National NCX3 Speaker Buzzing

[AB2EZ]

I did remove the 280v wire that comes in from the PS and heads for the 0a2 and associated voltage divider, the noise then stopped!

The voltage divider chain leading to the OA2 is drawing 80ma (I.e. a 60V drop across the 750 ohm resistor) from the capacitors across the 280V supply. This current would cause a 120 Hz voltage ripple having amplitude of 8.3V across a parallel pair of 40uF capacitors. The choke L2 should be providing a steady charging current to offset this discharging current, and therefore to reduce this ripple.

If disconnecting the voltage divider chain makes the hum go away, this would confirm that L2 is not doing its job. I.e. it doesn't have enough inductance to provide a smooth charging current of approximately 100mA to the parallel pair of 40 uF capacitors.

You could reduce the ripple by increasing the inductance and/or adding more capacitance (e.g. adding 100uF) across the 280V supply.

Separately, if the hum is at 60Hz (rather than 120 Hz), there is almost certainly  a defective diode, or an improperly installed diode in the 4-diode bridge.

Stu

[KA2DZT]

Didn't know that the set has interconnecting Cinch plugs.  I'm not familiar with the NCX3 but if the interconnecting cables uses a single ground wire for filament, B+ and the speaker return, this can be a problem.  The speaker should have a separate set of wires, to the spk and back to the output xfmr.

I have to look at the schematic, but from what Stu has mentioned, this set uses a single set of 4 FW bridge rectifiers.  The center tap of the HV xfmr is where 1/2 the HV is pulled from.  So, you get 700 volts from the output of the FWB rectifiers and the lower 350 volt B+ from the xfmr CT.  Each of these two B+ outputs needs to be filtered separately.

Fred

[MikeKE0ZUinkcmo]

Here are some schematics of the subject.

 

As mentioned, the two units are interconnected with about a 3' cable, partially visible in this picture

[wa3dsp]

It looks like there is a good amount of filtering - 40uf in the supply and another 40uf in the transceiver on the 280V line.

Both the low and high voltages are choke input. Specs are shown for L2 on the schematic but not L1 in the low voltage supply.

If you want to be sure that this is not being caused by the regulator pull when end of R107 which feeds B+ to it and see what happens. I doubt it is causing the problem but easy to try.  If the choke were shorted I would expect the B+ to be on the high side, possibly 330 volts or more. If you have a good digital meter you could measure the AC ripple on these lines. I would expect it to be less than a volt and preferably <100 millivolts on the low voltage lines.

[KA2DZT]

Mike,

Thanks for posting the schematics.  Take a close look at the Cinch connector on the PS.  Seems that the spk ground return is jump to the main ground return even though it shows having a separate pin connection on the other connector.  I have to take another look (can't do it while I'm typing) but there something there that needs a closer look.  Running the spk grd return with the filament return will cause a problem.

Fred

[KA2DZT]

Mike,

Just took another look,  maybe that jumped ground is at the return chassis end.  Could be the cables need to be checked for any problems or mods.

Fred

[Pete, WA2CWA]

That's why I posted to check the interconnecting cable plug on the power supply and the mate in the NCX-3. This type of wiring harness and typical connections were very common on a a number of HF and VHF transceivers back in the 60's and 70's. The cable stiffness exerts a lot of stress on the pins and the soldered connections. In almost all cases, there is no strain relief(clips, etc.) to hold the cable plug firmly into the mating connection on the rig. Poor soldered connections, oxidation or crud on the pins, and, since all the individual cable leads are stranded wire, slivers of stranded wire can bridge or break, causing ground connections to be marginalized. I suspect a redesign of the power supply is not necessary to solve the problem.

[AB2EZ]

Look at the schematics that Mike posted above.

The voltage across C3 should be about 350V (I.e. about half of the 700V B+, as Fred points out).

The voltage across C96B is supposed to be 280V.

Therefore the total current being drawn by the transceiver from the 350V supply is: (350V - 280V)/ 660 ohms = approximately 100mA

Likewise, the current flowing through R106 is supposed to be (280V - 220V) / 1000 ohms = 60mA.

Likewise, the current flowing into the voltage regulator is supposed to be (220V - 150V) / 2500 ohms = 28mA.

Therefore, if you want to partially unload the 350V power supply, to see if that reduces the ripple/hum, you should disconnect one end of R106 (rather than R107).

If the inductor L2 is not doing its job, then disconnecting one end of R106 should reduce the ripple/hum on the 280V (nominal) line.

Stu

[AB2EZ]

That's why I posted to check the interconnecting cable plug on the power supply and the mate in the NCX-3. This type of wiring harness and typical connections were very common on a a number of HF and VHF transceivers back in the 60's and 70's. The cable stiffness exerts a lot of stress on the pins and the soldered connections. In almost all cases, there is no strain relief(clips, etc.) to hold the cable plug firmly into the mating connection on the rig. Poor soldered connections, oxidation or crud on the pins, and, since all the individual cable leads are stranded wire, slivers of stranded wire can bridge or break, causing ground connections to be marginalized. I suspect a redesign of the power supply is not necessary to solve the problem.

Pete

This is good advice, but if a loose ground is the problem, it is hard for me to see why replacing L2 with a different choke would reduce the hum; or why some of the other tests that were done would produce the results that were posted.

Stu

[Pete, WA2CWA]

I just pulled my data on the NCX-A power supply. Looks like there was at least three different series numbers assigned to the product over its life but all circuitry looks the same. The basic circuit design of the NCX-A P.S./Speaker was used for the NCX-3, NCX-5, NCX-200, and the NCX-500.

For L1 Specs. - Swinging Choke - 60 ohms DC, 25/300 MA, 12/2 HY
For L2 Specs. - Filter Choke - 120 ohms DC, 200 MA, 8 HY

In the past, I've used a Clegg Venus P.S./speaker and the Swan 117-X P.S./speaker in place of the NCX-A. The designs were pretty much identical.

[Pete, WA2CWA]

That's why I posted to check the interconnecting cable plug on the power supply and the mate in the NCX-3. This type of wiring harness and typical connections were very common on a a number of HF and VHF transceivers back in the 60's and 70's. The cable stiffness exerts a lot of stress on the pins and the soldered connections. In almost all cases, there is no strain relief(clips, etc.) to hold the cable plug firmly into the mating connection on the rig. Poor soldered connections, oxidation or crud on the pins, and, since all the individual cable leads are stranded wire, slivers of stranded wire can bridge or break, causing ground connections to be marginalized. I suspect a redesign of the power supply is not necessary to solve the problem.

Pete

This is good advice, but if a loose ground is the problem, it is hard for me to see why replacing L2 with a different choke would reduce the hum; or why some of the other tests that were done would produce the results that were posted.

Stu

I'm only saying, based on my past experiences with a number of NCX-3's, 5's, 200's, and 500's, Clegg Venus, and various Swans, any speaker hum/buzzing, or similar anomaly, was found to be related to interconnection between the rig and the power supply/speaker. A simple way to test the integrity of the power supply/speaker and cable would be to derive the necessary receiver voltages from some other source (another power supply/another rig, etc.) using the classic clip lead approach directly to the NCX-3. The high voltage connection would not be necessary for this test. A separate speaker would also be necessary. The circuit design of the NCX-A power supply/speaker was around for years. While component breakdown is still a possible solution, sometimes a more obvious solution is sometimes overlooked.

[AB2EZ]

Pete

Yes... we are, I believe, in 100% agreement. Sometimes the simple solution is overlooked... Like a cable with a poorly soldered connection.

Separately, 8H sounds about right for 100mA and 350V. I wonder if the choke that is in the power supply is even the right choke (maybe replaced some time in the past with a smaller value choke)

Stu

[gerry_w1id]

I have been asked how I knew the filter choke had a loose lamination. I was referring to the LV choke and the HV choke. There was a persistent loud buzzing which I identified as a mechanical resonance of some kind. By listening closely, it was determined to come from the LV choke though I didn't know the cause. So I took the choke out and measured 2.4H and about 45 Ohms DCR. Visual inspection showed the I lamination was loose and was therefore the source of the loud buzzing. Rather than fix this choke I replaced it with a 10H potted unit rated for 150 ma continuous current. At the same time, I relocated the new choke away from the venting slots over the B+ bleeder resistors. The original choke location did not allow the HV bleeder resistors to properly cool and contributed to overheating the low B+ choke. Since the new choke had more DC resistance than the original, only one of the 330 Ohm 10W  resistors (either R5 or R6) was used. As I went through the bottom of the chassis, I also noticed some of the riveted terminal strips had gotten loose. Unfortunately, the ground lug of the terminal strip to which the ground wire going to the transceiver was soldered to was loose. At this point, I made a decision to drill out all the power supply rivets and secure the terminal strips with 6-32 hardware. As I mentioned previously, there are a lot of similarities between the NCX-3 and the Eico 753. A couple of articles have been published on improving the 753 but they do not address the fundamental problem: there is only one VFO inductor. Consequently calibration will not track for all 3 bands. You can calibrate the VFO to read accurately on either 80 and 20 but not 40 and vice versa.

[MikeKE0ZUinkcmo]

 The National P/Ns B50863(swinging choke), B50864 may be stamped on the top cover.   Don't have any idea who they bought them from but Carl (KM1H) probably knows.   if the P/Ns are on the chokes it would at least answer your questions about replacements.

[Pete, WA2CWA]

The National P/Ns B50863(swinging choke), B50864 may be stamped on the top cover.   Don't have any idea who they bought them from but Carl (KM1H) probably knows.   if the P/Ns are on the chokes it would at least answer your questions about replacements.

Actually, the swinging choke could be B508063 or B51781 or B52068 or something else.
The filter choke could be B50864 or B52067 or something else.

[KA2DZT]

I went back and re-read the details of what the OP has told us so far.  Two very important details, 1) he claims the buzz starts as soon as the set is turned on. 2)  he claims the buzz in there even with the output tube (6aq5) removed.

These two points tell me that the problem is not the power supply.  If you look close at the way the Cinch plug on the main chassis is wired.  The spk ground return and the ground pin are jumped together and then grounded to the chassis.  If the ground connection to the chassis is not 100%, what would happen?  Some ground current would travel through the spk coil to the secondary of the output xfmr, through the secondary winding to ground.  This would cause the buzz to begin as soon as the set is turned on and it would still be there with the output tube removed.

Fred

       

[MikeKE0ZUinkcmo]

Pete said;

Actually, the swinging choke could be B508063 or B51781 or B52068 or something else.   The filter choke could be B50864 or B52067 or something else.

As you know, the other numbers you mention depend upon which power supply you're looking at.   The AC-200 and AC-500 supplies were for the NCX-200 and NCX-500 transceivers respectively, and provided significantly more power than the NCX-A, shipped with the NCX-3s or NCX-5s.



Fred,

I think your hypothesis may very well be correct, and the input/output grounding points at both the radio and power supply, as well as the cable itself, should be thoroughly checked.

[KA2DZT]

>The output tube pulled results in no current through the output xfmr which should >leave no audio in the speaker.  So, if there is ripple on the B+ it still wouldn't cause >the hum to reach the speaker with the output tube removed.

Well not exactly true. B+ is on the bottom end of the audio output transformer and the two caps he mentioned are on the top[ side and would couple the energy across it. That is why removing the caps stopped the buzz.

So where is the scope? That would tell you whats up in an instant. 

What you're saying may also be correct.  The set uses solid state rectifiers, so the ripple (if any) would be present at start-up (no time delay waiting for tube rectifiers).  I also see where those caps could possible cause some energy to appear across the output xfmr with the output tube removed.  I'll have to take a closer look at the circuit around the 6aq5.  That cap between the plate and screen should be a rather small value, only used to limit higher freq audio.

Fred

[KA2DZT]

I looked at the schematic,  that .01 cap to the vox circuit would caused some ripple current to flow through the output xfmr all the time, even with the 6aq5 removed.  So, the problem may very well be poor filtering in the PS.

I guess we'll have to wait for the OP to tell us more.

Fred

[pa0ast]

How about a magnetic coupling betwee the trafo and the coil. Is the position of the coil or the trafo changed in the past ?
Anton

[Pete, WA2CWA]

Pete said;

Actually, the swinging choke could be B508063 or B51781 or B52068 or something else.   The filter choke could be B50864 or B52067 or something else.

As you know, the other numbers you mention depend upon which power supply you're looking at.   The AC-200 and AC-500 supplies were for the NCX-200 and NCX-500 transceivers respectively, and provided significantly more power than the NCX-A, shipped with the NCX-3s or NCX-5s.

Actually, the part numbers referenced in my previous post came directly from several original NCX-A manuals (tagged with different series numbers). And, if you review the NCX-A and AC-200 schematics and power supply output specs, they are virtually identical.  The NCX AC-500 uses pretty much the same parts as the other supplies except they moved both chokes (swinging and filter) to the low voltage leg, added an additional filter cap and bleeder to the high voltage side, and some other minor things. I ran a NCX-500 with an NCX-A for several years. The only difference was 700 volts on the HV side versus 1100 volts and a minor difference in the bias voltage. I even ran my NCX-5 with a Clegg Venus power supply for awhile until I found another NCX-A and it was only acquired for cabinet matching purposes.

[MikeKE0ZUinkcmo]

Pete said;

And, if you review the NCX-A and AC-200 schematics and power supply output specs, they are virtually identical.

Ooops, you're right.   The only significant difference is the 115/230 primary and switch in the AC-200 and 500.   Aside from the higher B+ the AC-500 didn't have an internal speaker either.



I also ran mostly National radios for some time, but had the associated power supplies for each. 

[AB2EZ]

N0WVA wrote:
"This radio has had the annoying buzz/hum since being bought 5 years or so ago. After putting it back on the operating bench, I am sort of picking up where I left off a few years back. Taking the power supply apart, I then realized that I did indeed replace those old filter capacitors with brand new ones, although they were computer type in series, and I put them in under the chassis. And I done a pretty good job, if I say so myself. All three supply caps were replaced ( the -45 , 280, and 700v lines). And I remember being frustrated after all that work that it didnt eliminate the problem back then, so I just run it like that ( annoying as it is on RX, hardly detectable on TX).

But I did NOT replace the three section electrolytic in the radio itself which goes to the OA2. So I did solder some good capacitors of value underneath, but it didnt change a thing."

Note:

If the old electrolytic capacitors are leaky, then putting new capacitors in parallel with them may not result in an improvement in their effectiveness as charge storage devices.
You might want to be sure that you have disconnected one end (or both ends) of each old capacitor that has been "replaced".

Stu

[WD5JKO]

If the hum comes on immediately with or without tubes present, or before the tubes warm up, then there is either magnetic coupling, or there is high ripple between the 220v DC screen supply (6AQ5) and the 280V source for the screen supply. Removing cap C97 eliminating the buzz confirms this. The buzz must therefore be high amplitude and narrow pulse width, and likely at a 120 hz rate. Too bad the scope is crapped out as this could confirm the problem.

My bet is the diodes have a different reverse recovery time then the ancient originals and they are shorting the transformer secondary for a tiny amount of time each half cycle, perhaps 100 us every zero crossing. When the slow to turn off rectifiers in the bridge turn off, the transformer generates a back emf pulse so narrow in pulse width, that it goes right through L2 unimpeded.

So as a test, try taking a series R-C snubber from the transformer secondary center tap to ground. Try something like a 3K 2 watt (not critical) and a .01 uf 1000 V capacitor...See if it has any effect. If so, perhaps try different values, or another R-C across the entire transformer secondary. The Cap would need to be rated higher there, at least 2 kv.

I cover the diode hash / buzz issue in more detail here in yesterday's post:
http://amfone.net/Amforum/index.php?topic=34793.0

Jim
WD5JKO

[AB2EZ]

So its an annoying buzz, and present with the 6aq5 pulled. The buzz starts as soon as the rig is turned on. It is being picked up through two capacitors c37 and c38. Lifting them makes it quiet. But then theres no VOX. Im about at the end of my knowledge here.

As has been previously posted, these capacitors are C97 and C98. They are both supposed to be 0.01uF capacitors. Therefore, the impedance of each should be 132k ohms at 120Hz and 265k ohms at 60Hz.

If the audio output tube is disconnected, either of these capacitors could (as previously posted) conduct current from the 280V B+ through the primary of the audio output transformer to ground (if the capacitor is connected between the top of the primary and ground)

However, given the high impedance of either capacitor at 60Hz and 120Hz, any hum between the 280V B+ line and ground should produce only a small current through the primary of the audio output transformer. The hum should be greatly attenuated because the input impedance of the audio output transformer (with the speaker attached to the secondary of the transformer) should be much less than the impedance of the capacitor.

Nevertheless, N0WVA reports that the hum goes away when both of these capacitors are disconnected at one end (with or without the audio output tube in place).

I suggest the following experiment:

Disconnect both ends of the audio output transformer's primary (i.e. disconnect everything connected to the top side of the primary, and disconnect the 280V B+ from the bottom of the primary). Place one of the 0.01uF capacitors across the primary.

This will result in a 133k ohm load across the primary (at 120Hz), and a corresponding much lower impedance looking into the secondary (where the speaker is connected). I.e. the impedance looking into the secondary of the audio transformer at 120Hz will be 133k ohms divided by the square of the transformer's turns ratio (ignoring the transformer's magnetizing inductance).

If there is a 60Hz voltage from the filament supply being coupled directly into the speaker leads (e.g. due to some problem in the cabling between the power supply/speaker and the transceiver)... and if that is the cause of the hum that is being heard, then the hum should still be present with the primary of the audio output transformer having only a .01uF capacitor across it (nothing else connected to either end). It should disappear (or be greatly reduced) when the .01uF capacitor across the primary is removed.

If there is no hum when the capacitor is across the primary of the audio output transformer (with nothing else connected to the primary) then try placing a short circuit across the primary of the audio output transformer (with nothing else connected to the primary). If there is still no hum, then it is unlikely that the hum is associated with 60Hz (e.g. filament voltage/current) being coupled from the filament supply into the speaker leads inside the cable between the power supply/speaker and the transceiver.

Given N0WVA's reported result that the hum goes away when L2 is replaced with a 1.2k ohm resistor, I would (again) be surprised if the hum is being introduced in the cable between the power supply/speaker and the transceiver.

 Stu