Wide-band doubler feedthrough suppression; 3 MHz - 3 GHz in

[ezalys]

Hey all,

I'm curious how the doubler inside the 85047A test set works, since it seems to go from 1.5 MHz up to 3 GHz. Since the doubler is engaged while measuring from 0-3, it must also have excellent feedthrough supression as well. Any idea how they may have accomplished this?

E

[hendorog]

I haven't tested one of the real HP doublers, and haven't seen inside one either so can't answer your question.

However I did manage to build one which the 8753 will use without showing errors. So it is possible to build one yourself instead of forking out $1K for an 85047A just to measure to 6GHz.

If the doubler doesn't suppress the harmonics or fundamental sufficiently then the 8753 will show lock errors. (I don't know the definition of 'sufficiently' unfortunately)

For mine I used a couple of minicircuits wideband baluns either side of an avago quad diode package - discontinued now.

[edavid]

I'm curious how the doubler inside the 85047A test set works, since it seems to go from 1.5 MHz up to 3 GHz. Since the doubler is engaged while measuring from 0-3, it must also have excellent feedthrough supression as well. Any idea how they may have accomplished this?

If you look at the specs closely, it appears that there is a typo, and the doubler is only used above 3GHz.

Other HP doublers, such as the 10515A and 11721A, use a simple center tapped transformer + full wave rectifier circuit.

[hendorog]

I'm curious how the doubler inside the 85047A test set works, since it seems to go from 1.5 MHz up to 3 GHz. Since the doubler is engaged while measuring from 0-3, it must also have excellent feedthrough supression as well. Any idea how they may have accomplished this?

If you look at the specs closely, it appears that there is a typo, and the doubler is only used above 3GHz.

Other HP doublers, such as the 10515A and 11721A, use a simple center tapped transformer + full wave rectifier circuit.

I don't think that is a typo. The start freq of high range is 3MHz as presumably that is the lower limit of the doubler. I guess limited by the frequency range of the balun/transformer. inside the doubler.

AFAIK the entire sweep changes to high range any time the sweep ends > 3GHz, it doesn't switch mid sweep between ranges. If it was switching in the middle of a sweep it would have to be a solid state switch and I am 99.9% sure that the transfer switch is just an ordinary mechanical coax switch. OTOH the output port switches are solid state unless you have a unit with option 009.

This is the doubler transfer switch part listed in the service manual:
https://www.ebay.com/itm/HP-Agilent-5180-8449-Coaxial-Switch-DC-to-6-GHz-/271245448404

[edavid]

AFAIK the entire sweep changes to high range any time the sweep ends > 3GHz, it doesn't switch mid sweep between ranges.

As an example, the insertion loss spec is:

4 dB +0.8 dB/GHz: 300 kHz to 3 GHz
17.5 dB +0.8 dB/GHz: 3 GHz to 6 GHz

If it didn't switch in the middle of the sweep, they would have to specify the higher insertion loss e.g. for a sweep from 2-4GHz, but they don't.

[hendorog]

AFAIK the entire sweep changes to high range any time the sweep ends > 3GHz, it doesn't switch mid sweep between ranges.

As an example, the insertion loss spec is:

4 dB +0.8 dB/GHz: 300 kHz to 3 GHz
17.5 dB +0.8 dB/GHz: 3 GHz to 6 GHz

If it didn't switch in the middle of the sweep, they would have to specify the higher insertion loss e.g. for a sweep from 2-4GHz, but they don't.

I'm seeing the insertion loss spec written like this  in the 85047A service manual on page 5 (this is for the solid state switch version):
RF In to Port 1,2
4 dB +0.8 dB/GHz:  3 GHz Range
17.5 dB +0.8 dB/GHz: 6 GHz Range

https://literature.cdn.keysight.com/litweb/pdf/85047-90001.pdf

[TheSteve]

In the service manual it shows the signal path for .300 to 3000 MHz and for 3 to 6000 MHz. It seems pretty clear that if you go above 3 GHz it uses the doubler path for everything and raises the lower minimum frequency to 3 MHz. This only applies to units that have a 3 GHz only source, if they have a 6 GHz source(newer units) then the doubler is never used and you can go .3 to 6000 MHz in a single sweep.

[edavid]

I'm seeing the insertion loss spec written like this  in the 85047A service manual on page 5 (this is for the solid state switch version):
RF In to Port 1,2
4 dB +0.8 dB/GHz:  3 GHz Range
17.5 dB +0.8 dB/GHz: 6 GHz Range

https://literature.cdn.keysight.com/litweb/pdf/85047-90001.pdf

Here's a datasheet that lists it the other way 

http://www.testequipmenthq.com/datasheets/Agilent-85047A-Datasheet.pdf

[hendorog]

I'm seeing the insertion loss spec written like this  in the 85047A service manual on page 5 (this is for the solid state switch version):
RF In to Port 1,2
4 dB +0.8 dB/GHz:  3 GHz Range
17.5 dB +0.8 dB/GHz: 6 GHz Range

https://literature.cdn.keysight.com/litweb/pdf/85047-90001.pdf

Here's a datasheet that lists it the other way 

http://www.testequipmenthq.com/datasheets/Agilent-85047A-Datasheet.pdf

Yep fair point. However I am sure that is the typo.