The AM Forum
March 28, 2024, 01:08:17 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1] 2   Go Down
  Print  
Author Topic: A Not Noticeably Scratchy Apache  (Read 15488 times)
0 Members and 1 Guest are viewing this topic.
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« on: March 04, 2007, 10:59:02 PM »

OK you guessed it another project - my first Apache TX-1.

I will need some help as I go. It is in pretty good shape but it is heavily modified in the audio section.

This one came from Peter N2IDU along with some notes and a manual. GREG N1EXI is also mentioned in the literature. What Peter was attempting is a classic shoehorn with a larger transformer in place of the original mod tranny. He started the Press Exchange mods which calls for KT-88's (6550's) in place of the EL-34's (6CA7's) and basically de-hellifying the audio chain. Something blew up and the project was abandoned.

The first three hours were spent basically figuring out what was done, what was half done and what blew up and trying to re-wire things.

I pulled the transformer which was simply resting upside down in the middle space. The mod tubes and the finals were missing, but he put in a nicer pair of ceramic sockets where the mod tubes would go.

MIlitary Transformer:

Looks to be about 100W
Primary 7500 CT
Secondary 3000 Ohms
Tertiary 30K Ohms
Feedback 600 Ohms

The original Iron is a 50 Watt
 Primary 11K Ohm CT
Secondary 3000 Ohms

The larger transformer may work. I will have to do some calculations and look at the tube charts. Mounting the transformer right side up seems feasible. See the picture. The pair of 807's is for size reference (they would actually just fit if you replaced the sockets and slid them just a tad). That would be a DX-100 setup!
I decided to remove ALL of the parts from the audio section except for the sockets.

My first task was to clean it up, get the mechanicals and roller functional. Make the lamps come on.  Get all of the high voltage going and check the fils. All repaired now.

The next task is to get the oscillator and RF running. There is no RF at the moment!

Mike WU2D
 

 Tongue Tongue


* ApacheTight2.jpg (140.92 KB, 1332x1113 - viewed 488 times.)

* ApacheSqueeze.jpg (262.66 KB, 1656x1461 - viewed 488 times.)
Logged

These are the good old days of AM
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #1 on: March 05, 2007, 10:13:24 PM »

Well tonight I put a couple of hours into the Apache's RF section. I rewired the key jack and hooked up B+ to the RF section. No output - nothing.

The linkages and metal belt and suspenders bandswitching is absolutely amazing in its complexity on the Apache. I can not believe that a kit builder could possibly get that right on the first try.
 
VFO problems. No VFO output on three bands. The weird switch that changes the VFO rotary from the 1.75 - 2.0 tank on 80m to another tank on 40M and yet another tank on 20M and 15M and then back to the 40M tank on 10M was misalligned. It took me a while to figure out what was wrong.
Anyway, I now have the VFO back together and good output and half OK dial calibration on all bands.

The only other thing I did was grid dip the driver tank switching to make sure that it is tracking on the right bands! Thank heavens it was correct. I was got nervous when the VFO was off. 

Mike WU2D
Logged

These are the good old days of AM
Steve - WB3HUZ
Guest
« Reply #2 on: March 06, 2007, 09:04:47 AM »

Quote
The seperate front panel "GAIN" control and the hidden "LEVEL CONTROL" seem to have confused hams of the day. Some guys really had them screwed up bad.

I think you nailed it. Many guys thought they were turning up the mic gain but instead were cranking up the clipping. Yuk!

I recall someone doing some measurements on the filter following the clipper stage. Seems when it was removed from the cabinet, it had a pretty flat response up to 3 kHz or so. But once it was in the cabinet, the curve had some nasty peaks and valleys, probably due to the inductors interacting with the cabinet. Unfortunately, most of the filter used in amateur radio equipment following clippers were ill suited for the job since no consideration for the impulse response was given. The output of the filters actually had more peaks due to overshoot than if no clipper filter had been used.

Anyway, the Apache can sound very good with the stock iron. The Timtron turbo mod (feeding the B+ to the secondary through the 500 Ohm tap) seems to fix the impedance mismatch and gets rid of the high level clipping. Tron also used a little negative feedback around the mod tranny. The one he did up when he was in Syracuse a number of years ago sounded excellent.
Logged
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #3 on: March 07, 2007, 10:34:14 AM »

Well, we will see when I get to the modulation part of this adventure. I do have a stock Apache XFMR.

Now back to my "no RF" Apache problem.

Something Frank GFZ said helped me to solve the big problem of no RF drive. "make sure you know where you are putting your tube"

I got in a hurry and was confused and put the 5763 where the 6CL6 goes - that is apparently bad because it did not work.  The fact that the buffer is closer to the finals than the driver threw me. Good thing I was using junkbox tubes!

I now have 30 Watts out on 80M - 20M and 20 W on 15M and 10M. I will check the bias levels next and put in some good tubes and see what she will do.

Mike WU2D
Logged

These are the good old days of AM
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #4 on: March 10, 2007, 01:18:03 PM »

I'm still not happy with the RF output situation on the Apache. I have a problem which is driving me nuts. The 6146's keep popping out of the sockets!

Heathkit used some very high quality sockets, the large ceramic kind with the riveted contacts. And - the RF finals cage is really a cage. The sockets are elevated just off chassis with standoffs. Assembly instructions required that the  whole deal had to be pre-wired before installation. The heaters, cathode meter tap, grid bias and screen voltages come up through very small holes which the PVC covered wire just clears directly under the socket pair. This is sure is excellent direct wiring but it seems a bit dangerous without grommets! The problem with the big old sockets is that the 6146 pins do not bite very far into the socket and Heath used round head screws which do not allow many types of 6146s to get effective insertion on the socket at all. Raytheons, for instance, simply do not go in.

Here is my rebuild approach:

I cleaned up a pair of conventional ceramic sockets and tried them with the 6146's. I had to work the tubes out of these when I tested the insertion strength - nice. I will stand these off with short hex male-female standoffs. I will attempt direct grounding to lugs on the chassis close to the lugs for bypassing rather than depending on one lug or the grounding ring on the sockets. Finally I will install some filter feedthrus and grommets where appropriate.

Mike WU2D



* Apache6146soctop.jpg (163.9 KB, 1218x588 - viewed 415 times.)

* Apache6146socbot.jpg (168.69 KB, 1392x609 - viewed 466 times.)
Logged

These are the good old days of AM
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #5 on: March 11, 2007, 07:57:42 PM »

Wow, that was more work than I thought it would be!

Mr. Rogers - I had no idea about the bases being different on 6146 variants but in pawing through my collection it certainly confirms that everybody had a different idea.

In doing the socket swap-out, I had to remove the cage shield and disconnect the loading capacitor (which hangs OK in mid air) and the feed choke and the Amphenol shroud which is the coaxial output interface.

This exposed a problem, the capacitor on the bottom of the feed choke, a 1000 pf doorknob, which was cold soldered and was loose. This may have been the largest contributor to the poor output. The other dorknob is the main coupling capacitor which is a 500 pf. This was not so tightly connected to the Johnson main tuning capacitor and I fixed that. The Heath Apache has a tremendous Johnson by the way. 

Where there were only 4 grounding lugs to chassis around the sockets, there are now 6. Also Heath only connected to one cathode pin on each tube (pin 1). I strapped all six cathode pins together (pins 1, 4 and 6 on each tube).


The results are impressive.  I could only muster 70 Watts out on 40M before the conversion and almost nothing on 10M.

80M 120W
40M 130W
20M 115W
15M 110W
10M 90W

And this is on a pair of RCA 6146B pulls which came out of my Marauder, not new tubes.

Mike WU2D


 Grin
Logged

These are the good old days of AM
The Slab Bacon
Member

Offline Offline

Posts: 3934



« Reply #6 on: March 11, 2007, 08:57:15 PM »

Just a FYI about 6146s, not all are created equal.  There have been several discussions on here in the past. Here is the deal: 6146, 6146-A, and 6146-W (ruggedized) are all basically the same and interchangable. The 6146-B is a little different animal though. It has a little higher plate dissapation rating and a little different scren grid structure. The others will operate in just about any class, but the "B" was really designed for class AB1 sideband linear service.

In most cases the "B" will interchange with the others, but occasionally it wont and you will have complications. This is application specific and varies from rig to rig. In some rigs they will work ok, in some they wont. If at all possible you should try to stay away from the "B"s for class C service.

You may or may not have problems, you'll have to see, usually they are ok.

I have also ran into the final socket interferance problems with these rigs. Usually the newer tubes with the "tapered" bases will work. If not you just have to take a "Dremel Tool" and grind small clearance notches in the tube's bases. I have built / rebuilt 5 or 6 Apaches in the last few years. They have all come out very good.

                                                                              The Slab Bacon
Logged

"No is not an answer and failure is not an option!"
WBear2GCR
Member

Offline Offline

Posts: 4135


Brrrr- it's cold in the shack! Fire up the BIG RIG


WWW
« Reply #7 on: March 12, 2007, 07:38:39 PM »

Fyi, when you get to the audio part, if you do a search here I posted some interesting "uses" for the existing parts in the filter part of the "clipper" circuit a while back...

I doubt that the 6550s will stand 750vdc on their plates, although if they do, then you'll be able to sink a lot more current than you would with 6CA7/EL34s or a set of 807s... probably... which would me more power available from the modulator given the lower Z plate iron you have there. I'm guessing that this was the "idea"?

              _-_-WBear2GCR
Logged

_-_- bear WB2GCR                   http://www.bearlabs.com
kc2ifr
Guest
« Reply #8 on: March 13, 2007, 11:06:18 AM »

Here is a dandy toob that I was going to use in my Valiant but they are very hard to find and expensive if u do find them. They were used a lot in England for both rf and audio. Basically a KT88 with a plate cap. I wish they were still around.
http://www.retrovox.com.au/c20store/TT21.html
Logged
The Slab Bacon
Member

Offline Offline

Posts: 3934



« Reply #9 on: March 13, 2007, 11:33:28 AM »

Kool bottle!! I might have to try a pair in my Valiant! the plate cap will eliminate the flashover problems between the plate and screen pins!
Logged

"No is not an answer and failure is not an option!"
Bacon, WA3WDR
Member

Offline Offline

Posts: 881



« Reply #10 on: March 13, 2007, 12:55:34 PM »

I use 8417s (sort of skinny 6550s) for modulators in my Viking II, and I added a couple of diodes and a capacitive-input filter section for them - about 975VDC, with the original 600 volts or so on the 6146 plates, and it hasn't arced yet.  I had replaced the blown original mod transformer with a Stancor A3894 125 watt unit.

Logged

Truth can be stranger than fiction.  But fiction can be pretty strange, too!
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #11 on: March 15, 2007, 12:49:33 PM »

All great suggestions on the tubes. I have been looking at KT-90's. This tube has a 50W plate and will do 750 VDC. It is tall. Around $70-80 per pair.

Mike WU2D


* KT90.jpg (68.92 KB, 507x1440 - viewed 409 times.)
* KT90specs.pdf (65.6 KB - downloaded 190 times.)
Logged

These are the good old days of AM
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #12 on: March 18, 2007, 10:25:52 AM »

Before I get into the fun audio part, I am still trying to figure out a few things like:

1. How does the neutralization work on this thing?

There is no doubt that it is unstable on the higher bands at least. You probably would notice nothing on 80 or 40M, so don't get excited, but is sure goes wild on some settings on the higher bands.. My Apache has been messed with in the neutralization hookup, but the schematic seems to show a wire which is bent into position to couple to the 6146 plates feeding back directly to the HOT side of the driver tank. That seems to be positive feedback - how can this work?

2. I want to solid state the HV to relieve the transformer of some power load and am considering a time delay circuit for the medium voltage to come on allowing the fils to stabilize. This is a way to relax things a bit for these old transformers and tubes.

3. PTT - there is a tremendous event which occurs when you go into transmit. A big clunk, dimming lights, a big red TX lamp etc..
Does anybody have a pet Apache PTT approach. I am considering using a plug in relay on one of the liberated 5R4 sockets. (after solid stating the HV).

Thanks Guys!

Mike WU2D
Logged

These are the good old days of AM
Steve - WB3HUZ
Guest
« Reply #13 on: March 18, 2007, 02:16:04 PM »

Quote
a wire which is bent into position to couple to the 6146 plates feeding back directly to the HOT side of the driver tank. That seems to be positive feedback - how can this work?

Doesn't a 180 degree phase shift occur through the final amp, so the feedback would be negative? Or am I missing something? Probably. Kiss
Logged
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #14 on: March 18, 2007, 06:56:28 PM »

Steve,

You are absolutely correct; there is a 180 degree phase shift - and if we were talking about audio amplifiers and untuned RF amplifiers, we would be OK. (Well OK most of the time...the audio guys are neutralizing audio amps!!) The problem is that we have these darned tuned circuits on both grid and plate, and without much difficulty, we can get into a positive feedback phase condition.

Now all that is needed is a path from plate to grid. Enter the plate to grid capacitance.

That is the principle of the TPTG oscillator. Actually with a TPTG, the grid and plate tuned circuits are deliberately tuned to slightly different frequencies, or it will not oscillate at all!

So with neutralization, all we are trying to do is cancel out the offending plate to grid capacitance with a bridge circuit. We need a way to introduce equal but opposite coupled energy into one of the tanks (grid or plate) to zero out this capacitive path.

With two 6146's, we may be talking about 0.5 pf.


Mike WU2D
Logged

These are the good old days of AM
Steve - WB3HUZ
Guest
« Reply #15 on: March 18, 2007, 08:51:45 PM »

Oh yea, the phase shift from the tank circuit. Most neutralizing circuits I've seen do go to the bottom/cold side of the grid tank.
Logged
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #16 on: March 18, 2007, 10:36:51 PM »

Steve,

The Apache has an advanced Pi-Type driver coupling network which is going to be interesting to deal with as far as grid neutralization goes.

Mike WU2D
Logged

These are the good old days of AM
Bacon, WA3WDR
Member

Offline Offline

Posts: 881



« Reply #17 on: March 18, 2007, 11:25:21 PM »

A resonant pi-net is a phase inverter, basically a capacitor-tapped tuned circuit.  Neutralization would therefore go from the PA plate to the driver plate circuit.  Unless the right multi-section capacitor is used to tune the driver pi-net on both sides of the coil, grid tuning will change the transformation ratio somewhat, and this could cause neutralization to be out of kilter on some bands and some tuning settings.
Logged

Truth can be stranger than fiction.  But fiction can be pretty strange, too!
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #18 on: March 19, 2007, 08:51:42 AM »

Bacon,

That has to be what the Heath engineers were trying to do! Those clever buggers. They did not read that out of the ARRL handbook!

I will attempt to neutralize after I make sure I am hooked up on the correct side of the network.

See how simple these old rigs are? They all have secrets I guess.

Mike WU2D
Logged

These are the good old days of AM
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #19 on: June 03, 2007, 02:19:58 PM »

The audio will have to wait for Fall, but before I put the beast away, I wanted to make sure that I documented the power supply modifications that I made earlier in the Spring.

The idea was to solid state the Apache's power supplies, thus eliminating the 5V4 medium voltage tube and the two 5R4 high voltage rectifiers. This would reduce fillament drain and heating (30 Watts saving). The HV should also increase somewhat. This would also free up a socket for a PTT relay.

The two HV modules were made from octal old relay cans. I used modern well matched 3 AMP 1000 PIV standard recovery rectifiers, 1N5408's. We shall see how good they are beacuse I used only 3 on the HV on each side and 2 on the Medium Voltage on each side and employed no surge protection and no balancing resistors or capacitors. So far they are working perfectly. 

The issue of solid stating at all is another concern. Does the sudden high voltage or increased voltage due to the solid state power supplies cause trouble with boatanchors?  This is not a problem with the HV in the Apache, which is controlled with the plate switch (usually actuated well after the tubes are lit!). The medium voltage supply however might cause a problem. I made a simple circuit that does not allow the medium voltage supply to come on until a time constant allows a relay to ground the center tap of the supply. This gives around 10 seconds of warm up with my relay.

Mike WU2D


* ApachePower.jpg (121.85 KB, 1242x785 - viewed 467 times.)
Logged

These are the good old days of AM
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #20 on: June 04, 2007, 08:47:39 PM »

OK let's hear it - Mike you can't do that! I have had the setup as shown working on and off for a couple months now BUT....

Tonight I wanted to give the Apache one last tuneup before I put it away and Bang!
The HV string went poop and the fuse popped. Two diodes were shorted in the HV string.

Obviously the diodes were not as well matched in terms of PIV and leakage, the transformer was not as balanced, they could not take the surge, the moon got full twice in one month etc.. and it did not work.

OK I call this good luck.

Anyway, after beefing up the strings as shown and testing I am a little more confident. Instead of two diodes, use three on the Medium Voltage and instead of 3, use 5 on the high voltage and use some 220K 1/2Watt resistors for insurance.

have redrawn the schematic - comments?

Mike  WU2D


* ApachePowerTry Again.jpg (138.04 KB, 1242x785 - viewed 476 times.)
Logged

These are the good old days of AM
John K5PRO
Member

Offline Offline

Posts: 1033



« Reply #21 on: June 05, 2007, 11:53:27 AM »

I usually have selected diode strings to withstand 2 to 3 times the PIV of the transformer/rectifier consideration in use. It gives a decent safety factor, for commerical gear.

I was wondering about the time delay relay that connects the center tap to ground after the transformer is energized. Would it not be more useful to just have a step start resistor that is inserted with the transformer primary, that is removed or shorted out after 1/2 second or less. Just to handle the inrush. The circuit you have, lets the iron energize first, then applies the empty capacitor across the output of the rectifiers. The peak current in your diodes is going to be high, as the charging current for the caps is limited only by the diode resistance, the transformer winding resistance, the impedance of the transformer. A simple step start with a few ohms will reduce this inrush at the same time for the transformer iron as well as the cap current. Diodes will appreciate that.

Even more simple is to have an inrush limiter, a PTC device that changes resistance in proportion to the peak current.I use one for my DX100, makes it much easier for the HV switch, the fuse, the rectifiers.
Logged
WU2D
Contributing
Member
*
Offline Offline

Posts: 1800


CW is just a narrower version of AM


« Reply #22 on: June 05, 2007, 02:21:16 PM »

Hi John,

WA1QHQ, Mark was telling me about those components a few months ago and I have yet to grab some from him and play with them. I think that would be a great idea.

Who are you buying these from? Digi-Key or Mouser?

My circuit is not to limit inrush, it is simply to let the tube fils warm up before voltage is applied and it clicks in after 8 seconds.

Mike
Logged

These are the good old days of AM
Pete, WA2CWA
Moderator
Member

Offline Offline

Posts: 8154


CQ CQ CONTEST


WWW
« Reply #23 on: June 05, 2007, 02:39:32 PM »

I solid-stated my Apache's low and high voltage sections back in the 70's. Found some tube bases, mounted diode strings and equalizing resistors on perf board; stuck it in the tube base; wired the correct pins; epoxy the board to the tube base and stuck one in the 5R4 socket and one in the 5V4 socket. They've been in there even since and still using the same tubes. No in-rush whatever type protection. Where's the need? The tubes that the 5V4 powers are nothing special or of any critical nature.
Logged

Pete, WA2CWA - "A Cluttered Desk is a Sign of Genius"
Bacon, WA3WDR
Member

Offline Offline

Posts: 881



« Reply #24 on: June 05, 2007, 09:30:29 PM »

I'm not sure how high the modulator screen voltage is in an Apache, but one day when you have it working and you are bored, you might consider my old trick of adding a resistor in series with each modulator screen grid.  Maybe try 1K 2W to each modulator screen grid - not one resistor to both screens, but two resistors, one in series with each screen.  This gave me audibly lower distortion with 8417s, which are basically similar to EL34s and 6550s.  6L6s gave similar results in an audio amp.

IMPROVING THE LINEARITY OF SCREEN GRID TUBES
AM Press/Exchange, Issue #110
http://amfone.net/AMPX/110.htm

   Bacon
Logged

Truth can be stranger than fiction.  But fiction can be pretty strange, too!
Pages: [1] 2   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.082 seconds with 19 queries.