The AM Forum

I need some advice for my HQ-170. I've realigned it by the book, but it's still not quite right. Specifically, it seems like an IF pass band isn't properly centered. He're's the symptom: when I center the BFO and then switch side bands, I should see symmetric behavior: the frequency I hear from the xtal calibrator in LSB at an S-meter reading of, say, S5 should be the same as what I hear on USB.  So, as I tune across the xtal calibrator signal, I should hear the same range of audio hetrodyne frequencies regardless of which sideband I select, but I don't. Put another way, if I switch to upper sideband and listen to a 2700 Hz tone, I see a much higher S meter reading than when switch to lower sideband and listen to a 2700 Hz tone. One of the IFs isn't properly centered.

Part of my problem may be suboptimal signal generators but I've been as careful as I know how to be. I've read the circuit theory, but it's pretty sparse and I'm not clear about what I need to do differently. This is easy to set on my TS-930S, but this beast is different. Any and all help is appreciated.

73,

Kim N5OP

Kim,

You do need to use an accurately calibrated signal generator for the 60 Khz. IF alignment and also make sure the BFO is properly set to zero beat this generator frequency with the BFO set to center.  As I recall there are some differing alignment instructions in different versions of the HQ-170 manual but I always do the 60 Khz. IF alignment with the selectivity set to .5 and the sideband set to USB as called for in the HC-10 manual.

With the Hammarlund system both the IF bandwidth and passband frequency passed are switched as you change selectivity and sideband.  I have a feeling that most of the HQ-170/180 (and HC-10/SPC-10) straight from the factory would not have provided perfectly equal sideband response at all bandwidth settings and this certainly will not improve with component aging.  Although I like my Hammarlund receivers I much prefer the same era Hallicrafters system where only the bandwidth was changed in the IF channel and sideband selection was via conversion crystal switching (either 50 Khz. above or below the prior IF frequency).  With the Hallicrafters system properly aligned you can tune an AM signal to zero beat and select either sideband and the audio will sound the same but none of my Hammarlund receivers will do this at all bandwidths although most are fairly close.  If one sideband is seriously muffled when aligned and tuned properly then you will have to take a look at the capacitors and resistors used in the bandwidth and sideband selection system.  But before doing this make sure that the alignment is correct and you are tuned perfectly to the signal.  Heathkit used a similar system to Hallicrafters with their RX-1 receiver and they sent out a note providing additional advice since it turned out many users weren't very good at properly tuning an AM signal on a receiver designed to receive one sideband at a time.  Until you get used to it the best way to tune AM is to start with the BFO on and tune the receiver for zero beat.  With a HQ-170 this should result in pretty similar sounding sidebands although not perfect like the Halli system.

+1 For what Rodger said.   What is required is a stable, accurate signal generator.

Many wouldn't think that these old receivers would require very accurate signal sources, but they really do.   Like Rodger, I too prefer the Hallicrafters design approach because its simpler and the response is inherently symmetrical.

My HQ-180 had differing response on the two sidebands as well,  and with its narrower band width wasn't particularly well suited to AM, so I sold it and bought an NC-300.   

Thanks to you both! With all it's flaws, I probably won't sell it as I have a strong emotional attachment to it-- I simply need to learn how best to align it. I too have found different alignment methods, all in Hammarlund's own documents!

In my case, the documentation says to align it at 0.5 kHz and LSB, but I'll try it with USB to see. As I understand it, the 60 kHz IF is the last IF (either third or second, depending on the band) and where all detection and selectivity takes place. L4 controls the 395 kHz mixer oscillator that converts from 455 kHz to 60 kHz, so I might be able to do some IF shift tricks. This may take a bit of experimentation...

Kim N5OP

For tough conditions the HQ-170 does quite well and I used one with my most used AM station (Ranger driving Desk KW) for quite awhile.  When I first set up this station I used a HRO-50 like the Johnson advertising but I found that the combination of QRM and weak signals were too much for the HRO when I was acting as net control for a vintage 75 meter AM net and the HQ-170 was a big improvement.  From there I used a SX-115 for awhile and now a SX-88 does duty in this position.  But used for its purpose (high performance vintage receiver) it will fill its role magnificently.

The Hallicrafters 5 Khz. selectivity position provides the same general fidelity as a typical receiver with a 10 Khz. bandpass and the audio is quite good in this position.  The HQ-170 either in its single or DSB position provides a maximum effective bandpass equivalent to a 6 Khz. traditional receiver which sounds pretty tight.  With AM you cannot offset tune very far without having insufficient carrier for low distortion demodulation with these steeper bandpass receivers although you can try using "exalted carrier" (use the local BFO to substitute for the transmitted carrier) while experimenting with your receiver. 

My current favorite Hammarlund setup is a HQ-160 with a HC-10 SSB adapter.   The HQ-160 detector/output stage provides nice quality AM when conditions are good and when necessary the HC-10 accepts the 455 Khz. IF output of the HQ-160 and provides the equivalent of the HQ-170 final mixer, 60 Khz. IF, detector, audio system, and slot filter.  I am keeping an eye out at hamfests for a parts SX-101 series because I want to experiment with a SSB adapter roughly equivalent to the HC-10 using the Halli final IF.  I plan to build in a mixer with a two crystal oscillator allowing 455 Khz. input and selection of either sideband. 

I cleaned all the switch contacts with DeOxIt then realigned the 60 kHz IF with a signal that was only 9 Hz high (60.009 kHz) as this was as close as I could set the sig gen. I verified that it stayed at the same frequency throught the 60 kHz alignment. No difference! I guess that's simply the way of the receiver in it's current state. It seems symmetric until I get to the 3 kHz bandwidth, where I'm pretty sure the LSB side is too narrow.

I tried playing with L4 (395 kHz oscillator) and that didn't result in a shift in the 60 kHz pass and buy rather a change in the received frequency. Wrong!

Because Hammarlund is rather sparse on it's theory of operation section, I'll have to ponder the beast a bit before I try to make any further progress. I gave the selectivity and sideband switches a hard look and, well... Replacing the switched components is not for the faint of heart.

Kim N5OP

Kim,

Generally receivers with variable selectivity systems like these are aligned at the narrowest bandwidth but if you plan to primarily use the 3 Khz. bandwidth you might try experimenting with an alignment there instead of at .5 to see how you like the results.  My guess is the gain and shape will be a little worse at .5 but the overall results may be closer to what you want.  This would be a worthwhile experiment before you get into component level substitution.

To avoid completely driving yourself crazy if you are planning to try to do in-depth repair first develop a good understanding of how this system works.  If the problem is mostly at a single setting of the bandwidth and sideband selections then you will be able to narrow it down to a small set of components.  If the problems are widespread then plan on making your response and gain variation measurements stage by stage since the components in the three stages will not have aged equally.

If you do end up with a Hammarlund 170 that provides perfectly symmetrical response across all bandwidths and passbands then you will likely have a very rare beast as it could be the world's only one that works perfectly :)

Thanks, Rodger!

Funny you should mention aligning it for a particular bandwidth... I began that process but bailed when I found how much I had to change things. I certainly could have gotten it to work pretty much as I want at the 3 kHz selectivity (6 kHz with both sidebands) but not at any of the others. I had to pick my poison, as it were and it seemed that I'd sacrifice a lot of sensitivity at all of the narrower bandwidths if I got it symmetrical at the 3 kHz setting. So, I put it all back in a "by the book" configuration and the began carefully examining how well things matched up between LSB and USB at the different selectivity settings. I found that each wetting was different. All tended to show a too-narrow LSB compared to USB, but they weren't all the same.

Since I've known this particular receiver (not just the model, but this very one) since 1970, I began to recall this I'd noticed this since, well... Forever. This is the way it's always acted.

The methods Hammarlund uses to vary bandwidth and sideband with only LC filters (no crystal filters) is pretty clever, but depends a lot on fixed component values. As others have noted, it's unlikely that it came from the factory symmetrical at all bandwidths, though it may have been better than it is now. I have no burning desire to make it a (temporarily!) perfect HQ-170 at each and every selectivity setting, though I probably could if I didn't have a day job. Such a condition would be ephemeral though, as components aged and drifted. I use it mostly with my Globe Champ 350 for AM, occasionally for classic CW, and it does fine for both as it is.

I can only guess that Hammarlund (or Heath) didn't use crystal filters because they were expensive in their day. This method was cheaper though a bit less flexible and less tolerant of drifting component values.

I launched this effort because the RF alignment was off and learned a lot. And, yes, the RF alignment is now spot on!

Kim N5OP

for anyone that's interested... This morning, before work, I took the time to play a bit with the HQ-170. Feeding a dummy load with my Orion II, I did some testing of the bandwidth at different selectivity settings. I run the Orion II into a dummy load at low power and the receiver gets whatever gets past the isolation provided by the coax switch.

I start out at S 5 (the alignment benchmark) and define the bandwidth as 1 S unit. Ideally, that's 6 dB, but on this receiver, S5 to S4 is a drop of about 4.25 dB. How did I define this? I fed a measured 40 W into the dummy load, set the HQ-170 RF gain to show S5, and decrease power until it shows S4. It's crude, but all I have.  So, 1 S unit (about -4 dB) defines the bandwidth numbers I show here. Here's what I get:

At 0.5 kHz: LSB 380 Hz, USB 380 Hz, Both 680 Hz
At 1 kHz: LSB 560 Hz, USB 560 Hz, Both 1100 Hz
At 2 kHz: LSB 640 Hz, USB 860 Hz, Both 1290 Hz
At 3 kHz: LSB 600 Hz(!), USB 1320 Hz, Both: 2100 Hz

No wonder it sounds a bit narrow to my ear!

Almost all of the receiver's selectivity is in the 60 kHz IF strip, so that's where I need to focus my attention. When I have the time, I'll realign it with the bandwidth in the 3 kHz position and see what I get. The receiver is plenty sensitive everywhere but 10 m, where sensitivity drops a bit. However, I intend to use it primarily for AM, so if I can enhance it for that mode, I'll decide to be happy.

Kim N5OP

Hi Kim;
I've done a bunch of videos showing detailed alignment on the HQ 170-A and 180-A and 129-X.  If you can stay awake thru them,(hi)you might find some info that will help you.  Just go to Youtube and search "K7PP".
The guys suggesting an accurate 60khz source are right on.   The RF, high IF and local oscillator can be done without the 60khz generator and in most cases you can make a simple test to determine if you need to do the low IF alignment.   There is also a video on how to do that available.  This test works for all series using the 60khz low IF and no test equipment necessary.
You might also be advised that the last 11 of these radios I've repaired or aligned have had at least one, sometimes more, bad 6BE6's and at least a couple of bad or noisy 6C4's.  
Also,  if T28 is not adjusted right on,  the balance between upper and lower sideband may be off.  This sounds like where you might have an issue.  In the plain 170, the BFO adjustment is available in SSB mode and upper/lower sideband balance can be adjusted right from the front panel but in the 170-A the BFO cannot be adjusted from the front in SSB mode.  Slightly adjusting T28 while in SSB mode and switching between upper and lower will allow you to balance between the two modes.  It may also be an indicator that your low IF needs to be adjusted.  
BTW,  you will find much passionate disagreement on how to do this receiver alignment and mostly due to the fact that there are several versions of the manual and, oddly enough, the first being the most accurate and informative in my opinion.  Some will tell you that the latest manual is the only way to go but,
I'll leave it up to you to decide....I currently have three 170-A's, a 180-A and 129-X and really love the Hammarlund line.

Very Best,
Pete, K7PP

Hi Pete,

Yes, I ran across your videos on the 170A. Mine's a straight 170, not the A model. I think my source was as accurate as could reasonably be needed: I have a pretty accurate HP counter that I've compared against others here at work that are regularly calibrated against traceable stndards and it compares very well. When I aligns the 60 kHz IF, my injection frequency was 60.009 kHz. I hope that's close as it's as close as I'm likely to get.

I aligned to other IFs by the book, but probably could have down as well using the internal xtal calibrator. T28 adjusts the BFO for zero beat whne the BFO is ceneterd and that's set pretty well. I've treid adjusting the BFO but that doesn't solve my problem. Your videos are good but a bot more comprhensive than i need. I need some specific advice about the 60 kHz IF on the 170. Would it be better to discuss the in-the-weeds details off line?

Kim N5OP

Sure, Kim;
I'm new to the board so I'm not sure if I can IM or not.
I am good in QRZ.  Pete

Hmmm... Me neither. Well, then, my e-mail is
cw underscore de underscore n5op at sbcglobal dot net.

Let's get set up that way. Once I figure this part out (with your help), I'll document what I had to do and post it here for archival.

Kim N5OP

The no reason not to discuss the intimate details here. Others may learn from the discussion. We generally don't bite.

OK,  Thanks,  Pete.
Don't wan't to make any nubie mistakes if at all possible.

Kim;
I'll drop you an e mail when I get home but in the meantime,  any chance you
could take a peek at this video and let me know how your receiver responds?

I didn't say so in the video but you are looking for an S meter increase to tell you if alignment is needed.
The rest is self explanatory and by no means the best way to test but it gives
you an idea of what the receiver might need.

Best,  Pete

https://www.youtube.com/watch?v=4l1hWNvPe2g

Hi Pete,

Nicely done!  I did the test with mine and I see that as I change bandwidths on either sideband, I see no changes, but I do see a change between LSB and USB. USB S-meter readings are uniformly lower that LSB S-meter readings. On 0.5 kHz, the difference is zero, at 1.0 kHz, the difference is 1.5 S units, at 2 Khz, the difference is 2.5 S units and at 3.0 kHz the difference is 3 s units (all starting at S 6).

So, I fail the "equality of sideband selection" test. I have what I believe is the original HQ-170 manual and followed the alignment procedure in it. What's next? As an aside, i tested the tubes in the receiver and all tested good with my Weston tube tester. I know that tube testers really test mainly emission and transconductance and so don't tell the whole story. I have spare NOS tubes, but I don't thingk that's where my problem lies based on the test you outlined.

As I said earlier, I aligned it with an injected signal of 60.009 kHz (9 Hz high). Is that accurate enough? That's about as close as I can get with the equipment I have. I don't have a 60 kHz crystal for a crystal oscillator source, though I wish I did!

Kim N5OP

Kim,  
Please bear with me a bit on this and perhaps consider trying a simple adjustment which might help me understand if it is alignment or component.

Turn your SSB selector to "Both".

Turn your mode selector to AM.

Turn on your crystal calibrator and find a signal you can peak for maximum.

Adjust your RF gain for an S7 reading on your meter.  (don't want a saturating signal)

Adjust your BFO KHZ position to zero.   (even though your BFO is off)

Switch your "select khz" switch to .5 or narrowest position.

Now switch your "sidebands" switch back and fourth between upper and lower while adjusting the frequency knob until you find a spot where there is no change in the S meter reading.  It will be critical and it may take you a bit of
playing to get it.

When you find that spot,  turn on your BFO and adjust T 28 for a Zero beat.

This will put your filters in a position to provide equal sideband detection on either side of the received signal.

This is the easiest way I can think of to determine if it's adjustment vs component failure.

It just may be that your BFO and 60khz IF's and Product detector are not
lined up....

When all this is done,  try the aformentioned test again and see how you do.
If the S meter change goes up when you switch from .5 to 3 khz bandwidth, you may still have a misaligned 60 khz IF or a component failure.
As you may have noted in the video,  I could change from .5 to 3 khz and from upper to lower to both without an S meter change.   If you can make that happen,  you're dead on.
If this test procedure works for you I can toss together another video doing it on my own equipment or if you are more comfortable having the video already done so you can reference it,  just give me a day or two to get my film crew, director and studios set up....Hi.

Good luck
Please let me know.

Best,   Pete

Easy enough, Pete! I'll have to haul it back to the workshop, but it's not a big deal. Especially since I work at a Federal lab -- though I'm not a Federal employee -- I should have time on my hands for the foreseeable future.

Kim N5OP

OK,  Kim;
In the meantime,  I'll get busy with another one of epic videos. 
I should have it posted this evening.

I cringe at the thought of folks doing a capacitor change out or some other major repair on one of these receivers when it just might be a simple adjustment.
My next video will show all four of my receivers hopefully acting the same way during the test I mentioned.
Sometimes it's a matter of being able recognize "normal", so you have a baseline.

Pete

Ha! You don't as much as I do when I consider actually doing such a thing!

Kim N5OP

OK,  Kim;
Hot out of the film room.  My latest "B" movie.

I did a comparison between four Hammarlund receivers and the results can be
viewed at:

http://www.youtube.com/watch?v=DD-kIXIUX-s


Good Luck;
Pete, K7PP

aka:   Cecil B. DePete

Well, I followed the procedure for adjusting T28 and I still have the seam problem: the LSB is much narrower than the USB. Any additional ideas, or is it time to 1) decide that this is just fine or 2) disassemble the selectivity switch stack and start replacing components?

Kim N5OP

OK,  Kim;
If your willing....

Switch on the Cal and find a Cal signal on any freq you chose.
Switch to .5 bandwidth and then switch to upper sideband. 
peak the signal on the S meter.   
Adjust the RF gain control for mid scale.
Take note of the reading.

Now switch to lower sideband and tune the receiver back and fourth for maximum S meter reading.   Take note of the reading.
Question:   Can you duplicate the S meter reading if you retune the receiver while in lower sideband or is the  reading "low" in the lower siband position no matter what you do?  I think you can see where this test is going.

I'm trying to ascertain if you have a leaky cap or perhaps a bad switch section.

All of the filtering for upper and lower sideband are done in the lower IF (60kz). The switching of different values of caps change the upper and lower sideband selection.  Your symptoms could include a bad wafer section or a leaky cap.
It would help to know these things in order to know where to start trouble shooting.
Also,  you mention that the lower sideband seems to be narrower?   

Regards,
Pete

Thanks Pete! Yes, the LSB is narrower on the 2 and 3 kHz settings. I'm an amateur musician (violin), and I'm fortunate to have very good relative pitch. -- I can recreate a given tone over a short period with very high accuracy. Zero-beating the BFO as in the first test, I find that the bandwidth of the 0.5 and 1 kHz settings are about the same (same tone yields same S-meter reading in each sideband), but there's a *big* difference at 2 and 3 kHz -- at higher pitches the same tone yields vastly different S-meter readings, and -- before anyone asks -- yes, I'm on the proper side of the signal.

Here are the results of your suggested test. My standard was S7:

Selectivity    LSB     USB
     0.5         6.9      7.0
     1.0         6.0      7.0
     2.0         5.8      7.0
     3.0         5.8      7.0

There's an asymmetry in the IF chain and the culprit appears to be in the LSB position. Next hint?

Kim N5OP

Kim,  let me do one more last video.  You may be describing a normal condition but what you are seeing is a bit confusing to me.   I sure don't want to see you tearing into your radio if it's an alignment issue.   Give me a day or so. 

Best, Pete

Assuming it will take about a week for parts to arrive, I ordered a full compliment of replacement sliver mica capacitors for the sideband selection switch. I may be able to get away with replacing only those on the LSB side, if I could figure out the switch position shown on the schematic. The notation says "Sidebands Shown in upper pos. (panel knob indicates lower pos.)" How's that again?

Kim N5OP

No worries, Pete. Heck, I got nuthin' but time right now...

Kim N5OP

I'd be surprised if what I'm seeing is normal. Recall that I did similar tests with an external source and found that the LSB position was quite a bit narrower than the USB position. That the difference be imes striking at awvtivity settings of 2 and 3 kHz is, to my thinking, anyway, partcularly telling.

These are rather easy tests to make because what's being done is a "sweep" of the rx passband essentially by ear. In a properly aligned sine sideband rx, sweeping a carrier through a single side band will result in a sweeping tone. The signal strength is a function of tone frequency, filter skirts and bandwidth. If the passband has "brick wall" skirts, is 2700 Hz wide and passes frequencies spanning 3000-300 Hz, then the same frequencies should be heard in both sidebands. A tone of 3001 Hz or 299 Hz would vanish because they come from a carrier outside theRX passband.  If the carrier starts low and sweeps upward through the LSB, the tone will start high and sweep low. The opposite will occur in thee USB.

At least that's how I figure it and it's what happens in my TS-930S and Orion II receivers.

Kim N5OP

OK, Kim;

Can you check this one out and comment?

Best,
Pete

http://www.youtube.com/watch?v=75nOZJIxPRc

Well done video, Pete. I like seeing the passband move across the signal :)

Like yours, I see a weak response in the opposite sidbeand. UNlike yours, I never hear any tones even close to 3000 Hz in the LSB position but I DO in the USB position. Your sidebands respond symmetrically while mine do not. Mine is NOT balanced on both sidebands. We're looking at MUCH worse performance than yours.

The asymmetry between sideband response becomes most striking at selectivities of 2 and 3 kHz. There is negligible asymmetry in the 0.5 and 1 kHz selectivity settings.

I looked at the sideband selection switch and replacing any of the three capacitors hanging off of it is trivial. To replace any capacitors on the selectivity switch will require disassembly. How do I do that?

Kim N5OP

I've been studying the switch assemblies. All of the contacts have plenty of spring tension and all are intact. So there appears to be no obvious mechanical problems.

These switch assemblies can be removed, or at least dismounted and moved around a bit and were certainly not constructed in place. The 1/4" shaft from the knob goes into some sort of brass coupler to the keyed shaft that turns the movable contacts of the switch. All of this is mounted on a small subframe that provides mechanical stability.

I could go two slightly different ways, here. I could dismount the subframe and move the entire assembly towards the rear, get the shaft extensions free and move the assembly around as needed. That will be about awkward because of the long extensions but tractable. The second way is similar, but made easier if I can remove the 1/4" round shaft extensions. The second way may require removal of the indexing cam, and getting it back together could be... Sporting. Anyone with experience doing this on a similar receiver, feel free to chime in.

Kim N5OP

Hi again,  Kim;
Just got back from an all day Car show and saw your post. 
Before talking about pulling out switches, can I ask you to check one more thing?
Hope you don't think I'm a nut case but can you look at the notch filter and make sure that you get your notch in the middle of the range (0) by using the calibrator and once you verify that,  move the notch adjustment fully clockwise.
I just want to make sure that the notch filter isn't miss-adjusted and causing your symptom.
If you are satisfied that it is not part of the problem then, perhaps, one more check.
Check the alignment of "L4".  This can be done is step 3 of the service manual.
If this is off,  it is possible it could cause your symptom.
Keep in mind that L4 has three different peaks.   The largest of the peaks being the correct one.
If L4 is not centered on the proper peak, you could tune all day and not get the symmetry on both sides of your received signal.

That aside,  I'm going to have to agree that you may indeed have a bad cap.
I'll take another look at my receiver and let you know if I have any worthwhile comments.   I think you already know that the low IF string is shifted slightly to detect the upper and then the lower sideband.
The ganged wafer switches add or subtract values that shift each transformer stage all at once.  It would have been easier to use a crystal filter but it is what it is.

Pete

I'm pretty sure the slot is aligned, but I'll double check. I was utterly unaware that adjusting L4 could result in 3 different peaks. I've never e pediments experienced more than one, but it could have been mis-aligned arly on, so I'll check for that. IC those pan out, I've figured out how to get room to work on the assemblies and it won't be too terrible if I can gt enough room to work.

Thanks, Pete!

Kim N5OP

I have really enjoyed keeping up on this thread!  With all this great material we need to capture it in one documents so others can use it.  I have a 170 and although mine is in very good shape cosmetically, I am sure it will need some under the hood time. 

Thanks,

Joe, W3GMS

I don't recommend archival until we've figured out what the problem is :)

Figuring out how the switch assemblies were installed was interesting two wires prevent me from getting it free. All others were left dressed such that the assembly could be greed from the front of the chassis. I suspect a shift change resulted in a slightly different dressing technique because a different tech finished the assembly wiring.

I'll have to disconnect two wires to two different IF transformers. To free the assembly enough to work on it. Even then, it will be tight.

I suspect Hammarlind didn't use crystal filters due to cost. There were no monolithic crystal filters back then -- all crystal filters were discrete and were built and matched by hand. It's actually pretty impressive what was done with LC circuits. It's not as good as crystal filters, but still pretty danged good!

Kim N5OP

It seems to me this is taking an extreme turn that might not be necessary. 

1.  Response switching in the HQ-xxx 60 Khs IF strip was never MIL-spec. 

2.  You need to do measurements of response with the AGC OFF!! and an RMS voltmeter on the output, making sure the input levels don't overload anything.  The S-meter is inaccurate to the point of being useless.   If you don't have an analog RMS voltmeter, get one .. many on eBay for non-absurd pricing.

The best way to align one of these (or anything else for that matter) is with a sweep generator and a scope.  Beg/borrow/steal one.  If not, then align it by the book, adjust the 455 -> 60Khz bandpass oscillator for the best match between LSB/USB and move on to the next project.  I've done a half dozen of these things (HQ-170/180/HC-10) and never had to fiddle with the switches, or the caps, or anything in the IF bandwidth circuits.  I'm not saying that you might not be able to marginally improve things. but you have to consider the return ...

Grant NQ5T

Well .. not so much.  LC circuits have some advantages. The real question I've always had is why Hammarlund/Halli didn't add just one more stage with a couple more tuned circuits to improve the skirts a bit.  You won't find a crystal filter, or even a hallowed mechanical filter that will out do something as pedestrian as a BC-453 on skirts (relative to the era).  And the real benefit of an LC circuit is reduced group delay and distortion.  It's why AM'ers generally prefer an R-390 over an R390A.  It's quite possible to build an LC filter that will put a crystal filter to shame (if you can find the parts today) -- but the LC filter might be MORE expensive.  Obviously, YMMV, depending on a lot of stuff ... :)

I'd have to agree with some of Grants comments in that there are better but not always easier ways to trouble shoot problems like this.
It's one of the reasons I spent so much time trying to trouble shoot the issue using the receivers own meter and filters to narrow things down. 
Of course there are better ways to proceed IF, you have the proper test equipment.
Many hams do not have the required test equipment needed to do advanced troubleshooting and that puts them in a position to either have to borrow it,  have it done and perhaps pay for it or to shot gun the components in the circuit.

If it's possible to avoid the shotgun method, you might save yourself allot of trouble.  Even after you shotgun,  you still have to revisit the alignment issue.

 I might attack this a bit differently on the bench at my shack as it is fairly well equipped with the necessary test gear and start, as Grant suggested, with a sweep of the low IF and then take a scope and work your way down the 60 khz amps and look to see where the response goes bad.
I've also had pretty good luck using my frequency selective voltmeter to supplement past troubleshooting efforts.
If it's not an alignment issue to begin with then your choices become very limited.

I might also add that the "by the book" alignment does not call for a sweep alignment of the low IF.  They have you inject 60 khz and peak each of the transformers with the AVC disabled and an DC VTVM or equivalent to measure
the negative going voltage on the AM detector tube.  This seems fine for receivers that aren't having a problem but doesn't really tell you anything about the conditions Kim is encountering.

So,  that leaves Kim at an interesting crossroad.  The tests he's done seem to confirm a technical issue in the low IF.  He can borrow or buy the test equipment needed to inject a sweep signal into one or more stages and look at the response on a scope which should uncover the culprit or he can assume a faulty capacitor and shotgun them all without the need for buying or borrowing test equipment.

As many Hammarlunds as I've fixed over the years,  I've never had to delve into a component fix on the low IF so I have no worthwhile culprits to point my finger at.
Kim, If you can obtain the necessary test equipment and chose to do a stage by stage evaluation to trouble shoot this problem I could do a video showing a suggested procedure as I do have a sweep generator and scope as well as a frequency generator source.
If you decide on a wholesale replacement course of action, I'm sure we'll all still be here if you need further input.

Perhaps some of the other folks on the board might be interested in doing something similar regarding posting a troubleshooting procedure as well. 
I use an Iphone 5 for all of my YouTube videos and post directly from the phone to YouTube right there in the workshop.  I'm kind of an old geezer so I had to have one of my kids show me how.....LOL.

Best,
Pete

The alignment was done by "the book" using Issue No. 5, for serial numbers beyond 3800. Unfortunately, I have a malevolent sister who destroyed the original manual, so this is a replacement manual. I don;t know the serial number of the receiver, as I can't seem to find one anywhere. I know it was purchased in June 1959. Regardless, Issue No. 5 is "the book" I used for the alignment.

As per the manual, AVC was off and I took my readings across C44, a convenient place to access the output of the AM detector (and the point the manual recommends). Signal levels were kept in spec (no more than -1 VDC across C44). The point being that I didn't take any shortcuts.

That said, I am skeptical of passive component failures, especially silver mica failures. I don't have a service monitor, nor do I have a sweep generator. But W5JO has graciously offered me access to his sweep generator, so I think I'll take him up on it. If we find things truly wacked out by some sort of passive component failure, we'll find out for sure that way.

I don't expect MIL-spec performance from the old beast, but I know it can be better than what I'm seeing. I don;t understand what the problem is and I'm hesitant to simply start replacing stuff. If there's a problem, we'll uncover it and THEN fix it.

Kim N5OP

Jim is justifiably concerned that such a procedure could take several hours and that we might have to leave it in mid-procedure, so he's going to loan me his tracking sweep generator. Pete, that means I'll need your video to show me how you think this should be done. So, I accept your instructional offer!

Kim N5OP

OK,  Kim;
I'm going to need some time to pull the 170 out of the shack and get it up to the shop.   It will be fun to do as this 60 khz IF is common to several models and I've never done a sweep on the 170 before, but I do have a basic 170 so I can duplicate the exact radio you have. 
I think you are making the right choice and I'll do my best to get it onto a video.
I have done a sweep alignment on a 129-X which is currently on YouTube.

Pete

Hi Pete! I've jumped the gun on you and looked at the 60 kHz IF passband with a sweep generator. W5JO loaned me one as well, but I learned that the Elenco GF-8025 function generator that I recently bought out of an estate has sweep capability. Not only that, but it also has a sweep output that I can apply to the X channel of my Tektronix 2225 'scope so that the horizontal sweep is synced with the frequency. It's not the fanciest sweep generator, though, because I can't narrow down the sweep range very well, so it sweeps a broader range than I want. However, it's still easy to visualize what's going on.

I've attached several images I took of the passband -- all are at the same vertical sensitivity and none look like what I expect to see based on the manual. The vertical scale is 0.1 V per division and the sweep is about four per second to avoid any spurious high frequency components induced by the sweep itself. Signal injection is at the 3rd mixer and I pick off the signal at the AM detector, just as instructed by the manual for signal injection for alignment. AVC is off as is the BFO.

I played a bit with the 60 kHz alignment and I can change the shape of the passband and make it symmetric at 6 kHz bandwidths by selecting 3 kHz selectivity and "BOTH" sidebands, but I can't make it flat across 6 kHz. Making the passband symmetrical at 6 kHz bandwidth also screws up the sensitivity at narrow bandwidths making the receiver deaf as a post.  If I make the 6 kHz passband symmetric it retains a good-sized dip in the center and I still see the asymmetry between LSB and USB output. I cannot increase the USB output level while also lowering the LSB output level. 

After a complete "by the book" alignment, I get the following: image 1 is LSB at 0.5 kHz selectivity, image 2 is USB at 0.5 kHz, image 3 is LSB at 1 kHz. More images to follow...

Kim N5OP

More images: image 4 (in the series) is USB at 1 kHz, image 5 is LSB at 2 kHz, image 6 is USB at 2 kHz.

Kim N5OP

Image 7: LSB at 3 kHz, image 8 USB at 3 kHz, image 9 BOTH at 3 kHz.

Besides "not right," what is this telling me about where to look for problems?

Kim N5OP

I also want to reassure anyone interested that I went through a complete realignment of the 455 kHz (and ensured that I set L4 for the biggest peak) and 3035 kHz IFs, as well. While I found that the 455 kHz IF was a bit off and that getting it better enhanced the receiver sensitivity, it made no difference to the 60 kHz bandpass response.

Kim N5OP

Oh, and finally: yes, I inverted the channel when I took those pictures. Those are, in fact, negative voltage traces.

Kim N5OP

Wow,  You are way ahead of me,  Kim.
I haven't even cleaned my bench off yet.

Nice display pix.
The scope display seems to bear out your description of your problem.

Seems like the passband is working properly in the .5 position on both sidebands.  It's starting to appear as if the sideband switching might be normal and the capacitor or capacitors providing the pass band shift in one of the IF stages might have a problem.

Is it possible to move down the IF chain and inject at the grid of each amplifier tube and inspect the same sideband information?
You may come to a stage where anything down stream appears normal.

Also,  do you have a scope probe with a detector?  You might be able to test each stage independently.
One last question.  When you did the book alignment,  What sideband position were you using?
And,  another thought.  Notice the capacitors used in the band pass shift have a common connection and the other end goes to the wafer switch and are not connected to anything when not selected?  If you can connect a capacitor tester or even a simple ohm meter and un-select the cap being tested, you can check it for continuity.  Doesn't evey body has a little bit of Irish luck they can draw on from time to time?  Might find it right off the bat.

BTW, Nice work!
Regards,
Pete

Thanks, Pete!

Hmmm, lessee here...

Actually, if I get all the capacitors I need, its not that much more work to replace them all vs. replacing them piecemeal. The main work is in freeing up the switch assembly so as to be able to get to the capacitors.

No, I don't have a capacitor tester, especially anything that goes down to the capacitance values Hammarlund used. I have a Fluke meter that is suppose to do that, but I don't think it goes down to values that are small enough.

If I move down the IF chain, I wonder if I'd know when I found something odd? I sort of have to know what it should normally look like to know it's wrong. I'm seeing the entire response here -- if I'm unlucky enough to have two bad stages, that could get so confusing that I can no longer tell what's normal and what isn't.

No, I don't have a detector probe.

To do the ohm meter trick, I still need to dismount the switch assembly -- there's simply no room to squeeze in a probe otherwise. To get enough room to work on it, I'll have to cut two wires. All the others were left long enough, but two were not. Between the two that are not long enough and the the rest, I suspect there was a shift change.

No magic wands, eh?

Kim N5OP

Hi Kim;
I'll have some shop time this Saturday.  If at all possible I'll look at my 170 and see if I can get some pix.

How about taking a chance and seeing what the sweep shows at the last stage.
Inject the signal at pin 1 of V7 and see if you can get a trace that appears to be the same amplitude on either side band and see if the symmetry follows the 1, 2 and 3 khz positions.  If it looks normal,  move up a stage for the same test,  etc.
If It doesn't buy you anything and you're not in a great hurry, I'll try to post some pix Sat evening.

Pete

I realize the last pic you show looks pretty radical, and, the response is certainly not as it should be, but its not as bad as it seems.   Your scope is displaying a "linear" response, while the illustrations in the service manual show a logarithmic response.   The two scales are very different.   Actually there is only a 6db difference between the highest and lowest amplitudes at the pass band edges, not as earth shattering as it appears.  Linear images on your scope can show a voltage ratio of about 10:1 at best, while the illustrations in the manual, though somewhat idealized, show a voltage ratio of 10,000:1, certainly not the same scale.   Using a simple LOG detector, you can easily compare apples to apples, because the images will be on the same scale.

(http://i670.photobucket.com/albums/vv62/mikeinkcmo/Radios/HQ-170/IFPassband_zpsbf185d70.jpg)

Here is the schematic, and pic of a LOG detector.   The IC is about $7, and you can use a standard 10X scope probe for the input test lead hooked to the radio's detector output or any of the IF Amp grids.  Here (http://pdf1.alldatasheet.com/datasheet-pdf/view/174920/AD/AD8307.html) is the data sheet for the IC.

(http://i670.photobucket.com/albums/vv62/mikeinkcmo/testgear/Log%20Amp/LogAmpAD8307.jpg) (http://s670.photobucket.com/user/mikeinkcmo/library/testgear/Log%20Amp?sort=3&page=1)

(http://i670.photobucket.com/albums/vv62/mikeinkcmo/Radios/Halli%20SX101/Aligned%20IF%20Sweep/ASweepA.jpg)

Although the pic here is of an SX-101 I.F., the overall results would be similar for most receivers. The pic here shows a 1,000:1 or 60db voltage ratio.

(http://i670.photobucket.com/albums/vv62/mikeinkcmo/Radios/Halli%20SX101/Aligned%20IF%20Sweep/ASweep5.jpg)

You're right, Mike! I thought about that as I looked a the responses and did some log comparisons in my head. Something along the lines of "I'm looking at voltage, so dB is 20 time the log of the ratio..." and still didn't like what I saw. There's 6 dB difference between the USB and LSB setting and they are decidedly asymmetric. Those things alone are enough to force me (kicking and screaming, I assure you) to start diving in.

I need to look into building such a log prescaler for the probe.

I think, though, that I will start looking at each stage independently -- I may find that a particular stage is acting up. However, there are three stages and I might have more than one problem. If two stages have problems, then the good stage is the one that will look "different." I have to remember that this is, after all, a labor of love. If someone had to pay for this, well, it would probably never get done.

Kim N5OP

You know, when I first dive into these types of projects they wind up, at least for me, being considerably easier than I first imagined.  Not that I wanna do it again real soon. ;D

I believe you said earlier you ordered some caps.   You might also get some resistors in case some of those are out of spec as well.  I presume there are two sets of caps/resistors per stage, one for upper and one for lower, and if such is the case, you should separate your caps and resistors into "matched pairs" for each stage, even though the "sets" won't be exactly the same for each stage, your asymmetry will be better than what you have now.   The net result will be that each BW will probably have "softer" corners.

Thanks, Mike! Since this receiver and I go back a long way, I plan to keep it around for a long time and treat it well. I was thinking the same thing. They are all composition resistors and metal film is a better bet. I think I have enough resistors on hand, but if not they're
easy to get. I'll see if I can find someone with a capacitor checker so that I can match them up. I doubt that Hammarlund (or any of the other manufacturers) worried much about hand matching such components for the civilian market. They may have done it for some military applications and they may really have tested everything that went into even these stages.

At this point, what I really wish I had were HC6U crystals for 60 kHz, 455 kHz and 3035 kHz as that would make aligning this receiver much easier. Unfortunately, such crystals are rather difficult to find anymore; I've certainly had very little luck.

Kim N5OP

They call out DM-15 silver micas.   Could be wrong but I believe those are 5%.  In any event the ones in the radio probably will have the tolerance stamped on them.

The silver micas I could see said either 2% or had an actual capacitance tolerance, as in "+/- 2 pF." I also looked at the switched resistors. While nearly all are in spec (based on the tolerance markings), nearly all on on the high side, which is what I'd expect with carbon compositions resistors.

I did some stage-by-stage testing today. There are three IF stages and two IF amplifiers. The first stage is right off the 3rd mixer, the second stage comes after the 1st IF amp and the third stage after the second IF amp. So, I injected the signal at the control grid of the second IF amp, showing me the response of the last IF stage. Perfectly symmetrical with the proper shift between sidebands. Injecting the sweep at the control grid of the 1st IF amp, I saw the sidebands move as they should, but a bit of asymmetry in the amplitude between USB and LSB (USB was lower). The asymmetry gets worse when I inject the sweep at the 3rd mixer. Were I looking at this on a log scale, the defects I see would be much less evident.

I've been reading about silver mica failures and, except for a particular kind in IF cans, failures are very rare. The failure mode is almost always very characteristic ("thunderstorm static" as the unit warms up), and I see no inkling whatsoever of that. The silver micas in this receiver are brown epoxy-dipped units. I'm becoming convinced that there's a very high likelihood that the silver mica capacitors are absolutely fine. If anything, I should be more suspicious of the carbon comp resistors and replace them first with all 5% ceramic units, re-test and decide from there.

What I'm seeing and have so carefully documented may be much the way the receiver really was the day it arrived in my dad's shack in 1959. 

Kim N5OP

Kim,

Excellent work.  This is very interesting, and your last statement may well prove to be the most logical conclusion.   When you move the injection point from stage to stage, are you moving the scope along as well, so you were only looking at single stage responses?

And, if you were looking at the individual stages, did you attempt to alter their response by adjusting the inductors?

No, I didn't look at single-stage response because I don't have a detector probe. The only single stage I could look at was the last one: I could inject the signal at the 2nd IF amp and see what only the last transformer response looked like. I did try hanging a 1N34A diode off of the output of the transformer (essentially the control grid of the next IF amp) but that didn't work. It acted as if the stage became very heavily loaded and I could see nothing on the scope.

I then injected the signal into the second stage, but I then could see only the combination of the second and third stages and not the second stage in isolation.

I've looked at what the switching does. Obviously, it switches capacitors that step the IF passband to one side or the other. Switching also changes the screen voltage of the 6BA6 amplifiers. Switching the screen voltage must address the different gain needed at different bandwidths. It also looks like the bias is changed, too. Looking at the capacitors, it's evident that these are used to not only step the passband to one side or the other, but also appear to stagger tune the stages. Resistors are also placed in series with the switched capacitors, which must address the Q of the individual stages and so broadens the response. 

I didn't attempt to alter the response by adjusting the slugs. I'd tried something a bit similar earlier, but while looking at the entire IF response at the 6 kHz bandwidth (3 kHz selectivity and "both" sidebands. If you saw that post it said that I could get a pretty good looking 6 kHz response at the cost of an utterly deaf receiver at the narrower bandwidths.

I do not suspect the capacitors in the "cans" because each and every transformer shows a very marked peak as I adjust it. If those capacitors were bad/open/shorted, adjustments would make no difference.

Kim N5OP

Pete, K7PP, was kind enough to haul his HQ-170 out of its exalted spot on his operating desk and into his work shop. There, he swept the 60 kHz IF just as I had. And, Lo! and Behold! His looks just like mine.

I have come to the conclusion that there's no trouble to shoot in my receiver. The reason I thought there might be problems was multifaceted. First, I was bolstered by theory but forgot a fundamental maxim: "In theory, there is no difference between theory and practice. But in practice, there is." I know exactly what the ideal should look and act like. I've nearly accomplished that with my TS-930S and I expect no less from my Orion II; because it's all done in software, I can make it look pretty much any way I want. I worked diligently to make my TS-930S as near perfect as possible. I expected to be able to get pretty close with my HQ-170. And, in fact, I got closer than I'd initially thought.

Had I displayed the passband response on a log scale, instead of a linear one, I would have seen that it was flat to within about 1 dB. I certainly would have seen some asymmetry, but it wouldn't have looked as bad. I know because I took actual readings and then plotted them on a log scale and it sure looks a lot flatter that way. So, Mike, KE0ZU, was onto something significant in that alone.

I was surprised to see 3 dB difference between the USB and LSB selection, but forgot that this is a hot receiver with gain to spare. In practice, I'd never notice the 3 dB difference. Obviously, I hadn't noticed it during operation: I only became aware of it when compared to the ideal.

I then failed to temper my expectations with a dose of reality. In truth, I'd never delved this deeply into the alignment before, nor had I read of any details from anyone else who had. I mislead myself based on what I saw depicted in the manual, which was on a log scale.

In the process, though, I learned a helluva lot about this receiver. I actually measured some resistor values and learned that, because it has always been kept in a controlled environment, the resistors that are used in the 60 kHz IF chain are all still all within their tolerance specification. Some research showed that I needn't worry about the silver mica capacitors in this receiver: the type that are used in the IF chain bandpass and sideband selection simply don't go bad unless they're damaged physically or subjected to too much voltage. I checked all the tubes and found them all fine (I have a Weston 981 tester; it says it's a Type 2, but because of some mid-type production changes it's nearly a Type 3). I learned that I suffer no evidence at all of failures of the capacitors in the IF cans. Overall, I learned that this old beast is in remarkably good shape. As Jim, W5JO, quipped, I've been given a tremendous head start with this gear because my Dad kept it in good shape and well protected for many, many years before I got it.

So, with all of Pete's expertise and experience to reassure me, and Mike's note that I'm not looking at the response in the way it's usually depicted, I have decided that the performance I'm seeing is probably very representative of the way it was the day my Dad put it into service in his shack back in 1959. So, I'm going to button it up and put it back where it belongs: next to the Globe Champ 350 that has been its constant companion since the day it was unpacked.

Then I'm gonna enjoy some radio time.

Kim N5OP

One other note I feel I should add: having done all of this on my Hq-170, I learned a bot more abpout operating it. The LSB is a bit narrower than the USB -- that's the way it is. But, as I fiddled with it, I began to learn about some of its flexibility that isn't explicitly mentioned in the manual. In fact, the manual says to change the BFO setting only for the purposes of adjusting CW pitch. Maybe what I learned it was common knowledge in the day, but I've never heard anyone discuss it. 

In particular, on SSB it is possible to emulate, though not completely duplicate, variable passband tuning (PBT). I can move the location of the received passband by offsetting the BFO. On LSB, if I offset the BFO down in frequency (and re-tune for intelligibility, of course), I can move the passband into the upper frequencies of the transmitted signal. If I move it up in frequency, I move the passband into the lower frequencies of the transmitted signal; vice versa for USB.

Since my Globe Champ 350 doesn't do SSB, I'll probably not use this very much on a routine basis, but it might come in handy if I have QRM troubles on AM and decide I can alleviate some of it by  receiving only one of the sidebands.

Kim N5OP

Kim,  other than what you just mentioned,  I've found my Hammarlunds are a real joy on AM mode as well.  If you have ever been on the "AM" frequencies on 75 or 160,  you may hear background QSO's going on in other parts of the country and not strong enough to cause a local QSO a problem but strong enough to be an irritant.   Many are not "netted" to the exact frequency and a beat note can plainly be heard.  Try moving the notch filter to the beat note and you may be surprised at the suppression you can achieve.   Also,  you can switch sideband filters with the BFO off.  Sometimes a great QRM fighter if you have SSB above or below you.
Lastly,  the Hammarlund "A" series has a separate "SSB" position where the BFO control has no effect.   You can still switch to CW and do the same thing with the SSB in the upper or lower position.
Pretty slick for an old boat anchor.

Got my 170 put back together with new dial light bulbs so we're already to go back in service.  Just starting an installation of a crystal calibrator in the old 129-X.   Should be a real help finding the band edges.

Pete

I haven't needed the notch, yet, but I know that is deep enough to completely eliminate the calibrator signal! As for using USB and LSB without the BFO, I know it's possible, I've simply not needed it, yet. I have used the sideband mode in a few AM QSOs to dodge QRM. It almost always helps.

Thanks again for all your help and efforts, Pete!

73!

Kim N5OP