Impedance matching 75 ohm sweep generator to 50 ohms

[MrVa1ve]

Hi,

This question is related to my recent Telonic sweep generator purchase, but probably deserves its own thread since it's not related to documentation.

So, having had a confirmation that the generators output impedance is 75 ohms, there may be an issue with my test setup. All of my equipment (relevant to RF) is 50 ohms.

My intention is to use the generator to test the frequency response of mainly passive devices, filters, etc. up to the top of the FM frequency band. I'm not intending to do much alignment work, I'm more of a "student" of RF principles and relatively new to all this. But, I'm guessing impedance matching and having the correct termination will be a concern.

With the research I've done so far, I've turned up the following solutions:

1) Put a 25 ohm resistor in series with the RF output from the generator
Obviously, this isn't a bi-directional match, but it would seem that since my generator is the only power *source* in my test setup, it's good enough that the generator sees 75 ohms. I should point out that this is an opinion (somewhat disputed) I read online (I'll try dig up the link later).

2) Use a min loss pad
Either built or bought and accepting that there will be about 6db worth of loss. I'm not sure what the impact of this loss will be. Maybe someone can fill me in?

3) Use a transformer or some reactive arrangement
As I understand it, the downside of these approaches is their frequency dependence. The upside is that there's very little if any power loss. Another disadvantage with the transformer is that it won't pass DC, but I'm not sure that's a problem in my case.

So, what's the common approach to this problem? Has anyone faced the same issue and what did you do to solve it (and how's it working)?

Regards. D.

[Wendymott]

Hi Mr "D". I have a matching "pad" 75/50 ohms, it was part of the accessories for my Marconi 2022 Signal generator..... I seem to think it had a 50R in series from input to output. Other members may have other ideas..

[MrBungle]

Need an MLP

43.3 ohms in series with the 75R port, then 86.6 ohms across the 50R port. I made one which was "close enough" with a 47 in series and a 100 ohms paralleled with 1K across the 50R port.

5.7dB power loss either direction. 75 to 50 was around 7.5dB voltage loss. 50 to 75 was around 4dB voltage loss.

I blew it up shortly after I built it

This was built to match a 75 ohm generator to a 50 ohm system.

[chriswood1900]

You might find this useful, this is the matching pads Sound Technology suggest for their sweep gen for 300 and 75 ohm to 50 ohm and designed to give a loss of 6db.
Chris

[Chaparal]

The easiest/cheapest way is the 2 resistor solution.

But take a look at the BNC Connectors !

The input BNC are 50 ohms version
The output looks like a 75 R BNC version.
They are different !
See pic.

Peter

[MrVa1ve]

Thanks for the responses so far.

@MrBungle I laughed when I read this comment "I blew it up shortly after I built it", that's a common experience on my bench! And the My Little Pony suggestion is inspired!
She may not be happy when her lovely magical locks catch fire as a result of one of my "accidents".

Thanks Chris, that's another +1 for the two resistor solution.

That's a great catch chaparal, I actually performed a quick test with standard 50 ohm cable and I never noticed the slight difference in the jacks. However, I knew after reading karez's document (see documentation thread) that at least some of the inputs were 50 ohm. Is there a reason why they would have the RF output as 75 ohm and have say the marker input at 50 ohms? Seems a little odd.

[Chaparal]

The manual says:
Ext. Marker in BNC 50 Ohms for ext. RF Source
0.5V Marker for -6dBm input

Maybe Rf-Sources have in general 50 ohms.

--
Peter

[karesz*]

As I wrote, its a "TV-Sweeper"; such systems have 75 ohms -otherwise the some 50 Ohm inputs are a little bit special, "house own philosophy",_I think_ from vendor, but the more generios instruments are with 50 Ohm system-connections, also we can say-its OK...
BTW: what about a changing all Input terminations (eventually outs too_I know; its not a simple way because needs such recalibrations) all for 50 Ohm?

[GMB]

I see a lot on confusion on this subject.
The answer depends on what problem you are trying to solve and what you are going to be doing with it.

So it is likely that if the generator claims a calibrated output into 75R then you need it to be seeing 75R to rely on that. Your 25R series solution into 50R is fine if it actually is 50R, and the other suggestions are fine to make a full blown both-ways match to 50R if that is really needed.

I actually suspect that neither solution is appropriate if you are going to feed the signal into a funny network to be measured which is likely not 50R at all. The usual answer is an attenuator of sufficient reduction and matching so that the generator doesn't really notice the odd impedance hanging off the end, but made so the drive appears to come from 50R for the benefit of the measurement - if that is what wanted to be useful.

The other thing that is often forgotten is that correctly terminating a transmission line is only necessary if a reflection is unwanted and if power is going towards the terminator. A cleanly terminated line can be driven in any way you like since there will not be any power going back towards the driver.

[G0HZU_JMR]

You can also allow some extra loss in the matching pad and have a -10dB match loss if this helps with the generator settings for amplitude. For example, see the circuit below: This uses E12 resistors to produce a 75R to 50R match pad with a fairly accurate 10dB loss. It should be within about 0.1dB across the HF>VHF band if you use 1% resistors and the VSWR should be very (very) low as long as SMD resistors are used and the layout is tight. But you will probably find that your generator output impedance won't be a very accurate 75R so there's probably no point trying too hard to get the best from this circuit. But if you want accuracy in this pad when verified on test gear, the closer you get to the resistance values in the circuit the better this pad will perform on test gear.

So, what's the common approach to this problem? Has anyone faced the same issue and what did you do to solve it (and how's it working)?

In my case the situation is usually the other way around. I have 50R test gear here and often have to measure filters that are not 50R filters. In this case, I often use a 50R 2 port VNA to measure the (75R?) filter and then export a 50R s2p model from the 50R VNA. It's then possible to import this model data into a computer (running an RF simulator program like RFSIM99) and predict the RF performance as if the filter was in a 75R test setup. It's a bit clunky but it usually works very well for me.

10-15 years ago this could usually only be done with expensive RF test kit in a professional RF lab. But these days there are so many low cost VNA options for the hobbyist that it is possible to do this on a fairly low budget. This technique works well even for filters that are quite a way from 50R. For example, it can be used with crystal filters that have I/O impedance up in the region of 1000R. But this does need a decent/modern VNA and a decent set of calibration standards if good results are required.

[frankmcvey]

A couple of months ago, I was distributing a lot of ex-Admiralty electronic spares for charity. Among the items was a couple of 50-75 Ohm coax converters.

See this thread post #17.

I think I may have a couple left if you think they would suit your purpose, for a charity donation and the cost of postage.

Cheers,

Frank

[MrVa1ve]

Thanks again guys for this great advice.

Jeremy, I really like the idea of designing the pad for a -10db loss which you could offset (inaccuracy notwithstanding) with the sweeper. That's definitely a runner.

Frank, I've sent you a PM.

Kind Regards. D.

[emeritus]

My Suhner connector catalogue says that 75Ω and 50Ω BNC connectors can be mechanically mated in either direction without damage. Conversely, 50Ω and 75Ω Type N connectors should not be mated as damage is likely to occur.

[Chaparal]

The BNC connectors (50/75R) are intermateable
but not the coax cable.

For 50R use RG 58, 59, 62, 178, 196, 210, 310 etc.
For 75R use RG 318, 174, 179 etc.

depending on BNC connector version

---
Peter