Question Re: 2 FET follower use

Hard question? I ask it wrong? Nobody cares? ;-)

Mikek

I think there are so many misunderstandings and loose ends that setting everything straight is just too much work.

Jeroen Belleman

d a

? output

tor

=0

No details specs, an answer. BW ATI, NF, Cable type , etc etc

Yup. That's about what I was thinking. Before I could start, I had to redraw the schematic and get a few things clarified. I cleaned up the schematic so I could read it: and started on a list of dumb question:

1. Is the "1" capacitor in farads, millifarads, nanofarads, microfarads, or picofarads?
2. The bifilar winding instructions in the upper right are for T3. There is no T3, but there is a T2 that's missing the instructions.
3. How did you manage to get -6dBm loss? It's a source follower, which doesn't have any voltage gain, but should have some power gain.
4. All the voltages in red are wrong. For example, if you power the amp from 12V, the J271 gate voltage should be 1/2 the 12V power supply voltage or 6V, not 5.2V. If you measured 5.2V on the gate, then the J271 gate is drawing current which suggests it might be fried.
5. Why a pot for adjusting the drain current on the J310, but no pot for doing the same for J271?

Only then did I re-read the question and discover that there are too many questions and oddities. I don't have the time.

Since this thing works only at 1MHz, it should be easy enough to build an LTspice model in order to see what the input and output impedances might be, and what effect input loading might have.

I don't believe that a pair of mismatched discrete FETs running at 10 mA has an IP2 of +88 dBm, or an IP3 of +41 dBm. +10 dBm / 0 dBm would be more like it.

Where in the world did you get that circuit?

Cheers

Phil Hobbs

T1 and T2 are the same. T3 is not relevant to this circuit.

I don't see where I said -6dbm. what I said was in the circuit I have the amp adds "9.8db of gain" I did use ~ maybe you thought it was a negative sign. Below, I clarify, Gain over the use of a transformer.

Ya, sorry about that, I recorded those voltages when I was troubleshooting the circuit.

The original design did have two pots, the author had the equipment to adjust for maximum IPs. I think he simplified it as most people can't set IPs any way.

The question is about the transformer, Do I calculate the inductance of the primary any different than usual since the secondary is unloaded? Should I just add the resistor on the secondary?

I already know and said, the input is about 330k? and the output os about 100?. Loading the input with 330? works fine and has a measured 9.8db gain, over the use of a matching transformer. I just realized I added that "over the use of a matching transformer", that should have been added previously, may have helped. To clarify, if I'm using the matching transformer on the antenna and have 0.1v output and then I swap the transformer for the amp, I then have 0.312V output. Thus my 9.8db gain. Thanks Guys, Mikek

Hi Phil, I can't vouch for any of the numbers, the designer is Dallas Lankford. He was well aware of of problems caused by intercepts, especially in the AMBCB, because there is a signal every 10kHz to cause havoc. I think he started building and experimenting with Norton noiseless feedback amps, in 93' or 94'. I have only started testing the antenna system, So far the amp acted as expected. In the little use I gave it, I didn't notice any intermod. I have other bugs I'm working on, Mainly feed line ingress. I'm making all the mistakes, so learning a lot.

Dallas was well respected in the AM DXing community and had dozens of MW antenna designs. The BCB is difficult because you have many stations on the same frequency and because of the wavelength, it is more difficult to make a directional antenna. He had a few designs with 4 antennas phased together. He had most of them on a yahoo group and then 6 or 8 years ago, he said, if you want anything on the site get now, I'm shutting down the site, he did and and I haven't heard anything about him since. I suspect the amp is better than you think, he was pretty meticulous about everything he did. Also a bit cantankerous. btw, here's the two trimpot version of the schematic.

Mikek

d a

? output

tor

=0

Where are your design specs and tolerances for the input/output?

Haven't you ever used a CD4000 series non-buffered gate as a linear amplifi er? Self-heating is highly dependent on Vdd so start low until it feels wa rm. Then it will self-regulate without runaway. It is obvious dependant on supplier and if it unbuffered or not. GBW also increases with Vdd until i t self-heats then it slows down as RdsOn decreases.

I've never tried it on 74HC series but that might work at 3.3V but too hot at 5V.

Link to SIM

Your original schematic at: shows "-6 dBm preamp gain" in the box at the lower right.

You can blow up a JFET with a gate voltage pot that goes the full range from +12V to 0V. Did you actually build the circuit per the schematic?

I still don't understand the question (or what problem you're trying to solve). Too much is unknown or ambiguous. What do you mean by "usual way"? Add what resistor on the secondary? Series or parallel? For what purpose?

Source followers just love to oscillate. Did you look for oscillations when measuring the output? The J310 is a UHF device that works nicely up to about 1GHz. If you build your BCB 1MHz amplifier as if it were a broadband amplifier (i.e. no 1MHz tuning), it's going to oscillate at some frequency because of the inductances of the long lead lengths. The ferrite beads might help, but in my never humble opinion, if you're going to use UHF devices at 1MHz, you'll need to bypass everything from 1MHz to 1GHz and probably do something to limit the gain above about a dozen MHz.

and a

? output

ll

up.

t.

istor

dl=0

fier? Self-heating is highly dependent on Vdd so start low until it feels warm. Then it will self-regulate without runaway. It is obvious dependant on supplier and if it unbuffered or not. GBW also increases with Vdd until it self-heats then it slows down as RdsOn decreases.

ot at 5V.

I don't know anything about him, for or against, but that IP3 spec is a good seven orders of magnitude off from credibility. Even with ideal parts, you can't set the balance pot any where near that accurately.

Cheers

Phil Hobbs

The ~1M wideband input impedance will have dc-to-daylight range, so anything from 60 Hz up will cross-modulate a small RF signal.

The usual way to match an antenna to a fet is a tuned network. That limits bandwidth and optimizes the impedance match.

I'd use one fet, with real gain.

If you add a transformer, the secondary *is* loaded... by the fets.

The 1K resistor makes about 4 nV/rootHz noise. But a CMOS gate ain't very quiet itself.

BF862

d a

? output

tor

=0

Maybe I'm showing my ignorance here, but the question I have is what is the purpose of this amplifier and the transformers you may or may not use?

Usually the front end of radios are designed for low noise and are pretty a dequate unless there is a problem of noise pickup in the feed. The battle is typically S/N ratio rather than voltage. The first stage of amplificati on needs to add as little noise as possible which means a design optimized for that, not gain. Even just 10dB of gain means noise added by any subseq uent stages will be minute in comparison.

So is this intended to be a low noise preamp? Is it needed because of some problem with the antenna system?

If you don't need a good match on the output side, just input matching and gain, why not use an amp with feedback?

Sorry Jeff, On my reader your post doesn't display, went to google groups for a copy and paste.

Your original schematic at: shows "-6 dBm preamp gain" in the box at the lower right.

range from +12V to 0V. Did you actually build the circuit per the >schematic?

******************************************************************** I have a transformer that is seeing a 330k? load, I can't put enough turns on to match that. So my primary will not see any load reflected back. That's my conundrum.

I just wound a transformer, it has a 4.3 to 1 voltage step up, 10 turns to 44 turns, it would match 330 ? to 6,250 ?. But I'm connecting it to a

I have not played it on the radio at this gain, that's for another day, I don't think it has any problems, but it might.

Source followers just love to oscillate. Did you look for oscillations when measuring the output? The J310 is a UHF device that works nicely up to about 1GHz. If you build your BCB 1MHz amplifier as if it were a broadband amplifier (i.e. no 1MHz tuning), it's going to oscillate at some frequency because of the inductances of the long lead lengths. The ferrite beads might help, but in my never humble opinion, if you're going to use UHF devices at 1MHz, you'll need to bypass everything from 1MHz to 1GHz and probably do something to limit the gain above about a dozen MHz.

Thanks for engaging with me on this, this project has taken a long time, I finally got signal from the antenna to the radio 3 or 4 days ago. My first RF was mostly from the feed line, then I added the 100? to 50? transformer at the radio. In my haste to hear it play, I just connected the CAT6 wires to the connector on the radio. That let all the common made signal through. I still have some feed line pick up even with the isolation transformer, I have some ideas to reduce it further. Here's a schematic drawing of the switching box. I added the voltmeter, might help understanding.

And a picture of the layout.

Thank, Mikek

There's a problem with that. You said you got a 1.8 voltage gain from matching a high impedance with the antenna low impedance. The voltage follower has 0 dB of gain. How could you get 3.2 voltage gain from using this amp???

Are you familiar with the theory of simple transistor amps? Did I misunderstand what you are doing? There is some sort of disconnect somewhere.

Yes you can place a 2970R resistor across the hi-z input so it presents a load for your matching transformer to reflect 330R at the primary. For a hint remember those RIAA phono preamps from the 19702-80s that had a hi-Z amp with a 47k shunt resistor?

I very much doubt the 330k input impedance of that follower, at 1MHz that would require stray capacitance to be below 0.5pF

piglet