Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 [2] 3 Next   Go Down
  Print  
Author Topic: HQ 170 Dial Linearity  (Read 75930 times)
G3RZP
Member

Posts: 1268




Ignore
« Reply #15 on: August 30, 2014, 12:42:23 AM »

If you feed the 60kHz into the mixer (6BE6? from memory) that converts 455 to 60 kHz and count the signal at the next stage grid, the tuned circuit will have cleaned up the waveform enough for the counter to count it.
Logged
K5TED
Member

Posts: 233




Ignore
« Reply #16 on: August 30, 2014, 02:31:38 PM »

I got the 170A for  Christmas 1969 (I was 11) It was impossible keep calibrated to any modern standard. Last HQ I bought was in in 1995. First mod was to install a freq counter probe and rear jack. Used a cheap programmable CB freq counter. Golden. Dial freq is relative. Freq display is standard.
Logged
KD1I
Member

Posts: 466




Ignore
« Reply #17 on: August 30, 2014, 03:45:41 PM »

Great idea, Ted, but how did you handle the offset to convert to direct frequency readout..... and do you remember where you connected it? Logically, it would be the first local oscillator but maybe not.   Guess I never paid much attention to actual frequency years ago as long as we stayed within our bands. I'm very spoiled by modern digital readouts.   I started to calibrate and align.   First thing was to set the right hand fine adjust, C 30 Vernier to zero and then adjust L4.  The frequency should have been 395 KHZ. Instead, it was 382, a long way off but maybe not so bad for a 50 + year old radio. I do expect that it really affected the pass band of the 3rd IF however.
You must have been pretty excited to receive that for Christmas 1969. That's a pretty expensive gift. It would have been about 3 weeks pay for me then.
For G3RZP, you are probably most correct. That would be after the various 60 KHZ transformers. I did use 455 for now. I'll try your method as well and compare.  Love to experiment !!!

Best 73 to All, Jim
Logged
K5TED
Member

Posts: 233




Ignore
« Reply #18 on: August 30, 2014, 06:35:39 PM »

Digalog CRC-100, has a dipswitch to program the range and offset. probe was a small ferrite loopstick shoved under the tuning section, with a short lead to a 1/8" jack mounted in the rear panel.
Logged
KD1I
Member

Posts: 466




Ignore
« Reply #19 on: August 30, 2014, 08:25:40 PM »

I'll have to look into that frequency counter in the future when this receiver is working properly.
Well, I dug into the 60 KHZ IF this evening and found first of all, that the local oscillator, with C30 fine tuning set to midpoint '0', was set to 382 KHZ.  I carefully set it to as close to 395 as I could and then realigned all of the 60 KHZ transformers. That went very well.     OK so far.

Next, after adjusting the BFO, I went to step 3, and connected the signal generator to pin 7 of V2.   The signal generator did not like that at all.   The signal level collapsed so much that the frequency counter would not count it and the AGC voltage did not budge.   I'm not sure why this is. I tried to lighten the coupling by feeding the signal to V2 through a small capacitor but that was also not successful.

I did manage to measure a signal through the 455 KHZ 2nd IF and found that it resonated at 467.555 KHZ.    That would account for not passing a signal at 455 through the 2nd IF transformers. Observation: with the 2nd IF at 467 and the 3rd local oscillator at 382, the 3rd IF was probably peaked at 85 KHZ. Just a thought.

Right now I am thinking that it was never set to spec. There is nothing sacred about setting the IF frequencies to other than 455 and 60 as long as they all track. It is possible that with the equipment of the day, using dial calibrations on a signal generator rather than a frequency counter, that the technician optimized the receiver at these other frequencies.

I'm really challenged by the signal generator not working when connected to pin 7 of V2. I can't really explain that other than a deficiency in the instrument itself.

I tried to inject a 455 signal downstream at T3 but that only showed the T3, T4 and T5 were set to 467.555 and I could not seem to get them aligned to 455. I think that is because they are double tuned with a top and bottom slug. The bottom slugs are especially difficult to turn and I may need to custom fashion a tuning tool to adjust them.

Best 73,    Jim       KD1I
Logged
G3RZP
Member

Posts: 1268




Ignore
« Reply #20 on: August 31, 2014, 12:02:36 AM »

Jim,

What band were you when trying to inject into pin 7 of V2 (which if I remember is a 6BE6)? You might do better by switching the band to 160m: any high band will have a low impedance across the grid circuit, but the higher the band, the worse it is. From memory, you would need to be switched to 160, 80 or 40 anyway to have a 455 signal out of the first mixer.

For the 60kHz to be tuned to 85kHz, you would need either the inductance or capacitance to be out by a factor of 2 - or both of them out by 1.4. That seems a lot.
Logged
AC5UP
Member

Posts: 4546




Ignore
« Reply #21 on: August 31, 2014, 01:07:43 AM »

That seems a lot.

Same thought here.  Can't help but wonder if there's a test equipment issue or something amiss between the desk and chair.

It was mentioned the signal generator was badly distorted @ 60 kc's.... If that's the ANN/URM-25D there should be no problem as the generator can go low enough to annoy dogs - if - the operator remembers to use the left side of the X-Mult rotary switch setting (10 to 300 kc) - and - the generator itself is in good condition. I have two of them and one lost its internal 400 cps tone modulation. Opened it up to find several variations of Black Beauty tubular condensers that had become physically softened. Think in terms of a rubber belt turned to goo but not as severe.

Re-condensering a URM-25D is not a task for the feint of heart as the internal workings do get tight but that's exactly what mine needed. Still accurate, stable (after warm-up) and perfectly usable with a clean output. The mechanicals are quite robust compared to modern standards and a little oil in the right places won't hurt a thing. No better time than when the case is open............  Wink
Logged

AC5UP
Member

Posts: 4546




Ignore
« Reply #22 on: August 31, 2014, 01:47:49 AM »

Right now I am thinking that it was never set to spec. There is nothing sacred about setting the IF frequencies to other than 455 and 60 as long as they all track.

Not to get on your case, but this is exactly how an alignment job goes south through the miracle of magical thinking.

You have an issue you don't understand.  Instead of troubleshooting you're rationalizing because the easy out is to assume the factory screwed up decades ago.  And it doesn't matter if the IF and mixer stages are peaked with a little extra windage because 455 and 60 kc could be only 'suggested' values.  Besides, you're working with modern test equipment and as long as the dial lines up...........

Do you really believe that Hammarlund used sub-standard alignment gear to set up a high-end model like the HQ-170 and the alignment errors you describe were within tolerance and QC acceptance?

FYI:  A production line alignment bench might have a conventional signal generator but more likely uses a custom built test oscillator with switched crystals. Less chance of operator error and best possible accuracy.

It is possible that with the equipment of the day, using dial calibrations on a signal generator rather than a frequency counter, that the technician optimized the receiver at these other frequencies.

Sure, it's possible, like it's possible Mitt Romney might be president some day, but which is more likely:  The alignment tech at Hammarlund dropped the ball on their 325th HQ-170 chassis or that you're challenged by the first HQ-170 chassis you've aligned?

I'm thinking the radio or test equipment needs some troubleshooting before you'll get a good alignment.
Logged

KD1I
Member

Posts: 466




Ignore
« Reply #23 on: August 31, 2014, 01:04:00 PM »

For G3RZP, You are correct, Sir, it is a 6BE6 and I had the band selector set to 3.5 to 4 MCS as step 3 in the instructions stated. (always best to follow factory instructions - most of the time) I'll try your idea of the 160 M band and see if that helps.
For AC5UP: First let me state that it can always be the guy in the chair and I am no exception. While I've been working on radios for 50 plus years, I don't claim to be a genius.    Why did I suggest that Hammerlund used 'dial accuracy'? Because we did just that at Sylvania Government Systems building military radios in the late 60's. We did not always refer to our Nixie tube counters. I really doubt that Hammerlund had any custom built test equipment to manufacture a $ 400.00 radio when we were making a radio that cost as much as a 1969 Cadillac with off the shelf test equipment. Honest.
I will agree with you that I am certainly capable of magical thinking. It is only natural to speculate how things came to be and I'm perfectly capable of wrong conclusions.
Now the problem may well be my test equipment. I'm using an old Fluke 1900A multi-counter, last calibrated in May 2000, a long time ago. However, I'd rather trust that than the dial calibration on the signal generator whose vernier scale does not line up with the main dial so when the main dial is on 60, the Vernier is not on zero.
Now, because posting here is supposed to be a learning experience, let me admit that I did not use the left side of the X-Mult rotary switch, just as you pointed out. My Error. So I thank you for pointing that out. And, YES, I am challenged by aligning the HQ 170 which is why I am doing it.  So, I'll go back and see if I can verify the accuracy of the test equipment before proceeding.
Let me thank you both for your ideas. It is always valuable to share thoughts and sometimes suffer a little humility.      73, Jim
Logged
G3RZP
Member

Posts: 1268




Ignore
« Reply #24 on: September 01, 2014, 12:00:18 AM »

When repairing, you can find funnies that got through test and QC years before. One rx I got was unstable on 15m only. Took some digging down through layers of components to find that the local oscillator coil had a capacitor across it, one end of which had never been soldered.

Another classic is the radio with 'new' 6BA6s plugged in that are really rebranded Russian tubes that have g3 and cathode connected internally. I've even seen some tubes branded 5749 (the high rel 6BA6) that were the same.

So I have a degree of cynicism at times that the radio was actually built or repaired correctly.
Logged
KD1I
Member

Posts: 466




Ignore
« Reply #25 on: September 01, 2014, 06:25:37 AM »

Peter, thank you for the words of encouragement. After restoring other parts of this receiver and refinishing trim, etc, it was certainly a disappointment to have these issues with a realignment. I've done many broadcast radios in my collection of prewar sets and even an R 388 Collins so this one was a real surprise.   
I'm sure that there are other problems afoot here. Last night I checked the frequency counter against my FT2000 and they agreed to the last digit. Where it falls apart is when the level is reduced. The frequency counter becomes unreliable and I really suspect the signal generator changes frequency as I reduce the level to keep the AGC voltage within spec.   I connected the frequency meter to the signal generator at high level - a volt or so - and then reduced the level and it can vary considerably. Was it the meter or the generator at fault here? Well, I cross checked it with my scope.  A scope is too broad an instrument to make a good frequency determination but it did show the trace expanding or contracting slightly with level. Disconcerting. 
Now that I am able to produce a clean 60 KHZ with the signal generator, I'm able to get a good peak with the transformers in the 3rd IF. However, where they peak depends up what bandwidth is selected and which sideband is selected. There could, I suppose, be issues with the transformers themselves or as someone mentioned earlier, coupling capacitors. Some bandwidths produce a higher output than others, depending upon which bandwidth was selected when the alignment was performed. The peak at different points for different bandwidths. Hope this rambling makes sense.
Very Best, Jim
Logged
G3RZP
Member

Posts: 1268




Ignore
« Reply #26 on: September 01, 2014, 06:55:26 AM »

Jim, you may well have the frequency shifting with level. The answer, at last with the levels you need to use, is an external attenuator on the signal generator. Then you run at a level that suits the counter. Getting down to microvolts becomes more of a problem because of leakage.

I can 't remember Hammarlund's method of alignment: I would expect to go the 500Hz selectivity position and align for maximum signal. I also cannot remember if you can have a 500Hz USB and 500 Hz LSB or if it is 500 Hz wide centred on 60kHz. Either way, if you align on 500Hz, you should see the response come right on the wider bandwidths. If they don't, you need to check the coupling capacitors which are switched in to widen the bandwidth.  I know on my father's radio, there was some ripple, which didn't tie in with the handbook. The problem with filters of that sort is that the response is critically dependent on circuit Q and capacitor tolerance: over the years, the Q may have gone down in the IF transformers or the capacitors drifted. Unless those coupling capacitors are silver mica, I would look to replacing them with polystyrene or silver mica, depending on what you can get. Some examination of the circuit diagram should make things clearer.....an interesting exercise would be to take the handbook values of the capacitors and from the specified bandwidths, work back to determine the design Q of the transformers. Then they can be tested to see if the Q is still OK.

Back in 1965, a guy I worked with bought an AR88, and the selectivity was poor and it wouldn't align well. Checks on the IF transformers at work showed low Q. A quiet word with a lady in the coil winding department, a list of the number of turns and 'pie' width and a bottle of gin produced lovely newly wound transformers with a good Q......with vacuum polystyrene varnish impregnation rather than wax.

IF strips of this sort are where a sweep generator comes in handy.....

Sounds like an interesting restoration project. One thing I never understood was why the HQ180 had a crystal filter at 3035 and the '170 didn't. Cost maybe?
Logged
AC5UP
Member

Posts: 4546




Ignore
« Reply #27 on: September 01, 2014, 08:52:29 AM »

...I really suspect the signal generator changes frequency as I reduce the level to keep the AGC voltage within spec.

If you're speaking of the URM-25D there should be little to no change in frequency with a change in output level. IIRC the output is buffered though a cathode follower and the cathode follower drives a constant impedance 10 dB rotary switch attenuator. Both the oscillator and follower should see a near constant load. The output level vernier does vary the gain of the follower circuit but I don't suspect that would have a significant influence on the frequency. Back in the day the URM-25 and HP 608 series were the boxes to lust after.

Tell me that you're driving the freq counter from the RF Out X-200K BNC (almost dead center on the front panel) and using the RF Multiplier BNC (bottom right) through a 'scope probe to the radio. Going the other way will give more signal level but you lose the buffering between the oscillator and attenuator. Remember to calibrate the meter to 10 whenever you change ranges then use the RF Multiplier switch to adjust the output level. The URM-25D is a flexible and accurate signal generator in terms of both frequency and level but does have a learning curve.

You would normally see some change on a 'scope display as you vary the output level because the auto trigger may well be compensating the sweep start point, but that's easy enough to check with a count of cycles per CM on the graticule.

And... Glad to hear that you're riding the levels as low as practical. Good practice and easy to forget as the AVC makes the peak harder to see.

...where they peak depends up what bandwidth is selected and which sideband is selected.

As a general rule you should peak the IF in the narrowest bandwidth setting because that's where a mis-alignment would be most apparent. Consult the service manual for confirmation. It was not unusual for the tuning of multiple width IF transformers to vary as the extra winding(s) or bypass components could de-tune the active windings. IF transformers are a compromise between bandwidth and coupling efficiency determined in large part by the size and spacing of the Pi's and the fewer parts inside the can the better.
Logged

G3RZP
Member

Posts: 1268




Ignore
« Reply #28 on: September 01, 2014, 10:06:26 AM »

The best approach to the design is to go for a kQ of about 0.7 or 0.8 - that gives a reasonable compromise between bandwidth and amplitude. k is the coupling coefficient. I can't remember if the HQ170 uses separate coils and switches the coupling capacitors: mixed coupling with part inductive and part capacitive is very difficult to get to give a symmetrical frequency response. The easiest approach is to use bottom capacity coupling, and switch the capacitors: the biggest one is for minimum bandwidth. This allows the use of a progressively shorting switch bringing in more capacity as the bandwidth is narrowed.

Radios got a LOT easier to design once you could just buy block crystal filters....
Logged
KD1I
Member

Posts: 466




Ignore
« Reply #29 on: September 01, 2014, 06:52:21 PM »

Well, AC5UP comes to the rescue for test equipment again.   Yes, I was using the Upper port, RF Out x-200K for both the counter and the radio. When I did it your way, it was much improved.   The frequency counter works fine now and the level does not affect the frequency. I had checked the frequency meter against my FT 2000 from 1800 KHZ to 50+ MHZ and various points in between and they agreed digit for digit as long as the level was sufficient. So this issue is resolved and it really was my lack of experience or knowledge in using this particular signal generator. It is no excuse but it is my first time using this particular generator and I have no manual for it.
The 60 KHZ IF is now well tuned and reasonably symmetrical. A web search has shown that these receivers are not well balanced between sidebands...lots of posts on this. The difference between sidebands with the calibrator peaked and BFO off is very little, perhaps 1/2 S unit. The big difference is between the single sidebands and both sidebands. At .5 it drops 2 S units, at 1 KHZ it does not drop at all. At 2 KHZ there is hardly a difference but at 3 KHZ it drops a full S unit when switching to both sidebands. I think I can live with that.
I still cannot inject a signal at V2 pin 7. I also tried to inject at V3 just to test but it had the same results. The test clips I was using came with the signal generator and transition from coax (BNC) to clips leads. I found a pair of capacitors inside so thinking they may possible be the issue, I purchased a 'scope probe, set it for X1 and tried again. Same result. I just cannot inject enough signal into the receiver at this point to move the AVC voltage or the voltage on the grid of the meter amp pin 2 of V13.   So it looks like the 60 KHZ IF is finally aligned as well as I can make it but the other two IFs remain to be done.
For G3RZP, I tried using 160 Meters on the band selector as well as 80 but it made no difference.  I can try swapping the 6BE6s I suppose.
and, Yes, a crystal filter at 3035 would most likely have made for much improved performance. Even my old HQ 129 X had a crystal and it seemed to work just fine.     
Logged
Pages: Prev 1 [2] 3 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!