Spectrum Anaslyzers; is a 50m to 75 ohm xformer really necessary?

Status
Not open for further replies.

videobruce

Member
Joined
Jan 5, 2004
Messages
90
Location
Buffalo NY
I have a Spectrum Analyzer with a tracking generator. The input and output are 50 ohm.
I mainly use it for receive, OTA TV which is 75 ohm.
What benefit would I get from using a matching xformer for the input and output when using this generator other than a more accurate reading?

Anyone suggest a source that has a 'adapter xformer' with a N male to a F female?

Thanks.
 

zz0468

QRT
Banned
Joined
Feb 6, 2007
Messages
6,036
I have a Spectrum Analyzer with a tracking generator. The input and output are 50 ohm.
I mainly use it for receive, OTA TV which is 75 ohm.
What benefit would I get from using a matching xformer for the input and output when using this generator other than a more accurate reading?

Anyone suggest a source that has a 'adapter xformer' with a N male to a F female?

Thanks.
There are some applications where precision measurements are needed and impedance matching is critical. It's not critical, if you're just looking at signals over the air. If that's what you're doing, then there will be no discernible benefit.

If you're trying to sweep tune 75 ohm filters or amplifiers, the transformers could be important, because even a slight mismatch could cause an error in tuning.
 

videobruce

Member
Joined
Jan 5, 2004
Messages
90
Location
Buffalo NY
The filters I use are single channel TV traps. Were/how would using a pair of matching xformers in line make a difference?
 

zz0468

QRT
Banned
Joined
Feb 6, 2007
Messages
6,036
The filters I use are single channel TV traps. Were/how would using a pair of matching xformers in line make a difference?
Filters are very dependent on the impedance that terminates each port. If there is a mismatch, that can impact frequency response, insertion loss, ripple, group delay, and just about every other characteristic that a filter has. Where it becomes important is things like multipole wideband filters that have critical needs to be particularly flat across the pass band, and ripple and group delay specifications must be known and kept under control.

Now, practically speaking, in your application, I doubt you'd notice.
 

gcgrotz

Member
Joined
Feb 21, 2006
Messages
2,093
Location
Charlottesville
I will agree with zz. Having used many spec analyzers, unless you are doing some kind of precise lab measurements, I wouldn't worry about 50 vs 75 ohm mis-match. Getting a good quality fractional ratio unbalanced/unbalanced transformer probably won't be easy or cheap. The one you listed looks pretty good, but I see they don't give a price. Kind of like an expensive restaurant? I'd love to know how it works out, and if you can see any difference.
 

prcguy

Member
Joined
Jun 30, 2006
Messages
10,531
Location
So Cal - Richardson, TX - Tewksbury, MA
50/75 matching transformers exist but they are not what you want in a measurement setup. A resistive matching network called a "minimal loss pad" is used instead and it provides a perfect match to both sides of the circuit. These pads have 5.7dB loss each which must be accounted for in the measurement. Check Mini-Circuits or Ebay.
prcguy
 

videobruce

Member
Joined
Jan 5, 2004
Messages
90
Location
Buffalo NY
I have seen those also, but was concerned about the loss. Yes, I know you can adjust for the difference, but it also dropps sensistiivity of the scope also. YTrying to measure weak signals would suffer.
Why are these better??

BTW, yes I know 'Analyzers' is spelled wrong, but the forum doesn't allow editing.
 

prcguy

Member
Joined
Jun 30, 2006
Messages
10,531
Location
So Cal - Richardson, TX - Tewksbury, MA
The minimal loss pad presents a resistive load to the generator and device under test. Depending on what you are testing the VSWR can vary wildly like a band pass filter out of band and a transformer can have frequency response problems when presented with high VSWR. This can show up when you calibrate the test setup and your filter or other passive device shows a little gain at some frequencies. If you need more signal you can amplify the generator output but its good to place a small value attenuator pad (3dB, 6dB) at its output to help isolate the amp from external VSWR influence. In fact, whenever I am making critical measurements with a scaler or vector network analyzer, I always use a 6dB pad at the end of every cable to the device under test to provide a better match to the device and reduce cable interaction. These losses are something you have to be aware of in the measurement setup and try to have at least maybe 15dB or so more dynamic range available than the lowest level you need to discern in the measurement.
prcguy
I have seen those also, but was concerned about the loss. Yes, I know you can adjust for the difference, but it also dropps sensistiivity of the scope also. YTrying to measure weak signals would suffer.
Why are these better??

BTW, yes I know 'Analyzers' is spelled wrong, but the forum doesn't allow editing.
 

videobruce

Member
Joined
Jan 5, 2004
Messages
90
Location
Buffalo NY
Ok;
1. Is a 50 to 75 ohm ( 1.5:1) mismatch really that much a problem? From a xmit standpoint I know it isn't.
2. Would just a "pad", say 3db, make a difference without any matching?
 

prcguy

Member
Joined
Jun 30, 2006
Messages
10,531
Location
So Cal - Richardson, TX - Tewksbury, MA
The mismatch is a problem if you want a precise measurement. For example, a filter's amplitude and frequency response can vary slightly when not terminated with the design impedance.

If you terminate a 50 ohm generator with a 75 ohm pad (I would use 6dB or more) for a 75 ohm device under test and calibrate with that, it may give a reasonable result.
prcguy
Ok;
1. Is a 50 to 75 ohm ( 1.5:1) mismatch really that much a problem? From a xmit standpoint I know it isn't.
2. Would just a "pad", say 3db, make a difference without any matching?
 

videobruce

Member
Joined
Jan 5, 2004
Messages
90
Location
Buffalo NY
I guess the problem is one's defination of "slightly".

By measuring a 75 ohm filter, doesn't that "terminate" the scope by having it attached? When you say terminate, is that both ends if I'm using the tracking generator or just the input of the scope?
 

prcguy

Member
Joined
Jun 30, 2006
Messages
10,531
Location
So Cal - Richardson, TX - Tewksbury, MA
You should terminate both sides of a filter with the specified impedance for it to meet its design specs.
prcguy
I guess the problem is one's defination of "slightly".

By measuring a 75 ohm filter, doesn't that "terminate" the scope by having it attached? When you say terminate, is that both ends if I'm using the tracking generator or just the input of the scope?
 

Cowthief

Member
Joined
Oct 10, 2003
Messages
102
Location
Texas
25 and 75 ohm resistors?

Hello.

How about a 75 ohm resistor across the 75 ohm side and a 25 ohm resistor in series?
This would give the 75 ohm device correct termination and I would guess the receiver is terminated to 50 ohms.
This was very much the norm with direct video distro' at the TV stations.
The older monitors were all 50 ohms and the newer ones 75.
The other solution was to remove the terminating resistor in the set and terminate external to the set.
The advantage of a high impedance termination device and external resistor is 2 fold.
First off, there is very little loading, one could run 20 monitors and one termination or 20 terminated monitors and splitters and amplifiers.
Second, since this is external, the termination is just a matter of the what value you use.
Remember, on a modern receiver, the actual circuit impedance is a high value, usually around 400 to 500 ohms, it is the maching network that brings it down.
In the case of tube or FET circuits, this can sometimes be several thousand ohms.
In the case of 300 ohm balanced to 75 ohm unbalanced there were 2 things going on.
Andeven than, there were 2 ways to do this,matching transformer or resistors.
I put this out as this may give you some good ideas in your application.
 
Status
Not open for further replies.
Top