ad: M2Ant-1

Experiments With A 6AR8 AM/SSB Modulator

Discussion in 'Homebrew and Kit Projects' started by W0RIO, Sep 9, 2019.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: Subscribe
ad: Left-2
ad: MessiPaoloni-1
  1. W0RIO

    W0RIO Ham Member QRZ Page

    I let this project sit idle for over a month, with the output imbalance issue gnawing a me the whole time ;-)
    A couple of days ago I put it all back together and came up with a better way to produce a balanced output
    signal. This version of the circuit adds two resistors to the output side of the modulation transformer, those combine
    with the 10K balance pot to provide enough deflection plate bias range to properly null out the carrier.

    I also experimented with lowering the bias voltage on the accelerator element, which adds a tiny amount
    of limiting action to the tube. That changed the cathode bias, so I adjusted the cathode resistor accordingly.
    I changed the cathode and deflector bias points down slightly after looking over the tube data sheet
    some more. (I just noticed that the 2.5V p-p RF carrier raises the grid above the cathode, which should be avoided, I may
    change to a 750 ohm cathode bias resistor to increase the cathode bias a small amount).

    The circuit produces a nice sounding signal at this point, it's still far from being something that I'd put on the air.

  2. KB1GMX

    KB1GMX Ham Member QRZ Page

    The 6AR8 is a sheet beam tube just like the 7360 and the design purpose was
    color chroma decoder for color tv. balanced modulator is a natural use for that as well.

    I've used the 6ar8 very successfully with no balance issues like you discuss. I have
    used it as mixer, product detector, and balanced modulator. However
    mechanical balance and short leads is an absolute must for 7360/6AR8/6JH8.

    The output coil should be bifilar wound and the two turn link wound
    at the center tap point. A breadboard like that is likely to fail based on experience.

    The 6AR8 circuit shown is not what I'd consider a good one. Find the
    GE SIDEBAND HANDbook as known good circuits are there. It can be found on the 'net.

    Generally the Beam deflectors (pins 1 and 2) have a positive bias around 8 to 10V.

  3. KX4OM

    KX4OM Ham Member QRZ Page

    The 7360 was used in the (in)famous EICO 753. The and only 7360 in my stash came from my disassembly of a 753 carcass. A very good copy of the 753 manual and schematic is available free from It might be a good reference for the balancing issue of the 6AR8.

    If you want to see a beautifully restored 753, go to Dale, W4OP's vintage restoration pages at:

    Ted, KX4OM
    Last edited: Oct 23, 2019
  4. W0RIO

    W0RIO Ham Member QRZ Page

    The first circuit I posted is *not* a good one, the second circuit (with some minor refinements shown on my web page)
    solves the balance problem and works quite nicely. If it works, why is it not "good"? Note that my circuit is
    more of a lab experiment and is not cost-optimized for manufacturing. I agree with your comments about short leads
    and mechanical balance, better shielding is also important for reducing carrier leakage.

    Last night I tried replacing the large ceramic coil form with a bifilar-wound T50-2 toroid, that works OK
    although the Q of the large coil seems to be much higher than the Q of the toroid.

    The real test of the circuit (IMHO) is how it sounds when modulating music, although that's not what it is intended for in the long run.
    If it can produce good sounding modulation with complex music waveforms then it will also work well with simpler voice waveforms.
    I'm a big fan of one of Bob Heil's design tips, which boils down to: how does it sound?

    It's a lot easier listening to a modulated music source compared to saying "testing one two three" into a mic and listening
    to a receiver with headphones. DSB and SSB modulated music can really sound odd if the tuning is not exactly
    on frequency, but it's still very useful for finding and removing sources of distortion.

    BTW, there are some differences between the 6AR8 and 7360, the 6AR8 likes to have the deflectors biased at -8V
    and the 7360 deflectors should be biased at +25V. The deflector voltage swings are also quite different.
    Other sheet beam tubes have different bias requirements.

    From the GE 6AR8 data sheet:

    I've been having a lot of fun learning how this type of tube works, that's really the only objective I've had
    in this project. I'd also like to experiment with using the 6AR8 as a product detector.
  5. W9WQA

    W9WQA Ham Member QRZ Page

    stuff you cant do with transistors. theres magic in tubes. love to see them light up.
  6. W0RIO

    W0RIO Ham Member QRZ Page

    Just for fun, yesterday I "flipped" this circuit around by moving the audio transformer to the plate side and the RF transformer to the
    deflector side, turning it into a product detector. The results were promising. When feeding the circuit with two RF generators close
    in frequency, I was able to get a strong audio tone on the output above and below point where the input frequencies were equal.

    Taking it further, I connected the deflector input link coil to my 40 meter dipole and set the control grid generator to 5 Mhz,
    a very faint WWV signal was heard on the audio output! The setup definitely needs an RF amp ahead of the detector before
    it can really be considered a receiver.

    Doing these RF experiments on a proto board is definitely not optimal, but it can be made to work at lower frequencies.
    It's much easier to modify the circuit with the proto board compared to soldering/unsoldering parts in a metal box.
    I use several tricks including very short alligator clip-leads and a proto board with a metal ground plane.
    The RF lines are never routed on adjacent 5 pin "busses" to lessen the effects of stray capacitances.

    Next step: put some of this stuff into metal boxes with shielded tube sockets.
  7. KM1H

    KM1H Ham Member QRZ Page

    The 6AR8 was a disaster in TV service and quickly replaced by RCA which is why so many NIB still show up with no takers or at very low cost.

    I have not used it but several 6JH8, 6ME8, and 7360 projects have worked out very well since I converted both 6BA7 mixers in a 75A4 to 7360's in 1965 and still use.

  8. W0RIO

    W0RIO Ham Member QRZ Page

    Carl, could you explain what the problems with the 6AR8 were that caused it to be such a disaster?
    Are the 6JH8 or 6ME8 better tubes in any way?

    It's interesting to hear the history behind the sheet beam tubes.
    I've found the 6AR8 to be fairly stable, if somewhat microphonic.

    I was able to take my product detector experiment further, I borrowed a 6AU6 RF amp and preselector from another project and
    inserted that between the antenna and the detector. I changed the output transformer to a 220VCT/56V device for higher audio levels
    and have been able to pick up 40 meter SSB and CW signals nicely day and night. I've experimented with tube direct conversion
    receivers before and this setup is much closer to being a useful device, even while it is spread across my workbench.

    I'm using an HP8640B signal generator as the LO, I'd like to try using the 6AR8 cathode/grid as its own oscillator.
  9. KM1H

    KM1H Ham Member QRZ Page

  10. W0RIO

    W0RIO Ham Member QRZ Page

    I've been playing with the 6AR8 product detector project further, and am having some good results making a 40 meter
    tube direct conversion receiver.

    In short, what I now have is: antenna->pre-selector->6AU6 RF amp->unbalanced to balanced toroid transformer->6AR8->
    output transformer->LPF->audio amp. I'm still using the HP8640B as the VFO, driving a toroid
    step-up transformer to the 6AR8 grid #1. I moved everything from the proto-board to a metal box.

    The receiver works surprisingly well, the fidelity of SSB reception is really clear and nice to listen to. CW reception is also quite good,
    although it is still a bit wide, even with the LPF cranked down.

    I bought a pair of pin-compatible NOS 6JH8 tubes for $10 shipped off of eBay and those seem to be a bit more sensitive than the 6AR8
    tubes. The 6JH8 may be an improved version of the 6AR8, based on the dates from the RCA tube manuals.

    The best deflector bias (most sensitive) for the 6JH8 seems to be a few volts positive compared to a few volts negative for the 6AR8.
    One thing I discovered is that filament hum gets through to the audio, that was nicely fixed by running the 6AR8 filament on DC.
    Like all direct conversion receivers, a ton of audio gain is required, which explains the hum.
    Putting everything into metal boxes also removed another source of hum which was caused by stray pickup of the oscillator signal,
    another common problem with direct conversion designs.
    Interestingly, the balance issues I had with the DSB modulator circuit don't seem to matter as much with the product detector circuit.

    I still have a few a few more tweaks to do on the circuit, I'll post the schematic when the circuit is more mature.
    Unlike my previous experiments with a regenerative receiver, I could actually see this receiver being good enough to use in a QSO.
    The regen is still superior at receiving AM SW broadcast stations since it doesn't have the zero-beat problems of the direct conversion rx.
    KE5UFY likes this.

Share This Page