You are supposed to tune to a dead part of the FM band and adjust to zero at that point. Often the manual will specify a frequency to use, but you can't if there is a station very close to that.
Plastic tools always !!!!! Cracked cores are a real issue, and the next guy who does know what they are doing can miss it until they have problems in that area. I even get tuners in with no cores!
Plastic tools always !!!!! Cracked cores are a real issue, and the next guy who does know what they are doing can miss it until they have problems in that area. I even get tuners in with no cores!
You can't adjust the discriminator without an actual station - how is that supposed to work?
The entire point of a discriminator is giving out a voltage proportional to frequency offset between tuned frequency and input signal frequency. This may be offset from 0 V when the thing is not tuned to the actual IF. But without an input signal that on average is where it's supposed to be (and regulations in the broadcasting space happen to be fairly tight in terms of frequency accuracy), odds are it'll always output zero or thereabouts.
The entire point of a discriminator is giving out a voltage proportional to frequency offset between tuned frequency and input signal frequency. This may be offset from 0 V when the thing is not tuned to the actual IF. But without an input signal that on average is where it's supposed to be (and regulations in the broadcasting space happen to be fairly tight in terms of frequency accuracy), odds are it'll always output zero or thereabouts.
Hi sgrossklass,
You cannot adjust the discriminator on a station. Not unless you have adjusted everything else perfectly. That never happens.
It is adjusted with random noise and the mixed down IF output from the tuner pack through all the IF filters. A good bench has equipment more accurate than the output from a radio station (back in the day anyway). My stuff uses the GPS frequency reference, so stupid accurate. Even back in the day the high stability oscillator in an HP piece of test equipment was on the order of 10 exp-9.
That is why you have to tune to an area in the band where there isn't another station. You need broadband noise or it may pull your adjustment. What I do is create an RF environment by mixing a few RF generators, so known frequencies. I can then efficiently adjust everything including dial tracking. This also greatly attenuates any normal stations, but I still go for the hole in the band.
You cannot adjust the discriminator on a station. Not unless you have adjusted everything else perfectly. That never happens.
It is adjusted with random noise and the mixed down IF output from the tuner pack through all the IF filters. A good bench has equipment more accurate than the output from a radio station (back in the day anyway). My stuff uses the GPS frequency reference, so stupid accurate. Even back in the day the high stability oscillator in an HP piece of test equipment was on the order of 10 exp-9.
That is why you have to tune to an area in the band where there isn't another station. You need broadband noise or it may pull your adjustment. What I do is create an RF environment by mixing a few RF generators, so known frequencies. I can then efficiently adjust everything including dial tracking. This also greatly attenuates any normal stations, but I still go for the hole in the band.
Perhaps we could agree that DC centering must be done with a broad-band noise source (the front end noise should be broad enough and big enough with no external signal) in order to DC center zero signal in the IF passband. Then, distortion, with the usual quadrature detector, is minimized by tuning minimum distortion from a DC centered signal.
This choice is based on fixed bandpass (ceramic) filters and 1970s/1980s tech, but earlier chained RLC bandpass filters staggered between limiting stages that dynamically varied Q (and slightly, center frequency) wth signal level might need some iteration and futzing with test signal levels vs. optimum tuning. If it really matters, especially if you're charging somebody money for the work, this old fashioned stuff can take some time and plotting results.
Really really old stuff like vacuum valve FM IF strips, with large interaction between signal level and IF tuning have different schools of thought. Some can be peak tuned, just like the front end, but all are fiddly - document your efforts to be certain you're following your Hypocratic oath.
"Life is short, Art is long, Opportunity fleeting, Experiment treacherous, Judgement difficult."
I've got mains power back after six weeks. Time to get back to work.
Chris
This choice is based on fixed bandpass (ceramic) filters and 1970s/1980s tech, but earlier chained RLC bandpass filters staggered between limiting stages that dynamically varied Q (and slightly, center frequency) wth signal level might need some iteration and futzing with test signal levels vs. optimum tuning. If it really matters, especially if you're charging somebody money for the work, this old fashioned stuff can take some time and plotting results.
Really really old stuff like vacuum valve FM IF strips, with large interaction between signal level and IF tuning have different schools of thought. Some can be peak tuned, just like the front end, but all are fiddly - document your efforts to be certain you're following your Hypocratic oath.
"Life is short, Art is long, Opportunity fleeting, Experiment treacherous, Judgement difficult."
I've got mains power back after six weeks. Time to get back to work.
Chris
Hi Chris,
Following the service manuals are a good start. Except with some Japanese manuals that have you running in circles and you never end up with things aligned properly. Then I let experience take over and do it right.
I can't tell you how often I fire up the HP 3585A to set up or diagnose an IF strip. Especially if someone else was twiddling cores! If you know what things should be like (the basics) you can clean up most messes. We are into time at that point, you just do your best.
Ceramic filters. Yeah. I have a jig for checking / matching these things. I still have some I can match a set from. The new ones have far lower insertion loss (but don't go replacing stuff on that basis!).
Following the service manuals are a good start. Except with some Japanese manuals that have you running in circles and you never end up with things aligned properly. Then I let experience take over and do it right.
I can't tell you how often I fire up the HP 3585A to set up or diagnose an IF strip. Especially if someone else was twiddling cores! If you know what things should be like (the basics) you can clean up most messes. We are into time at that point, you just do your best.
Ceramic filters. Yeah. I have a jig for checking / matching these things. I still have some I can match a set from. The new ones have far lower insertion loss (but don't go replacing stuff on that basis!).
Unfortunately, threads like these turn into far more than nesseccary, and minimal academic interesse for the OP`s.
The short answer to OP´s question would have been:
"The tiniest turn either CW or CCW on the detector coil would have solved his problem"
The short answer to OP´s question would have been:
"The tiniest turn either CW or CCW on the detector coil would have solved his problem"
Agree.......
But how hard can it actually be??
If I recall, OP has the manual.
A small turn the wrong way just put you display 50/100KHz further off.(assuming you found the Detector)
Then just do the math 🤣
But how hard can it actually be??
If I recall, OP has the manual.
A small turn the wrong way just put you display 50/100KHz further off.(assuming you found the Detector)
Then just do the math 🤣
Hi Boydk,
Remember being a baby technician? There was someone to help with experience, and the right test equipment with you. I could do it without much test equipment - maybe (if that was in fact the problem). That's with decades of experience.
What is easy for an experienced person might turn out to be a disaster for another person. Heck, I know of "technicians" that would mess this up, and it wouldn't be their first time. It's difficult to remember what it was like just starting out, and you already had some understanding and the thinking behind it.
Service manuals really only show you the specifics of a design. They are not intended to teach you how to be a technician, or how to do a procedure. Not to mention many have errors, some really big errors. Every single service manual out there assumes you have basic knowledge, experience and the proper test equipment required.
So how hard can it be? Depends, was the troubleshooting even correct?
Remember being a baby technician? There was someone to help with experience, and the right test equipment with you. I could do it without much test equipment - maybe (if that was in fact the problem). That's with decades of experience.
What is easy for an experienced person might turn out to be a disaster for another person. Heck, I know of "technicians" that would mess this up, and it wouldn't be their first time. It's difficult to remember what it was like just starting out, and you already had some understanding and the thinking behind it.
Service manuals really only show you the specifics of a design. They are not intended to teach you how to be a technician, or how to do a procedure. Not to mention many have errors, some really big errors. Every single service manual out there assumes you have basic knowledge, experience and the proper test equipment required.
So how hard can it be? Depends, was the troubleshooting even correct?
That sounds like the right approach for a classic analog tuner. Centering the discrimininator in the passband like that is pretty clever, actually. (That obviously assumes that the filter response is nice and symmetrical, which with ceramic filters it may very well not be.) As a broke student with no equipment 20 years ago, I went with the approach of minimizing audible signal amplitude on the scope output's V channel when adjusting the AFC discriminator on the Kenwood KT-1100, which I think would do something similar (put the station at the flattest part of filter passband).
Now the OP has a PLL synthesizer affair with 50 kHz steps and twin discriminators, one being for actual demodulation (T202, T203) and the other for station detection only (T204). I hope we can agree that the output of the latter (as measured at R236 vs. ground) should be 0 V when the tuner displays the nominal frequency of the station tuned.
I don't think the Hitachi's µPD1707G synthesizer sports any FM IF fine-tuning as found in some other PLL tuners, so it's pretty much 10.70 MHz or die. If your filters are off-center, tough luck. I don't see a reason to center the discriminator in the filter passband then, if anything you might even want to offset it to the other direction if it helps the distortion department. Again, my default strategy under these conditions would be aiming for 0 V at R232/233 when adjusting T202 on the nominal frequency of a station. Without the necessary equipment to do accurate distortion alignment at the specified 0.02% level, I would leave T203 well enough alone.
Should the NARROW filter be so far off as to give best reception one step above or below nominal frequency, deliberate mistuning may be called for, even if it screws up the cosmetics. (Tune around a strong station that has nothing much else in its immediate vicinity and observe audio quality at ±50, 100, 150 kHz offset.)
With most digital PLL tuners and ceramic filters, you pretty much get what you get. That's why I originally made the test jig, worst fears confirmed many times.
Still, you need to troubleshoot the issue before you go adjusting anything. You need to be able to confirm correct adjustment before changing anything as well. Otherwise you may well be adding issues for someone else to try and figure out. Some Japanese tuners have the most stupid, some incorrect alignment instructions too. Kenwood tuners can be a real pain to align.
One thing for sure, never comply with the instructions in some manuals to set adjustments to positions differently from where you found them. Never do that! This will mess you up so badly! Not unless someone else has really messed around with it. I still prefer to start where I find it and prove each adjustment is correct. Tuners once worked where they were set. Maybe not aligned well, but they did work. Same for CD players and other things.
Still, you need to troubleshoot the issue before you go adjusting anything. You need to be able to confirm correct adjustment before changing anything as well. Otherwise you may well be adding issues for someone else to try and figure out. Some Japanese tuners have the most stupid, some incorrect alignment instructions too. Kenwood tuners can be a real pain to align.
One thing for sure, never comply with the instructions in some manuals to set adjustments to positions differently from where you found them. Never do that! This will mess you up so badly! Not unless someone else has really messed around with it. I still prefer to start where I find it and prove each adjustment is correct. Tuners once worked where they were set. Maybe not aligned well, but they did work. Same for CD players and other things.
This makes a lot of (practical) sense for modern-ish PPL + ceramic filter tuners, but I have no personal experience:
https://k6sti.neocities.org/offset
All good fortune,
Chris
https://k6sti.neocities.org/offset
All good fortune,
Chris
Hi Chris,
Yes, he is correct.
That reference oscillator he talks about can be off, so I have used a trimmer to correct it. Then match a set of filters. Narrower filters will always increase distortion. DX is simply going to be noisy and higher distortion, so balance selectivity and distortion if you're going to try this.
Yes, he is correct.
That reference oscillator he talks about can be off, so I have used a trimmer to correct it. Then match a set of filters. Narrower filters will always increase distortion. DX is simply going to be noisy and higher distortion, so balance selectivity and distortion if you're going to try this.
I remember these Murata ceramic filters being sold in centre frequency selections. What happened to the non 10.700 MHz selections is a mystery.
The very long term aging of these filters is also not documented anywhere that I have seen.
Tuners from this century don't use 10.7 MHz IF
The very long term aging of these filters is also not documented anywhere that I have seen.
Tuners from this century don't use 10.7 MHz IF
Well, they didn't get warm, and they were ceramic and very stable. I'd say they did not require any aging.
There are a lot of devices that used a 10.7 MHz IF frequency since FM tuners made these parts available and cheap. I wouldn't be surprised to see other radio and communication devices that used it.
There are a lot of devices that used a 10.7 MHz IF frequency since FM tuners made these parts available and cheap. I wouldn't be surprised to see other radio and communication devices that used it.
The 1970s/1980s ceramic filters came factory-matched (or maybe the tuner manufacturers had to do it? - seems unlikely) into batches, color coded with paint stripes. It wasn't as important then to hit 10.70000 center (whatever that means) because the VFO was analog. It would be interesting to know their matching criteria, as sgrossklass points out. How do you do it?
All good fortune,
Chris
All good fortune,
Chris
After testing many 1970's, 1980's filters, the shape and slops were pretty variable even when marked the same. Later production in the 2000's was much, much better. It's the same thing as matched transistors, as manufacturing progressed, the matches were far better and yield from batches much higher. I hate matching transistors, and do a lot of it.
I believe the factory applied the matching colour codes (they are in the data manuals). Close matching would be terrible expensive to do, so they aren't as close as a technician would do it. Transistors were matched by the end user.
I believe the factory applied the matching colour codes (they are in the data manuals). Close matching would be terrible expensive to do, so they aren't as close as a technician would do it. Transistors were matched by the end user.
Can you discuss how your matching jig makes its judgements? If this is proprietary, of course feel free to be general, but folk like me would be interested in your criteria. I would expect that a factory match would derive from a broadband noise source and be amplitude (only) measurement, but I certainly have no certainty.
Much thanks, as always,
Chris
Much thanks, as always,
Chris
https://www.alldatasheet.com/html-pdf/83991/TI/TRF6901/99/4/TRF6901.html
I've been in two projects around the turn of the century that still used ceramic filters, but after that, everything became low-IF with integrated filters (first analogue, later partly analogue, partly digital).