EMRFD Message Archive 8174
Message Date From Subject 8174 2013-01-30 10:48:21 Nick Tsakonas RF measurements guidelines
the other days I built a 20m DTC from EMRFD p3.14 and while I was trying to measure the response I was getting erroneous readings on the insertion loss (something like 17-25db) [1]. I have not concluded the task yet, but I realized that although I have read and consulted many times Ch.7 of EMRFD I start to feel less confident of my skills on RF measurements.
I would like to ask for suggestions on any good book (or Application Note) regarding RF measurements at the HF/VHF range (I am not a VHF+ experimenter).
I am interested on the techniques, methodologies,guidelines (proper setup for a given measurement) and DO's/DONT's of RF measurements.
And since we are here, I would like to ask, a short 50 Ohm cable with a 50 Ohm resistor at the oscilloscope end, doesn't qualify as a 50 Ohm termination for a circuit requiring a 50 Ohm load?
Thanks,
Nick, SV1DJG
note[1]: I found some old threads on the group today regarding proper filter measurement setup and I will repeat my measurements according to those.
[Non-text portions of this message have been removed]8175 2013-01-30 11:52:02 victor Re: RF measurements guidelines
To measure the filter insertion loss you have first to take the filter out of the setup, connect the signal generator directly to the load and read the reference voltage on the oscilloscope. Then you connect back the filter into the setup and read its output voltage. The ratio of the two measured voltages is the filter inserti8176 2013-01-30 12:08:58 Jerry Haigwood Re: RF measurements guidelines
I like to tack a 51 ohm resistor directly across the output of the
filter rather than using the O'scope's termination. I am not sure it is any
better but it eliminates one more variable. It also gives me a place to
hang my scope probe on. ;-) If you have a good L/C meter such as the AADE,
you can measure each L and C before soldering them down. Also, I find stray
capacitance and the distributed capacitance of the inductors tend to add a
few pF of capacitance so you might want to account for that.
Jerry W5JH
"building something without experimenting is just solder practice"
8177 2013-01-30 13:04:29 victor Re: RF measurements guidelines
In real world the scope input capacitance also has some influence but it is much smaller than the capacitance of a open end 1 meter of coax.
victor - 4Z4ME
8178 2013-01-30 13:46:22 Jerry Haigwood Re: RF measurements guidelines
Jerry W5JH
"building something without experimenting is just solder practice"
8179 2013-01-30 14:05:07 victor Re: RF measurements guidelines 8180 2013-01-31 01:13:17 Nick Tsakonas Re: RF measurements guidelines
thanks again for the info.
I do not have the book with me at the moment, but I built the 20m filter which uses the 34p input/output capacitor and the 6.3p coupling capacitor. The inductors are 10t on T50-6 as described. The only difference is that I used whatever I had in my component box, which was 39p input/output and 6.8p coupling. Instead of the fixed 271p tuning capacitor I used a parallel combination of 220p+plastic trimer of 2-65pf for tuning.
I simulated the circuit in LTSpice first to see if it was ok to use different values and I notice a small change in the passbad shape, but it was not so important. I used the trimmers to peak the filter and it peaks a little low, but still good for the 20m band.I understand this as component tolerance and inductors actual values may be responsible and it does not bother me this at the moment.
What I was trying to do is to manually sweep the filter in 1Khz steps and record the response, but after seeing the excessive loss I realized that either I built something wrong or I am doing the measurements wrong.
I checked (at peak frequency) with the 10x probe the voltages on the filter nodes (input, before the coupling capacitor, after the coupling capacitor and ouput) and the voltages were not following the pattern I was seeing at LTSpice. For example, before the coupling capacitor the voltage reported by LTSpice is larger than the input voltage by a factor of 3 or 4 (if I recall well) but in my case it was just 1.2 and so on.
So, besides my lack of deep theoretic background on filters, I think that the way I am trying to take the measurements may be totaly out of context and wrong.
I already started reading some theory on filters, but the only thing I can find on measurements are some application notes from Agilent ("Back to basic series").If you are aware of anything more complete I would like to add it in my library.
73,
Nick,SV1DJG
________________________________
8181 2013-01-31 03:20:15 Ashhar Farhan Re: RF measurements guidelines
First, the generator should be followed by a 6db or higher attenuator.
This will ensure that the source, as seen from the other end of the
attenuator is robustly around 50 ohms.
Second, make a cable with bnc connectors at both ends, but a 50 ohms
(I use two 100 ohms in parallel) resistor across the bnc connector
that connects to the scope. No, we don't use a compensated probe for
the measurement.
Now, setup the filter with bnc female connectors at both ends, but run
a very small wire directly between the two female connectors, entirely
bypassing the filter for the time being. The filter should be
disconnected at both ends.
Sweep the range, noting the readings at each frequency. Note them in
your lab book. These are your reference readings.
Next, remove the shorting wire, connect back the filter and repeat the
sweeps. Note the readings at the same frequencies as the reference
sweep.
We will assume that the readings you took are peak voltages. This
needs clarification as well. Suppose a sinewave starts from zero, goes
up to 1.5v, comes down, dips to -1.5V, then we will say that peak
voltage is 1.5V. Peak-to-peak will be 3 volts. However, what we must
note is just the peak voltage.
Now, to calculate the power loss at any frequency, divide the filter
reading by reference reading (peak voltage) and square it up.
- farhan
On 1/31/13, Nick Tsakonas <magicnickus@yahoo.com> wrote:
> Hello Victor,
> thanks again for the info.
>
> I do not have the book with me at the moment, but I built the 20m filter
> which uses the 34p input/output capacitor and the 6.3p coupling capacitor.
> The inductors are 10t on T50-6 as described. The only difference is that I
> used whatever I had in my component box, which was 39p input/output and 6.8p
> coupling. Instead of the fixed 271p tuning capacitor I used a parallel
> combination of 220p+plastic trimer of 2-65pf for tuning.
>
> I simulated the circuit in LTSpice first to see if it was ok to use
> different values and I notice a small change in the passbad shape, but it
> was not so important. I used the trimmers to peak the filter and it peaks a
> little low, but still good for the 20m band.I understand this as component
> tolerance and inductors actual values may be responsible and it does not
> bother me this at the moment.
>
> What I was trying to do is to manually sweep the filter in 1Khz steps and
> record the response, but after seeing the excessive loss I realized that
> either I built something wrong or I am doing the measurements wrong.
> I checked (at peak frequency) with the 10x probe the voltages on the filter
> nodes (input, before the coupling capacitor, after the coupling capacitor
> and ouput) and the voltages were not following the pattern I was seeing at
> LTSpice. For example, before the coupling capacitor the voltage reported by
> LTSpice is larger than the input voltage by a factor of 3 or 4 (if I recall
> well) but in my case it was just 1.2 and so on.
>
> So, besides my lack of deep theoretic background on filters, I think that
> the way I am trying to take the measurements may be totaly out of context
> and wrong.
>
> I already started reading some theory on filters, but the only thing I can
> find on measurements are some application notes from Agilent ("Back to basic
> series").If you are aware of anything more complete I would like to add it
> in my library.
>
> 73,
> Nick,SV1DJG
>
>
>
>
> ________________________________
>8183 2013-01-31 11:36:32 Bob Re: RF measurements guidelines
I've been following this discussion with interest until curiosity got the best of me so I decided to build this circuit and measure the results. I am a FIRM believer in TMITK To Measure Is To Know. As far as I'm concerned, one can simulate `till the end of time I'd rather KNOW what a circuit actually does!
The FILES folder K3NHI-DTC contains some results. My observations follow here:
The circuit was built with all capacitors measured out to be within 2% of the design value. Also pairs were within the same 2% limit. The inductors were wound as described. In sweeping the filter I found the Fc to be less than 12 MHz confusing since I had measured all the caps as above. So - I liberated the L's and found them to be about 0.66 uH and not the expected 0.4 uH as in EMRFD Table 3.4. (TMITK!) I blamed the batch of cores (T50-6)I had on hand and proceeded to remove turns using my AADE for guidance. The cores wound up with 7 turns spaced to yield the desired value. The spacing was further tweaked while sweeping the filter to yield the results in FILES.
As a side note, my measurements were done using some HB sweeper stuff with a 8307 based power meter as a detector. Signal out of the latter was displayed on a DSO to yield the data herein. The test system is a 50 ohm system throughout i.e. no scope probes.
Note the results presented are "as built" plus the turns spacing tweaking. Were I to seriously build such a filter, I'd probably ease my life with some adjustable Ct to better set Fc. I'd probably allow some variability in C12 as well since this is a low value cap and assy. strays could compete for attention here. In fact, I found that just playing with the physical spacing of the toroidal L's themselves could make a noticeable difference in the filter BW and thus to some degree the IL.
In conclusion, I think in general one is best off working within an all-50 ohm environment for this kind of RF measurement need.
Cordially,
Bob, K3NHI
8184 2013-01-31 12:10:49 Jerry Haigwood Re: RF measurements guidelines
Great post. Like I said in one of my posts before, I always measure
inductors and capacitors with my AADE meter. I have found that
manufacturers of toroid cores list the AL value more as an estimate rather
than an absolute number. I expect the AL value to be something like +/- 10%
or even more in some cases. And, as you saw, the stray capacitance always
plays a big part in detuning LC circuits. Recently, I found a surplus house
that is selling ceramic trimmers at a reasonable price. These are high
quality silver plated units that appear to be very stable (maybe NP0?). I
purchased a couple of dozen and have been using them in my tuned circuits.
Jerry W5JH
"building something without experimenting is just solder practice"
8185 2013-01-31 12:19:29 Jerry Haigwood Re: RF measurements guidelines
I failed to mention that I usually try to spread the turns on the
toroid to about 75%. Did you purposely leave the turns tight to achieve
the L value?
Jerry W5JH
"building something without experimenting is just solder practice"
8186 2013-01-31 20:08:22 Leonard Meek Re: RF measurements guidelines
you mentioned here?
Leonard
----- Original Message -----
8190 2013-02-01 08:53:50 Jerry Haigwood Re: RF measurements guidelines
I didn't expect to see you over here on this list since tube circuits
are not discussed all that much here. ;-) Anyway, the surplus house is B&G
Micro in Garland, TX (suburb of Dallas). They have two values: 5-65pF and
5-95pF (that is what I measure). The price is 40 cents each or if you buy
10 the price drops to 34 cents each. Shipping is reasonable because you can
select USPS First Class mail. Here is the URL.
http://www.bgmicro.com/trimmers_1.aspx
BTW, how is the Allied Radio regen doing?
Jerry W5JH
"building something without experimenting is just solder practice"
8191 2013-02-01 09:37:42 William Carver BG micro trimmers
W7AAZ
>Hi Leonard,I didn't expect to see you over here on this list since tube circuits
are not discussed all that much here. ;-) Anyway, the surplus house is
B&G Micro in Garland, TX (suburb of Dallas). They have two values:
5-65pF and 5-95pF (that is what I measure). The price is 40 cents each
or if you buy 10 the price drops to 34 cents each. Shipping is
reasonable because you can select USPS First Class mail. Here is the
URL.
http://www.bgmicro.com/trimmers_1.aspx
BTW, how is the Allied Radio regen doing?
Jerry W5JH8202 2013-02-01 19:45:16 Leonard Meek Re: RF measurements guidelines
contact with them. It's true, I'm a tube guy, but I have been known to stoop
to using semiconductors on occassion. This group has some interesting banter
on RF measurements, etc., an area that I'm admitedly weak in. Most of my
working experience has been with analog computers and servo systems. I built
regens when I was a kid, but only got back into them about 10 years ago. Had
a ball ever since.
Leonard
----- Original Message -----
8214 2013-02-03 13:38:55 Leonard Meek Re: RF measurements guidelines
reception on my kluge circuits. Gives me a good bench mark.
Leonard
----- Original Message -----