Need help with 2 meter receiving converter

[metasequoia]

Hello all,
In 1973 I built a 2 meter receiving converter from the ARRL Handbook.  At the time I had inadequate test equipment and knowledge and never could get it to work and set it aside until this week.  Now I have better test equipment but am still not smart enough, hence my plea for help on this forum. I have attached the schematic and hope it is clear enough.  I started my troubleshooting at the 58MHz oscillator section- no oscillation.  The original MPF102 JFET was shorted so I tried a 2N5457- no help and then a 2N3819- no help.  I can clearly see the parallel and series resonant crystal frequencies when I connect the crystal to my spectrum analyzer/tracking generator.  I have also adjusted L19 so L19/C55 resonate at 59MHz as recommended in the Handbook.  The drain of Q9 is stuck at 9V.  I also tried increasing the value of R28 from the suggested 100K to 2M-- no help.  I would like to keep the circuit as close as possible to the original schematic and would greatly appreciate any suggestions.  Thank you.

[LM21]

First, Q9 drain should be at 9 volts because there is 9V zener CR8 to keep it there. The oscillator looks very simple, perhaps it works, perhaps not, but try tuning L19 and C55 to crystal frequency 58Mhz, not 59MHz. Have you kept the circuit tight, no long wires between components.

[chilternview]

The only way that circuit will oscillate is via feedback from the drain to the gate of the FET. So it relies on Cgd of the FET, and possibly even the layout, to get that feedback.

Now on first sight it would not seem to oscillate, as the FET provides 180 degrees phase shift between input and output, and the Cgd will give 90 degrees. So it would appear a bit short of the 360 degrees to maintain oscillation.

But in practise the 180 phase shift is only true for a purely resistive load; we have a tuned circuit in the drain side. So the drain voltage can lead current by more than 180 degrees, and the same can happen in the gate tuned circuit due to phase shift of the crystal. With the Cgd phase shift and at the frequency of oscillation, the total adds up to 360 degrees. So that parasitic gate-drain capacitance is critical. Try adding more e.g. using a pair of twisted wires between gate and drain.

Alternatively try the crystal between gate and drain, in series with a capacitor say 1nF.

[metasequoia]

Thank you for the replies.  No long wires as shown by the attached pictures.  I forgot to mention that I had already tried 18pF and 100pF caps between Drain and Gate.  I will try your suggestion of moving crystal location.

[dmowziz]

Hello all,
In 1973 I built a 2 meter receiving converter from the ARRL Handbook.  At the time I had inadequate test equipment and knowledge and never could get it to work and set it aside until this week.  Now I have better test equipment but am still not smart enough, hence my plea for help on this forum. I have attached the schematic and hope it is clear enough.  I started my troubleshooting at the 58MHz oscillator section- no oscillation.  The original MPF102 JFET was shorted so I tried a 2N5457- no help and then a 2N3819- no help.  I can clearly see the parallel and series resonant crystal frequencies when I connect the crystal to my spectrum analyzer/tracking generator.  I have also adjusted L19 so L19/C55 resonate at 59MHz as recommended in the Handbook.  The drain of Q9 is stuck at 9V.  I also tried increasing the value of R28 from the suggested 100K to 2M-- no help.  I would like to keep the circuit as close as possible to the original schematic and would greatly appreciate any suggestions.  Thank you.

Hi... I think the oscillator should work according to the schematic.  Assuming I know what I'm doing

Are you using an original MPF102?
Looking at the datasheet, I filled in a small signal model. (Ciss ? = Cgs ?) which is the first image

The oscillator works (assuming I know what i'm doing) by creating negative resistance at the gate
The 2nd plot shows this


How are you checking for oscillation? If you load the drain or gate with 50 ohms (input of spectrum Analyzer), the negative resistance will be lost. You can use a 1 pf into 50 ohms

I can clearly see the parallel and series resonant crystal frequencies when I connect the crystal to my spectrum analyzer/tracking generator

For analysis, it is good to use this data to form the classic RLC model of a crystal


It will mean the crystal is operating in inductive region and it is also on harmonic (The ESR will likely be greater than the negative resistance) , adding a series inductor to it will be helpful.. around 325 nH in this case ?

https://www.vintage-radio.net/forum/showthread.php?t=147265&page=2
This nice thread analyzes these kind of oscillators



I'll have a go later today at making the oscillator to confirm I have a little idea.

[metasequoia]

Thanks for your reply.  Original MPF102 has an open Gate so I have tried 2N5457 and 2N3819.  I'm going to place an order for more MPF102 today.  I am checking for oscillation with X10 probe connected to my scope.

[dmowziz]

Just to confirm again that the original circuit is ok and should work...


I have a crystal marked for 100 MHz... Built a circuit using a MPF102 from Aliexpress for 100 MHz, it did not work.
Please where are you buying the MPF102 from?


I thought to use an original smd MMBFU310, built the circuit in the image below for 100 MHz, and realized it oscillated at 60 MHz with very low power.
Initially thought the closest I had to 58 MHz was 100 MHz, I forgot about (20 *3). So adjusted the inductor value and got the spectrum below..


This circuit accuracy now is poor, and 24.7 mA for just -3 dB isn't ideal (?)  This just proves the circuit works.


All these took about 4 hours but there was confidence that the circuit must oscillate!   If you care to know why :


Breaking the circuit at the gate and looking into the gate (using a drain resonator of 39 pF and 62 nH) and assuming the spice model for MMBFU310 is accurate :
It looks like a series reactance of about -26 ohms (60.5 pF) and series resistance of about -105 ohms..  At 100 MHz
All we need to add to the gate to get oscillation is a series RL  circuit with L = 42 nH  and |R| < 105 ohms

So the crystal has to look like an RL (operating in inductive region)
The crystal below ESR is 45.546 ohms when operating at series resonance (The crystal looks like a resistor).

In inductive region, the ESR =  45.546 *(1 + (3.7592/60.5) )^2    = 51. 38    < 105
So surely, the crystal must oscillate at 100 MHz
But it oscillates at the first harmonic (60 MHz)... There's usually a parallel LC circuit at source to prevent this




The circuit is fine.
If you can, maybe try and build only the oscillator, dead bug style, tight layout and see ?
Your capacitors look original.
and are you certain the resonance is at 58 MHz?
adjust L19 with eyes on the scope. A 1pf feeding your analyzer will be good



Will have a go tomorrow using the aliexpress MPF102 at 60 MHz, dead bug style


Thanks for the circuit

[szoftveres]

Not sure if it helps, but if the issue is too low milller capacitance then maybe try adding a 1-2pf cap between the drain and the gate of the oscillator FET.

Also, how much current is going through the oscillator JFET and what are the voltages on it in its non-oscillating state?

Also, this style of crystal oscillator is easy to build with bipolars - I've built a similar with 2N3904 which is easier to find and the different makes are more consistent - the oscillator doesn't necessarily need to be made of JFET.

And one more thing: if anything on the output side loads the LC tank down then the oscillator won't start. Totally mistuned LC tank, too low resistors, forward-biased gate-source PN-junction or totally shorted doubler JFET, etc...

[metasequoia]

dmowziz, I am sorry you have had to put so much time into helping me, but I appreciate it!
I don't remember where I purchased the original MPF102- that was 50 years ago!
I just ordered NOS (new, old-stock) MPF102 from a seller on eBay who has a static shielding bag full of them from Digikey with a date code of 2003.  They were only $.50USD each with $4 shipping which seems very reasonable.
I am quite sure that the crystal is resonant at 58 MHz and the LC tank is tuned to 59MHz- I just checked this again with my spectrum analyzer and tracking generator.
C56 couples the 58MHz signal to the rest of the circuit and for all of my testing I have disconnected one end of C56 so even if downstream circuitry is shorted, the oscillator won't care.
With the circuit powered by 12V supply, oscillator FET is a 2N3819 and the drain current is 4.8mA
Thanks for all the help!