Just getting Started


Tim
 

Have hardware working and trying to get handle on Windows program

I am looking at Parameters file and it seems to be one long line.

I do not see any line separators in file? However if I change Comport setting it seems to recognize change.

Sorry for easy question but I am not seeing something.

Tim W4YN


William Kimber
 

Hi,

I seem to recall that came up before or maybe in the Nat group. It is
result of using wrong type of word processing program.

You want a programmers type rather than an authors type.

It has to put in the Cr and Lf characters rather than run lines together.

Cheers,
Will
.

On 05/06/15 15:35, w4yn@... [PHSNA] wrote:


Have hardware working and trying to get handle on Windows program

I am looking at Parameters file and it seems to be one long line.

I do not see any line separators in file? However if I change Comport
setting it seems to recognize change.

Sorry for easy question but I am not seeing something.

Tim W4YN


Tim
 

Tnx Will
I was using the parameter editor in the Windows program.
Tim


N5IB
 

I just re-uploaded a pair of example parameters files into the "Windows PHSNA" folder. Be sure to rename to "parameters.txt" and make whatever edits are appropriate for your unit.

These **should** open with Notepad, or preferably with the editor inside the WinPHSNA app, and display with proper line breaks.

Jim, N5IB


Nick Kennedy
 

Well, you are correct, Tim.  I don't know how this is happening but once again the parameters.txt file on the site has 0x0A line terminators only, instead of 0x0D, 0x0A.  (In decimal, that's 13, 10.)

I just uploaded a version that has the correct terminators.  Looks like Jim is looking at this at the same time as me.

After breakfast I'll play with it a little more to make sure I didn't foul up something else.  I did start up PHSNA and did a "write" of the parameters.txt file to verify that it did not alter the end of line codes.

73-

Nick, WA5BDU

On Thu, Jun 4, 2015 at 10:35 PM, w4yn@... [PHSNA] <PHSNA@...> wrote:
 

Have hardware working and trying to get handle on Windows program

I am looking at Parameters file and it seems to be one long line.

I do not see any line separators in file? However if I change Comport setting it seems to recognize change.

Sorry for easy question but I am not seeing something.

Tim W4YN



Tim
 

Tnx Nick and Jim

Another probably dumb question.
I was doing the calculations for compensate for out change over frequency.
I noticed in the PDF the DBm values were all negative.
Using windows version and power meter combo the values I see for uncompensated readings are all positive DBm readings and resulting graph has opposite slope as example.
When taking readings my uncompensated values increase with frequency.
Sorry for newbe questions.

BTW I found an online program for those like me that do not want to spend retirement funds on Microsoft products: http://polynomialregression.drque.net/online.php

 

Open office does not have same capability as Excel does but that online resource works well.

Tim W4YN
 


N5IB
 

Whether you see plus or minus dBm values just depends on the output level of your DDS (and calibration of your 8307 meter). The plot of uncompensated values may decrease, increase, or do both over the frequency range depending on the performance of the low pass filter following the DDS.

You should not be concerned as long as the total change from around 1 MHz to the top of the range is not more than a few (well under 10) dB - at least up to 30 MHz (9850) or 60 MHz (9851).

Be sure to carry the coefficients for the polynomial out to seven significant digits at least. Those fifth power terms are very sensitive to small changes.

My latest build starts off at around +2 dBm, sags a bit then has a little hump at about 0 dBm before rolling off to about -3 at the top of the range. The polynomial compensation flattens it within 0.3 dB.

Post the plots you got and we can take a peek to see if all is copacetic.

Jim, N5IB




---In PHSNA@..., <w4yn@...> wrote :

Tnx Nick and Jim

Another probably dumb question.
I was doing the calculations for compensate for out change over frequency.
I noticed in the PDF the DBm values were all negative.
Using windows version and power meter combo the values I see for uncompensated readings are all positive DBm readings and resulting graph has opposite slope as example.
When taking readings my uncompensated values increase with frequency.



Nick Kennedy
 

The set of readings you take for computing the polynomial  coefficients might be all negative, or might be some negative and some positive, depending on how your built your hardware.  Some have a maximum of 0 dBm or maybe -10 dBm so their readings will all be negative.  Mine goes slightly above 0 dBm so I might have some positive readings.   These readings are taken after you have done the slope and intercept calculations and it shouldn't matter if some are positive.

As to having rising power on rising frequency, uncompensated ... I guess a bit of a rise over a range of frequencies could occur depending on your output filter characteristic, but overall going from 2 MHz to 30 or 60 MHz for example, I'd expect a downward trend.

What are your slope and intercept numbers?

A little box tells me that Jim has responded, so maybe he understood the question better than I did.

73-

Nick, WA5BDU


On Fri, Jun 5, 2015 at 8:43 AM, w4yn@... [PHSNA] <PHSNA@...> wrote:
 

Tnx Nick and Jim


Another probably dumb question.
I was doing the calculations for compensate for out change over frequency.
I noticed in the PDF the DBm values were all negative.
Using windows version and power meter combo the values I see for uncompensated readings are all positive DBm readings and resulting graph has opposite slope as example.
When taking readings my uncompensated values increase with frequency.
Sorry for newbe questions.

BTW I found an online program for those like me that do not want to spend retirement funds on Microsoft products: http://polynomialregression.drque.net/online.php

 

Open office does not have same capability as Excel does but that online resource works well.

Tim W4YN
 



Nick Kennedy
 

Oh yeah, the alternate tool for computing the coefficients sounds good.  I'm just starting to learn a bit about Octave (again!). Maybe it could do the job too.

73-

Nick, WA5BDU

BTW I found an online program for those like me that do not want to spend retirement funds on Microsoft products: http://polynomialregression.drque.net/online.php


Tim
 

TNX Jim
After building your bridge I found a slight difference between open and short readings around or a bit less than 1 Db, is that acceptable? I am only interested in tuning Band Pass filters.

Tim 


N5IB
 

Should be fine. Is your 50 ohm terminated reading about 12 to 13 dB below the normal DDS output?

What do you use for a short? A wire jumper might not be good enough.

I made a "careful" one.
Used a piece of copper shim soldered to a wire from the BNC center pin, then  soldered it around the rim of the BNC shell.

Jim, N5IB






---In PHSNA@..., <w4yn@...> wrote :

TNX Jim
After building your bridge I found a slight difference between open and short readings around or a bit less than 1 Db, is that acceptable? I am only interested in tuning Band Pass filters.

Tim 


Tim
 

I have some calibrated open shot loads from another project.
I believe 12-13 below is about what I saw, will re-verify when all is calibrated.
Tnx Tim


Nick Kennedy
 

Speaking of this, return loss should be the same with the port open or shorted, so I took the easy way out and went with open.  Then Ignacio EB4APL clued me in on the fact that small errors might on occasion cause a measured return loss that's negative (not allowed!) or zero (where SWR is infinite). So I built a shorting BNC like Jim just described. Now I check both ways and use the number that's larger (more positive). 

If compensation is good, the open / shorted return loss shouldn't vary with frequency, but it might, a tiny amount.  Ignacio goes through the extra step of sweeping the frequency range of interest and selecting the largest value, if it does vary.  I did incorporate a software fix for invalid values though, as described below.

Here's some info I added to the User's Guide (the second paragraph anyway) after working through this issue:

Note that the information box shows the current dBm value for the open or shorted detector condition.  This value is used in the calculation.  This value is determined by connecting the PHSNA’s generator output to the RLB and the RLB’s detector port to the AD8307 circuit and leaving the “unknown” or “load” port open or shorted.  Then read the value in the dBm(f) window. Verify that it is close or identical to the value read from the parameters.txt file and if not, make it so!

Negative return loss values are invalid and a value of zero will result in an SWR of INFINITY.  The program substitutes 0 for negative return losses and uses 0.1 for the SWR calculation (SWR = 173) to prevent the infinite result (divide by zero) in such cases.  It prints a warning regarding the substitution.  These invalid readings may be caused by having a shorted / open bridge reading that’s too small (too negative) in the parameters.txt file. 

73-

Nick, WA5BDU


Tim
 

I up loaded a sweep after I did my first calibration run look in my folder W4YN
Tnx Tim


Nick Kennedy
 

It's very strange to see your uncompensated value go steadily up with frequency like that.  But the compensation does flatten it out the way it should.

Are you using an eBay AD9850 or AD9851, with outboard filter?

73-

Nick, WA5BDU

On Fri, Jun 5, 2015 at 9:42 PM, w4yn@... [PHSNA] <PHSNA@...> wrote:
 

I up loaded a sweep after I did my first calibration run look in my folder W4YN

Tnx Tim



Tim
 

9850 e-bay but I did not build the board, I bought it assembled from another Amateur.
On front of enclosure the output is marked +3DB is that normal output ?
Tim


Tim
 

It was getting late last night but I seem to remember seeing very little change on scope when feed directly from generator output. However when connected to meter then I saw the values change with frequency.
Will verify later today.
Tim


N5IB
 

Be sure when observing with the 'scope that the DDS output is terminated in 50 ohms, since the scope input is high impedance. I use a BNC Tee at the 'scope input with a 50 ohm terminator on one side of the tee.

The power meter can also serve as the termination. In fact a good test is to make a scope run with a dummy load, and another with the power meter to see if there are any differences.

Your 1.5 dB rise shown on the plot would amount to about a 20% increase in peak-peak voltage.

Your RLB plot looks fine. You could flatten it out on the low end with a few more turns on the transformer, but that would probably degrade the upper end. Best is probably to make a pair of RLB - one with the -43 ferrite and a few extra turns to perform well at the low end, and another with -61 ferrite for better performance at the high end.

Jim, N5IB




---In PHSNA@..., <w4yn@...> wrote :

It was getting late last night but I seem to remember seeing very little change on scope when feed directly from generator output. However when connected to meter then I saw the values change with frequency.
Will verify later today.
Tim


Tim
 

Yes same setup, 50 ohm load on tee at scope.
Scope measurements :
1Mhz 1.44 dBm
14Mhz 1.54dBm
30Mhz 2.64dBm

When using meter, uncompensated:
1Mhz 2.02 dBm
14Mhz 2.48 dBm
30Mhz 3.48 dBm

Tim


N5IB
 

Is that second data set PHSNA readings or scope measurements with the power meter serving as the termination?

Jim, N5IB



---In PHSNA@..., <w4yn@...> wrote :

Yes same setup, 50 ohm load on tee at scope.
Scope measurements :
1Mhz 1.44 dBm
14Mhz 1.54dBm
30Mhz 2.64dBm

When using meter, uncompensated:
1Mhz 2.02 dBm
14Mhz 2.48 dBm
30Mhz 3.48 dBm

Tim