HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

General discussions
sdrom33
Posts: 128
Joined: Sun Dec 09, 2018 4:38 pm

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by sdrom33 » Wed Jan 23, 2019 5:24 pm

Hi glovisol, thanks for all the info for the filter, but i find difficult to get the toroid. do you have a seller pse? Very intersting!

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Wed Jan 23, 2019 6:33 pm

Sdrom33,

You can find these toroids, as described in previous posts, from ebay vendors. There is where I buy.

====

Tony,

I forgot to mention that type 2 material has the best temperature stabilty, 96 PPM/°C as compared to type 15, 190 PPM/°C or to type 1, 280 PPM/°C. This means that in a filter built with type 2 material its inductance values will drift two times less than one built with type 15 material and three times less than one built with type 1 material.

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Wed Jan 23, 2019 10:01 pm

ERRATA CORRIGE

In the "Experimental prototype illustration" there is a mistake in the NOTE about L3-L4. The correct text follows:

L1=L7 NOTE: initial measurements with coils again with type 26 material, but losses were too high and cutoff transition not sharp (see posted initial PC screen). Then put in large diameter air core coils (on Kodak film former, Q>150) and performance improved as shown in the PC screen posted next. Then used toroids with type 2 material and performance was the same as with air core coils.
.......
L3=L5: Toroid core etc.....
........
All enamelled copper wire used is 0.5 mm, not 0.35 mm.
..........
Toroid core T80-2 has Al=55, NOT Al=49

Apologies for all these mistakes!,

glovisol

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Thu Jan 24, 2019 5:47 pm

My HP 4342A Q Meter is back!!
Attachments
HP Q meter.JPG
HP Q meter.JPG (75.83 KiB) Viewed 48064 times

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Sat Jan 26, 2019 11:13 am

WORK IN PROGRESS - HIGHER PASSBAND RIPPLE PROTOTYPE - ALL COILS WITH HIGH Q TYPE 2 MATERIAL

A second prototype has been built and tested to assess:

1) The advantage of employing all coils with type 2 High Q Toroid material.
2) The effect of choosing a higher ripple Chebyshev prototype.
3) The use of coil taps in place of links.

Spectrum analyser measurement shows that higher prototype ripple is not the way to go. The advantage of choosing a higher ripple is that, all parameters being equal, lower inductance coils are required, easier to wind and lower copper losses, but, as we see, at the cost of an unacceptable attenuation change within the passband.

Use of type 2 material has brought about a big improvement in selectivity, insertion loss and ultimate attenuation. Insertion loss, at the ripple peak at 2.2 MHz is below 1 db. Attenuation at 1.5 MHz is now 40 db, very near the teoretical value of 44 dB and ultimate attenuation is better than 48 dB, near the system measurement limit.

The next step will be to test a third prototype chosen with a significantly lower ripple parameter.
Attachments
Chebyshev N=7 Experimental prototype #2.jpg
Chebyshev N=7 Experimental prototype #2.jpg (128.21 KiB) Viewed 47967 times
HI PASS 6.jpg
HI PASS 6.jpg (139.26 KiB) Viewed 47979 times
Second filter prototype test.JPG
Second filter prototype test.JPG (61.43 KiB) Viewed 47979 times

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Sun Jan 27, 2019 3:47 pm

THIRD CHEBYSHEV N=7 HIGH PASS FILTER PROTOTYPE - FINAL RESULTS

Performance of the previous prototype was objectionable because of high ripple in the passband and higher than expected insertion loss which was thought to depend only on the 0.5dB ripple of the tabular data. The new prototype was chosen for 0.1 dB ripple, but measurement results were not satisfactory. It was discovered that in-band ripple causing a wide through in the passband was due to insufficient coupling of the impedance trasformation link. The link was removed and both external coils equipped with a tap, with a very noticeable improvement in insertion loss and passband ripple, as can be seen by comparing the PC screen with that of prototype #2.

Changing the ripple from 0.5 to 0.1 db in the tabular data causes the inductance of all coils to increase uncomfortably. For this reason internal impedance level was decreased from 500 to 300 Ohm and total number of turns & tap turns re-calculated accordingly. Type 2 material toroids have an extremely uniform and accurate Al: once the calculated number of turns was applied, measurement on the HP 4342A Q meter (+/- 3% accuracy) verified the calculated inductance to within measurement tolerance.

Once the filter is put together, if capacitors (which have standard values) are measured to be accurate, each coil is adjusted for best performance by adding or removing one turn, to obtain the best compromise. To achieve this, one terminal for each coil must be kept a little longer and trimmed once the right turns combination is found.

The question of coupling to toroid inductors requires further investigation.
Attachments
Chebyshev N=7 Experimental prototype #3.jpg
Chebyshev N=7 High Pass Filter prototype data
Chebyshev N=7 Experimental prototype #3.jpg (121.65 KiB) Viewed 47932 times
HI PASS 11.jpg
Direct connection noise generator - RSP-1A to check for insertion loss by comparision
HI PASS 11.jpg (141.24 KiB) Viewed 47932 times
HI PASS 10.jpg
Chebyshev N=7 third prototype filter performance
HI PASS 10.jpg (139.38 KiB) Viewed 47932 times

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Mon Jan 28, 2019 7:48 am

THIRD CHEBYSHEV N=7 HIGH PASS FILTER PROTOTYPE - ADDITIONAL MEASUREMENTS
Attachments
HI PASS 8.jpg
Chebyshev N=7 prototype attenuation in stopband
HI PASS 8.jpg (139.96 KiB) Viewed 47860 times

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Mon Jan 28, 2019 5:20 pm

FOURTH CHEBYSHEV N=7 HIGH PASS FILTER PROTOTYPE - TOROID COIL COUPLING

The relevant difference between link coupling and tap coupling to the exterior coils warrants further investigation. To this end a fourth prototype at Z=50 Ohn in/out has been calculated (schematic uploaded below) and will be tested to compare its performance with that of the third prototype. The tabular parameters taken are identical to that of the previous prototype, so it will be possible to obtain a meaningful comparison. In the schematic all test data, including coil Q is omitted as measurements have yet to take place.
Attachments
Chebyshev N=7 Experimental prototype #4.jpg
Chebyshev N=7, r=0.1 dB, fourth prototype
Chebyshev N=7 Experimental prototype #4.jpg (104.57 KiB) Viewed 47812 times

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Wed Jan 30, 2019 1:30 pm

FOURTH CHEBYSHEV N=7 HIGH PASS FILTER PROTOTYPE - TOROID COIL COUPLING PROBLEMS

Further tests and comparison between third and fourth prototype revealed that unfortunately inserting a link or a tap in the external toroid coils to match lower impedances is not possible because coupling changes with frequency and insertion loss at higher frequencies increases out of proportion. It is like having an unsymmetrical bandpass, with a very step low frequency skirt and a gradually sloping high frequency skirt.
This happens because the low permeability of iron powder material causes magnetic field leakage resulting in poor coupling between windings and increasing high frequency attenuation.....From RFA Micrometals RF catalogue
On the other side of the coin it was possible to buid and test the 50 Ohm in/out filter, once high stability, 1 nF mylar capacitors were obtained. Working at low impedance requires relatively low inductance coils, that can be easily built as air core, with no toroids, as posted below. There is a very good correspondence between calculated (by Coil 32) and measured number of turns (by HP 4342A Q meter). At this frequency calculated Q (Coil 32) is very pessimistic, compared to measured Q, even after the prescribed -20% correction.

We are now on safe ground to finally calculate and test the final Hi Pass prototype going back to N=9 for improved selectivity, as originally planned. This will be covered in the next post.
Attachments
HI PASS 12-7-50.jpg
Prototype #5- Identical design to #4 but air core coils
HI PASS 12-7-50.jpg (137.81 KiB) Viewed 47687 times
Chebyshev N=7 Experimental prototype #5.jpg
Chebyshev N=7, r=0.1 dB prototype @ Z=50 Ohm in/out
Chebyshev N=7 Experimental prototype #5.jpg (89.27 KiB) Viewed 47687 times

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: HIGH PASS FILTER FOR OPTIMUM HF RECEPTION

Post by glovisol » Wed Jan 30, 2019 2:53 pm

PRACTICAL TEST OF THE CHEBYSHEV N=9-r=0.1- Zin=Zout=50 Ohm (1)

The final filter N=9 has been calculated, built and tested. Here below results when filter is inserted between antenna and the RSPduo 50 Ohm input. There is a stepdown (550 to 50 Ohm) balun between antenna & filter input. In practice all signals from DC to 1600 KHz are wiped out. This performance is surprising, considering the filter is a breadboard prototype with no shielding enclosure.

Filter data & details in next post.
Attachments
MW band no Filters.jpg
MW broadcasto band from Beverage antenna
MW band no Filters.jpg (193.69 KiB) Viewed 47678 times
MW band Internal Notch in.jpg
RSPduo internal notch MW band attenuation.
MW band Internal Notch in.jpg (185.68 KiB) Viewed 47678 times
MW band HI PASS FLT IN.jpg
Chebyshev N=9, r=0.1 dB, 50/50 Ohm HP filter. MW band attenuation.
MW band HI PASS FLT IN.jpg (186.72 KiB) Viewed 47678 times

Reason: No reason

Post Reply