Splitting/daisy chaining coax ?

[KM4FER]

Caution: Hobbiest at play.


Suppose I have device A, which generates an RF signal with a 50 Ohm output, connected to device B with a 50 Ohm input using 50 Ohm coax.  Life is good.

Now suppose I want to attach device C, maybe a frequency counter, with a 50 Ohm input between A and B. 

One option would be to insert a T adapter between A and B with a branch going off to C.  But then wouldn't A see a load of 25 Ohms or is there some magic going on in the T adapter?

Another option would be to have A connected to C and then C repowers a separate connection to B.  But this would introduce a phase shift which may or may not be important to the system.

So how is this handled in the real world?

Thanks,

earl

[xrunner]

So how is this handled in the real world?

With a power splitter -

http://www.ebay.com/itm/HP-50-ohm-Power-Splitter-04192-61001-/222100349005?hash=item33b636604d:g:MqgAAOSwGYVXBwWb

[KM4FER]

OK, a power splitter makes sense, I didn't think about those.

But then what are these used for?
http://www.ebay.com/itm/10-X-BNC-Male-To-Double-BNC-T-Female-Adapter-Connector-USA-Seller-Since-1991-/272113548078?hash=item3f5b3b3b2e:g:16gAAOSwd4tUE9uh

Oh, maybe I just figured it out.  If I put one of these on the input channel of my scope in high impedance mode then essentially I could tap a coax line.  Since the connection between the adapter and the scope is so short there shouldn't be any problem with reflections.  Right ?

Thanks

[xrunner]

OK, a power splitter makes sense, I didn't think about those.

But then what are these used for?
http://www.ebay.com/itm/10-X-BNC-Male-To-Double-BNC-T-Female-Adapter-Connector-USA-Seller-Since-1991-/272113548078?hash=item3f5b3b3b2e:g:16gAAOSwd4tUE9uh

Well, you can connect several instruments to a low frequency signal, such as a DMM and o-scope, with those, because their inputs are a very high impedance.

[Delta]

So how is this handled in the real world?

With a power splitter -

http://www.ebay.com/itm/HP-50-ohm-Power-Splitter-04192-61001-/222100349005?hash=item33b636604d:g:MqgAAOSwGYVXBwWb

'ow much??!!! 

http://www.ebay.co.uk/itm/Triax-2-Way-F-Type-5-2400-Mhz-Splitter-SCS-2-349802-/111337263039

[xrunner]

'ow much??!!! 

That was simply an example of a power splitter - not the specific one I or he would buy. 

[tggzzz]

Oh, maybe I just figured it out.  If I put one of these on the input channel of my scope in high impedance mode then essentially I could tap a coax line.  Since the connection between the adapter and the scope is so short there shouldn't be any problem with reflections.  Right ?

Nope

The typical "high" input impedance of a scope isn't 1Mohm. It ought to be written on the front panel; if not it should be in the operations manual; if not, get another scope!

Usually  the input impedance is 1Mohm // 20pF.

You don't mention what frequencies interest you, but let's assume 100MHz to get a feel for the figures. The reactance of the capacitance is 1/(2*pi*f*C), or 80ohms!

In addition there will be a "spur" line in the t-piece between the coax and the scope input. This will act as a comb band-reject filter at frequencies dependent on the length.

[MrSlack]

RF sampler i.e. toroid threaded on the transmission line in a shielded box?

Also you can measure things after the termination using an RF probe if you need to. Well that's what I do with my dummy load (10W 50 ohm with detector strapped on it's bum).

[MikeLogix]

The splitter will have 3 or more dB of loss. A directional coupler would be another way to go where the A to B (thru) port would have less than a dB loss, and the coupled port would be some n- dB down from the thru port. I recently built a cheap coupler using a Mini-ckt part I snagged on ebay. I needed a quick and dirty splitter the other day so I grabbed a toroid and wound my own 180 deg splitter, it was only about 3.5 dB loss at 8 MHz which is the frequency I was using.

I have some LTSpice files and some actual measured test data for the splitter if you want to wind your own. If your working in the HF-VHF band then I can tweak  it for your frequency. If it is much higher in freq than HF then I would need some smaller OD toroids which I would need to order in which case your on your own 

[Z80]

Hi Earl,  If you are dealing with low frequency (< 100MHz) and low power (milliwatts) then you can build your own resistive splitter very easily and cheaply.  For HF frequencies, 3 resistors soldered together should work ok, if you do a careful construction job in a good shielded box then low VHF should be useable.  If you want to measure UHF or above then a commercial unit is the only real way to go to ensure it is accurate and linear.  Here is a web page which describes the splitter quite nicely:
http://www.radio-electronics.com/info/rf-technology-design/coupler-combiner-splitter/rf-resistive-splitter-combiner-divider.php

[madires]

For 10Base2 (10Mbit/s Ethernet) the T-pieces worked fine.

[AF6LJ]

There are so many answers considering the that more information is really needed to give a proper answer to fit your needs.

[KM4FER]

Well Sue, what I was thinking is kind of a Lego concept for the MF/HF bands. 

If I had a varible oscillator box and an amplifier box and a mixer box and a detector box and filter boxes and so forth all with 50 Ohm  in and out connections I could have a lot of fun configuring them in different ways.  I know I wouldn't break any records for sensitivity or selectivity but it would still be fun.

Back in the middle of the previous century in high school electronics class we had components with button snaps on the leads and you just snapped the components together to make a circuit.  We even made an AM broadcast receiver, with vacuum tubes no less.  With 250 V floating around you had to be very carefull.

[MrSlack]

That's actually pretty much what I'm working on at the moment...

So far I've got a voltage controlled VFO, PA and load. Going to add a HFA3096 based gilbert cell mixer and then a 455KHz IF block. Edit: and looking at my list, an AD9850 based synthesizer, step attenuator, another mixer, BFO and all sorts.

I started with BNC connectors and through hole but this is getting a bit large looking so I'm going to move to SMA and surface mount when I get bombed with most of aliexpress shortly. Standardised on 12v power supply as well.

Ultimately I want to have enough modules to knock out a spectrum analyser.

[PA0PBZ]

The splitter will have 3 or more dB of loss.

It's not fair to the splitter to call it loss. If you split a 10 dollar bill between 2 people they each get 5 and nothing is lost. 

[MikeLogix]

The splitter will have 3 or more dB of loss.

It's not fair to the splitter to call it loss. If you split a 10 dollar bill between 2 people they each get 5 and nothing is lost. 

LOL, Good way to look at it. 

[DimitriP]

The splitter will have 3 or more dB of loss.

It's not fair to the splitter to call it loss. If you split a 10 dollar bill between 2 people they each get 5 and nothing is lost. 

Well ...sort of....If YOU do the split, between two OTHER people, you are correct.
But if someone else takes YOUR 10 dollar bill and splits it between you and someone else, you just lost 5 dollars

[AF6LJ]

Well Sue, what I was thinking is kind of a Lego concept for the MF/HF bands. 

If I had a varible oscillator box and an amplifier box and a mixer box and a detector box and filter boxes and so forth all with 50 Ohm  in and out connections I could have a lot of fun configuring them in different ways.  I know I wouldn't break any records for sensitivity or selectivity but it would still be fun.

Back in the middle of the previous century in high school electronics class we had components with button snaps on the leads and you just snapped the components together to make a circuit.  We even made an AM broadcast receiver, with vacuum tubes no less.  With 250 V floating around you had to be very carefull.

I get what you are saying...
you have a couple of options.
A 3DB splitter that would give you two ports each slightly over -3DB of your original signal, and it would be broadband and provide about 6DB of isolation.
You could wind a hybrid transformer and play with that can of worms.
I would go with a resistive splitter.

[AF6LJ]

Oh and if this involves VHF/UHF you could use a wilkinson divider.  https://en.wikipedia.org/wiki/Wilkinson_power_divider

[MikeLogix]

Well Sue, what I was thinking is kind of a Lego concept for the MF/HF bands. 

If I had a varible oscillator box and an amplifier box and a mixer box and a detector box and filter boxes and so forth all with 50 Ohm  in and out connections I could have a lot of fun configuring them in different ways.  I know I wouldn't break any records for sensitivity or selectivity but it would still be fun.

Back in the middle of the previous century in high school electronics class we had components with button snaps on the leads and you just snapped the components together to make a circuit.  We even made an AM broadcast receiver, with vacuum tubes no less.  With 250 V floating around you had to be very carefull.

I get what you are saying...
you have a couple of options.
A 3DB splitter that would give you two ports each slightly over -3DB of your original signal, and it would be broadband and provide about 6DB of isolation.
You could wind a hybrid transformer and play with that can of worms.
I would go with a resistive splitter.

A two-way (3-port resistive splitter has 6 dB loss. For resistive splitters The isolation of a resistive splitter is equal to its insertion loss.
Knowing the actual application would help in determining the requirements for loss and isolation.   

[MikeLogix]

Oh and if this involves VHF/UHF you could use a wilkinson divider.  https://en.wikipedia.org/wiki/Wilkinson_power_divider

The Wilkinson is typically used at Microwave frequencies due to the fact that it requires 1/4 wave stubs (usually done with PCB traces). At VHF/UHF the 1/4 wave stubs would be to large to be practical. At least in my experience>

[Howardlong]

'ow much??!!! 

I use these, although I don't pay EUR 160 for them. The HP equivalents to 26GHz are even more expensive. I accidentally mangled one of the receptacles on one a few weeks ago, it's not hard to do if you're not properly engaged in the job in hand. Those rated to 18GHz, and even worse the SMA compatible APC 3.5mm to 26GHz are terribly easy to mash as there's no solid dielectric to make the centre receptacle fully mechanically rigid.

http://www.ebay.com/itm/Weinschel-1515-1-Broadband-Resistive-Power-Divider-with-mounting-bracket-/291675716306?hash=item43e93a3ad2:g:TcUAAOSwpzdWsMnz

How about this one?

http://www.ebay.com/itm/Agilent-HP-11667B-Power-Splitter-DC-to-26-5-GHz-CALIBRATED-/251999508327

[MikeLogix]

So how is this handled in the real world?

With a power splitter -

http://www.ebay.com/itm/HP-50-ohm-Power-Splitter-04192-61001-/222100349005?hash=item33b636604d:g:MqgAAOSwGYVXBwWb

'ow much??!!! 

http://www.ebay.co.uk/itm/Triax-2-Way-F-Type-5-2400-Mhz-Splitter-SCS-2-349802-/111337263039

That splitter you linked to on Ebay is for 75 ohm impedance. There performance is quite poor as well.

[MikeLogix]

RF Bay Inc. (Ebay Store in US) Makes a real nice low cost ($12.00) PCB for some SMT Mini Circuit splitters, and these come out real nice. Only takes 5 mins to put together (not including time to order and get parts). See image below:
They also make PCB's for many single RF components, like Amps, directional couplers, mixers, etc. I have purchased quite a few boards from these guys and they work perfect. They come out real clean too. So this is a good option in my opinion (for whatever thats worth) 
Here is a link for there full list of PCB's.
http://stores.ebay.com/RF-Basic-Store/Develop-PCB-/_i.html?_fsub=10&_sid=98139492&_trksid=p4634.c0.m322



Hope that helps.