This is a copy-paste post of a long post from the old TheRadioBoard forums. The purpose of this test post is to see how a long post will be formatted in this new forum, to see if it is easily readable or not.
BEGIN PASTED CONTENT
-----------------------
by vladn » Thu Sep 26, 2013 11:27 pm
The original 6AS6 detector was a proof of concept design for a hybrid feedback in oscillators/regens. For newcomers to this board the development and evolution of the concept and the receiver can be found in this long thread:
http://theradioboard.com/rb/viewtopic.php?f=3&t=3714 Any regeneration receiver has an inherent oscillation amplitude stabilization mechanism. The 6AS6 detector was interesting in two ways: (i) it has more gain in the amplitude control loop (AAC) compared to classic detectors and (ii) the loop is external and explicit, while the regeneration amplifer section operates in a linear class A mode. This made it an ideal experimental testbed for learning the effects of AAC parameters on the oscillator/regen behavior. The results were interesting and not obvious (at least to me) and are listed in the theory thread here:
http://theradioboard.com/rb/viewtopic.php?f=1&t=4680 In short, the most important practical outcome of this study is that:
1. Using a scale-invariant control law in the AAC loop eliminates the bandwidth expansion with the increase in the oscillation amplitude.
2. Narrowing the AAC loop bandwidth with a variable RC filter allows for a wide range Q-factor control above the oscillation threshold and a full recovery of the Q-factor back to the level available at the threshold. Eventually, the 6AS6/6HA5 testbed turned into something loosely resembling a radio with potentiometers hanging in all directions. The motivation is to combine the results of both studies (the hybrid feedback and the parametric AAC control) in a coherent and practical way. Below is the circuit diagram of my testbed detector (the diagram below has a slightly better, albeit untested, implementation of the diode detector bias, but this is a very minor deviation):
6as6_3controls.GIF (11.57 KiB) Viewed 17611 times
6AS6 is a dual control pentode. The regenerative feedback is taken from the screen grid g2. The regenerative gain is controlled by the suppressor grid g3 DC potential. The pentode works at full (nominal) cathode current as a linear class A amplifier. R5 is not a grid leak ! The AAC loop kicks in long before the grid current appears at the g1. Note that the regenerative feedback gain increases with lowering the DC potential at the g3 and the g3 transconductance reaches maximum around -3..-4V relative to the cathode (not ground). R7 is the "original" regeneration control. It shifts the DC level at the g3 down until the oscillation starts. Then the detected signal amplitude is added to the preset level, the g3 potential returns back to the threshold DC value and the AAC loop stabilizes. Due to a fairly high gain in the AAC loop (pentode RF gain x Gm(g3)) the amplitude is almost perfectly proportional to the DC value at the R7 tap. However this is not a scale invariant control and the detector bandwidth monotonically and nearly linearly expands (and Q lowers) as you increase the amplitude. R3 is the "new" amplitude control that follows the scale-invariant law. As the R3 wiper moves down (as drawn) it decreases the AAC loop gain with a matching increase in the oscillation amplitude stabilization point, maintaining the constant relative sensitivity to the amplitude variations. A 1% change at 0.5V oscillation amplitude will produce the same delta gain applied via g3 as a 1% change at 5V amplitude. There is no measurable change in Q (and bandwidth) when adjusting oscillation amplitude with R3. Anti-log taper for the R3 makes the control more linear. Note that R3 has no effect on the oscillation threshold ! You can only use it above the threshold. This is a cause of a major inconvenience - you have to operate two inter-dependent controls (R7 and R3) with the possibility to get outside the acceptable range of amplitudes. I can meaningfully operate the receiver only with a scope attached to the detector output I guess one can add a high impedance DC voltmeter to this configuration . R1 is the bandwidth control of the AAC loop. It has a direct effect on the compression effects (the details can be found in the theory thread), the regenerated Q and hence the overall detector gain when the detector operates under the control of the AAC loop (above the threshold). R1 has no effect on the oscillation amplitude. Reducing the AAC loop bandwidth allows full recovery of the regenerated Q at any reasonable amplitude (as long as the pentode stays deep in class A). From the control theory perspective R1 determines the ratio of the integral and proportional terms in the AAC loop. Due to several reasons there are practical stability limits in this single tube detector that determine the range of adjustment for R3 and R1, so the R2 and R4 trims are needed. The simplest possible solution to the inter-dependent regeneration controls is to use the receiver as a pure synchrodyne (R7 removed, C4 trim set slightly above oscillation, say 0.3-0.5V at the detector across the tuning range). Then R3 is used as a frequency "lockability" control - the higher is the magnitude the harder it is to injection lock the oscillator. There is some rationale in favor of a pure synchrodyne:
1. With the new AAC loop bandwidth control it is possible to recover Q (and gain) back to the threshold levels at any reasonable oscillation amplitude. 2. One can have a continuously running digital frequency readout scale. 3. Diode detector biasing not needed and some circuit simplification is possible. ... n. And I usually use my regens as synchrodynes 99% of the time anyway . Here is a possible circuit diagram for such synchrodyne. This is the simplest variant of the 6AS6 AAC detector I can think of:
6as6_synchrodyne.GIF (11.32 KiB) Viewed 17611 times
However in few email conversations with RW4CMG he almost convinced me that in some specific cases it may still be useful to use the RX below the threshold. After struggling with various options (like potentiometers with integrated switches for interlocking the two controls) I came up with the compromise solution: adding a small DC potential to the R3 via the detector bias network. This makes R3 into a joint threshold + amplitude control that changes both the threshold and the amplitude. It will not be mathematically optimal (like the separate controls) but if the added threshold tilt is small it will hopefully have a negligible effect on Q, yet allow to kill the oscillation at the lowest setting. Here is one possible implementation.
6as6_comb.GIF (11.02 KiB) Viewed 17609 times
If desired this circuit can be converted back to a pure synchrodyne with a single SPST switch (by shorting the LED) pretty much without the threshold readjustment. Finally, my original circuit uses a 6HA5 "super-triode" for the RF amplifier with AGC. I like the AGC action in that radio. However the original 6AS6 detector was reflexed for the AF amplification and the new one is not. I had to think over the RF stage so I can somehow add AF reflexing to it without noticeable dynamic range degradation. After some thinking and modelling in LTSpice I settled on a dual triode cascode stage, where GC and GG sections are AC coupled for RF. The GC section has an AGC applied to it, while the GG section runs at a fixed bias and a nominal plate current and is reflexed for the AF:
cascode_agc_reflex_1.GIF (10.4 KiB) Viewed 17611 times
C9,R6,D1,D2,C2 is the AF-derived AGC circuit. R2 serves the dual purpose of a grid leak bias for U1 and the AGC release. The coupling between the GC and GG stages can alternatively be done with a broadband xformer instead of RFC1,RFC2,C5. I had some difficulty with the AGC implementation in my 1t HFRR. Wide variation of the plate current in the RF stage due to the AGC action slightly detuned the regenerator, most likely via the change in the pentode plate impedance and/or space charge related capacitance. It took a lot of effort to compensate for it (RW4CMG suggested using a small compensating varactor controlled by the AGC voltage and that finally did it). The same detuning happens with a 6HA5 RF stage but to a lesser extent due to very low matching impedance between the 6HA5 plate and the detector tank. This explains the decision to use the cascode configuration and run the GG section with a fixed bias in a new design. I did quite a bit of simulation on dual triode cascode circuits with an AGC. As long as the load impedance for the GC section is much lower than the triode own internal plate impedance, even the general purrpose triode provides a decent range of gain adjustment. But there are dedicated semi-remote triodes optimised for this function (like 6ES7, 6BC8, 6BZ8 etc.). Since (i) the GG section runs at fixed nominal plate current, and (ii) the GG section input impedance is much lower than the GC output impedance for the RF (if a conservative low impedance matching is used between the GG section output and the regenerated tank) the GG stage is very linear and exhibits minimal intermodulation, which is very important in a reflex operation. There should be no direct detection issues that I experienced with the 1t HFRR receiver RF stage.
Last edited by vladn on Wed Feb 05, 2014 9:59 am, edited 1 time in total.
OK, a couple of things I noticed when copy/pasting that old post into this new forum:
When using the arrow keys to navigate through a very long text, the cursor sometimes jumps to unexpected locations in the text.
When copy/pasting images, sometimes the wrong image was pasted.
When submitting this post (a long text with several images), the forum asked me to verify I was human and I had to do a captcha test and click on various images. This was rather annoying.
It is possible to lose the entire post if you accidentally press backspace at the wrong time, causing the browser to go to the previous page. Normally I would expect an "unsaved data" warning from the browser.