The AM Forum
March 19, 2024, 06:09:35 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: 1 ... 4 [5] 6 7   Go Down
  Print  
Author Topic: Building my first oscillator  (Read 151750 times)
0 Members and 1 Guest are viewing this topic.
W4AMV
Member

Offline Offline

Posts: 672


« Reply #100 on: September 13, 2011, 08:07:55 AM »

You did not mention the trifilar transformers... the material used to construct... at these low frequencies to get a working trifilar xmfr is going to require a pretty high permeability, say 850-2200 or higher. With only a few turns of trifilar wire, you will be hard pressed to achieve suitable inductive reactance relative to 50 ohms. The mixer will be very lossy.
Logged
DMOD
AC0OB - A Place where Thermionic Emitters Rule!
Contributing
Member
*
Offline Offline

Posts: 1768


« Reply #101 on: September 13, 2011, 02:08:04 PM »

Quote
i got this brilliant idea to replace the 2n3866 voltage divider circuit with 15k pots.

My experiences with the 2N3866, 2N5109, 2N3553, or NTE 278 or NTE 311, or NTE 473 buffers/voltage amps and linearity at 12-15 V ARE:

I use a base bias of approx. 2.5 mA composed of a bypassed 3.3K (.05 uFd) from the collector, then go to a 560 ohm to base; and then a 1k from base to ground. This way, you get both AC and DC feedback for stability and improved linearity. Any lower values of base voltage divider resistances and the base impedance gets so low it will load down the previous stage.

I also prefer a 40-50 mA collector current, so I use a 33-39 ohm resistor in the emitter. And use a top hat heat sink.

In addition, if you isolate the buffer/voltage amp and it tends to want to oscillate at VHF (only had this situation with the 2N5109), connect a series circuit from the collector to ground consisting of a 470 ohm 1/4w watt CC resistor in series with a 15 pf cap.

I know you have a lot of information to digest, but keep up the experimentation.  You will learn a lot from simulating and then building. Kinda like OJT.  Cheesy

Phil
Logged

Charlie Eppes: Dad would be so happy if we married a doctor.
Don Eppes: Yeah, well, Dad would be happy if I married someone with a pulse.NUMB3RS   Smiley
W4AMV
Member

Offline Offline

Posts: 672


« Reply #102 on: September 13, 2011, 03:08:48 PM »

Good inputs Phil. Thanks. This form of DC bias is improved over the standard 4-R bias that I believe he is currently using. The AC feedback incorporated into the DC bias system is also a good thing. It would be useful for him to place this configuration into Spice and see the differences between the circuits, both AC and DC.
Logged
ssbothwell SWL
Guest
« Reply #103 on: September 13, 2011, 04:59:04 PM »

as far as the transformer goes: i used FT-37-43 torroids with 11turns on of #28 enamel wire as recomended here: http://www.qrp.pops.net/xmfr.asp



i just checked the diodes on my DMM and replaced one that wasn't matched up very well. i think this made a huge difference.

here is an image with RF=1.8MHz and LO=1.8MHz
http://i.imgur.com/i9JoQ.jpg
here is an image with RF=1.0MHz and LO=1.8MHz
http://i.imgur.com/eYjyP.jpg

is this second image giving me a hint of the sum and difference signals?

thanks DMOD for the biasing tips. i will definetely try your method of biasing once i work through this mixer stuff.
Logged
DMOD
AC0OB - A Place where Thermionic Emitters Rule!
Contributing
Member
*
Offline Offline

Posts: 1768


« Reply #104 on: September 13, 2011, 05:14:38 PM »

Right out of the mixer you should get 2.8 Mhz and 0.800 Mhz (or 800  Khz), the additive and subtractive frequencies.

So you have two possible IF frequencies from which to select and filter.
Logged

Charlie Eppes: Dad would be so happy if we married a doctor.
Don Eppes: Yeah, well, Dad would be happy if I married someone with a pulse.NUMB3RS   Smiley
ssbothwell SWL
Guest
« Reply #105 on: September 13, 2011, 05:28:08 PM »

i just increased the voltage of my RF signal and now i have this waveform: http://i.imgur.com/NKFdx.jpg
in that image i have taken the mxier output and ran it through a 3section 1mhz butterworth lowpass.

i moved the cursor manually to two the zero crossing points of the outside edges of the signal and its about 800KHz.

is this image showing the mixing that i am looking for? does it just need to be better filtered? would a bandpass filter set around my desired IF frequency get me the the difference signal i am looking for?
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #106 on: September 13, 2011, 05:42:24 PM »

From the images, yes, I would say the mixer is functioning. You are just going to need more filtering or as you suggest, a BPF. However again a simple experiment to confirm. Fix the LO frequency. Very carefully walk the RF frequency up and down about the LO frequency. By carefully, I mean if you have the capability say every 10 Hz steps! Oh, it can be more than that, but what you are looking for is a clear BEATING of the 2- frequencies. It will be obvious to you when you see the heterodyning, that the mixer is functioning.  Smiley
Logged
ssbothwell SWL
Guest
« Reply #107 on: September 13, 2011, 06:46:00 PM »

i'm using a HP 3312A function generator which only goes up to 13MHz. On the mhz range setting you really dont get much fine tuned control but here is a video of me sweeping the RF oscillator from 0.5mhz to 3mhz and back to 0.5mHz: http://www.youtube.com/watch?v=W4fbKeGuH18

Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #108 on: September 13, 2011, 08:48:31 PM »

It is not as clear as I had hoped. I believe it is due to your inability to really fine tune control your signal and this is further aggravated by O'scope triggering issues. I suggest you set the LO and RF near each other, say 100 kHz or so difference, LO at 1.8 MHz, RF at 1.7 MHz, then a nice R-C and/or  LC filter following should permit the 100 kHz signal to stand out and the sum signal should be nil.
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #109 on: September 13, 2011, 11:05:22 PM »

Here is a well behaved DBM. The generators are quite clean and settable in 1 Hz steps. Actually, in hindsight, this is not an easy waveform to capture in the time domain, this old scope must be set carefully to trigger so the intensified difference frequency which is clearly seen. The larger sinusoid signals are the LO and RF both set to -6 dBm. That is SMALL signal. Perhaps that is one of your issues, namely distortion. You might consider reducing your signal levels and also LP your signal sources. If they have significant harmonics you will see all the inter mix products. That is really going to cause quite a visual mess! 


* Mixer.jpg (157.45 KB, 640x480 - viewed 752 times.)
Logged
ssbothwell SWL
Guest
« Reply #110 on: September 13, 2011, 11:37:08 PM »

oh wow that is really helpful. at least now i know what i am trying to achieve. Smiley

i think i found the difference frequency! i put a 455khz LPF on the mixer output and fed it with a 1.6MHz LO and 1.7MHz RF. i had to crank up the voltage on the RF input to 2.8VRMS but i was able to find a 100KHz signal in the filtered mixer output.

i've attached images of the filtered and unfiltered IF output. (the webhosting ih ave been using is down right this second).

my unfiltered output does not look the same as yours. why is that?


* image (31).jpg (204.41 KB, 1280x960 - viewed 726 times.)

* image (31).jpg (222.9 KB, 1280x960 - viewed 755 times.)
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #111 on: September 14, 2011, 05:17:40 AM »

Your mixer appears to be functioning fine. Although the conversion loss may be higher than normal and the isolation less than desirable. How did you construct this mixer? Even at 1 MHz, it is imperative that good RF layout and construction practice be used. I'll take a pix of a homemade ring diode DBM and post. To do the mixer justice it would be nice to have a spectrum analyzer. Then you can measure accurately conversion loss and isolation as well as mixer balance. I suspect the first main difference between your screen shot and mine, is my difference frequency was much lower, if I recall 5 kHz! These generators are locked to a stable clock,so their is no wandering in frequency which aggravates the scope synch and finally how clean are your sources. If their is significant harmonic content that will further mess with the picture. I assume you are delivering both signals RF and LO from commercial generators? Balance and LO/RF leakage to the IF is also key. Ideally, when the LO or the RF is TURNED off, there is ZERO if output?
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #112 on: September 14, 2011, 05:10:26 PM »

Hi, I went back and looked at your video again. Indeed you can see ZERO beat quite well.  Smiley Watch the last 5 seconds or so towards the end of the video. You can see a sinusoid whose amplitude DISAPPEARS to zero. Then re-builds. That is the difference frequency which is going to DC.  Cheesy
Logged
ssbothwell SWL
Guest
« Reply #113 on: September 14, 2011, 06:31:55 PM »

okay now i am really confused. i was looking the mixer last night and i noticed that one of the ring was not assembled properly. it was really 3 diodes in a ring with the 4th one's cathode not connected to anything. this is one of the many problems with using breadboard.

so i fixed the diode ring but now i cant for the life of me find the difference frequency.

to give you a clear picture of what i am doing here is a series of photos which show exactly what kind of signals i am inputting and what i am getting at the IF:

RF Input (from an HP 3312A functon generator):
http://i.imgur.com/wUwpZ.jpg

LO Input (from my homemade oscillator):
http://i.imgur.com/StVn0.jpg

unfiltered IF output:
http://i.imgur.com/rgald.jpg

IF output with a 455KHz butterworth filter:
http://i.imgur.com/5L05j.jpg

heres some questions:

1. what voltage should i set my RF and LO signals to? should they be around the same voltage or should one be higher?

2. my mixer is on the same piece of breadboard as my oscillator and has tons of jumper cables. could this be a part of the problem?

3. my mixer is exactly like this: http://www.qrp.pops.net/images/2008/2008-larger/big_dbm.gif the schematic you posted a while ago has capacitors and a few chokes. what are those doing and should i be using them?

i feel like i am really close to making this thing work but i am running against a couple of different issues none of which is enough to break the circuit but in combination they make it unreliable and lossy.

just for fun i tried reconnecting it as a modulator with audio in the IF port and the carrier on the RF port and i was able to hear audio on a radio receiver. it was super noisy and undermodulated but it did work.
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #114 on: September 14, 2011, 08:42:14 PM »

Hi, well let me keep it brief. I'll try to get to your question in a latter post. Although you had the ring 3/4 correct, it still performed albeit not very well. Find attached a set of pix on construction, one is homebrew ( a knockoff of the one on the right) and one commercial, on the right. Emphasis is on construction technique. Can't over emphasize how key that is to performance. Please note, the HP mixer on the right has two additional xmfrs and will leave them out of the discussion for now. Note the tidy connections and minimum crossovers of connections. Hope this helps.


* Mixer 003.jpg (143.97 KB, 640x480 - viewed 764 times.)

* Mixer 002.jpg (145.65 KB, 640x480 - viewed 732 times.)
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #115 on: September 14, 2011, 09:54:26 PM »

okay now i am really confused. i was looking the mixer last night and i noticed that one of the ring was not assembled properly. it was really 3 diodes in a ring with the 4th one's cathode not connected to anything. this is one of the many problems with using breadboard.

so i fixed the diode ring but now i cant for the life of me find the difference frequency.

to give you a clear picture of what i am doing here is a series of photos which show exactly what kind of signals i am inputting and what i am getting at the IF:

RF Input (from an HP 3312A functon generator):
http://i.imgur.com/wUwpZ.jpg

LO Input (from my homemade oscillator):
http://i.imgur.com/StVn0.jpg

unfiltered IF output:
http://i.imgur.com/rgald.jpg

IF output with a 455KHz butterworth filter:
http://i.imgur.com/5L05j.jpg

heres some questions:

1. what voltage should i set my RF and LO signals to? should they be around the same voltage or should one be higher?

2. my mixer is on the same piece of breadboard as my oscillator and has tons of jumper cables. could this be a part of the problem?

3. my mixer is exactly like this: http://www.qrp.pops.net/images/2008/2008-larger/big_dbm.gif the schematic you posted a while ago has capacitors and a few chokes. what are those doing and should i be using them?

i feel like i am really close to making this thing work but i am running against a couple of different issues none of which is enough to break the circuit but in combination they make it unreliable and lossy.

just for fun i tried reconnecting it as a modulator with audio in the IF port and the carrier on the RF port and i was able to hear audio on a radio receiver. it was super noisy and undermodulated but it did work.

Based on your input signals, which look quite good, your IF would be 200 kHz. So your LP is really to broad. Also, your RF input level is much to large. Attenuate it. I would vary its level until you begin to see some occurence of an IF at 200 KHz. Design an RC low pass with a cutoff of say 100 kHz to suppress the LO and RF signals. Again, if your layout of all this is not clean, you are going to get cross talk and intermixing of signals that will really confuse the result.  Mixer balance is quite helpful and its not clear that your balance (due to layout?) is very good. Yes, osc on same card with tons of cables, UGH, sounds awful. Forget about the mixer design I posted, it was for starved LO, and that is not your issue. Again, clean up your layout, reduce your RF level, reduce your cutoff freq for the current IF and try again.
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #116 on: September 14, 2011, 10:10:42 PM »

ah, a light bulb moment! Since your isolation between LO and RF may not be very good and your buffer for the oscillator not up to par (an assumption), you are aggravating another electrical problem called injection lock. I will spare the details, but do look at your oscillator (via your O'scope) when connected to the mixer. Your oscillator can be moved or pulled in frequency by the RF signal, so reducing the RF level is imperative. This would explain why the sources look fine without connection to the mixe, but there is NO IF to be found! 
Logged
ssbothwell SWL
Guest
« Reply #117 on: September 14, 2011, 11:03:46 PM »

i rebuilt the mixer on copper clad board in an attempt to clean things up: http://i.imgur.com/xjiSS.jpg
i still have jumper wires but theres way less of them.

heres a difference frequency i was able to isolate: http://i.imgur.com/djiiS.jpg
LO was 1.6VRMS and RF was 3.5VRMS. if i drop the RF voltage then i cant find it anymore.

i can reliably find a difference frequency of whatever value but the signal always looks weird like in that last image and i have to keep the rf voltage cranked up really high.

now that i have rebuilt the circuit on copper clad board, what else can i do to try to achieve a balanced signal?



when the RF osc is connected, the LO output has a similar sort of 'modulated' look to it as the IF output. is that maybe the injection lock you are talking about?

edit:
it is really hard to capture this in photographs but are these bright spots in the unfiltered mixer IF output the difference frequencies? http://i.imgur.com/1CANL.jpg
here is a poorly highlighted version of what i mean: http://i.imgur.com/5K1Iq.jpg

the bright spots expand and contract as i adjust the RF frequency around the LO frequency and seem to correspond to the difference between the RF and LO frequencies.
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #118 on: September 15, 2011, 05:07:20 AM »

Hi, Gasp Shocked. I looked at your layout. Alligator clips! Fine for DC. Wires! Fine for DC. Patch boards, good for...well I sold all of mine and gave them away. Also, it is not clear in pix, but it looks like only half the ring is connected? Unless the connect is via the copper, cannot tell. Let me see if I can pix an RF breadboard built with Xacto blade knife and/or Demel tool...(moto-drill).
Logged
ssbothwell SWL
Guest
« Reply #119 on: September 15, 2011, 06:08:48 AM »

the transformer on the right side has insulated wire attached as jumpers for part of the secondary coil whereas i used scrap component leads as jumpers on the left transformer. the component leads done show up in photos from my junky cellphone camera.

i want to build the oscillator on copperclad board but i am hesitant to solder the parts down while i am still adjusting and tweaking things constantly. it would be nice to have some kind of modular system where each stage (oscillator with a buffer and amp to get to a standard dB gain, mixer unit, various rf amps with standardized gains, demodulator,etc) was in its own enclosure and i used coaxial cable to patch everything together. would that be a good system?
Logged
ssbothwell SWL
Guest
« Reply #120 on: September 15, 2011, 06:59:34 AM »

i'm having a lot of problems with my filters. i can't seem to suppress the LO and RF signals. i've tried various cutoff points, different torroids (37-43 and t68-1) and rc filters. they all drop the LO and RF signals down to around 100-300mV but no filters get rid of them entirely. i think i am seeing the difference frequencies in there but i cant isolate them.

Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #121 on: September 15, 2011, 04:43:40 PM »

http://www.dartmouth.edu/~sullivan/prototyping.pdf

HI. This is a good site to look over. Again, if I have time I will post pix of some cards built. You can build sub modules  like the mixer, buffer, osc,, etc... and link them via coax cable, the mini 50 ohm, RG 174 or its varations. Next time you are at the Ham Fest, pick up epoxy-glass PC panels, FR-4, 1 oz copper 32 or 64 mil card, double sided and/or single sided. You can build boxes and circuits literally in 3-D, so they become compact amd minimize all the problems you currently face.
Logged
ssbothwell SWL
Guest
« Reply #122 on: September 15, 2011, 08:45:06 PM »

i would really appreciate photos of your construction technique.

do you think it might be possible that the ft37-43 is the wrong toroid for my application?
http://toroids.info/FT37-43.php

i'm not sure if these specs are relevant but toroids.info says:
Quote
Application Freq Range
Wideband Transformers 5 - 400 MHz
Power Transformers 0.5 - 30 MHz
RFI Suppression 5 - 500 MHz

would the fact that the wideband transformer min frequency is 5mhz mean that this toroid is unsuitable for a 1-2mhz transformer?

also, is the transformer's resonant frequency relevant to this application?
i tried checking the resonant frequency of the trifilar transformers by running my function generator through the primary winding and connecting my scope to the secondary winding. the resonant frequency appears to be above 12mhz (the upper limit of my function generator). the rms voltage was 3.5VRMS at 2mhz and 6VRMS at 12MHz.
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #123 on: September 15, 2011, 09:22:43 PM »

No, 43 material which has a permeability of 850 is fine. What is useful to check is the total start-to-finish inductance of the secondary, which has the center tapped connection. The inductance of this part of the winding should have a reactance of 500 ohms at the lowest application frequency, say at 2 MHz. 39 uH would be nice. If say a third of that, 12 or so uH, still ok. The intent, try to minimize the mixer loss. I really think your main issue is construction technique and stray coupling and significant un wanted radiation. Coax interconnect is great solution, and connectors are not required, just use the coax cable, properly dressed for short run inter connect.   
Logged
ssbothwell SWL
Guest
« Reply #124 on: September 15, 2011, 10:59:27 PM »

you are right. my mess of wires was confusing everything. i just ripped apart all my alligator clips and rebuilt them using coaxial. this simplified things massively and i am now able to get a clear difference signal if i filter the IF output.

here is the mixer with coaxial alligator clip connections:
http://i.imgur.com/bK1Y5.jpg
the oscillator and if filter are still on breadboard. i plan to put them on copper and put each component into its own little enclosure with so-239 connectors in the near future.

here is the difference frequency:
http://i.imgur.com/qAxIR.jpg

i am able to turn the RF input down to under 1vRMS and still have a clear signal.

the mixer still is not balanced but at least i am getting a clear difference frequency. i assume that a tuned IF amp will further cleanup the signal, is that correct?
Logged
Pages: 1 ... 4 [5] 6 7   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.217 seconds with 18 queries.