Celwave circulator/isolator question

[GSXR1181]

I have a Celwave circulator/isolator Model # CD460-B. I've been all over the internet trying to find any info on this. And I can't find anything. Not even on RFS's website. Here is my issue.

I'm trying to adjust the circulator with the adjustment screws. I know to adjust the first 4 screws, you adjust for max power forward. Problem is, when i get the screw to it's peak, and I over adjust it, I can never get it to come back to where it peaks at. It's almost like its pushing something down, and when it's down, it will not come back up.

Same thing happens when I adjust the others for a null. When the signal goes down to it's lowest point, then starts to increase, I can't get it back to the lowest point. And the more I mess with it, seems the worst it gets.

Does anyone have any idea what I might be doing wrong? Or is this circulator just messed up?

 

[zz0468]

What are you using for a signal source? You'll have better results if you put attenuator pads between the source and the circulator, and between the circulator and the load. What's likely happening is that as you tune the circulator, it's affecting the source. Also, is it a circulator, or an isolator? What's terminating the third port? If you're using a transmitter, there could be some influence between the circulator and the transmitter.

The proper way to tune these is with a network analyzer, where you can observe both forward and reverse transmission paths. You don't necessarily tune them for lowest loss, you tune them for best isolation in the reverse direction, as well as lowest loss forward. Since everything interacts with everything else, it can get kinda hairy.

 

[prcguy]

Don't forget all ports must be terminated during adjustment, you need to move the load between ports. Some circulators/isolators have only 3 adjustments and you adjust for a null at port 1 when feeding port 2, null port 2 when feeding port 3 and null port 3 when feeding port 1. If there are more adjustments you map out which ones reduce insertion loss in the forward path (port 1 feeding port 2, etc). You don't always adjust the null for your transmit frequency, it may be best to find what frequency within the isolator's band is coming down the coax the strongest and mixing your transmitter and notch that.
prcguy

What are you using for a signal source? You'll have better results if you put attenuator pads between the source and the circulator, and between the circulator and the load. What's likely happening is that as you tune the circulator, it's affecting the source. Also, is it a circulator, or an isolator? What's terminating the third port? If you're using a transmitter, there could be some influence between the circulator and the transmitter.

The proper way to tune these is with a network analyzer, where you can observe both forward and reverse transmission paths. You don't necessarily tune them for lowest loss, you tune them for best isolation in the reverse direction, as well as lowest loss forward. Since everything interacts with everything else, it can get kinda hairy.

 

[GSXR1181]

What I'm doing is using a Motorola R-2670, and generating a frequency that I'm transmitting, and inputting that into the input port. I use about a -10dB that attenuated at 40dB.

On the output end, I'm inputting that into my 2670, and looking at it on the spectrum analyzer. The other two ports I have terminated with 50 ohm dummy loads.

When I have it connected this way, I was tuning it for a peak output. But like stated before, it seems what ever the tuning adjustments are doing, they seem not to want to go back when I try to back off the adjustment. And the more I screw them in, the further off they go, and I seem that I can't recover them. It's almost like its pushing something down, and once it's down, they don't want to return. There are 4 adjustments for the first two ports, and this happens on all of them.

When I'm tuning for a null I use the adjustments for the next two ports. I put in my freq on the output port, and the next port I use a dummy load, followed by the input to my 2670. Then another dummy load on the next port. Then I adjust for the null using the adjustment for the next port. But the same thing happens again. When I hit the lowest point, and over shoot it, it won't return.

I know I probably sound confusing, but I'm not experienced in using these things. I've tuned them like this before with good results. I was actually tought by a guy who works for Motorola, and he showed me this method. But the circulator's we used would actually adjust the way they should.

Oh and I guess you can say I'm using it as an isolator. This will have two dummy loads to prevent anything trying to get back into my power amp. The output will go to a tuning cavity.

Is there a better way of doing this with a 2670? By the way I also have a TIMS and a wattmeter at my disposal if needed. And here is a pic of the circulator I'm using. Matter of fact, this is the complete setup I plan on using.