ad: elecraft

AM to PM conversion in plate modulated transmitters

Discussion in 'Amplitude Modulation' started by KA9Q, Oct 10, 2019.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: Left-2
ad: Left-3
ad: L-MFJ
ad: abrind-2
  1. KA9Q

    KA9Q Ham Member QRZ Page

    I've been studying WWV's signals for some time, and I observe significant AM-to-PM (Amplitude Modulation to Phase Modulation) distortion on their 5, 10 and 15 MHz transmitters (especially 10 MHz). These three transmitters use high level plate modulation and produce 10 kW. Their 2.5, 20 and 25 MHz transmitters use 2.5 kW linear amplifiers and are clean.

    Seen on an I/Q (inphase/quadrature) display, an AM signal should appear as a flat line on the +I (horizontal) axis. The unmodulated carrier appears as a dot in the middle of the +I axis, and modulation spreads that dot left and right into a line that never goes past the origin. That's what I see on local AM broadcast stations and on WWV's 2.5, 20 and 25 MHz linear transmitters. But the three plate-modulated transmitters (especially 10 MHz) have I/Q displays that look like bananas; as the amplitude increases the phase advances as much as 45 degrees or more.

    The fact that this happens on all three transmitters makes me think it's an inherent property of AM plate modulation of vacuum tube amplifiers. I need to brush up on vacuum tube physics, but I think the varying B+ changes the distribution and density of electrons within the tube, varying the plate-to-grid capacitance and thus the phase shift of the amplifier.

    Anybody out there with a plate-modulated AM transmitter and some SDR equipment want to test this?

    I first noticed this phenomenon as noticeably increased audio distortion when receiving WWV on 10 MHz with a synchronous AM detector vs an envelope detector. Have any of you noticed this kind of distortion?

    Screenshots of I/Q displays of all six WWV signals made with a local Kiwisdr in Ft Collins CO are on my website here:
    Last edited: Oct 10, 2019
  2. N2DTS

    N2DTS Ham Member QRZ Page

    I thought that was caused by the interaction of the final and the VFO (frequency reference device)
    You can sometimes see AM signals with more audio on one side of the carrier (old tube gear) which I have been told is phase modulation of the VFO.....
  3. SM0AOM

    SM0AOM Ham Member QRZ Page

    I would say that it is very difficult to change the frequency of the phase-locked sources used in the WWV transmitters.

    The kind of asymmetrical AM that once was common in transmitters of yesteryear was primarily caused by VFO frequency pulling from modulation. In my first 144 MHz transmitter this was noticed due to the use of unstabilised B+ for an overtone crystal oscillator.

    AM to PM conversion was sometimes an issue in amplifier klystrons that were common in microwave relay systems of yesteryear, and was caused by detuning of the cavities by the modulated electron stream.

    WD4IGX and KA9Q like this.
  4. KA9Q

    KA9Q Ham Member QRZ Page

    Yes, the 5/10/15 MHz transmitters at WWV are driven by a very clean and stable 1 MHz sine wave multiplied up to those frequencies. NIST SP 250-67 contains a pretty detailed description of all the transmitters at WWV/WWVB/WWVH, and it's how I know that the 5/10/15 MHz transmitters use high level plate modulation. There can't be many other places where the distortion happens. Those three transmitters were made by CCA and installed in 1990, but they don't give the actual model numbers so I don't know the actual tube type or schematic of the final amplifier. They produce 10 kW with 54% DC-RF conversion efficiency, so it probably has something like a 4CX15,000 (a 4CX10,000 in AM service cannot handle the 5.5 kV stated to be the plate voltage). I'd bet this phenomenon depends on the number of tube elements (triode, pentode, etc). Pentodes seem most common in large HF transmitters.

    AM-to-PM conversion is indeed a common phenomenon in microwave amplifiers. I know about it mainly in the traveling wave tube amplifiers that are still commonly used in communication satellite transponders. Digital signals incorporating AM as well as PM (e.g., QAM) sometimes have to be pre-distorted at the uplink transmitter to compensate for the effect.
    N2EY likes this.
  5. WA3VJB

    WA3VJB Ham Member QRZ Page

    A few years ago, NIST added a phase component to their signals as part of a new data detection scheme for receivers. I have one of the first consumer models of "atomic clock" using the newer method, and it really is vastly superior to other clocks I have using only the older method of synchonization.

    Some details are in this main story, and the additional links embedded in that story near the end.
  6. N2DTS

    N2DTS Ham Member QRZ Page

    CCA used to be local, in a few South Jersey locations before they folded.
    Cherry Hill was the last location I think.
    We used to dumpster dive there and I may still have some panel meters from there.
    Their stuff seemed low quality....
  7. KA9Q

    KA9Q Ham Member QRZ Page

    That's on WWVB's 60 kHz signal. DCF77, their German counterpart, has had something like it for a while and it has worked well.

    It's really hard to modify a signal like that while retaining complete backward compatibility with older receivers. It was also done with FM stereo and NTSC compatible color, both were real achievements in their day.
  8. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    If the carrier of an AM signal is shifted 90° relative to the sidebands, the signal is converted to pure PM. That's how Armstrong built his first experimental FM transmitter back in the 30s. He generated a DSB suppressed carrier signal at some low frequency below the broadcast band, and then re-inserted a 90° out-of-phase carrier. The resultant PM signal was then converted to FM by passing the modulating signal through a simple RC filter with a 6dB / octave roll-off. The deviation of the resultant FM signal was very low, but when multiplied up to the operating frequency, somewhere in the vicinity of 40 MHz, the signal was wide-band FM.
    WD4IGX and KA9Q like this.
  9. KA9Q

    KA9Q Ham Member QRZ Page

    Right. PM is also used in deep space communications, though not as much as it used to. The Apollo Unified S-Band (USB) system used PM at a low modulation index, leaving lots of residual carrier for Doppler tracking. As you say, the sidebands look just like AM except that they're shifted 90 degrees. This is easy to see if you look at PM as the sum of a constant carrier at 0 degrees plus a suppressed-carrier DSB signal at 90 degrees. The modulation index has to be low enough that the two signals sum to one with an approximately constant amplitude.

    Now this has me wondering just what you get when you simultaneously modulate amplitude and phase in a more controlled, cleaner manner.
  10. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    SSB. That's how phasing exciters work.

    Sometimes this happens accidentally with an AM transmitter, resulting in one sideband being noticeably louder than the other. Because of poor isolation of the VFO from the modulated output stage, some RF inadvertently gets fed back, causing phase modulation of the AM carrier.
    WD4IGX likes this.

Share This Page