|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
29th May 2017, 6:12 pm | #1 |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
RF signal path in this circuit...
Hi folks,
I have managed to talk myself into confusion over the RF signal path here... Isn't the 56pF capacitor in the output transistor effectively in series with half the capacitative voltage divider (8,000pF or 8nF) at at time across the magnetic loop antenna? ...I can't 'get' this! It can't be - since the series capacitance of even a large capacitor with another makes the sum smaller than the smaller capacitance, ie, less than 56pF, which would totally mess up the resonance calculations for this antenna. (Mine is resonant at around 10nF per half of the capacitative voltage divider and around 2.5uH). On the other hand, I can't see how the RF signal path can avoid this being the case. Doesn't the RF component evolve like this? 1) from the crystal oscillator (1MHz) 2) via the RF buffer, where it is also amplitude modulated by the audio amplifier transistor (does this count as a cascode but with an RF source and an AF source?) 3) do an 180 degree phase shift through the output transistor, and specifically via the 56pF capacitor... 4) and hence to the resonant magnetic loop antenna? Hence the 56 pF is in series with half the capacitative voltage divider, so around 55.7pF series total? Please note, this is not a query about capacitor series calculations, completely OT, please. It's about an RF signal pathway in this specific circuit. If anyone can be kind enough to talk me round this, I'd be very pleased. Thank you!
__________________
Al Last edited by Al (astral highway); 29th May 2017 at 6:37 pm. Reason: clarity |
29th May 2017, 6:37 pm | #2 |
Octode
Join Date: Sep 2004
Location: St. Albans, Hertfordshire, UK.
Posts: 1,477
|
Re: RF signal path in this circuit...
Isn't it effectively in parallel with one of the 8000pf caps? (via the 0.1uf decoupler).
__________________
Regards, Richard, BVWS member |
29th May 2017, 6:38 pm | #3 |
Octode
Join Date: May 2011
Location: Lancashire, UK.
Posts: 1,349
|
Re: RF signal path in this circuit...
RF design isn't my field but I would have said at face value that the signal fed to the antenna is pretty much solely dependent on the action of the output transistor, the 56pf being deemed necessary to simply optimise the operation, perhaps preventing instability or harmonics. The 56pf isn't 'in the antenna feed imo.
|
29th May 2017, 6:45 pm | #4 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
|
Re: RF signal path in this circuit...
That 56pF capacitor is in a feedback position around the right-hand 2n4401, so its capacitance is magnified by the gain. THEN there is the 100nF to ground from the base, and another capacitor in feedback on the left hand 2N4401.
At this point my selective aversion kicks in. It looks like someone had problems and kept adding capacitors to try to tame it. Some of them are in most peculiar positions. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
29th May 2017, 7:03 pm | #5 | ||
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: RF signal path in this circuit...
Quote:
I imagine the 27pF in the RF buffer was to prevent parasitic oscillations? I had problems with a higher-gain (RF-) transistor as the buffer to start with, not dependent on layout deficiencies. (I'm using the Manhattan style with glue-on pads on a ground plane). But the question in my OP remains a head-scratcher for me... Quote:
__________________
Al Last edited by Al (astral highway); 29th May 2017 at 7:20 pm. |
||
29th May 2017, 8:11 pm | #6 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: RF signal path in this circuit...
Isn't this just adding miller capacitance to tail the fT of the transistor off to prevent VHF oscillation? Also does the phase relationship between collector and base have any nulling effect?
|
29th May 2017, 8:17 pm | #7 |
Octode
Join Date: May 2011
Location: Lancashire, UK.
Posts: 1,349
|
Re: RF signal path in this circuit...
I just simulated this (rough and ready) and the cap appears quite critical in tuning the circuit to get the maximum amplitude and lowest distortion output. Here is 120pf and 56pf. I've attached the sim if anyone wants a play. (There is no RF load on the output).
|
29th May 2017, 8:25 pm | #8 | |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: RF signal path in this circuit...
Quote:
Now we just need a reason why it's so critical.
__________________
Al |
|
29th May 2017, 8:30 pm | #9 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
|
Re: RF signal path in this circuit...
Miller is the key work if you want to look it up. It isn't exclusively an RF thing, it works all the way down to low frequencies.
It doesn't quite change the Ft of the transistor, the transistor Ft stays the same, but the added capacitor adds an extra pole in the response which acts to filter the signal being passed, but being an extra pole it adds an extra 90 degrees of ultimate phase shift. If it had just moved the existing Ft value, it would not. Imagine we wiggle the voltage of that base sinusoidally by a little amount. Imagine that 56pf went to ground not to the collector. Knowing the wiggle voltage and frequency, we could calculate the current in the 56pf C. and the phase shift will be 90 degrees between voltage and current. Now, if we connect the capacitor back to the collector, when the base wiggles, the other end of the C shoots off in the OPPOSITE direction by a factor of the transistor circuit's voltage gain. Oooo-Errr! that means the voltage on the C is now the gain times the wiggle, and the current is thus scaled up by that factor. From the pint of view of the signal applied to the base, the signal is what it's always been, but the capacitor is now taking a lot more current.... the gain times more current. From the point of view of the signal at the base, the capacitor behaves like one to ground, but of gain times more Farads. This trick is used on ICs to make the effect of bigger capacitors without taking up too much silicon area. A famous example is the 'amplified capacitor' used to make the 10Hz loop compensation roll-off in the 741 opamp.... definitely not RF territory! David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
29th May 2017, 8:34 pm | #10 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
|
Re: RF signal path in this circuit...
What d'ya think I'm typing this on?
LT spice has a Mac version, but the user interface is awful. So my macbook pro has XP on it running in a virtualised window and LTSpice (windows version) runs on that quite nicely. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
29th May 2017, 8:35 pm | #11 |
Octode
Join Date: May 2011
Location: Lancashire, UK.
Posts: 1,349
|
Re: RF signal path in this circuit...
You're welcome
You have to bear in mind the simulation really is 'rough and ready' and parasitics aren't included plus the actual transistor models used will make a real difference as well but having said that the effect of varying the cap value seems pretty conclusive. For the real circuit (a real build) you really would have to look at the output and adjust accordingly I suspect. In other words you won't be able to pick optimal values from the basic simulation. Its critical because it is tuning the resonant collector load but all the parasitics and actual component values (such as coil resistance and inter winding capacitance) also play a part. |
29th May 2017, 8:54 pm | #12 | |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: RF signal path in this circuit...
Quote:
But not really satisfactorily with the resonant loop, as designed. Far from it. All sorts of horrible distortion and low output. Could you possibly please run the sim with almost the load shown but 2.5uH, a 5nF x 2 for the voltage divider and a centre frequency of 1 MHz?? That would be great if you can!
__________________
Al |
|
29th May 2017, 9:10 pm | #13 |
Rest in Peace
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
|
Re: RF signal path in this circuit...
The 56pF goes straight into a 0.1uF to ground so it won't feed much signal back. It is unclear to me what it does. It would appear in parallel with 8000pF, so have little effect. That whole circuit seems somewhat peculiar. Is the circuit diagram actually correct?
|
29th May 2017, 9:22 pm | #14 |
Dekatron
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
|
Re: RF signal path in this circuit...
|
30th May 2017, 12:52 am | #15 | |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: RF signal path in this circuit...
Quote:
Hah! Sounds like I should have another go at getting LT Spice running here! Seems to be some consensus for this view Dave... what else is attracting your attention apart from the spurious RF signal path to ground from the base of the output transistor via the larger cap you mention? The circuit is as I found it - Lawrence is quite right - although other variants of it have a tank circuit in the output collector. These variants dispense with the resonant magnetic loop, however, and show coupling to a 3m antenna via impedance matching and tuning components.
__________________
Al Last edited by Al (astral highway); 30th May 2017 at 12:58 am. |
|
30th May 2017, 7:50 am | #16 |
Dekatron
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
|
Re: RF signal path in this circuit...
It does seem a peculiar circuit. Assuming that the copper loop is resonant with the two 8,000pF capacitors, why is there a 10k resistor in parallel? It won't affect the DC, obviously, and at resonance I can see the loss of the tuned circuit being dominated by the other components, so adding 10k to limit the Q will have negligible effect!
|
30th May 2017, 8:10 am | #17 | |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: RF signal path in this circuit...
Quote:
i am resolutely determined to get this to work with a resonant magnetic antenna , however! Hence my unpicking of basics like the RF signal path first...
__________________
Al |
|
30th May 2017, 8:20 am | #18 | |
Octode
Join Date: May 2011
Location: Lancashire, UK.
Posts: 1,349
|
Re: RF signal path in this circuit...
Quote:
That seems better, now 56pf seems pretty optimal and in fact we can also go much lower with the cap. Even down to 10pf. Amplitude is essentially constant. I'll have a think about simulating the load. The copper loop is obviously anything but at short at the frequencies involved... that's the problem simulating that area... what do we substitute for the loop, some form of L,C,R combination I suppose. |
|
30th May 2017, 8:47 am | #19 | |
Dekatron
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
|
Re: RF signal path in this circuit...
Quote:
Try: source voltage 55V peak-to-peak, f=1Mhz sine wave Source Z 100k?? Or a bit higher ? Inductance 2.6uH Capacitance 5nF for each half of voltage divider No damping resistor That should be close
__________________
Al |
|
30th May 2017, 9:08 am | #20 |
No Longer a Member
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
|
Re: RF signal path in this circuit...
Hi Astral Highway.
Any talk of the Miller effect and Miller capacitance is not relevant to the right hand transistor or the 56pF capacitor, because the transistor's base is well grounded from the AC perspective with the 0.1uF capacitor. The power supply is also at ground from the AC perspective, assuming it is properly bypassed for AC with appropriate filter capacitors. Therefore the 56pF capacitor, from the electrical perspective, can be regarded as basically in parallel with the transistor's collector choke, or the 8000pF capacitor to ground. This is unlike the other 27pF base to collector capacitor on the left hand transistor that does add to the Miller effect, turns the transistor stage into an integrator, rolling off that stage's high frequency response. So while the circuit appears as a differential amplifier with the gain controlled by the modulator transistor, one other way to regard it; is that the right (output transistor) is merely a grounded base amplifier with drive applied to its emitter and output from its collector. I suspect that the circuit was the result of a number of cut & try practical experiments. For example if there was some inductance in circuit tracks in the area around the right transistor it could have caused instability and then the 56pF may have had some Miller feedback effects knocking out parasitic oscillations if the grounding of the base by the 0.1uF capacitor was poor, but it would have to have been a pretty poor layout, or bad components. This basic method of making an amplitude modulator by modulating the combined emitter currents of a differential pair is a very old technique. |