Ok, I didn't do the math, but I see the gate/source impedance is not so high. But the resulting impedance of C3 is still very low even at 3.3 nF so the change should be very small from the gate to ground. A 2 dB drop is around 25%. This seems far too much to explain this way... except is the impedance of the bias network on the source amplified by the beta of the FET? It has been a long time since I've studied these things. I need to dig out some text books.
Where, at the output? I get 82 dB gain with the existing circuit. The goal is to keep the supply current as low as possible.
I meant to include the model I was using...
Or you can probe around. For example with 3300p you see 38dB at the source of T2 at 60kHz, with 0.033u you see less than 18dB. That difference is huge and with 10p for Cgs representing less than 300k at
60kHz that has to translate back through to the load of T1 (which sets its gain). It's like hanging a chain and ball to the source of T2. [...]
I think you mean the MC1330.
The MC1430/1530 is my first OpAmp design (~1963-64) and is still being sold by Lansdale... and it's MIL-qualified. ...Jim Thompson
To start with, it looks to me like the first stage is saturated.
On Wed, 06 Apr 2016 08:40:50 -0700, Jim Thompson Gave us:
moot point. The military uses COTS now.
Only aerospace projects use the high end stuff any more (typically, with a few exceptions).
I measure 625.74mV across the drain-source. Is that not enough? The input signal to this stage is low microvolts. Raising R3 to 300k still seems to work fine with the drain-source voltage closer to 500 mV.
My only concern with saturation is that there may be interference that needs more headroom to prevent distortion and intermods.
I'm not sure why I put this off for so long. I've got some BF862s on order and will try constructing this when they come in next week. Then I can see what sort of signal I'm picking up with the loop.
My concern is that the whole 80dB amp is wideband. Usually it is best to place the crystal filter as far up front as possible. Else you might see snuffing effects by all sorts of external noises. Fat transients when the vacuum or fridge compressor is turned on, noise from rogue switch-mode supplies, CFL, LED and so on. There are nights where I can barely decipher WWVB with a 300Hz crystal filter in my lab receiver. Then there are nights when even a
I am at the edge of the 100 uV/m region and my antenna has an effective height of about a quarter meter, so 25 uV if everything meets the theoretical optimum. East coast, near DC. Many commercial products won't work correctly here.
The antenna is *not* wideband. The bandwidth is expected to be less than 1 kHz. Even if there is noise at times, proper reception only requires 1 minute of continuous signal. With some intelligence multiple messages can be used to verify the current time setting in the clock. Likewise, multiple second boundaries and minute markers can be used to set the precise time.
This is not so different than the NRO using many atomic clocks to establish the "correct" time via a Kalman filter.
Yes, it was a typo. I was thinking of the 1350, 50 dB TV IF amp IC. The MC1330 was the video detector I.C. that followed the IF amp. I am having problems adjusting to changes to my medication for Diabetes, and I didn't bother to verify the number. :(
The new modulation scheme instituted by the NIST added phase modulation which makes it incompatible with a lot of designs for the original -10 dB level change in the 60 KHz carrier time code detectors. It also makes it less than useful for a frequency standard.
See WWVB-Schematic+Data.pdf on the S.E.D/Schematics Page of my website for a level-change WWVB signal receiver. ...Jim Thompson
I did mine in software, back in the Commodore 64 days. :)
Such digital tricks can work as long as you don't have any in-band noise of the rat-tat-tat kind. This is one reason why I got rid of all our X-10 home automation stuff last weekend. They thought that a 1kHz plus BW AM transmission protocol would be fine. It isn't.
The model predicts 80dB at 10KHz falling to about 48db at 60KHz.
Fig.6 of the data sheet "Forward transfer admittance as a function of drain current; typical values".
Shows the gm at low currents -> 0 . The Spice model may not be accurate down there. Spice models are for amusement only.
I could be wrong but the graph makes me nervous
Brian
I don't think there are many that are incompatible. What it messes up is the carrier lock that is used for a frequency output. Most receivers only look at the AM time/date code in the data stream which still works just fine.
X-10 isn't locked to a frequency is it? I had one many years ago and I seem to recall it was about 100 kHz, but set by an RC, so not very consistent between units or stable. It also didn't work so well.
Dual gate MOSFETs are a ready made, all in one package casc_o_de, and they're not quite as bad as JFETs for all over the place parameter spreads.
In most cases, you can set Id to whatever you want with the potential applied to G2.
MOSFETs liberated from a TV tuner have very low capacitance - especially from a UHF tuner.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required