开云体育

I am currently designing a receiver for the 40 meter band and will be using an NE602 for my mixer. I have most of the individual circuits built (pre-amp, bandpass filter, crystal filter etc) and they are all 50 ohm impedance.

As the NE602 has around 1.5k ohm input impedance I have have wound a 4:22 toroid for the RF input to convert from 50 ohm to 1.5k ohm.

Using my NanoVNA I want to measure the impedance (and return loss / VSWR) of my RF input circuit so I can adjust the windings to get a good match but the NanoVNA outputs around -13 to -9 dBm, this is too high for the NE602 RF input which has a maximum input of about -25dBm.

How would I go about attaching the NanoVNA to the RF input of the NE602 without causing an overload?

Thanks in advance for any suggestions or advise,

Kerr

Would a 50 ohm pad between the VNA and the input (50 ohm) to the transformer?

The input of the 602 can take more RF but the output circuits are in overload.
That would make gain compression if you were doing S21, for S11 it should
be fine.

The problem is there is a level where the base to base resistance may go
out of range but I believe its greater than 0dbm than suspected without damage.
Either way you will not break it at any level under 0DBM, why are we sure?
The nano VNA has several of them in there! Look at the schematic.

An aside the 602 is more gain than needed at 40M and input coupling loss
will help not hinder that. Most 40M 602 designs do not lack for sensitivity with
even with obvious input loss.

Allison
-----------------
No direct email, it goes to bit bucket due address harvesting in groups.IO

Please be aware the stated input impedance is per input so if you use the NE602 in balanced input mode (and you should) it's twice the resistance and half the capacity.
The impedance is fairly stable till above 100MHz so you better connect the transformer to a 3k resistor

Even better is to search for NXP AN1994 where the formula for calculating the optimal matching network is explained

--
NanoVNA Wiki: /g/nanovna-users/wiki/home
NanoVNA Files: /g/nanovna-users/files
Erik, PD0EK

Thank you very much for the information - it was just what I was looking for.

I will directly connect the NanoVNA to my input circuit to see how it looks now I know that I will not damage the IC by applying too much power - I am mainly wanting to do an S11 measurement so I can try to get a good match on the RF input.

Currently my circuit shows a 15dB gain when I apply a -30dB signal to the RF input.

My current test circuit (set up on a breadboard until I get a rough design in place) is in balanced mode and I was wondering about the 1.5k and 3k input impedance as I had read about this a few times. I think I will need to re-work my toroids as currently they do a 50 to 1.5k ohm transformation but I may also try redesigning this section and use the method described on page 10 of AN1994 (I will need to change a few details in the calculations as the PDF is for an SA605 which appears to have a higher input impedance at around 5k ohm):



Thanks again for the suggestions - they are really helpful.

I have been reading a bit more about these matching networks and have found another one I think I will try as well.



Starting on page 15 there is a worked example of how to create a 'bandpass matching network to transform 50 ohms to the input of an SA602 mixer'.

Noting the suggestion from the previous posts about balanced inputs I think this particular circuit is unbalanced as all the components are grounded just before the input. I think using this circuit I would put its output to pin 1 of the NE602 and ground pin 2 via a 47nF capacitor (I have seen this configuration in several other examples).

Initially I have modelled this network in SimSmith and adjusted the components to create a match at a frequency of 7.15MHz (as I think this is unbalanced I have used 1.5k ohm and 3pF for the load), the simulation looks good and will be a great starting point for building the circuit for real. I will now test these new ideas out in practise and see which one I think works the best for me.

My reason for doing these experiments is that I find it a great way to learn and I also really enjoy making radio circuits.

If anyone has any additional comments or suggestions I would be extremely happy to receive them.

Kerr,

Wes Hayward (W7ZOI) did some tests on the NE602 (lower temp range than SA602) and published them in his book Experimental Methods in RF Design. The balance input configuration will give you better performance. Here is an excerpt from his book showing the test results and how to wind the transformers....

Thanks for the book page image - this is essentially what I have currently breadboarded up although my toroids are wound for 1.5k rather than 3k. I am going to try removing a turn off the primary side and adding a turn on the secondary side initially as this will give me a 50 to 3k ohm match (well 50 to 2938 ohm).

The other difference is that the output transformer in the image uses an FT-50B-43 and I am using an FT37-43 for both input and output transformers as these are the ones I have plenty of.

Once I have made a few measurements I will then wind new toroids using the Wes Hayward numbers and see how they compare, this should be a 50 to 3828 ohm match. It will be interesting to see what difference the two windings make to the return loss / VSWR and output.

Thanks again for this suggestion.

Kerr,

I have wound these high ratio transformers using FT50-43 and FT37-43 toroids and always found leakage inductance which affected VSWR, Bandwidth and Insertion Loss. I then read about W8JI's experience with BN73-202 and BN43-202 binocular cores. You get better coupling with these and less leakage inductance. The initial permeability (2500) of the BN73-202 is so high that you only need a few turns to meet the minimum 4X reactance at the lowest frequency. So for a balanced input with a 3000-50 ohm match 1 turn on the primary and 8 turns on the secondary would work well @ 7 MHz. The 73 type material has low resistivity so I recommend that you use kevlar wirewrap wire or Cat5/6 insulated wire becasue enamel wire might nick and contact the core.

On Sun, Mar 8, 2020 at 08:00 AM, Kerr Smith wrote:

Using my NanoVNA I want to measure the impedance (and return loss / VSWR) of
my RF input circuit so I can adjust the windings to get a good match but the
NanoVNA outputs around -13 to -9 dBm, this is too high for the NE602 RF input
which has a maximum input of about -25dBm.

How would I go about attaching the NanoVNA to the RF input of the NE602
without causing an overload?

The NanoVNA output with older firmware is around the levels you posted. With Hugen's latest firmware it goes up to 0 dBm.

Let's assume you are using the older firmware and you want to get a test level lower than -13 dBm. There is a technique you can use to reduce the level being input to the Device Under Test (DUT). What you need is a good 20 dB SMA attenuator like the ones posted below. Connect a short SMA cable to CH0 of the NanoVNA and the other end to the input of the attenuator. Then set your frequency range to the band of interest (like 6 to 8 MHz for your 40M tests). Go into the Cal menu and do a Reset to clear all the cal parameters. Then do a calibration in the normal way placing the short, open and load on the output of the attenuator. When you have finished save to one of the memory locations (1-4).

Now you have calibrated with the attenuator in place and you can measure Return Loss and VSWR by setting your Traces accusingly. The maximum signal into the device, in this case your impedance transformer, will now be -29 to -33 dBm. This "trick" has some inaccuracies compared to the no attenuator method but I find it works well with a VSWR from 1 to 4. It works even better with 10 dB of attenuation but much worse with 30 dB.

Disclaimer: This is not my discovery. I found this technique posted by someone else on this group several months ago. Maybe they will follow up this post with their comments...

Hi Roger,

Thanks for the suggestions, I have the attenuator shown in the picture and did wonder if I could use this but was not sure about the calibration procedure. I will try this as well, currently I have my NanoVNA connected directly to the transformer and have been testing various winding combinations to get a good match. So far I have got a 1.5 to 1 VSWR at 7.15MHz but the circuit is still on a breadboard so I expect the numbers to change a bit when I solder it up on a veroboard.

As kb1gmx mentioned above, the NE602 has a very good gain at 40m and I am sure that what I have now would work quite well but I am quite curious if I can make a really good match now I have a VNA to work with. The other radios I have built for 40 and 20 meters have all been quite successful and I did not have a VNA for these - now I do have one I am finding it very interesting to experiment and really see what the changes I make actually do.

I have had a look around for the BN73-202 binocular cores you mentioned and have found them at the same place I buy my FT37-43 cores - I will order a few of these and try them out - hopefully they will arrive in a day or so.

this is slightly off topic but the ne602 is prone to overload. I recently built an upconverter and I used one of these that works quite well
.ADE-1-24 ADE-1-24+ DOUBLE BALANCED Frequency Mixer RF
I found that its lack of gain made no difference in the 1 -30mhz area. I see them on ebay

I have received my new toroids (BN-73-202) and have been having a look at the difference between them and my FT37-43 ones. In this reply I have attached three images.

The first is the NE602 with an FT37-43 on the input and output - with -30dB in at 10MHz the output is -15.79dB
The second image has a BN-73-202 on the input and an FT37-43 on the output - here the output is -23.69dB
The third image has a BN-73-202 on both input and output - here the output is -33.98dB

From these tests I can see that lots of the harmonics disappear when using two BN-73-202 toroids but the gain drops quite a bit.

I did try adding and removing turns on the BN-73-202 but the best ratio seems to be 0.5 to 8 (1 full turn on the primary reduces the gain and matching).

I will post the NanoVNA results in the next reply.

To this reply I have attached the images from my NanoVNA, there are two VSWR plots and two smith chart plots.

When using the FT37-43 toroids on the input and output I get about a 1.58 to 1 VSWR at 7.15MHz
When using the BN-73-202 toroids I get 1.093 to 1 VSWR

(When using a full turn on the primary the VSWR is 2.6 to 1 at 7.15MHz and increases rapidly, at 15MHz it is about 6 to 1)

From these two tests I can see the match is much better when using the BN-73-202 toroids and these work well up to 30MHz (VSWR 1.45:1).

Looking at my results so far I can see I get a good match when using 2 x BN-73-202 toroids but the gain out of the NE602 is reduced quite a bit.

hi Kerr,

I really don't understand the 0.5 to 8 ratio and the "full turn" comment. In toroidal cores, one pass through the hole counts as a full turn, there are not things as half turns, so probably your "full turn" is in fact 2 turns.

Regards,

Ignacio EB4APL

--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.

The BN cores aren't actually toroids, they are "balun" cores or "binocular" cores that have two holes through them. So, if you wind through only one hole, you have a half turn.

Ok, understood, but a picture would help to clarify it.

--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.

Kerr,

Something is off with your results. I don't know how you are winding the binocular cores. You need full turns and one turn is a complete loop through one hole and back again. Try measuring VSWR and insertion loss by making two transformers and connecting them back to back with a 50 ohm load on the second core.

Kerr,

Here is a link with actual tests of transformers made with 73-202 binocular cores. Make sure you use insulated wire because 73 Mix material has low resistivity compare to Mix 43.



Roger

Hi Roger,

When I do one full turn on the primary (through the first hole and back down the second) I get the results as shown in the attached images.

This is why I removed part of a turn to see what would happen.

When I measure the inductance of one turn I get 7.38uH which is as close as I can get to my calculated value of 4.5uH (using 4 x 50 ohms for the impedance).

If I have a full turn I get about 12.5uH inductance which is about three times too large when compared to the calculated value.