Super TGSL With H Bridge TX Diff RX Quad Sampling

[Oessi]

C7=470p R15=6k8

f=1/(2*3.14*C7*R15)=
=0,16/( 0,00047*0,068 ) =5000 ????

[Davor]

As I remember the R15 value was 3k3 before I increased it to 6k8, and obviously I failed to scale C7 and C9. I concluded that the noise improvement was not worth the extra current it required.
So yes, C7/C9 should be reduced.

[Oessi]

Tx=12,5Khz C7=C9=47p C6=C11=47n
It is impossible to apply such face values of condensers at which a fracture of phase-shift-frequency is near working frequency detector. In strengtheners it is necessary, what the phase of the accepted signal suffered a minimum of changes.

[Oessi]

Boys, in your chart very much many errors, I see them, but them so many, whatever I even want you about them to tell. If to do a detector on your chart, nothing there super will not be, it will be very badly to work. Count the face values of condensers even

[Davor]

As stated at several places before, this is a place to exchange ideas on how to build a next generation of TGSL-like device, not the recipe itself. I for example can't afford much time to this at the moment, so I simply give a suggestion or two now and then. As for my taste, there are several things that need redo, but again, this is a work in progress.

Regarding the Rx running at resonance, it is intentional. If you pay close attention to the front end, you'll notice that configuration is series, not parallel, and unlike the parallel tank this one has minimum phase shift at the very resonance. Due to a low Q factor of this configuration, the resonance is very broad, and phase shift varies very little. Unlike parallel resonance this one does not increase the voltage of the induced signal, and the low noise depends on the transistors and their low Rbb, In this configuration it should be less than 1.5nV/sqrt(Hz), which is lots better than a typical op amp front end.

Now, you seem to have a keen eye for detail, and I appreciate your contribution so far, so if you are into some creative electronics, feel free to suggest your visions as well. Silver Dollar is running the show.

If you wish to build a well established device, go for a TGSL or IDX projects that also have PCB-s ready, and eventually you'll be able to include some of the solutions you find here as mods for your working device. Fun does not stop when you build your gadget

[Oessi]

Dear Davor, I well see that inductance of Rx does not have parallel resonance, and it very well. But the capacity of condensers of C7 and C9 is small, it is necessary to increase in 2 times, otherwise there will be a change of phase on working frequency of detector. Operating strengtheners are also intortus placed, OP7 must be changed placed with Tl72. Then I yet will write that needs to be corrected.

[Davor]

As I said, C7 and C9 are supposed to be reduced, and it will be done in some revision for sure.
The phase is already taken care of in a way that this configuration will accommodate a wide range of front end phase shifts, so this is not an issue.
A choice of the front end op amp is almost arbitrary at this point, as long as it provides a good THD, because the most of the voltage gain, and the noise performance, is handled by transistors. Every op amp with input voltage noise less than ~30nV/sqrt(Hz) wil do fine, provided it does not ruin the signal by intermodulation distortion. Most of the opamps share the same pinout, so replacing the opamp is easy.

BTW, there is a hint of aggression in your posts that I do not entirely understand.

[Thomas]

It might be his use of the word "Boys" that comes across as condescending, but I don't believe that he uses English often, so it could be a translation thing. Or maybe not.

[Oessi]

Dear Davor, I beg pardon the not sufficient knowledges of English. I not very much well own them, from for it you can me not quite to understand. I live in Ukraine. I use computer translation for intercourse.

[Oessi]

[Davor]

Yes, it can certainly work that way, but I prefer the "superbal" configuration, or alternatively AC grounding the inverting path with a large capacitor. A superbal, or Birt configuration if you like, is better because of the equal loading of both inverting and non-inverting sides for both common mode and differential signals.

There's not much difference in the end-result with using a regular differential amplifier with the inverting one following it, and a superbal. In both cases you use 2 op amps, and in both cases you obtain symmetrical output, yet the superbal equalises input impedance on both branches, and a regular differential amplifier does not.

[Oessi]

I counted the face values of condensers, it must provide a zero phase of the accepted signal on working frequency of detector

[Davor]

More or less, yes. My whole idea of this configuration is providing the front end that can accept frequencies within an octave of frequencies with no difficult re-tuning. Some phase deviation in both directions is acceptable, as the phase shifting network compensates this deviation prior to extracting phases for discrimination etc.
So the idea is to set a coil/frequency and first tune the ferrite reference. Once it is tuned by a piece of ferrite, the rest of controls are set at virtually the same positions as for the previous coil/frequency.

[Oessi]

Will such configuration of advantage give on screening? Is it necessary to screen a graphite or alfol of Rx?

[Davor]

The idea of using a well balanced Rx front end with a centre tapped coil is in fact avoiding a shield completely. You can apply shielding, of course, and it will not harm detection, but avoiding it completely makes coil making a much more enjoyable activity.
If either Tx or Rx, or both are in balanced configuration, the coil balance improves. This is because in a balanced configuration the capacitive coupling is cancelled by opposing phases. Inductive coupling remains the same.