Date   

Re: Sweep output debugging [1 Attachment]

Marcus Gustafsson
 

Steve,

thank you for helping out.

Here are my measurements done with 7.040 MHz output and with oscilloscope with 1MOhm input and 1x probe. I also had a 50Ohm termination on the PHSNA output.


Vpp 1 dBm 1 delta dB Vpp 2 DBm 2
DDS output measure on the DDS board pins 0,223 -9,1 0,0 0,190 -10,4
After T1 0,400 -4,0 5,1

After 1st 6dB pad 0,213 -9,5 -5,5

Filter output 0,193 -10,3 -0,9

After 2nd 6dB pad 0,088 -17,1 -6,8

After ERA-3 and C8 0,435 -3,3 13,9

After 3rd 6dB pad/PHSNA output 0,208 -9,7 -6,4


The DC voltages:
Before R7: 12.6V
After R7:    5V
At ERA-3 output: 5V
At ERA-3 input: 3V

So from this it seems that DDS output is low and also ERA-3 amplification is also on the low side.

The resistor to the RSET pin on the AD9851 measures 3.81kOhm. I notice DDS-60 board uses 5.6kOhm.
Should I try to switch that?

And should I try lowering the R7 resistor value?

/Marcus, SA5PMG

ps. Ignacio your experiment board project sounds cool, keep us updated.


Re: PC board needs survey

Steven Greenfield AE7HD
 

I am in a maker club, about half-dozen of us are hams. So I may wish to order more than a few of these, if that is alright.

I'd rather over-order than be caught short later.

Steve Greenfield AE7HD


Re: Sweep output debugging [1 Attachment]

Steven Dick
 

First thing I would check is whether the DDS board itself has the proper outputs.  On a Type II DDS with its built-in LPF components removed and bypassed, I measured output 1 had 550 millivolts peak to peak and output 2 also had 550 millivolts peak to peak at the power-up 7.040 MHz, using a Coilcraft transformer at T1. Voltages could be higher with the specified transformer (non-bifilar wound), perhaps 800mV each since the DDS is a current output device and it's seeing a higher impedance with the non-bifilar wound transformer. You can easily check these at each side of the T1 transformer primary or carefully at pins 9 and 10 on the DDS output connector.  If you see two different voltages across the primary you have a problem. Also make sure you unsoldered R9 on the DDS board. It will be across one of the outputs if you don't remove it. If it wasn't unsoldered, you'll see two different output voltages. On my working PHSNA, I measured 500mV p-p at output of the lowpass filter. My final output was 0.65 dBM as measured on the power meter. I think your LPF is working OK since the output response is pretty flat from your plot. The other mostlikely problem could be a problem in the output amp circuit.  Measure voltage at the junction of R7 and RFC1. It should be about half way between ground and the +12V.  If it reads zero, you have a short

Hope this helps.
"Digital Steve", K1RF


Re: Measurement Receiver Inductors

John Kolb
 

I was a manufacturing engineer for Southcom International back in the dark ages. We had the coils wound by a subcontractor (our production ladies working at home) and then adjusted them tighter or loser using Boonton Q meters. I was going crazy writing engineering change notices on the coil drawings with each new batch of cores till I was allowed to specify the winding instructions as + or - a number of turns. It was probably +/- 4 turns for a coil of 120 turns.

So yes, for critical applications, toroids have to be tweaked by squeezing or expending the winding around the core or even changing the number of turns.

We once received a coil from a commercial coil manufacturer where their wire spool had run out in the middle of the winding and another wire was joined with nice square knot on the insulated wire. Couldn't figure out how a single toroid could not have continuity from one end to the other :)

John
KK6IL

On 2/19/2016 9:23 PM, k5ess.nothdurft@... [PHSNA] wrote:
Wonder if others have found that the inductors called for in the
Measurement Receiver need tweaking? I found that the inductors in the
impedance matching networks at each end of the crystal filter needed one
less turn (11 vs 12) and needed for the turns to be all scrunched up to
arrive at a value close to 8.05µH. Value was determined using the
Crystal Test Fixture and measured known capacitors.(This is another good
use for the Crystal Test Fixture). Still, I can't see any difference
when tuning the variable caps associated with the matching networks at
each end of the filter. I modeled the matching networks in LTSpice with
a 600 ohm 6dB pad between them and the response is pretty broad. So are
the variable caps even needed? I'd be interested in others experience.


Re: PC board needs survey

David Collins
 

Hi Jim,

I am wondering why the Measurement Receiver PCBs are not included in the poll.  I have finally gotten around to adding MR support to the MSNA firmware (planned for V5.00) and I could use another MR.  The one I have seems to work fine but I didn't package it with any consideration for shielding and the noise floor seems a little high.  I would prefer to have a second one for that investigation leaving the one I have for firmware development and support.


73,
Dave Collins - AD7JT


Re: Sweep output debugging [1 Attachment]

Marcus Gustafsson
 

It's from the final output Steve. The schematics are a bit ambiguous, saying +1dBm before the final connector and then -3dBm Out after the connector. Anyway you are right the output is low. 

Any idea what is wrong?


Re: Sweep output debugging [1 Attachment]

EB4APL
 

Hi Marcus,

It seems to that that you nailed it.  Congratulations.
Since I already have a PHSNA I plan to use one "Experimental board" for making a HF signal generator. (I have several but making this one could be fun and instructive).  I will pay due attention to the transformer issue.
I will use a Type II DDS, an Arduino Nano and a 31 dB solid state variable attenuator module (18 € from China).  I will add a couple of relay controlled 30 dB fixed attenuators for increasing the range and an ALC circuit for compensating the DDS slope and also to perform AM modulation.  I also plan to add a LCD and an encoder and, of course, USB computer control.

73 de Ignacio EB4APL


El 20/02/2016 a las 19:57, mankan@... [PHSNA] escribió:
 

Hi all,

I have now mounted a bifilar wound transformer. I did 13 turns and cut one wire on turn 7, unwound the ends half a turn and grounded them. Picture of sweep is attached. The curve might look jagged but note the scale on the Y-axis.

Do you think I can end my debugging?

Marcus, SA5PMG

Attachment(s) from mankan@... [PHSNA] | View attachments on the web

1 of 1 Photo(s)


Posted by: mankan@...


Este correo electrónico se ha enviado desde un equipo libre de virus y protegido por Avast.
www.avast.com


Re: Sweep output debugging [1 Attachment]

Steven Dick
 

Marcus, where are you taking the output from?  Is that the final output from the output amplifier stage or directly from the transformer?  If it’s from the transformer, it is varying only a small amount and should be fine.  If it’s from the final output, the level looks low at –9db or so at the low frequencies. Should be closer to around –0.5 to +0.5 0dB or so at low frequencies at the final output. Don’t understand the last data point at the end of the plot which dropped way off
 
“Digital Steve”, K1RF
 

Sent: Saturday, February 20, 2016 1:57 PM
Subject: [PHSNA] Re: Sweep output debugging [1 Attachment]
 
 

Hi all,

I have now mounted a bifilar wound transformer. I did 13 turns and cut one wire on turn 7, unwound the ends half a turn and grounded them. Picture of sweep is attached. The curve might look jagged but note the scale on the Y-axis.

Do you think I can end my debugging?

Marcus, SA5PMG


This email has been sent from a virus-free computer protected by Avast.
www.avast.com


Re: Sweep output debugging

Marcus Gustafsson
 

Hi all,

I have now mounted a bifilar wound transformer. I did 13 turns and cut one wire on turn 7, unwound the ends half a turn and grounded them. Picture of sweep is attached. The curve might look jagged but note the scale on the Y-axis.

Do you think I can end my debugging?

Marcus, SA5PMG


Re: PC board needs survey

Steven Greenfield AE7HD
 

Thanks!

Steve Greenfield AE7HD


Re: Sweep output debugging [2 Attachments]

EB4APL
 

Marcus,

Your double peaked plot reminds me of one that I did with my PHSNA, which shows a deep dip around 30 MHz.  The reason was a cable that connected  it to a high impedance RF voltmeter and was in parallel with the PHSNA detector.  It was 123 cm long and in fact it was working as a parallel open stub and the PHSNA shown its resonance.
I also attach the plot without the open cable and a sweep that I did during the calibration, showing how well the polynomial correction worked.
My PHSNA does not use The Type II DDS Chinese board, but a DDS-60 so it does not need the transformer. 
Some toughs about the problem:
Did you removed the filter components onboard the DDS?
I think that you may be double loading the transformer, hence the DDS.  If you are loading it with 50 Ohm and at the same time using the power meter (either internal or external) you have 2 50 Ohm loads in parallel.  If you are using an external high impedance meter at the end of your test cable, you are seeing the cable.
If you have access to a signal generator, manually sweep the transformer with a load at the proper side.  I have had problems with wrong cores.
I hope some of this may help.

73 de Ignacio EB4APL



El 19/02/2016 a las 21:19, mankan@... [PHSNA] escribió:
 

Some debug done, I terminated the output with 50Ohm and connected a BNC cable to RF detector with clips in the other end and measured my way back by sweeping with PHSNA.exe and found that is already after T1 I have the dip in the signal.

Steve, you might have been spot on from the start :-). I'll try the bifilar winding but how would I do the center tap which I assume that I should ground? Do I wind 12 bifilar turns and then just cut the primary winding in the middle and ground it?

Br
Marcus, SA5PMG

Attachment(s) from mankan@... [PHSNA] | View attachments on the web

2 of 2 Photo(s)


Posted by: mankan@...



Este correo electrónico se ha enviado desde un equipo libre de virus y protegido por Avast.
www.avast.com


Re: Measurement Receiver Inductors

vasilyivanenko@...
 

Hi Mike


Toroid turns calculators and tables give approximate L values with tolerances that may vary with toroid AL variations, wire spacing and to some extent frequency in a functional sense ( we usually put them in circuits with stray L and C )  Wire gauge may effect turns spacing and also perhaps how the wires orient and stay put on the toroid core. Even the target L can influence these as more windings might mean thinner gauge wire and wire turns can bunch up and not stay in place compared to a small uH coil wound with thicker wire etc.


With nH coils @ VHF, the lead length may affect precision. I don’t use turns calculators and just measure them with a LC meter and scrunch or spread the windings to get the desire L.  When I did use turns calculators, I always wound 1 less turn than calculated and then applied the scrunch/spread method to get the desired L.


In most L-networks, a tuning cap proves critical. Most of mine apply 2 caps [ L-C-C] network because 2 caps gives you 3 tweeks. The other tweek is to again scrunch or spread the inductor while peaking the trim cap(s) for the best impedance match. SPICE might not account for stray L and C and the reactive components at 1 or both ports.


If you experiment with a return loss bridge and a simple L or L-C-C network to match a network into 50 ohms --- while measuring RL on the 50 ohm port, you will learn just how “touchy” the tuning can get.


Discovery - Transformer T1 on Type I PHSNA vs Type 2 PHSNA

Steven Dick
 

In parusing the data in the Files section, I found the following important difference between T1 on the Type I and T1 on the Type II PHSNA:  On the Type I, transformer T1 is bifilar wound (12 turns). This is described in file "Winding and installing transformers T1 and T2.pdf"  in folder  "Type 1 PHSNA - Version using Type I PHSNA.  The type II PHSNA T1 is NOT bifilar wound but has two separate windings. This is shown in file "PHSNA Type II details 21 Mar 2014.pdf in folder "Type II PHSNA using Type 2 DDS module. This configuration will have much poorer flatness and wideband performance compared to the bifilar version of T1.

I strongly recommend winding T1 bifilar for the Type II version though adding a center tap on on winding makes it a bit more complex to wind. The center tap on one of the bifilar pairs must have exactly the same number of turns on each side of the center tap for balanced push-pull operation.  Or use a commercial transformer like the Coilcraft WB1-6TSLB. cost $2.61 but 8.00 shipping

Regards, "Digital Steve", K1RF



Re: How to measure Baun Common mode impedance

Silva
 

I think You will need a RLB to do it...

73,

Silva.

Em 20-02-2016 09:46, mendiola_loyola@... [PHSNA] escreveu:
 

Dear All:

How can I measure a Balun common mode impedance using the PHSNA ? I need Rs and Xs, the impedance values could reach 10k of impedance.

Do I need an Impedance Bridge?

73
Alfredo.
OA4AJP



-- 
****************************************
Brazilian Ham Radio - ARRL Member
Callsign: PY8ELO
Name: Silva
Grid Locator: GI25ar
Address: POBOX 151 CEP 68180030
ITAITUBA - PA - BRAZIL
****************************************

"The quieter You become the more You are able to hear"


How to measure Baun Common mode impedance

Alfredo Mendiola Loyola
 

Dear All:

How can I measure a Balun common mode impedance using the PHSNA ? I need Rs and Xs, the impedance values could reach 10k of impedance.

Do I need an Impedance Bridge?

73
Alfredo.
OA4AJP


Re: Completely lost

William Kimber
 

Hi Tony.

You have managed to post message so you are obviously subscribed.

Yahoo is known for its eccentricities. That might account for some of
your problems.

Have you signed in to Yahoo then not been able to find the group when
you do a search for groups?

I'm not surprised that PHSNA doesn't show up in manage your groups.
Yahoo again, couple of years back they did an upgrade that wasn't
needed and now things just don't work any sense.

What do you expect from the group now you are subscribed?


Cheers,
Will

On 20/02/16 21:47, 'Tony Jaques' anthonyjaques94@... [PHSNA] wrote:


Hi whoever,
Sorry about that, but I have no idea where this finishes up!
Having been through the "Subscribe" process (I think), and
been advised that I am "subscribed" I find that I can do
nothing more than I could before. That is I can look at old
messages, and that's it. I could do that before!
The link provided in an email just goes to a page which
claims "No such group".
PHSNA does not appear in the list when I "Manage my
groups", so I can change nothing. For example I want the
e-mails to go to anthonyjaques94@... rather than
the address used to subscribe, but there appears to be no
way to change it - or indeed anything else.
In the end I was left with no choice but to "unsubscribe"
(Then meaning to start from scratch in the hope that it might
work a second time). But that doesn't work either.
I would really have liked to have a go at this project, and
I still would, but not I think if I have to re-invent my own
wheel from scratch.
(Tony
G3PTD..)


Completely lost

Tony Jaques <anthonyjaques94@...>
 

Hi whoever,
Sorry about that, but I have no idea where this finishes up!
Having been through the "Subscribe" process (I think), and
been advised that I am "subscribed" I find that I can do
nothing more than I could before. That is I can look at old
messages, and that's it. I could do that before!
The link provided in an email just goes to a page which
claims "No such group".
PHSNA does not appear in the list when I "Manage my
groups", so I can change nothing. For example I want the
e-mails to go to anthonyjaques94@... rather than
the address used to subscribe, but there appears to be no
way to change it - or indeed anything else.
In the end I was left with no choice but to "unsubscribe"
(Then meaning to start from scratch in the hope that it might
work a second time). But that doesn't work either.
I would really have liked to have a go at this project, and
I still would, but not I think if I have to re-invent my own
wheel from scratch.
(Tony G3PTD..)


Measurement Receiver Inductors

K5ESS
 

Wonder if others have found that the inductors called for in the Measurement Receiver need tweaking?  I found that the inductors in the impedance matching networks at each end of the crystal filter needed one less turn (11 vs 12) and needed for the turns to be all scrunched up to arrive at a value close to 8.05µH. Value was determined using the Crystal Test Fixture and measured known capacitors.(This is another good use for the Crystal Test Fixture).  Still,  I can't see any difference when tuning the variable caps associated with the matching networks at each end of the filter.  I modeled the matching networks in LTSpice with a 600 ohm 6dB pad between them and the response is pretty broad.  So are the variable caps even needed?  I'd be interested in others experience. 

Also, the inductors in the Bridge Tee Diplexer needed three fewer turns to arrive close to the specified value of 2.43 µH. 

In both cases using the AL value given by Amidon, the number of turns stated in the schematic give the specified inductance.  Somewhere i saw that Amidon stated the accuracy of AL as ± 10 to 20% so there is a wide margin on the inductance you'd get for a given number of turns.  Apparently the cores I bought vary considerably from these values.  I bought my cores from Kits & Parts.  Don't know if they were manufactured by Amidon or not.

Anyway, I'd like the hear any comments on these questions and how best to determine proper inductance values.

I did see in one of K1RF's recent posts that the inductors in the 70MHz LPFs needed reduced turns to achieve the proper cutoff frequency.


Mike K5ESS



Re: Sweep output debugging [2 Attachments]

Steven Dick
 

Just to be clearer, there should be an equal number of turns around the center tap so push-pull operation is balanced.  So there really needs to be a total of 11 turns or 13 turns. One at the center and 5 or 6 on each side of the center turn. 5 better if you want to have the best high frequency response, 6 is better is you want best low end response.


Re: Sweep output debugging [2 Attachments]

Steven Dick
 

I would first get two lengths of different colored wire.  Then take the middle of one wire and make a small loop in it about an inch long.  You can twist the loop itself a bit to keep it in a loop. Then put the two wires together and twist them together.  Wind a single turn through the toroid at the middle of the wire pair with the loop at the center sticking out. Then complete winding 6 turns around each side of the middle giving you the required 12 turns. Make sure the 6 turns on each side of the “loop” are wound in the same direction. I.E. of the loop wasn’t there, you would wind 12 turns of twisted pair.  Then tie the loop to ground. The loop becomes center tap of the bifilar wound transformer.
 
Regards, “Digital Steve”, K1RF
 

Sent: Friday, February 19, 2016 3:19 PM
Subject: [PHSNA] Re: Sweep output debugging [2 Attachments]
 
 

Some debug done, I terminated the output with 50Ohm and connected a BNC cable to RF detector with clips in the other end and measured my way back by sweeping with PHSNA.exe and found that is already after T1 I have the dip in the signal.

Steve, you might have been spot on from the start :-). I'll try the bifilar winding but how would I do the center tap which I assume that I should ground? Do I wind 12 bifilar turns and then just cut the primary winding in the middle and ground it?

Br
Marcus, SA5PMG


This email has been sent from a virus-free computer protected by Avast.
www.avast.com

761 - 780 of 3088