RF coax cable attentuation - cable selection

[JohnnyMalaria]

I have an application that uses an RF amp to drive an acousto-optic modulator at ~1.6W/50ohm/40MHz. After gutting my instrument and rebuilding it, I set about adjusting the amp output and discovered something odd (to me, at least). With some cables I saw ~3dB loss in power whereas other cables behaved well. I set up a test configuration where I fed the amp output into my 'scope together with a 50ohm terminator. In both cases, the terminator warmed up but much more so with the attenuating cable. In fact, it was almost too hot to touch. I've repeated this with other identical cables and see the same thing.

There are three differences between the cables:

1. RG58/U vs. RG58C/U
2. Brand - unknown (RG58/U) vs Cal Test(RG58C/U from Mouser)
3. Length - 2ft (unknown) vs 6ft (Cal Test)

It's the Cal Test from Mouser that show the attenuation.

Also, when I feed the function generator source (amp input) directly to the 'scope, I do not see this difference)

What's the likely culprit/mismatch? I'm concerned that I am driving the amp and modulator harder than necessary to overcome the attentuation. I can use the Cal Test cables for low frequency stuff but I need long cables for the RF connection.

Thanks.

[w2aew]

Does the RF amp have a DC bias on the output perhaps?

[JohnnyMalaria]

I've just checked. None that I can detect. I have the output of the amp going directly to my 'scope with a 50 ohm terminator at the 'scope end. With or without the amp on, it reads ~300 - 400mV. I get the same reading without anything connected to the 'scope as well. So I'm confident that if there is any bias then it would be no more than a few 10mV.

I actually have two such amps which are manufactured by the same company as the modulators and are intended for use with them. I see the same behavior whether I use the modulators as loads or 50 ohm terminators.

[G0HZU]

I'm a bit confused by your first post because you refer to the long cable as the attenuating cable but you say it delivers the most heat to the termination?

However, one thing to bear in mind is that your scope + terminator may have >22pF combined capacitance and a 6ft length of 50R transmission line will transform this complex load about 3/4 of the way around a smith chart at 40MHz. So the difference between the short cable and the long cable could be like comparing a 40R load against a 65R load when comparing a 2ft cable against a 6ft cable at 40MHz. i.e. the load presented to the amp will be on a different part of the smith chart in each case. I wouldn't expect this to be enough to cause such a marked difference in level though. Are you driving the amp into saturation? How much combined capacitance is there in your terminator and your scope?

Also, are you putting the terminator at the scope end of the coax cable (recommended) or at the amp end of the cable? (obviously not recommended as the cable and scope will introduce a complex load that varies a lot with frequency). This would cause all sorts of confusing issues with respect to what you see on the scope and how hot the terminator gets in each case. I assume you aren't doing this but I thought it was worth a mention anyway.

[JohnnyMalaria]

I'm a bit confused by your first post because you refer to the long cable as the attenuating cable but you say it delivers the most heat to the termination?

Yes, that's right.

However, one thing to bear in mind is that your scope + terminator may have >22pF combined capacitance and a 6ft length of 50R transmission line will transform this complex load about 3/4 of the way around a smith chart at 40MHz. So the difference between the short cable and the long cable could be like comparing a 40R load against a 65R load when comparing a 2ft cable against a 6ft cable at 40MHz. i.e. the load presented to the amp will be on a different part of the smith chart in each case. I wouldn't expect this to be enough to cause such a marked difference in level though.

I'll have to Google what you have said  That's the kind of information I was hoping for so that I can dig in further.

Are you driving the amp into saturation?

No, I can increase the output amplitude at least another 50% before I see any evidence of saturation.

How much combined capacitance is there in your terminator and your scope?

I have no idea. I see the same behavior with two different scopes. A Rigol DSO and a '74 vintage Tektronix 475.

Also, are you putting the terminator at the scope end of the coax cable (recommended) or at the amp end of the cable?

Initially at the amp end but I thought better of it and put it and the scope end.

(obviously not recommended as the cable and scope will introduce a complex load that varies a lot with frequency). This would cause all sorts of confusing issues with respect to what you see on the scope and how hot the terminator gets in each case. I assume you aren't doing this but I thought it was worth a mention anyway.

Regarding the terminators, I also used the actual modulators in place of terminators and saw the same behavior.

Right now, I am swapping the "problem" cables with the LF ones I use in other parts of the instrument to see if I can find an acceptable combination for the RF side of things. There are a lot of cables so it's quite confusing and time consuming.

[G0HZU]

One thing I forgot to mention was that the velocity factor of RG58 cable is about 0.66 and this is why a 6ft length of it can transform so far around a smith chart at 40MHz.
You could try experimenting with longer cables to see if it affects the behaviour. This may cause even more confusion at first but a longer cable could be arranged to be 1/2 wavelength longer at 40MHz than the 2ft cable. So it would present the same load as the 2ft cable at 40MHz. Maybe try an RG58C/U cable 10ft long rather than 6ft?

i.e. try 2ft + 2ft + 6ft to make up a 10ft cable? It might achieve nothing or make it even worse but it might also help diagnose what is going on.

[JohnnyMalaria]

Thanks - I'll give that a try.

[G0HZU]

Quote from: G0HZU on Today at 09:38:19 am

I'm a bit confused by your first post because you refer to the long cable as the attenuating cable but you say it delivers the most heat to the termination?

JohnnyMalaria: Yes, that's right.

Whoops, I missed this confirmation. I can make sense of this if the termination was (wrongly) fitted directly at the amp end rather than the scope end of the 6ft length of coax. If the termination was at the amp then the scope at the (unterminated) far end of the  6ft cable would show a low RF voltage level despite having lots of power in the load. So it would appear that the cable was lossy even though it isn't.

With the termination at the amp and a 2ft cable into the unterminated scope, this would actually cause a bad mismatch to be presented to the amplifier at 40MHz. It could easily look like 50R in parallel with 140pF at 40MHz. VSWR = 4.6:1? It would be much worse in this respect than the 6ft cable which would probably present a load to the amp with VSWR of about 1.8:1 at 40MHz. So I think this would stack up with all your symptoms if you had the termination directly at the amp output rather than at the scope end of the coax? I'd expect to see a hotter termination with the 6ft cable (partly because of the lower load VSWR presented to the amp) but the scope will also show a lowish voltage level making the 6ft cable appear to have high attenuation. So this would appear confusing.

Note that it's late and I'm off to bed so my numbers above need to be rechecked when I'm less tired