Central Electronics 20A PTT

[sndtubes]

I am getting a Central Electronics 20A that I plan to use with my DDS VFO.   The 20A apparently uses VOX only.  Is anyone aware of a simple PTT conversion.

Thanks 73
Mike WB0SND

[WD5JKO]

The 20A has vox capability, and also has a turn the knob capability to transmit. You can add a PTT relay to that signal if you like. Being a tube guy you might find it interesting that a 12AT7 is used to pull in that big relay coil. Oh, I mentioned a 12AT7....I can see WB2WIK's blood boil! (see his recent QRZ post about AM).

Jim
Wd5JKO

[WD5JKO]

Mike,

   I have a number of 20A's, and over the years I've tried a number of things.

Here is a thread you might find interesting:
http://amfone.net/Amforum/index.php?topic=27305.0

With your DDS VFO, remember you need somewhere around 8v peak drive from the VFO. The original VFO was a modified BC-458. With 5-5.6 Mhz VFO you get 80m and 20m when beat against the 9 Mhz rock in the 20A. For 160m the VFO is doubled, and for 40M the VFO is tripled. Take this into account with your VFO.

The 9 mhz crystal in the 20A is pretty old, and they all drift...some drift up to 600 hz over the first hour!! Getting a new crystal, and putting it in an oven will help a lot. Sometimes you can find a NOS old style CB crystal (I forget but channel 3 or 4 transmit is close) that when divided by 3 is real close to 9 Mhz. These are 3rd overtone crystals.

The audio phase shifter can be modified quite easily to be good up to 4+ Khz. I talk about it in the link above.

Both Larry NE1S and myself have went into 20A QRO projects. Larry gets over 100W PEP from his on the lower bands using a pair of 7984's in the final. I have used a pair of 7591's to get 80W PEP, and a single EL-34 (E34L) for 40W PEP.

There is an easy way on AM to bypass the audio phase shifter. Makes real clean AM that way.

Have fun with your 20A!

Jim
Wd5JKO

[sndtubes]

Jim

Thanks so much for your reply.  Sorry for the delayed answer.  I have had a really hectic week and it's going to get worse..

First, I received the 20A today.  I am not sure what revision it is.  There is no letter in the serial number, so I'm afraid it's early.  I can only hope that the previous owners installed the updates.

As advertised, it does seem to work.  I plugged in a crystal for 40 and it works fine on CW.  I have to rig up a microphone for it still....  It may sound like crap on SSB/AM, hi hi.

I will probably recap and align it and install the updates before trying to use it on the air.  I have a beautiful 600L that I plan on pairing with it.  I don't think I'll go for the QRO mods.

I have a wonderful DDS VFO that I plan on using with it.  It has the buffer amplifer in it which drives tube rigs great.  My Viking 2 loves it.

It hadn't occurred to me that the 9 MHz crystal would also drift.  I'm thinking that these DDS vfos are so stable and relatively cheap, so why not have a second DDS VFO running on 9 MHz in place of the crystal.  I have to think that the DDS VFO's are going to be as stable or more stable than any crystal in an oven.  Just a thought.....

When I get a chance, I will join the Central Electronics reflector.  I also have a nice 200V which is in good operating condition except for the stupid PTO.  It, of course, is rough and sticky and needs to be rebuilt.  I'm a little afraid of that job. 

Thanks again for the links and hints.

73
Mike WB0SND

[WD5JKO]

mike,

     A ham band crystal is good for CW only. For phone you must beat with 9 MHz. One exception is to use no crystal, and tune the RF stages to 9 MHz. Try band 40 or 20 and see if you can drive 9 MHz through. This way you can null carrier, use drive control, and then add some audio.
     The 9 MHz crystal drift might not be so bad, and should be low priority until other things are taken care of. The B+ runs high, and the rig really cooks. I'd consider SS the rectifier, then use the 5v winding to buck the primary. Then switch to choke input filtering. Any supply ripple will be seen in the nulled carrier. Sometimes SS rectifiers cause back emf spikes to come through causing a nasty buzz in the nulled carrier. Those spikes must be filtered out.

The 6U8 runs pretty hot. A good uprated sub is the 6MU8. This will drop in with no changes needed. The pentode section has a higher PD rating. The pentode is the 9 MHz oscillator.

Jim
Wd5JKO
 

[sndtubes]

A bit of an update.  First, I had a blast testing the thing out.  I thought I'd act silly and call CQ on 40 CW.  Someone answered me.  TWICE.  Had 2 nice QSO's for over a half hour with 579 to 599 reports with this thing barefoot.  CW note sounded good.  Plus I worked W1AW/7 in which there was a small pile up.  Got through with only 3 or 4 tries.  So, it's working great on CW.  I used the DDS VFO for 2 of the QSO's. 

I tried injecting the suggested frequencies in the manual into the crystal socket.  The CW note sounded funky and there were many harmonics.  However, I was able to get a carrier null.  Maybe I should be injecting the signal at the VFO socket on the back instead.  I will try that tonight. 

I should have done my homework before buying this beast.  I didn't spend much on it, though.  Apparently I bought a fairly early unit (S/N 4436 no letter).  It does have the 6U8 in the osc. and it looks WHIPPED!  I will try the 6MU8's.  I have lots! The transmitter looks to be factory wired and also is untouched.  Lots of old electrolytics and paper caps just waiting to fail.  I have no idea of what if any updates have been done.  Does anyone know serial numbers and when the updates were applied?  I didn't really want a big project.  I'd love to find one in "ready to go" condition.

Oh, I also signed up for the CE reflector

73 again, and thanks

Mike WB0SND

[WD5JKO]

Mike,

   Bravo on the success you have had thus far. Hopefully that will motivate you to get it all working. My 1st 20A was in much worse shape, and it had peculiar problems of an intermittent nature. It went back on the shelf a few times, but it was always calling me, "fix me, fix me". :-) These rigs can be kind of addictive.

  If you have a 600L, then forgoing the QRO mods makes sense. The 6AG7's can be a very linear tube. In the 20A though that will take some work. One caution is that the old tube sockets can have dirty contacts. One consequence is the bias rectifier tube (6AL5) loses filament contact. This creates the Chernobyl effect as the 6AG7's and the power supply have a big fight to survive. Word to the wise, put a 1N4007 across the plate and cathode of that 6AL5 bias rectifier. Then leave the tube in there.

  I had a list of modifications from CE, but right now I cannot find them. There was a major mod in the output tank circuit, which as I recall effected the upper bands only.

   The text below is from an email I sent to Larry NE1S and George W1LSB. Those guys are both 20A fellows who got the 20A bug like I did. After this I was able to get AM THD to -40db or 1% at 75% modulation. Also two tone IMD was -40 db as well for the 3rd order product. This is dam good for a rig like this.

quote from email:
I did a few things to my 7591A tube based 20A today, and I kind of lucked into a big gain today..

   So, I was figuring that the presence of the audio phase shifter when running AM, only screws up the frequency response. Running square waves, or triangle through the phase shifter gets all screwed up..only a sine wave between 200-3000 cps gets through unscathed. So why do we need this when running AM when only one audio channel is used?

  First, here is phase 1 of what I did several years back:

1.) At the audio driver 6U8 (I use 6MU8), make a series R-C consisting of 470K and .01uf, and place between pins 1-9. Then remove C4 in the same area. This NFB really cleans up this stage without sacrificing too much gain.

2.) To open up the top end of the phase shifter to about 5 Khz, make another series R-C of 390K and 180pf, and place under PS-1 socket at pins 2-9. Then bridge R15 at the balance pot with a 1.8K. Then re-do the phase shifter setup the usual way. You will see as ripple starts to emerge at about 3 Kc, it then diminishes going up and then at 5 Kc it starts to lose the shift like it did stock at about 2500 cps.

Then last year I lost an audio modulation transformer. After putting in a new one, a used a 22K 1W load resistor for the 12AT7  drivers going to LV B+,and capacity coupled the audio paraphase style with 1uf polypropylene coupling capacitors. No more DC current in the modulation transformer primary.
 

  So here is what I did next:

1.) There is an unused set of contacts on the mode switch. The common is all the way against the chassis side, but with a small iron it can be soldered to. Connect a twisted pair of wires to the mode switch (shorts on AM only), and route to wires to under PS-1 socket, and connect to pins 3-4. This effectively bypasses PS-1 when on AM!

2.) Install a 2.2 meg resistor at audio preamp tube (12at7a) between pins 1-6. This adds about 6db NFB to the preamp stage. More importantly this raises the overload point when driving from line level audio. I never had enough to use the other audio jack, and I always had to watch the incoming level for overload signs. No longer a problem.

3.) Magic eye tube. I replaced mine years ago with a 1629 after making the switch to an Octal socket, and add a 1/2 wave voltage doubler for the filament + a small 12V fan to cool the PA tubes. I was out of eye tubes except for some Russian 6E5C's I got several years ago on Ebay. The 6E5C has a 6.3v filament along with an Octal base. The pinout is almost the same as a 1629..Had to move a wire and resistor at pin 4 to pin 6. Works!

So with these changes, using a crystal D-104 Mic, the audio level for full output is about 2 O'clock on the knob...So gain is fine.
 

More important, the AM mode sine wave frequency response at -3db (.707 volts) is 40 hz to 10 kc after the NFB and getting rid of PS-1. Playing with triangle and square waves is much improved..Not perfect but damn good midband.

Running SSB it is as before except the overload issue with the preamp is gone.

Then the big discovery...On AM, the tendency to overmodulate into DSB is greatly reduced. I took out my diode-resistor limiter..not needed. I have no idea why. Dumb luck I guess.
End of quote from email.

Jim
Wd5JKO

[sndtubes]

Hi Jim,

Thanks again for all of the info.  I did a little more testing last night, but not much time to play.  I am getting some audio through and a carrier null, so the unit is basically working.  There isn't much power output when using mixing with the 9 MHz osc. so, I'm thinking there are weak tube and alignment issues.  I will start with replacing the tubes and the paper/electrolytic caps and then an alignment.

Are there supposed to be one or two 6U8's in there?  Mine has 1.  I hope that my oscillator isn't the dreaded early 12AT7.  The serial number is 43xx.  Is that are real early one? 

I would surely be interested in any mods you have.  I have the factory changes mods, already.  I hope most or all of them have been done.  I will surely bypass the PS-1 for AM mode.  Sounds pretty easy to do.  I don't need to change the eye tube.  The one in it is bright and I have bunches of them in stock, anyway. 

I have discovered that my DDS VFO only goes to 30 MHZ.  The 20A wants 37 MHz for 10 meters, so I will have to address that later.  I'll probably build a couple more VFO's when I can as I really like them.  No chirp.  No drift.  Nice. 

Thanks again,
Mike

[WD5JKO]

Mike,

   The real early 20a's had NO 6U8. Since you have a 6U8 you should be all set.

  A peculiar problem with the 20A is that the frequency pulls a bit when the drive control is maxed out. By pull, the 9 Mhz crystal shifts a bit. I think this coincides with the 6AG7 output tubes being driven into grid current (Ab2). Since the bias supply is high impedance, drawing grid current has all kinds of unintended consequences. It is very easy to add an adjustable zener diode (TL-431) to set the 6AG7 idle current. Of course that bias goes to -100 when in standby.

  I wanted to add to what I stated earlier about the modulator 12AT7 using Para-feed coupling to the modulation transformers. The balance control varies one of the 12AT7 cathode resistors. This changes the DC operating point, and when going to Para-feed the plate voltage moves around a lot. The circuit change I post in the image below fixes that. With this change the DC parameters are constant while the AC gain can vary to get a nice balance. Tony Vitale W2EWL in his "Quick and Easy SSB" article back in the late 1950's also used Para-feed to the modulation transformers.

  I also attach harmonic distortion and IMD plots from this rig running 12W AM at 75% modulation. It would be a worthwhile adventure to take a stock 20A and to get very clean AM with a 2 watt carrier. This drive level is about right for driving the grid of the 813 in your 600L amplifier.

Jim
Wd5JKO

[sndtubes]

I have just acquired another 20A.  It's a Nick Tusa restored one.  I am just not getting the power output I get on the other unrestored unit.  Furthermore, neither of them put out much when using the DFS vfo on 40 meters.  Other bands are much better.  The DDS VFO has the buffer amplifier for driving boat anchors.  When using the DDS VFO at 16 mhz for 40 the 600L only is driven to 100 watts output. It's even worse with the 418 VFO.   Is this normal for the 20A or does it sound like it needs alignment or?

Mike

[WD5JKO]

Mike,

40M is the worst band for a 20A due to the high VFO injection frequency. Since you have a DDS, try injecting at the other end...instead of 16-9=7; try 9-2=7... Either way, you need 8-10v peak to get full output.

Could plug in a 40m ham band rock to see what the RF output would be...A CW only option.

Jim
Wd5JKO

[w1vtp]

Larry, NE1S has done some nice work on his 20A including some QRO work.  Bet he has addressed that PTT issue.

Al

[sndtubes]

Thanks for the info on the PTT issue.  Actually I've worked that out.  I should have posted that.  For CW, I key the blocking bias while the mode switch is in STANDBY.  For PTT phone, the PTT switch is connected to the KEY JACK on the front panel.  With the MODE switch in MANUAL, you have PTT.  Works like a charm. 

As for the power output, I still have a problem on 40 meters.  Jim, I tried injecting 2 MHz which works, but with only a small improvement in output.  Still not enough to drive the 600L fully.  Also, the manual says you shouldn't use 2 MHz injection for 40 meters due to image problems.  Injecting 7 MHz for CW works with further improvement in output, but still not full output.  My old unit will give me full output on 40 CW which makes me think the restored unit may need some alignment touch up....maybe.  Every time I buy something that is supposed to be "plug and play", ....it isn't.   Maybe I just can't expect full power on 40.  I haven't measured the output of the DDS VFO but with the buffer amp, it's plenty for the Viking 2, etc.  The other issue with the DDS VFO I have is 10 meters.  My VFO only goes to 30 MHz.  We need 39 MHz for 10 meters...

BTW, Jim, if you are still using 7591's, I suggest you change the sockets and use 6GM5's instead.  They take a standard 9 pin miniature base but are identical in every other way to 7591's or 7868's.  The advantage is price and availability. 

As soon as I get settled, I plan on performing the AM mods you suggested.   I'm a little confused.   Do I need to do the "Phase 1" mods and then the PS1 bypass along with resistor change for negative FB.  Or just the latter?

[WD5JKO]

As for the power output, I still have a problem on 40 meters.  Jim, I tried injecting 2 MHz which works, but with only a small improvement in output.  Still not enough to drive the 600L fully.  Also, the manual says you shouldn't use 2 MHz injection for 40 meters due to image problems.  Injecting 7 MHz for CW works with further improvement in output, but still not full output.  My old unit will give me full output on 40 CW which makes me think the restored unit may need some alignment touch up....maybe.  Every time I buy something that is supposed to be "plug and play", ....it isn't.   Maybe I just can't expect full power on 40.  I haven't measured the output of the DDS VFO but with the buffer amp, it's plenty for the Viking 2, etc.  The other issue with the DDS VFO I have is 10 meters.  My VFO only goes to 30 MHz.  We need 39 MHz for 10 meters...

BTW, Jim, if you are still using 7591's, I suggest you change the sockets and use 6GM5's instead.  They take a standard 9 pin miniature base but are identical in every other way to 7591's or 7868's.  The advantage is price and availability.  

As soon as I get settled, I plan on performing the AM mods you suggested.   I'm a little confused.   Do I need to do the "Phase 1" mods and then the PS1 bypass along with resistor change for negative FB.  Or just the latter?

Mike,

   OK, the output can vary a lot depending on the parts used in the mixer plate circuit. If you look at the plate tank and tuning condenser, the circuit is pretty strange. First there is a capacitor in series with the mixer tuning air variable. This is to keep from shorting out the mixer B+ in case the rotor-stator clearance gets too close. The tank 'Q' is high, so that series capacitor has to carry some RF current. Then there is another capacitor at the mixer B+ to the cap that goes to ground. This cap also must pass the RF current of the mixer L-C when tuned to resonance. Then to lower the 'Q' there are fixed resistors that vary from band to band that are placed across a portion of the multiband mixer coil.

   Part of the reason they lower the 'Q' is to add to the stability...consider this is a multi-stage RF strip with a lot of gain, and no neutralization.

   So here is the problem. The plate bypass capacitors used (B+ to gnd) on the mixer plate can be either an old fashioned silver mica, or a disk ceramic. I think they are .005 uf or very close to that. The 20A's with the Mica bypass have more gain, so as a consequence the shunt resistor(s) across the coil have a lower value. The rigs with the ceramic disk have less gain, and therefore the shunt resistor(s) have a higher value. Either way works, BUT, if some recapper dude came by and changed the Mica with a ceramic disk, and did not adjust the shunt resistor(s) upward, then the gain would drop. The most marginal band between 160-20m is 40m, so you would see it there first.

   Look see what resistor you have for 40m shunting the coil. If you clip it out, the gain should soar, perhaps to the point of instability. See attached file where I outline the RF current path in yellow dots, and show critical components in red outline. The capacitor ESR is critical. With my QRO projects I had to learn this the hard way to get more drive. On one rig I used the silver mica with no shunt resistor. To lower the gain, I put 2 ohms 1/2W CC in series with that cap. The gain dropped too much, so I paralleled it with another (1 ohm) and that was about right. I was seeking maximum drive at maximum linearity as long as it was stable. That Mixer plate is no doubt the crux of your problem.

   The 7591 tube family is interesting. These have very low Cgp (.25 pf for the 7591) and two screen grid pins on the socket. These are a very good RF tube. Problem though is the modern day versions don't compare at all. The JJ type is a true pentode and I think the EH is also a pentode. The old vintage versions are beam power tubes. Take one apart, there is no suppressor grid. Instead they have beam forming plates like a 6V6. Funny that all the old books call then power pentodes. So the JJ version is at least 20% down on power from the vintage ones, and they don't like AB2 to get more power. The EH's might be 10% down on power with full Ab1 drive. These will run into Ab2 to get more power, but the onset of grid current causes a slope change in the trapezoid. I posted a trapezoid earlier in this thread, and that thing has perfectly straight walls. Those are from 50+ year old Fisher branded 7591's with thousands of hours run time. I run them with 560V on the plate and regulated 375 on the screen grids. The Grid1 bias is about -22v with a tl-431 adjustable shunt zener. The bias points are VERY STABLE.

Pat mentioned Larry's 20A stuff. This guy is amazing. He put a RF gain stage following the mixer (removed the cathode follower), and used a pair of 7984's as finals. His works 160-10m! On the low bands he gets ~ 100W PEP RF out. He completely rebuilt the RF output cage with HD parts.

As to the 20A mods I suggested, consider them all stand alone changes:

9 Mhz oscillator oven
12AT7 NFB
6U8 to 6MU8 switch
6U8/6MU8 triode NFB
Audio phase shifter mods to get ~ 100hz to 5khz response with good SB rejection
Modulation transformer Para-Feed, and balance ckt mods to keep each tube at same bias point
Bypass phase shifter when on AM only

Good Luck,
Jim
Wd5JKO