that fault but okay, it is working now! Continuing with the alignment. Hopefully no more issues. Thanks a lot!

 

[Handy Andy]

The PLL uses a channel selector but uses "gates" to work out the channel programming.

That's why the reference - seen kids goof these chips so they can jump channels but wound up skipping to other frequencies WITHIN the channels - not A Channels - they went from Nickles to Dimes and back again.

They may have goofed up one of the gates - so it's output is not right and it's "skipping" over and leaving you off 1kHz in PLL tuning. I just replaced those MC14008 IC ships - it's simple insurance - the diode matrix can offer a bit of a challenge as well.

Ok, to help you with the PLL stuff. Had a radio that used a 10.7MHz Xtal instead of the 10.240 - replaced with a Uniden 10.2417 seemed to be more on frequency than the original 10.240 - if that helps.


And you're reference to the 1kHz, there is a shift in TX and RX - again - based upon channels and programming. Not sure what you're up against, but if the PLL is shifting the Oscillator input down 1kHz or up 1kHz, then that may be from someone playing with the channel selector programming and the 10khZ slide stuff. I'd have to see your board - because it sounds a lot like the one I worked on before.

Then while I was busy I caught this...

It seems the issue was C84 27pf ceramic. Not everyday I see that fault but okay, it is working now! Continuing with the alignment. Hopefully no more issues. Thanks a lot!

Glad to know - and that was one of the parts I was pointing to in the graphic.

Otherwise, if you have that problem or wish to see what happens - replace the 10.240 with a 10.2417 once and see if it gets you where you want to go. I found that these older Uniden Xtals - either by internal mounting or otherwise, just offered better performance - either from trimming / paring up or down fine tuning - they took well to series capacitance with less headaches.

:+> Andy <+:

 

[sonoma]

glad you found the culprit. good luck on your next trial.
yes Andy you did point to that part in your posting. that was why I mentioned to him to see if that part was good or bad. like I mentioned to him I have some variable pf caps I have used a couple of times. your drawing was very good. saved a lot of time trying to find the parts that are able to act up.
sure glad you posted them. just had to save them to my hard drive for future listings or for my own problems. good job Andy.

 

[Tolteca]

Indeed, thanks a lot Andy.

Onward with the alignment I'm noticing I cannot get the right amount of power out. This thing transmits like a Cobra 148. About 6 watts AM 12 Watts PEP SSB. I haven't done the receiver alignment and I've had radios have low power because the receive alignment wasn't done yet so hopefully after that it will work again though I am skeptical.

 

[sonoma]

this may not be the problem but almost sounds like 1 final is not putting out. lift the legs on 1 and see if the out put drops. if not resolder the legs of the one and unsolder the other. just to check if one is dead or not. at least that is how I check a dual final radio. some other might have a easier way to do it.

 

[nomadradio]

Another method that won't require a soldering iron is to pull out the bias-test jumper board. It's held in place by the friction of three 'fork'-shaped terminals. Connect a meter set to the 10-Amp range from the right-hand terminal on one lead to the center terminal on the meter's other lead. This will pass all the current feeding from the power supply to the two finals through the meter.

Turn the mike gain to zero, select either USB or LSB (won't matter which) and key the mike. there is a trimpot just to the front of each final transistor. Turn them both the direction that gives you a zero reading. Now slowly turn up one of them (won't matter which) until you have a reading of 0.060. Now turn the other trimpot to double this reading to 0.120 Amps.

If one of the two trimmer pots shows no response, this suggests that the final transistor just behind THAT trimpot should get the leads unsoldered and checked. Good chance that one will test bad.

There are other things that can produce this "zero" effect from one trimpot. If the radio blew a final, it could have been replaced by someone who doesn't check the other parts on the input side of the blown transistor. When a final goes bad in this kind of radio it can damage parts on that final transistor's input circuit. If there is a blown resistor feeding into this final, that can produce the 'zero' effect from that trimpot even if that final transistor is good. Reduces power roughly in half, as well.

More often, we find that someone turned those trimpots for max wattmeter reading. This tends to run a lot more than the 60 mA of steady current that you should have. Makes that final overheat and fail.

The color of the metal tab on these transistors can be a clue. If the metal surface of the mounting tab is dark and discolored, it was probably running way too hot.

73

 

[Tolteca]

Turns out the jumper to my wattmeter had a bad connection. I've got 10 watts on AM now however still not entirely good on side band. When I yell into the microphone on USB I can just about get about 12 watts out of it. On LSB I can get maybe 18 watts if I hit a sweet note. When the roger beep kicks in however, on LSB I get the full 28 watts, on USB I still only get about 15 watts.

I don't think this is entirely the microphone as I'm seeing similar results with the roger beep. Any ideas as to what could be causing this and what the deal is with the lower wattage on USB as compared to LSB?

My other issue is that SSB is also low on reception, even after aligning everything. When I try and align the S meter for SSB, with the pot turned all the way I can barely get it to measure close to 8 S units.

EDIT:

Almost forgot. During the transmitter alignment I didn't do the "SSB APC" part because I can't for the life of me figure out where TP7 is.

 

[sonoma]

this could be why the out put is low. tp7-tp8-tp9 are the 3 pins that hold the mirror board in place. you pull the little board out to test the bias feed to the finals and driver.
for some reason it does not show them numbered in the 2517 radio. here is a link to a different radio but same board and they are numbered.

http://www.cbtricks.com/radios/galaxy/dx959/dx959_adj.htm