BC-1G,T Final Grid Circuit

Discussion in 'Amplitude Modulation' started by AC0OB, May 16, 2021.

ad: L-HROutlet
ad: l-rl
ad: Left-3
ad: L-MFJ
ad: abrind-2
ad: Left-2
ad: L-Geochron
  1. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    With the popularity of the Gates BC-1G,T for amateur radio conversions, one would think there would be plenty of conversion notes with schematics on the Internet, but few apparently have survived or have been preserved. If anyone has further information or schematics of grid and plate tuning conversions it would be interesting to analyze these circuits for posterity.

    Changing the Final stage tuning to a single Pi-Net tuning system is easily done today with the help of online network calculators.

    The Driver Tuning Circuit is of the “Rice” design for neutralization and consists of Lres (L4 or L8) and Cres, C4 or C7, a 250pF variable capacitor. An optional, extra padding capacitor for frequencies at or below 1050kHz, CFD, can be added if necessary. There is a tuning chart in the manuals to tell you what capacitance to add across C4 or C7 for those lower frequencies. Cres, C4 or C7, resonates the circuit from the neutralization connection to the selected inductance tap in most cases. The bottom neutralization tap can be moved upward if the Final's grid current indication doesn't result in the proper tuning indications. Unfortunately Gates did not expound in any detail on the Lres coil tap positions, which is why some amateurs opted for other drive circuit configurations.

    Lres, L4 or L8, G or T respectively, is 2.8125" Long, has a Diameter of 2.5," and is mounted on a 5.75" long coil form with 8 taps. Measured DC resistance was 0.5 ohms. The measured total inductance was 86 uH and the calculated inductance was 85.x uH. A tap near the center of the coil provides for grid leak bias.

    Cres, C4 or C7, G or T respectively, is a 250pF variable capacitor, is isolated and mounted on 1" ceramic standoffs. With this station at 1.57MHz, the taps were set to 72uH and the Cres cap was set to approx. 145pF. One can connect the RF drive input, the Cres capacitor, and the grid drive line to the inductance tap needed and short out any inductance not needed (with a shorting switch) at the top of the coil to prevent induced stray currents for > 1.6MHz operation.

    With the optional CFD absent, one only needs about 152pF of C7 to resonate at 1.9MHz with the 50uH tap, and 36pF to resonate at 3.75MHz, with the 50 uH tap, all well within the tuning range of Cres alone.

    The Amperex 833 specifications suggested -300Volts of bias and 150mA of grid current with a grid voltage of 352Volts rms. 60 Watts of drive was also suggested. The suggested grid leak resistance was 2k for two tubes in parallel. Using R
    EQV = 352Vrms^2/60W = 2.065k, this would be the equivalent resistance seen by the output of the exciter circuit when two 833 grids are being fed in parallel.

    Gates apparently decided that a grid bias of -360Volts and a grid current of 144 mA was better since one might also be using RCA or GE tubes. They used a grid leak resistance of 2.5k implemented by two 5k 50W resistors in parallel, so V
    RG = .144mA*2.5k = 360Volts dc bias.

    Using the same set of equations, the 3-400Z page 4 describes that bias system.


    Attached Files:

    Last edited: May 16, 2021
    1 person likes this.
  2. W3SLK

    W3SLK Ham Member QRZ Page

    Phil, would it have been any easier to "drive" the existing driver in the -1T? Your circuit does make it easy to use a conventional driver such as a DX-100 or possibly a Ranger. I almost ended up with a single RCA 833 BC rig that was already converted.
  3. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    One of the problems with doing that is that the driver needs to be modulated along with the final, using the tertiary winding of the modulation transformer. The fact is, a pair of 807s or 6BG6s, isn't sufficient to fully drive a pair of 833As in class-C plate modulated service, so the modulation of the driver stage is used to give the PA grid drive an extra "kick in the arse" when it needs it, on positive modulation peaks. It's unlikely that a pair of 6146s as in Viking II or DX-100 would do much better than the 807s, although it shouldn't be too difficult to run a link from the tertiary winding back to the plate circuit of the outboard driver transmitter, to plate modulate the pair of 6146s in sync with the 833As. The plate voltage on the DX-100 is pretty close to what Gates puts on the internal 807 driver stage. No way would a Ranger make sufficient RF drive for this transmitter, although it might work with one of the "500-watt" versions that use a single 833A in the PA. A valiant, with three 6146s in the final, might eek out enough to drive a pair of 833As without modulating the 6146s.

    The pair of 807s in the BC-1T and versions to follow, is an example of "value engineering" adopted by Gates after they were acquired by Harris Intertype. Earlier BC-1 series transmitters, like the BC1-F, used a single 813 to drive the 833A. The Raytheon RA-1000 went a step further with overkill, using a pair of 813s in push-pull to drive a pair of 833A in the PA. Harries evidently figured it was cheaper to use a pair of 6BG6s and later 807s with the tertiary winding on the modulation transformer, than to use a full-fledged RF driver like one or two 813s to adequately drive the 833As with no need for modulating the driver stage. The 807/6BG6 driver stage is marginal at best, but even with a fully adequate driver stage like what's in the BC1-F and RA-1000, the modulation of the driver still wouldn't be a bad idea, to assure the absolute best modulation linearity.
    K0UO likes this.
  4. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    Yes you can as long as the driving rig or exciter is terminated properly into 50 ohms (100 Watt non-inductive resistor) and then decoupled with a 0.001uF cap before the grid tuning circuit.

    In fact, many have driven the BC-1 series and other BC transmitters with a DX-100 or a Johnson Viking I or II, or similar with good results. This method is not the most efficient, but it does work.

    Using a lower power vintage transmitter ahead of the grid circuit offers another advantage and that is, further harmonics are reduced by the lower power vintage transmitter's Pi-Net tuning circuit.


    Attached Files:

    Last edited: May 16, 2021
    K0UO and W3SLK like this.
  5. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    Forgot to add the coil is 47 turns of #16 with adjacent turns spacing of approx. 1 wire diameter.

    AND THIS SENTENCE SHOULD READ: They used a grid leak resistance of 2.5k implemented by two 5k 50W resistors in parallel, so VRG = 144mA*2.5k = 360Volts dc bias.

  6. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    This is a little confusing; do you mean using the DX-100 to drive the pair of 807s that drive the 833As, or for driving the grids of the 833As directly? The stock pair of 807s at 180 -200 mA total cathode current just barely drives the 833As to about 140 mA grid current, with minimal RF losses in the coupling circuit. Loading a DX-100 down with a 50Ω non-inductive resistor would suck up half or more of the available RF driving power, and wouldn't leave enough to fully drive the 833A grids directly.

    You could use an external transmitter to drive the 807 stage, replacing the RF crystal oscillator unit inside the transmitter with the outboard unit with VFO control. But doing this with a DX-100 would be like using a 12 gauge shotgun to kill a fly, since in the stock circuit a single 12BY7 gives plenty of grid drive to the 807 grids. A better choice of RF driver unit in this case would be a low-power solid state plastic radio with adjustable carrier output settings, or perhaps a "vintage" QRP transmitter, since 10 to 20 watts would be more than adequate to drive the 807 grids even with an inefficient coupling configuration.
  7. W7TFO

    W7TFO Premium Subscriber QRZ Page

    Ummm...Clash of the Titans...

    WZ5Q likes this.
  8. W3SLK

    W3SLK Ham Member QRZ Page

    K4KYV said:
    To be truthful Don, that was the premise of my question. I was curious if it would be easier to drive the driver, as in this case the 807 with something along the lines of a T-368VFO or reasonable facsimile thereof. Granted you would probably need a tuned input to the 807. But I would imagine something small with low Q would suffice. Phil's circuit allows you to 'brute force' the grids of the 833's. I can recall where a DX-100 was used as a driver with an 833 box but I'm not sure if it was for 2 or only 1.
  9. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    A T-368 master oscillator unit should easily drive the 807s, since it drove a single 4-400 in the T3. The only thing, the m.o. would have to be pretty close to the transmitter, since the output BNC connector is connected right to the plate of the 6000 tube - not designed to be a matched transmission line, but simply serve as a shielded lead from the plate of the 6000 to the grid of the 4-400. The capacitance of a significant length of coax would swamp out the resonating cap in the tuned circuit, which is merely that - a parallel tuned circuit using a slug-tuned coil as the variable element. You might get it work with a few feet of coax, but it would likely screw up the tracking of the 6000 plate; maybe you could align it mid-band and it would still track well enough to cover the ham band, since the rest of the tuning range is useless for amateur operation. I modified mine to link coupled output from each plate coil instead of the simple blocking cap, but that's a whole other story of a seriously involved project that I worked on for months before I got right.
    The tuned output coil in the T-3 m.o. unit would serve as tuned input to the 807s. All you would have to do would be to get it to resonate to frequency with the added capacitance of the coax run.
    It would take more than a DX-100 to brute-force the grids of a pair of 833s to 150 mA grid current across a 50-ohm swamping resistor. The full RF voltage applied to the 833 grids would have to be developed across the 50Ω terminating resistor. According to the RCA tube manual, an 833 running CCS plate-modulated class-C service at 2500 plate volts, requires 460 peak grid volts, which equals 325 volts r.m.s. Using Ohm's law, the power dissipated in the resistor would be V² ÷ R, in this case 325² ÷ 50, which comes out to 105625 ÷ 50, = 2112.5 watts!

    To brute-force drive the 833s directly, eliminate the 50Ω terminating resistor and load the output of the transmitter directly into the grids, if the Pi-net in the DX-100 would be capable of working into that high a load impedance. The easiest way, if direct feed didn't work, would be to wind a few turns of link coil about the mid-point of the grid coil. Adjust the number of turns on the link coil to produce something in the ballpark of a 50-75 ohm load and feed the link with the pi- network output of the DX-100. Note: This would entail also running an additional line from the tertiary winding on the Gates modulation transformer to the DX-100, wired to plate modulate the 6146 finals in sync with the 833A plates.
    Last edited: May 17, 2021
  10. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    I took Mike's (W3SLK) question (post #2) as driving the 833 grids directly with an external vintage transmitter which is why I included the schematic.

    If Mike was also asking about driving the 'Second Stage IPA' with a vintage transmitter, then yes, one could also drive the BC-1 to its full potential by driving the Second IPA Stage directly with a vintage transmitter with approx. 45 Watts or less of drive from the Vintage transmitter, according to the GE 6BG6G tube data curves and the BC-1 T's Voltage chart.



    Attached Files:

    Last edited: May 17, 2021

Share This Page

ad: vanity