DX-60A output much lower on 20 meters

[ka1tdq]

I bought a DX-60A off eBay a little while back. I did the typical stuff to it like adding a fuse, adding a 3-prong cord, replaced the electrolytics and rectifier diodes.

It loads fine on 75 and 40 meters. I get a little over 50 watts output with 120ma plate current and 2.5ma of grid drive.

When I go to 20 meters output drops dramatically. With still 2.5ma of grid drive still, output drops to a little over 10 watts. To get more, I really need to increase loading and get plate current up to 200ma (just to test; I don't keep it there).

I'm running on crystals and using a dummy load.

Any idea what would cause this?

Jon

[Pete, WA2CWA]

On 20 meters, are you using a 80 or 40 meter crystal?
Have you tried 15 and 10 meters? Results?
I would check L2 and L3 connections to their respective switch wafer.
Check voltages in the 20 meter position.

[ka1tdq]

I'm using 40 meter crystals (except when on 75 meters, of course).

When I first bought it, as I remember, 15 and 10 meters had zero output.  I've only recently tried it on 20 meters and the results are as posted above.

I pulled the covers and checked a lot of the wiring.  Everything looks correct, solder joints and all.  The wafer switches are making contact on the pins as I can see them move a little when the wafer comes around.  I gave them all a little "push" for good measure though.  

The wiring to L2 & L3 is correct.  

I did find two resistors that looked heat damaged and they're out of tolerance:

R4 - screen dropping resistor on the oscillator tube. Should be 15k and is now 19.5k.

R6 - grid leak resistor for the buffer tube. Should be 18k and is now 22.5k.

I should mention as well that the "Drive Tune" knob is fully unmeshed to get the maximum drive current on 40 and 80 meters.  The "Drive Level" control is almost at maximum for the necessary drive.  

Jon

[AB2EZ]

Jon

The DX60 employs a combination of fixed negative biasing and self biasing of the grid of the 6146.

Check the voltage on the output of the fixed negative bias supply... and the resistances of the associated voltage divider resistors.

If the fixed negative bias is too large (too negative), then the grid drive will be insufficient to produce the correct grid current... and also insufficient to cause enough plate current to flow.

Stu

[ka1tdq]

Hey Stu,

I'll check that in a little bit, but I found something really interesting on L2 (the grid tank inductor).

I think either the builder or Heathkit made a mistake on this wafer switch.  Normally, in the 80 meter position as shown on Heathkit's schematic, none of the coil is shorted.  In mine, the 80 meter position shorts out the large portion of the coil.  

It continues that way all the way down the band switch.  So in the 10 meter position, the entire coil is shorted out, instead of leaving the last small segment of the coil in the circuit.  Its like the copper portion of the wafer switch is one position too long.  If it were possible, I'd need to get a knife in there and lop off about an 1/8" of copper so that the switch would short portions of the coil correctly.

Alternatively, I could add a fixed amount of capacitance to the variable capacitor to bring the whole thing into resonance since I'm short on inductance for each band (no inductance on 10 meters).  Or, add a fixed amount of inductance in series with the inductor.

...or, I can sue Heathkit for every dime they've got!  

Jon

  

[Pete, WA2CWA]

There's no way to short the entire L2 coil. The hot end of L2, closest to the 10 meter tap, goes directly to the plate of V2(6CL6) and C9 and C11. It could be that maybe the wafer is mounted backwards or the inner rotating contact was rotated 180 degrees before assembling the wafer on to the main switch shaft. Or maybe someone replaced the original wafer with an incorrect one.

[ka1tdq]

I made a 90 second YouTube video to try to show what I'm talking about:

https://www.youtube.com/watch?v=tUL5oY3qmqE

It looks like the original Heathkit wafer.  I checked the keying, and it aligns with the other wafer on the band switch.  It is literally just a 1/64" too long and barely makes contact with the next terminal. 

Jon

[AB2EZ]

Jon

I would not do any surgery on the switch wafer... until you check out other things more thoroughly.

You mentioned that on 80m and 40m the maximum output tube grid current occurs with the driver tube's output tuned circuit capacitor in the fully unmeshed position.

If there were too little inductance in the driver tube's output tuned circuit, then you would need more capacitance (not less) to achieve resonance.

Stu

[ka1tdq]

The bias unloaded voltage is -152 vdc.  I replaced the original two electrolytics with one 100uf/450vdc.  I didn't put a bleeder resistor across it, and I'm just using the neon bulb and original 470k resistor to bleed off voltage.  It goes away in a few seconds.

Heathkit put a resistor in series after the diode and another one in series with the next parallel electrolytic.  I'm guessing they did that to minimize surge current for the old style diodes.  I replaced the diodes with modern, over-rated ones so I got rid of the resistors.

As for the coil...

I wanted to see if that was shorting or not, without doing major surgery.  I carefully clipped the lead going to the 40 meter tap on the coil so that I could ohm check between the 80 meter side of the coil and the 40 meter terminal.

Lo-and-behold, it's not shorted.  But, to the contrary, when I switch the band switch to 40 meters (to short the coil taps) it doesn't short either! The leading edge of that copper strip doesn't make contact.  It's all grimy, dirty and black.  When I move it to the 20 meter position, THEN the 80 meter coil shorts out.  The leading edge of the wafer isn't working, but the rest of it is.  (If you follow...) I even nudged in the 40 meter tap on the wafer switch to see if that would help, but it didn't.  

So, that's why I was getting fully-unmeshed.  The coil was actually not shorting out and giving me too much inductance.  

Result:

As a test, I took an alligator jumper and shorted out the coil as it should be in normal operation for 20 meters.  It worked a lot better!  At 120ma on the plate and 2.5ma on the grid, I was getting 48 watts.  However, the grid capacitor was still fully unmeshed.  The grid drive control however was set a lot lower at only 3 instead of 10 as before. 

I'm guessing I'm going to have to do some major repair cleaning on that wafer switch or replace it altogether (if they even exist today).

Jon

[KA2DZT]

Good,  looks like you're making progress.  I watched the video but couldn't see much of what the problem could be.

Fred

[WA2SQQ]

Hey Stu,

I'll check that in a little bit, but I found something really interesting on L2 (the grid tank inductor).

I think either the builder or Heathkit made a mistake on this wafer switch.  Normally, in the 80 meter position as shown on Heathkit's schematic, none of the coil is shorted.  In mine, the 80 meter position shorts out the large portion of the coil.  

It continues that way all the way down the band switch.  So in the 10 meter position, the entire coil is shorted out, instead of leaving the last small segment of the coil in the circuit.  Its like the copper portion of the wafer switch is one position too long.  If it were possible, I'd need to get a knife in there and lop off about an 1/8" of copper so that the switch would short portions of the coil correctly.

Alternatively, I could add a fixed amount of capacitance to the variable capacitor to bring the whole thing into resonance since I'm short on inductance for each band (no inductance on 10 meters).  Or, add a fixed amount of inductance in series with the inductor.

...or, I can sue Heathkit for every dime they've got!  

Jon

  

So back in 1971 I built my DX-60B. I could not get the plate current to dip no mater what I tried. Gave it to my Elmer, W2MJD who found the problem. Heath had assembled one of the wafer switches incorrectly, it was 180 degrees out. The error was not mine! After correcting that problem it worked great. Since, I've heard of two similar cases. The radio was probably built and never worked correctly.

[ka1tdq]

I originally bought this rig to operate 20 meters CW. I was disappointed when it wouldn't work on that band, so I've been on 40 meters CW since then.  But, now that I've found out that the wafer switch was causing the problem, I'm just going to solder in a jumper to make 20 meters work. I could try to repair the wafer, but I don't care about the other bands with this transmitter.  

I have class E rigs for 75 and 40, and a Chinese 10 meter AM transceiver for when sunspots come back.  I'll get a couple crystals so that I can operate AM too on 20 meters using the DX-60.

I've heard stories of the function wafer switch going bad on these too. It switches AC power and HV. There's a YouTube video of someone who removed it and put switches all over the faceplate to replicate functionality. I don't want to destroy the look of the transmitter, so I'll cross that bridge if/when I get there.

Jon

[AB2EZ]

Jon

Keep in mind that... with a 40 meter crystal selected... the plate current in the 6CL6 should be a periodic sequence of narrow pulses at a repetion rate equal the frequency of the crystal. This will be independent of the band you select.

This periodic sequence of narrow plate current pulses is rich in harmonics. If the 6CL6 tank circuit is resonant on the fundamental frequency, then the voltage applied to the grid of the 6146 will be a sine wave at the frequency of the crystal + a negative D.C. bias.

If the 6CL6 tank circuit is resonant at the 2nd harmonic (i.e. 14.xxx MHz) then the voltage applied to the grid of the 6146 will be a sine wave at twice the crystal frequency + a negative D.C. bias.

Either way, if the sine wave component has a large enough amplitude... the 6146 will draw grid current

However, even if the 6146 is drawing 2.5mA of grid current... that doesn't tell you if the 6CL6 tank circuit is tuned to (i.e. resonant at) the crystal frequency or a harmonic of the crystal frequency.

If the 6CL6 tank circuit is properly tuned to the 2nd harmonic of the crystal frequency, the associated tuning capacitor should not be completely unmeshed. If it is completely unmeshed, the 6CL6 tank circuit is probably tuned to a higher harmonic.

Stu

[ka1tdq]

Hey Stu,

I read that twice, but I got it. I'll play around with it to make sure the tank is tuned properly to 14 MHz. I placed a Digikey order for those resistors and to replace the old tubular capacitors in the audio section. Once I get all that, I'll experiment more.

Jon

[WA2SQQ]

I've heard stories of the function wafer switch going bad on these too. It switches AC power and HV.
Jon

So many years ago a good friend's first radio was a Geloso, made in Italy. The "Transmit" switch was a wafer switch which also switched HV. The design nightmare was that when you took it out of transmit it used a separate set of contacts to discharge the HV filter caps. I did not take too long before the contacts were vaporized!

[ka1tdq]

I replaced those two resistors and the caps in the audio chain. I also hard-wired the grid inductor to 20 meters.  I'm now getting a peak in the grid tuning around "3", so the variable capacitor is no longer fully unmeshed.  The required drive level is lower as well.  The knob is set to about "1.5" for 2.5 mA grid current key-down.

Now my problem is that I'm not getting a dip in the plate.  No matter where I rotate the plate tuning capacitor, it remains at 200 mA.  I've checked everything in the plate circuit and all seems well.  I even went directly from the antenna jack to a dummy load, and it's still the same story.  

Jon

[AB2EZ]

Jon

Is there another position of the drive tuning capacitor (more capacitance) that produces a peak in the grid current. I.e. Is the RF component of grid voltage a sine wave at the 2nd harmonic of the crystal frequency (and not a higher harmonic)

Stu

[ka1tdq]

Ah, thanks!

Actually, there was only one peak in grid current for the full rotation of the grid tuning cap. But, that led me back to the wafer switch contacts. Apparently, that copper slider was making contact with the 15 meter tap (in the 20 meter position). I was peaking on the 15 meter band. I cut the 15 meter tap going to the wafer switch and I was able to get a 20 meter peak.  It resonated with a little more capacitance.

Now everything works fine. I'm getting 55 watts output with 120 mA plate current and 2.5 mA grid current. 

Those wafer switches are an absolute piece-o-crap.  In the not too distant future I'm going to replace the function wafer switch too. I'll put the power switch and AM/CW switch on the back of the rig, and use a switch on the front (using the existing function switch hole) to switch between "tune," "standby" and "operate."  This should minimize the impact to the transmitter cosmetically.

After that, this rig should be up to snuff.

Jon

[AB2EZ]

Jon

Congratulations!

Stu

[wa1knx]

Jon
       Congrats on getting the dx60 working!  And from reading this the other day, I have now left my dx60 function
switch to AM and I am switching the AC to the rig.  I've been using mine daily and noticing how sloppy the function
switch. is.  No more.  And the band switch is bound up, I have to use vice grips to change bands.  I've lubed it etc. can't
see what the problem is with my bad eyes. so I leave it on 40.

       I'll post the audio mods (yet another take) another night.  Hope I hear you working someone on 20, we are
prolly to close for local contact.  lots of activity on 7293 AM out here in  AZ

73 deano wa1knx

[ka1tdq]

Hey Dean!

We might be able to do 20 meters between here and Tucson. I hooked up my DX-60 to a DDS VFO so frequency mobility won't be a problem.

The last time we talked, I was using three 3-500's on 40 meters. I have a lot less power now with a 50 watt class E rig. The good thing about it is that the shack doesn't turn into a sauna, especially in the summer.

From what it looks like, my DX-60 never worked right with that band switch. Maybe that's why it got little use and everything else looks good. It's only a matter of time before my function switch goes too, so I'm going to replace now.

The benefit of replacing the function switch is that, when going from "off" to "stdby," you don't turn on the HV briefly while traveling through the switch rotation in the "tune" function. So, the filaments will have time to heat up before any HV is applied.

Also, 120 vac is switched on that tiny little wafer along with the HV. It's such a small amount of copper for important stuff. The switches I have ordered are beefy and I'll sleep better.

Jon

[wa1knx]

Jon,
     I remember that amp, I know I have pics of your 3-500z amp on one of my 5 laptops I rotate through (long story)
you sent them to me back then.


     And speaking of sweltering, we lost *power* this afternoon.  The shack soared up HOT and quick.  I had power
to my Icom via a 13.8v solar regulator I finished a cpl weeks ago.  I run it directly off panels, ie no batts.  The
class E is prolly the way to go.  ice cold on hot days. Curious how the dx60 switch replacement goes where did you find them?

dean (regulator pic if it comes through)

[ka1tdq]

That's something to run directly off the solar and not use batteries as a buffer. I guess you can here though since the sun is just freakin' screaming every day!

There's nothing special about these switches.  The way I'm doing it, you'd need:

3PDT (for the function part)

1P (on/on)

1P (off/on)

I've attached a hand sketch to show how it's all wired up.

Jon

[kb3rdt]

Here a DX-60 I just redone the power supply and modulator.

[DMOD]

Here are some DX-60 enhancements I made to mine:

http://amfone.net/Amforum/index.php?topic=40809.0


Phil - AC0OB