EMRFD Message Archive 10740
Message Date From Subject 10740 2015-01-20 11:46:29 Dana Myers Si5351A output level vs power into 50 ohms and mixer levels Hello Jason,
In "SI5351A INVESTIGATIONS PART 3", you wrote:
"An ADE-1 diode-ring mixer is fed directly from one of the Si5351 ports
set for the lowest output level (which is about +8 dBm)."
I'm not sure how you're measuring that +8dBm. My understanding of the
diode mixer specifications is that the LO drive is specified as the drive
that produces a given power in a 50 ohm load. The mixer LO port may not
present the same load (I expect it is non-linear), so I don't think it
is correct to measure the LO drive at the LO port of a mixer. Instead, the
LO drive is determined into a 50 ohm load, then that is replaced by the
mixer LO port. Further, the drive is fundamental power - at least I think so.
With that in mind, I measure a ~400mV peak-to-peak square wave from
an Si5351A @ 2mA driving a 50 ohm resistive load at 12.7MHz (happened
to be a frequency of interest and reasonably measurable for me).
The amplitude is linear with drive setting; at 8mA, I see 1.6V peak-to-peak.
Converting from voltage/resistance to power:
Vpeak = Vpeak-peak / 2
Vrms = Vpeak / sqrt(2)
Po = Vrms^2 / 50
Po = ((Vpeak-peak / 2) / sqrt(2))^2 / 50 ->
Po = Vpeak-peak ^ 2 / 400 (right?)
Note this calculates the power of the fundamental in the square
wave, which, I believe, is what is desired for the mixer injection.
I might be confused here, and would love to be corrected.
So I think the Si5351A @ 2mA is producing ~.4mW or -4dBm of
fundamental energy. As low as this is, I do recall Wes Hayward
investigating square-wave drive to diode mixers and publishing
numbers that show good mixer performance at much lower drive
levels than usual.
At 8mA drive, I estimate ~ +8dBm of fundamental drive :-)
I'm really just forming some thoughts here before I embarrass
myself on the EMRFD list :-) Anything you can do to get me on
the right track if I'm not is appreciated. It might even be "RTFM"
BTW I do rather like your Si5351 library you've posted; looking to
adopt it for my PSoC 5LP-based experiment, though it's already pretty
platform-neutral and I'll probably end up using an MSP430 part for
an actual project.
73 de Dana K6JQ
10741 2015-01-20 12:13:42 Dana Myers Re: Si5351A output level vs power into 50 ohms and mixer levels 10747 2015-01-20 22:06:15 victorkoren Re: Si5351A output level vs power into 50 ohms and mixer levels The local oscillator input of a passive double balanced mixer is in essence two diodes connected in series which are in parallel with a second similar diode pair but reversed in direction, so for each driving voltage polarity it is two diodes in series.If the diodes are plain switching silicon diodes you need to drive with a signal that is larger than 2.4 Volt peak to peak, or almost 1 V rms.If the mixer diodes are Schottky then a lower voltage will suffice.Victor - 4Z4ME 10748 2015-01-21 14:39:22 Dana Myers Re: Si5351A output level vs power into 50 ohms and mixer levels Updating my own measurements.
10753 2015-01-26 17:33:22 Dana Myers Re: Si5351A output level vs power into 50 ohms and mixer levels 10754 2015-01-26 20:07:06 vasilyivanenko Re: Si5351A output level vs power into 50 ohms and mixer levels Hi Dana -- due to the wonders of the (newer) Yahoo format -- your posts seem truncated as I read within this forum thread and proves a little difficult to follow. I rarely "troll" here --- so that might be a factor.
I love your displayed work and diligence to solve this riddle.
I respectfully wonder about your statement -----> .... also, I believe the AD807 presents a better load than the 100-ohm resistors in parallel. <----
I made some measures last year and two 100 ohm resistors gave a return loss of up to 10 dB better than the AD8307 device depending on how you prepare the input circuitry. My best circuit ran SMA connectors, "microstripline style paths" plus SMT parts and 2-sides Cu board with via wires [especially as you move into the VHF-UHF end]. Variables seem to include the ground plane style , frequency and input parts style + layout. I also ran the SMT version of this fun little IC in my 500 MHz wide return loss sweeps.
From my measures, parasitic C may affect the AD8307 S11 measures without due care and two 100 ohm 1% SMT (0603-0805) resistors gave a better termination to the source impedance than the AD8307 at any frequency up to ~ 500 MHz. Two 100 ohm resitos gave a little better S11 than a single 49.9 ohm resistor at > 300 MHz.
This is 1 case where the Lord Kelvin / Bob Kopski mantra : to measure is to know rings true.
10755 2015-01-26 20:41:24 Dana Myers Re: Si5351A output level vs power into 50 ohms and mixer levels 10757 2015-01-27 10:34:38 Jason Milldrum Re: Si5351A output level vs power into 50 ohms and mixer levels
Apologies for the late reply. I've been very busy with the Si5351A Breakout Board campaign. I just wanted to let you know that you are correct in your analysis. The problem is that when I wrote that blog post, I had some (unknown to me at the time) non-standard Si5351As that delivered quite a bit more power than stated on the official datasheet (they also had a different I2C address). A standard Si5351 does deliver exactly the amount of power that you say. I will update that post to make a note of that.
10758 2015-01-27 11:24:15 Dana Myers Re: Si5351A output level vs power into 50 ohms and mixer levels