The Evolution of a lonely AF-67

Discussion in 'Amplitude Modulation' started by WZ5Q, Jun 23, 2017.

ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: abrind-2
ad: Left-3
ad: Subscribe
ad: L-MFJ
  1. WZ5Q

    WZ5Q Ham Member QRZ Page

    Howdy Y'all,

    Here I sit once again sleepless, listening to white noise and static crashes on the Hallicrafters R274/FRR as my mind wonders endlessly around in infinite little circles...

    I've made mention of my temporary AM Station on this forum a few times, sometimes posting pictures of the old buzzardly layout. I thought I would add to that with some schematics and some history.

    As the slow process of getting the new shack built proceeds (it seems like it will never get accomplished), there had to be somewhere to play radio in the meantime, so one of the spare bedrooms in the main house had been designated as the temporary shack. It was a "girls" room and had been painted pink with little frilly thingies on the light switches and such. Being in a hurry to get something on the air, this room was not remodeled and remained in it's feminine state as the shelves and benches were stacked and assembled. It is now called "The Pink Room". Posh!

    Now that a location for the temporary shack was secured, a decision had to be made on what to put in it. I remembered there was a Multi-Elmac AF-67 Exciter shoved up on a shelf somewhere that would match up nicely to one of the R274/FRR Receivers, but it had no power supply. Easy enough solution, just go out into the barn and get the parts to cobble one together. This was accomplished but there were no "small" parts available, so everything is HUGE and was built breadboard style being a temporary shack and all. The classic brute force design was followed using a pair of Murky Vapor Rectifiers followed by a choke input filter and voltage divider/bleeder resisters.

    With the Power Supply built, the stock AF-67 was hooked up and tested. Oh my... it was just horrid. The Modulator was full of distortion, limited in frequency response, and lacked enough balls to fully modulate the rig. Not to mention that it looked like a complete horror show on the Scope. The 6146 RF Final had horrible Modulation Linearity no matter what you did with the drive and the screen. Improvements were definitely mandated.

    First things to go were the 6AG5 Buffer/Multiplier, 6AQ5 RF Driver, & the 6146 RF Final. The RF Section was redesigned to use the 6AQ5 as the Buffer/Multiplier driving the 6146 as an RF Driver using hand wound, plug in coils for it's output tank. The 6146 drives the piss out of the Medium Mu 809 Triode RF Final into Deep Class C. Using Grid Leak, the drive hungry 809 Triode is operating at 8x cutoff just under it's max grid specs, which gives me an ultra narrow conduction angle with extreme modulation linearity. Operating conditions of the 809 are 370V on the plate @ 60mA which gives me close to 20 watts carrier output.

    Next was the carnage on the Modulator. I never did like the 6AU6 Pentode Preamp Multi-Elmac used as the first audio amplifier. This stage could never be gotten linear and symmetrical without using tons of local feedback which of course reduced the gain a substantial amount. The extra gain wasn't needed as line level audio was being injected, so if it was going to be lost anyway, I would rather tie the Suppressor and Screen Grids to the Plate in a Triode configuration and use no feedback. That is exactly what I did and it looks terrific on the scope from 20cps thru 20Kc.

    The 12AU7 Driver was changed to a 12B4A for its low plate resistance. The socket was rewired and the cathode resister changed to 1.8KΩ then bypassed with a 100uF/450v cap. The stock 3:1 Class B Driver Iron was changed to a UTC Linear Standard LS-21. This Iron cannot take any DC on it's Primary at all, so a Parafeed Input configuration was implemented. The 12B4A plate is choke loaded using a 100H @ 40mA Plate Reactor with a 3uF polypropylene coupling capacitor to the LS-21 Primary. 520VDC is supplied to the reactor and the 12B4A idles at 38mA. Yea its screaming, but it can take it!

    The stock 5881 Modulator Tubes were changed out to screen driven, Triode connected 6AV5's. The G1 Control Grids are tied to the Cathodes which are then grounded, and the G2 Screens are being driven by the LS-21 Secondary. I added 20Ω Resisters in series with the 6AV5 Plates for stability against parasitic oscillation. This gives enough Audio Power to run the modulation at 98% Negative with 185% Positive Peaks. My natural voice asymmetry runs the modulation at 130% Positive/98% Negative Peaks so there is a lot of reserve left over. Operating conditions are 520V on the Plates with 40mA resting current, swinging about 150mA on voice peaks.

    The finishing touch was to replace the stock Modulation Iron with a much bigger Thordarson T-19M16 100w Mod Transformer. The windings were configured for a 1:25 Turns Ratio, 1:56 Impedance Ratio. With the 809 modulation Impedance of 6200Ω, this should reflect around 9687Ω for the 6AV5's. The DC was removed from the Secondary using a 10uF coupling capacitor with a 40H Modulation Reactor.

    The spread out is definitely "Old Buzzardly", but neatness was not my prime objective for this temporary lash up. I still have some tweaks and a few more mods to do, but at least it sounds decent now.

    I have attached a schematic of the modifications as well as a stock AF-67 schematic for comparison. There are also a few pictures.


    af67 809 mods.gif stock af67.gif

    1_little_frankie_far.jpg 2_little_frankie_near.jpg

    3_little_frankie_audio_iron.jpg 4_little_frankie_chassis_top.jpg

    KC8YLT and N6YW like this.
  2. AF6LJ

    AF6LJ Ham Member QRZ Page

    This is going to be good, I have an AF-67 in the garage. :)
  3. K9STH

    K9STH Platinum Subscriber Platinum Subscriber QRZ Page

    I have found that the best thing to do with an AF-67 is to replace the internal bias battery with external bias. There is an unused contact on the power plug that can be used for the external bias. The only modification to the transmitter itself is a single wire to the old battery negative terminal and adding a bypass capacitor.

    Then, by varying the bias voltage from an external source, the optimum value can be found for the best sounding audio.

    I first did this at the request of a client who wanted to use the AF-67 with a PMR-8 receiver for a "portable" AC operated AM station. The results were beyond his expectations.

    One can completely rebuild the transmitter if they want to. The equipment belongs to that person and they have every right to modify the equipment as they see fit. However, if someone doesn't want to do all that, and if that someone wants to retain the originality, then replacing the internal bias battery with an external bias supply is something to consider.

    Glen, K9STH
    N2EY, WZ5Q, N6YW and 1 other person like this.
  4. KL7AJ

    KL7AJ XML Subscriber QRZ Page

    Doesn't look too lonely to me!
    WZ5Q likes this.
  5. KA0HCP

    KA0HCP XML Subscriber QRZ Page

    How many radios have a "Gear shift lever"? :)

    I've got two of them. It was a package deal, "great price but, you must take them both".
  6. W7TFO

    W7TFO Platinum Subscriber Platinum Subscriber QRZ Page

    Chain driven internals. Good stuff.

    Reminds me of the R-1051 Navy RX.

  7. KA0HCP

    KA0HCP XML Subscriber QRZ Page

  8. KL7AJ

    KL7AJ XML Subscriber QRZ Page

    Central Electronics 100V has chain drive too!
  9. KE5YTV

    KE5YTV Ham Member QRZ Page

    I'm very impressed with that manly power supply. That's the way John Wayne would build it! :cool:
    WZ5Q likes this.
  10. WZ5Q

    WZ5Q Ham Member QRZ Page

    Howdy Glen,

    Yes, you are correct. There are definitely many different levels of modification that can be performed ranging from minor to very invasive, all depending upon what level of performance you are willing to accept.

    I didn't mention all the interim mods that were performed that led to the final previously posted configuration of my AF-67.

    Just on the Modulator Section, all the simple tricks were performed first like a full alignment, then the bias mod you described. I still wasn't satisfied with the performance I was seeing, which led to passive component changes in the Speech Amp, changing the driver tube type and its operating parameters, changing out the restrictive Class B driver transformer and adding impedance loading to it's secondary for AB2 grid current draw, dicking with the nasty beam tetrode's bypassing and screen voltages, wrapping local negative feedback around the different stages, trying various modulation transformers, and the list goes on.

    The same thing happened with the RF Section. Simple things were tried first for the best modulation depth and linearity like an alignment, component value checks, and changing the impedance of the Final Pi-net for correct band loading. This did not satisfy me in the least so the natural progression was to start dicking with the screen modulation and grid leak values. Various different screen modulation "tricks" and configurations were tried including WA1KNX, Deano's modification. IMHO, none of these worked to my satisfaction, the best modulation linearity was with feeding the 6146 Screen grid with the tertiary winding from the secondary of an ART-13 Mod transformer. Ultimately the evil 6146 Beam Pentode was ripped out roots and all and replaced with a hard driven 809 Triode. All sins were then forgiven.

    Like you eluded to, not everyone is the same in what they want or looking for. I am performance oriented and nothing is sacred to me when it comes to making it perform the best it can, whatever it is.

Share This Page