Technical Topics > The RF Workbench

5 watt crystal oscillator transmitter

<< < (2/3) > >>

makeRF:
OZNRH, thanks for the input!  I added two 100K ohm resistors as a voltage divider on the IRF510 gate for biasing.  This seems to have done the trick.  I also tried the 0.1 uF cap and this too worked just as well as the 1000uF capacitor.  I now measure 4 watts RF going into a dummy load.

I'm more pleased with this circuit, because it's less of a hack.  I was scratching my head when originally playing with this wondering why I couldn't bias it and make it work.  I must have had something plugged in wrong at the time.

ff:
Hi MakeRF.  Let me join the dogpile here.  One thing I would add is a 0.1 uF cap to ground on the V+ side of your 3.3 uH choke.  That will work with the choke to help keep RF out of your voltage feed.  Also, using two 100K as a divider sets your gate bias point at 6V.  In this situation there's no reason to make the IRF amplify linearly.  The higher the bias point, the hotter the IRF will run because it will be conducting for longer portions of the cycle.  A single gate resistor to ground will set the gate impedance.  I normally use one between 1K-10K, but that's just my fetish :).  And Ole makes great points - use a 0.1 uF coupling cap in the drain circuit followed by a C-L-C LPF.  This will improve your output waveform by reducing the harmonic energy. I love your site!

makeRF:
Hey ff, thanks for the comments, and I appreciate the kind words about my site!  I'm planning on getting an improved version of the article out with these suggestions (with credits to this forum!).


--- Quote from: ff on July 31, 2014, 1354 UTC ---One thing I would add is a 0.1 uF cap to ground on the V+ side of your 3.3 uH choke.  That will work with the choke to help keep RF out of your voltage feed.
--- End quote ---
I added a 100uF capacitor at the RFC as a voltage stabilizer.  Would the 0.1uF be enough capacitance?  I guess it depends on the stability of the power supply.  I've usually used a higher value electrolytic as voltage stabilizers.  One structure I've seen in similar circuits, is a 100uF capacitor and a 0.1uF in series at the RFC.  I assume there is a reason for this, but it just looks like a 99.9uF capacitor to me.


--- Quote from: ff on July 31, 2014, 1354 UTC ---Also, using two 100K as a divider sets your gate bias point at 6V.  In this situation there's no reason to make the IRF amplify linearly.  The higher the bias point, the hotter the IRF will run because it will be conducting for longer portions of the cycle.  A single gate resistor to ground will set the gate impedance.  I normally use one between 1K-10K, but that's just my fetish :).
--- End quote ---
For some reason my IRF510 stops working if I just add a single resistor bias on the gate.  If I lower the voltage with the divider it starts working.  I'm really not sure why this is, as I believe it's suppose to operate with a 12v input.  I might edit the divider though to bring the voltage up to +9 or +10v and see if it still operates.  Maybe I have some knock off FET's, although I bought them from a reputable source.


--- Quote from: ff on July 31, 2014, 1354 UTC ---And Ole makes great points - use a 0.1 uF coupling cap in the drain circuit followed by a C-L-C LPF.  This will improve your output waveform by reducing the harmonic energy.
--- End quote ---
Yes, for sure if I was to use this in any serious way.  I've plugged it into the altoids low pass filter I've previously created just to see how it cleans things up, although that filter is for 40m, so it's not really ideal.  It is improving the waveform on the harmonics in that range though.

At the moment I'm just using the canned crystal oscillator as a simple RF source to amplify.  I don't plan to keep it in a final design, but I do find it interesting to be able to max the power output of these things.  I usually build things out in stages, and then combine into something more useful -- It's all a learning experience for me.  One project I have in queue, is to better understand some standard crystal oscillator designs, and once I'm happy with a configuration, combine the two into a QRP transmitter.  Baby steps  :)

ff:

--- Quote from: makeRF on July 31, 2014, 1422 UTC ---For some reason my IRF510 stops working if I just add a single resistor bias on the gate.  If I lower the voltage with the divider it starts working.  I'm really not sure why this is, as I believe it's suppose to operate with a 12v input.  I might edit the divider though to bring the voltage up to +9 or +10v and see if it still operates.

--- End quote ---

Yes, my bad.  Your oscillator probably doesn't have enough signal voltage output to drive the IRF into conduction.  Lacking a driver stage, the divider is probably the easiest way to go.  The 510 should be driven into saturation with about 8V on the gate (they do vary a bit) so you should keep the bias point south of there.  Personally I'd try about 2V initially (like Dave Martin's Corsair) and increase it gradually to determine what tradeoff of output vs waste heat you are willing to have.  And yes, your 100uF is probably enough filtering.  However good construction practice calls for another cap on the high side, and physically close to the choke, with a timing constant more suitable for RF, like 0.1uF.  That one is your call, but redundancy is good policy when playing with RF circuits.  And I always take baby steps when playing with my circuits too... 73!

OZNRH:
Hi again here :).

Here is a diagram more after my head :D.. As 'ff' write, the signal from the oscillator is too 'small' to driving the FET. Right now it work nearly as a kind of class C with a 'to small' input, so yes a 'biasing' will probably help. Take a look on my diagram as inspiration.

Ole :)..

Navigation

[0] Message Index

[#] Next page

[*] Previous page