Alignment...

As I mentioned in my last post, my FRG is in need of alignment - tuning at the bottom of each band is cramped and non-linear, it falls about 50kHz short at the top end of each band (when it should go about 50kHz past the end), and the whole thing feels a bit insensitive (which may just be psychological on my part).

That's not unusual; as far as I know, the set has never been aligned since it left the factory 30-odd years ago, and being kept for a few years in a hot shed won't have done it any favours.  At least it hasn't been "adjusted" by someone who's skills would be more in demand in the meat production industry than in an electronics workshop. Alignment isn't hard, and is quite well documented in both the instruction and service manuals, but it does require a few tools and a certain touch…

I won't go into the whole process here - read the manual for the details, and Wim Pender's excellent FRG-7700 Survival Guide for a few extra tips - but I'll give a bit of a explanation of what's needed, what the steps are, and what I found when aligning mine.

Tools Required:

• A set of non-metallic radio alignment tools. Read that again. Non-metallic! Trust me that no, you can't just use a screwdriver. No, not even a one with just a metal tip. No, not even if you're very careful and promise not to break anything. Not only will metal de-tune the coil you're trying to adjust, you will chip or crack a core. Take it from someone who has (a) thought exactly like you that he could use a screwdriver if he was really careful, and (b) broken cores every single time. Spend the $5 to buy a cheap 10-piece set from the fart-toy shop or wherever crap radio tools are sold.
• A small flat-bladed screwdriver. Preferably non-metallic (e.g. ceramic), but metallic will do at a pinch. Use this only on adjustments labelled VRxxxx - for anything marked Txxxx or TCxxxx, use the proper radio alignment tools described in the first point!
• A decent digital multimeter. The manual actually specifies a VTVM (Vacuum Tube Volt Meter) - a hangover from the days when valves roamed the earth, typical hobbyist analogue voltmeters/multimeters had an input impedance of 1000Ω/volt or so, and a top-quality expensive meter maybe had an impedance of 20kΩ/volt. A VTVM typically had an input impedance of 10MΩ; high enough to avoid loading down the low level/high impedance signals typically found in RF stages. These days even a cheap digital meter will have 10MΩ input impedance, and better ones may have the option of even higher impedances e.g. my Fluke 19 can be configured to >4GΩ on the 400mV range!
• A decent frequency counter. It needs to read 0 - 50MHz with 5 digits resolution and reasonable accuracy (i.e. calibrated against something accurate) - so it doesn't need to be super-good, but Arduino or PIC specials, or the cheap $35 Chinese 'portable radio frequency counters' from eBay need not apply. You'll need a high impedance 1:1 or 10:1 probe for it.

That's all that's needed for a general alignment. To do the full alignment as per the manual you'll also need:

• An RF millivoltmeter or RF probe. As the name suggests, this is a specialised voltmeter or adaptor probe for measuring AC voltages in the RF range (most multimeters run out of puff above a few kHz on the AC ranges). When aligning the FRG-7700 it's only used to peak a few voltages, so an oscilloscope or a simple uncalibrated home-made probe will suffice.
• An RF signal generator capable of 8.01MHz with calibrated output level from -10dBµV up to 90dBµV. Not owning a proper signal generator I use a cheap eBay/DX DDS module controlled by an Arduino, with a handful of fixed attenuators plus a variable attenuator (all home made). With that, I can set the level by measuring the p-p voltage on an oscilloscope and doing a few calculations.
• An RF wideband noise source. This is getting a bit specialised, but the basic design is fairly simple and various DIY versions can be found all over the 'net. Whenever I've needed one I've simply breadboarded up something like the 2-transistor design found here using whatever bits happen to be floating around my workbench at the time.

Alignment is pretty straightforward. Spend some time first looking over the board layout pictures in the manual, so you can easily find the test points and adjustments. Follow the steps in the manual, with reference to these (& Wim's) notes. As I said, it does require a certain touch - both mentally and physically. Realise that you'll never get a perfect adjustment, particularly in a set of this style - there's just too much jitter and short-term drift in most of the settings - so aim to get as close as possible. Use the correct tools - non-metallic alignment tools for cores (Txxxx) & trimcaps (TCxxxx), with the screwdriver only for trimpots (TRxxxx). Never force anything - if something is stuck, it's stuck for a reason; if you don't understand the reason then you've got a good chance of breaking it. It's better to be gentle and own a maladjusted radio than be strong and own a broken radio…

 

Alignment Steps:

 Main Board (top)

Clock/counter oscillator frequency (in kHz)

1. Clock counter frequency adjustment - this is simply the 3.2768MHz clock for the display IC, which doesn't affect the set's performance as such but does affect the accuracy of the frequency display and clock/wake up/sleep timer. In theory you should be able to adjust this either side of the nominal value; mine couldn't be adjusted any lower than 3.2768293MHz (yes, I'm showing off there!). However, that's only ~30Hz, or ~9ppm, out - bloody good for a standard crystal oscillator which has a typical aging rate of 5-10ppm per year! Ideally I'd tweak the capacitor values or maybe replace the crystal to get it back into line, but that can wait until I have a reason to pull the main board out.
2. SSB carrier frequency adjustment - this step adjusts the two BFO oscillators (upper and lower) used for single sideband reception. These are simple ceramic resonator oscillators, nothing much to go wrong there. Both of mine were a bit out, but had enough adjustment range to pull them onto frequency.
3. First and second IF adjustment - this aligns the transformers in the first (VCO -> 48.055MHz) and second (47.6MHz -> 455kHz) IF stages for maximum sensitivity. Skip this if you don't have a signal generator; otherwise it's pretty straightforward. The only real trick is that, because there's several transformers in each stage and some slight interaction between them, it's worthwhile repeating this step one or two times to get everything lined up and peaked together. Mine needed minor tweaking only.
4. S-meter sensitivity & full-scale adjustment - again, skip this if you don't have a signal generator; at worst, your S-meter readings will be out. As far as I could tell (since you have to do some adjustment to find out!), mine was spot-on and didn't need adjustment.
5. Noise blanker adjustment - once again, you can skip at least the first part of this if you don't have a signal generator. If you also don't have a noise source, you can approximate the second part by tuning to a quiet spot (with some white-ish noise) around 8.01MHz and doing the adjustment to minimise noise from the speaker. Although, since the analogue noise blanker in these old sets is about as useful as tits on Elvis, there's no great disadvantage from skipping the whole section altogether…
6. Trap adjustment - this is basically to filter out any frequencies close to the 48.055MHz first IF LO, which might 'beat' with it and result in spurious noise. For example, in most (all?) of the world 40-70MHz is allocated to fixed & mobile 2-way radio, and in some places analogue TV channel 1 straddled 48MHz. If you can rustle up a signal generator this step is definitely worthwhile, otherwise skip it. Again, as far as I could tell mine didn't need adjustment (but got some anyway!)

PLL Board (bottom)

1. PLL reference oscillator adjustment - as per the title, this adjusts the 3.2MHz reference oscillator for the phase-locked loop, which is multiplied up to produce the 47.6MHz fed to the 2nd IF mixer. As such, it's a fairly critical adjustment which affects the signal chain in the middle of the set, so get this one as right as possible. In fact it would make more sense to do this step (and the next) before step 3 of the main board adjustment, so I don't know why it's left until this late in the process? Regardless of that, Wim suggests connecting to TP2004 (rather than probing pin 9 of Q2041) and adjusting TC2002 for 6.4MHz (because you're now measuring the reference oscillator after it's been frequency doubled). That reduces the fiddlyness of the connection and potential for short-circuits, which seems like a good idea to me. Mine was found to be several kHz out - which in turn affects the next step …
2. PLL local oscillator alignment - skip the first part of this if you don't have an RF millivoltmeter, RF probe, or oscilloscope, although since it affects the level of the IF signal at the mixer it's worthwhile doing if you can. The manual says the level should be 100-200mV RMS - and, as far as I can tell from the circuit diagram, it should ideally be at the high end of that range. The second part of this step though is critical, as it sets the 455kHz out of the 2nd IF mixer - so again, it's one to get as right as possible. Once this step was done I went back to the main board and readjusted the 2nd IF strip
3. VCV line adjustment - this is the step that would fix the linearity and bandspan problems I'd found, so I spent a bit of time here getting it exactly right. Read Wim's description, and note the voltage differences he found at the low end of the scale - I found the same discrepancies. All the transformers/chokes adjusted at this step are inside shielded boxes and stabilised in wax. Unless the set has been worked on previously, you won't be able to see the cores or their adjustment slots clearly. They take the same tool as all the other Txxxx transformers, so use that tool to feel through the wax until you find the slot (there's often a bubble of air trapped in the slot, so the tool 'pops' in with a little waxy 'crunch' feeling). There appears to be some degree of interaction between the adjustments on each group of bands, so after finishing the whole step go back and repeat all the adjustments a few times. Aim to get as close to 7.0v on TP2005 and 7.4v on TP2003 as possible at the high end of the each band group. Mine were all way out to start with - which is why tuning ran out before the top of each band - but adjusted to spec quite easily.

Memory Unit

I don't have the memory unit in mine so I can't really comment, except to say that the alignment procedure looks pretty straightforward. It should be easy enough if you've managed to align the main and PLL boards.

Result:

With a full alignment done, the radio now works as it should. Tuning extends 50kHz past the ends of all bands (e.g. the 8Mhz band tunes from 7950MHz to 9050MHz), and sensitivity and selectivity are spot on. On top of that, it sounds beautiful. The only minor issue is that, when tuning, the actual frequency is right 'on the edge' of the displayed frequency - that is, to tune exactly to (for example) 15.245MHz, you need to tune upwards from 15.244MHz until the last digit just tips over to 5, rather than the correct frequency being in the middle of the last digit. This may actually be normal for the FRG-7700 - I have a vague memory that it was like this when I was using it before - or it may be due to slight misalignment of the VCO. In fact it can be something of an advantage, particularly for AM modes; you can tune to within a few 10's of Hz quite easily. However it's a bit of a pain for SSB modes, since due to the method of operation the display is ±1.5kHz off-frequency to start with.

Overall, though, a good result!