Wenzel Low Frequency Circulator

[uofmrob]

I think I'm going to make a big order for all the stuff tomorrow and give it a whirl. Thanks so much for all the help everyone. This forum is such an amazing resource with so much knowledge and generosity!

[coppercone2]

So without any pads, do you just glue the ICs upside down on ground plate and solder to the connectors?I can 3D print supports for the BNC connectors pretty easily and glue them down.

Sometimes I super glue parts upside down and bend the leads down and solder them to ground but it depends on what part you have. IMO this kind of construction gets old after a while.. and it helps if you have precise hand tools.

The problem is if you have a thick ground plane, the chip and glue gets real hot when you are tacking the nearby lead to ground, so it could help if you pre-tin it. And you can glue copper to something, cut a hole out under a chip, then glue the chip not to the copper, so you don't have heat issues..

I found that its kind of harder then it looks, and you end up with messy construction unless you do alot of thinking.

[uofmrob]

Oh, and another question I had. For this type of circuit, should the ground be an earth ground?

[coppercone2]

i would power it off batteries. if you have a real circulator or isolator, they don't have any connection on them other then the ports. They are usually a passive magnetic element. I would design it as a module you can use on a RF bench with batteries so you can use it the same as a regular circulator to keep things simple and reliable on some AAs.

Once you setup some microwave assemblies you will be happy there is less wires.

those RF jacks connect to the RF equipment, which is usually connected to ground, by the connector on the chassis.

in a RF assembly typically you see a can with little solder feed through connections that the DC power goes into. I don't think its typically earthed?

[RoGeorge]

This was my first prototype using these op amps  (with most of the bare wire replaced with insulated wire before I powered it up once I got the connectivity right).

https://www.mouser.com/ProductDetail/NJR/NJM2137D?qs=54bfbthyeuO6TvPIDHWJog%3D%3D

 Initially, I just powered it directly from my bench power supply, then I watched a bunch of videos and learned more about op amps and figured out that I needed to use my dual tracking supply in serial mode. I'm only shooting for a 25 MHz spectrometer. I've never built something like what you had in the photo. Is the copper plate the ground and then everything else is wired directly above the plate?

You may want to look at the following prototyping techniques:







For high frequency, pick a method that uses a ground (copper) plane.  Shielding might help a lot, too.

Solderless breadboards are to be avoided for high frequency circuits.  Those breadboards produces unpredictable results, mostly because their internal traces have a lot of stray inductance, capacitance and high contact resistance.  In breadboards, internal traces are in fact parallel sheets of metal.

[uofmrob]

Thanks!

[uofmrob]

I had actually just subbed to this channel. There's some great content! I had just watched all of the NanoNVA videos. I have a NanoVNA 2 coming soon to help with the coil design and matching. I have a super expensive Network analyzer at work, but I wanted to have one at home. Part of this project is to see how cheaply I can make an EPR. It's going to be baby steps, but eventually, I want to drive the source and magnet with a raspberry PI. There are a lot of these out there as people have linked above, but those are all based on the original instruments that detected the effects. I'm shooting more for a magnetic field sweeping spectrometer based off of the modern instrument designs, just operating at lower fields with off the shelf components. If I can get it to work well with relatively simple and inexpensive parts and methods, I'm probably going to present it at the Rocky Mountain EPR conference. You know, if it ever meets again.

[uofmrob]

OK, I ordered some individual op-amps and some copper clad boards. They should all be here Friday. It looks like I have a weekend project!

[David Hess]

So without any pads, do you just glue the ICs upside down on ground plate and solder to the connectors?I can 3D print supports for the BNC connectors pretty easily and glue them down.

What I generally do is anchor the package right-side-up above the ground plane using the decoupling capacitors between the ground plane and supply pins.  Then I use either wire wrap wire or enameled copper to do point to point wiring.

However in a real layout, the equivalent of not using pads is to cut away the ground plane to raise the impedance which the above method does not allow whether pads are used or not.

[uofmrob]

I was thinking that for the places in the circuit where multiple connections come together, I would use my rotary tool and cut an island in the coper board

[David Hess]

I was thinking that for the places in the circuit where multiple connections come together, I would use my rotary tool and cut an island in the coper board

That  is a common method; I still prefer to glue or solder a little pad down if I am not wiring things in the air.  Strips of copper clad board can be used as controlled impedance transmission lines.

High value surface mount resistors can also be used as terminals and to anchor other parts.

The photograph below shows an example where I used the leaded multilayer ceramic decoupling capacitors to anchor DIP sockets.  The RF bridge detector using 1N270 germanium diodes worked to 100s of MHz.

[KE5FX]

If you try that with modern current-feedback amps, you'll end up with a competitive entry in the "Who can build the highest-frequency oscillator on a breadboard?" thread.  They will complain because you didn't actually use a solderless breadboard, but the spirit of the rules will certainly have been satisfied.

But hey, knock yourself out.

[SilverSolder]

I was thinking that for the places in the circuit where multiple connections come together, I would use my rotary tool and cut an island in the coper board

That  is a common method; I still prefer to glue or solder a little pad down if I am not wiring things in the air.  Strips of copper clad board can be used as controlled impedance transmission lines.

High value surface mount resistors can also be used as terminals and to anchor other parts.

The photograph below shows an example where I used the leaded multilayer ceramic decoupling capacitors to anchor DIP sockets.  The RF bridge detector using 1N270 germanium diodes worked to 100s of MHz.

Artwork, love it  -  Soldering the battery in is very hardcore!

[uofmrob]

I have another question for all of you experts. I've been thinking about this circuit design and I think it might not actually work for what I need. At the coil, I need about a watt of power, but will I actually get no power carrying forward in this circuit? Sure, I'll get a voltage carrying forward from the op amps, but will I get no significant current because of how the op amps work?

[uofmrob]

With the ferrite-based circulators in our instruments, they carry forward the 200 mW microwaves in our continuous wave instruments and even the 1000 W microwaves in the pulse instruments.

[uofmrob]

Thanks for the explanation!

[KE5FX]

I have another question for all of you experts. I've been thinking about this circuit design and I think it might not actually work for what I need. At the coil, I need about a watt of power, but will I actually get no power carrying forward in this circuit? Sure, I'll get a voltage carrying forward from the op amps, but will I get no significant current because of how the op amps work?

In combination with the available supply voltage, the power output is determined by the sum of the output impedance of the opamps, typically a few ohms, plus whatever series resistance is added at their outputs.  The output current capability of the opamps is closely related to their output impedance; often the data sheet will show both. 

The circuit as published uses opamps rated for output currents up to 70 mA, so that's about a quarter-watt into 50 ohms.  Power in the 1-watt range is achievable, but you'll need to go with beefier opamps.   The THS3491 would be a great choice since it can run from +/- 12V rails.  At lower voltages you'd need to scale down the series termination resistors at the outputs, and you're better off not doing that for various reasons (notably the ability to tolerate a reasonable amount of load capacitance and compatibility with 50-ohm cabling and components).

[uofmrob]

It looks like the ones that I ordered have an output current up to 100 mA. Then I think I should be OK. 1 W would be great, but even a quarter watt should be enough to see something if I can get a well-matched saddle coil with a high Q. Thanks again!

[David Hess]

I have another question for all of you experts. I've been thinking about this circuit design and I think it might not actually work for what I need. At the coil, I need about a watt of power, but will I actually get no power carrying forward in this circuit? Sure, I'll get a voltage carrying forward from the op amps, but will I get no significant current because of how the op amps work?

Power levels are limited by the output voltage and current capability of the operational amplifiers.  Source termination further lowers the output power capability.

I think practical designs with integrated parts are going to be limited to less than 100 milliwatts.  36 volt and 200 milliamp parts can do a little better than this.

[uofmrob]

Thanks for the advice! I think that I'll try out the design and the ugly construction techniques that you all suggested this weekend with the ones that I ordered from Mouser and see if I can just get one that works at all and then look into more powerful op-amps for a second version if I can get this one to work.

[uofmrob]

So something like this might be better for the final product:

https://www.mouser.com/ProductDetail/Texas-Instruments/THS3091DDAR?qs=ZmJdcv7QZ9oZqiofJS38Ww%3D%3D

[uofmrob]

Really, I should just bump the operating frequency up and buy something like this. It would certainly be easier. But not as fun and educational. 

https://www.ebay.com/itm/400-420MHz-RF-coaxial-circulator-Isolator-SMA-Joint-frequency/254530755929?hash=item3b43371d59:g:K2oAAOSwsiBeX2En

[uofmrob]

Well, I learned another lesson. Buy more components than you need. 🤣. I snapped a pin off of one of the op amps, and clipped another one getting it out of the package. Those buggers are tiny! I've got another order in, but for now two of them are wired up with the filter capacitors holding them up and they are wired for power!

[KE5FX]

A good start!   

And yes, never order just one or two of anything.

[uofmrob]

Hi everyone. I have it a whirl again tonight. The small surface mount op amps that I got are just too small. Well I can deal with the size, but the leads are just too fragile. Once you bend them once, any further flexing causes then to snap. I got them all wired up to power, but when I started soldering the resistors on, the leads kept snapping. This is what I have:

https://www.mouser.com/ProductDetail/Analog-Devices-Linear-Technology/LT1395CS5TRMPBF?qs=ytflclh7QUWCZinjfMpnVQ%3D%3D

Can anyone recommend a part number for a good op amp for this application that would come in a DIP IC that would be easier to work with?
Thanks!