开云体育

Hi,

I have an HP 11729B that looks an anwful lot more like an 11729C, complete with the 640 MHz input (BNC) and output (SMA) ... as you may know you can apparently link these two together to use an "interanal 640 MHz" to drive the comb filter generator.

It's not actually an oscillator output to feed the amp input, but the aux output of the 640 MHz power amp, at about +12 dBm .. fed through a suitable attenuator back to the input it self-oscillates (or is supposed to) at around the peak of the input tuned filter.?? I've tried with a 10dB pad and a 5dB pad ... both of which should be enough to cause it to oscillate,? its would get around 2.5 dBm at the input, which is sufficent drive, assuming the phsae shift is there or therabouts. any other clues as to what the approx value of the HP magic cable attenuator is?? The part number is? HP 11729-60096 or 11729-60098? but I can find no further info on it.? Maybe it is just a series capacitor or something?

"Ozan" <ozan_g@...> writes:

Hi Robin,
I built such a cable. As you said output is 11dBm-13dBm, input is 1dBm-4dBm level. Cable needs to provide ~ 9dB attenuation so that the output after going through attenuator provides >=1dBm and <=4dBm.

Length (phase) of the cable is critical. There is a very sharp 640MHz filter and phase should be exactly right otherwise it won't oscillate. To get an idea I fed in 640MHz with a splitter and measured the phase of input signal vs output signal at the oscillation frequency. It was ~ 1.6ns (wasn't a super accurate measurement).

I built a 9dB attenuator, and cut a coax to approximate length, then with trimming the coax I tuned it for 640MHz. I also built a tap off from the attenuator so that I can measure the frequency. I can send pictures if it is useful.

Hi Ozan,

I'd be interested in pictures if it's easy for you to provide them.

How were you able to trim the cable? Presumably, to test it you need
connectors on both ends but that would make trimming it difficult.

Thanks,
Matt

Hi Matt,
I don't know if I have a picture of what is inside the copper box but I built an attenuator with SMD resistors. In one leg of the pi-network I also added a 50-ohm matched tap to take out the oscillation frequency to a counter. It is the port with the 50-ohm termination in the picture.?

I soldered the coax to the board, I cut and soldered to try. You can see I estimated the length wrong and cut the coax too short to start with, adding another small section. Construction was on a scrap board without much consideration of impedance match to test the idea so the delays were not well controlled.?

I also looked at old pictures and 11729C delay was ~ 1.9ns, not 1.6ns as I wrote in the earlier message.?

I will put more pictures to the albums area as I find them.
Ozan

Ah, thanks for that!? I shall try messing with the cable length then until it takes off ...? 1.9nS is arounf 400mm of cable at vf=0.7 ... the one shown in the manuel is much shorter, so maybe it is a bit variable, but I shall try with 1.9nS andsee how it goes, thanks again!

Hi Robin,
Phase is a very strong function of the frequency (for a good reason). I re-measured the phase this time checking the frequency of the signal generator with a counter, it was few kHz off. With exactly 640MHz now I measure 0.95ns. This is with a 2467B, a 400MHz scope measuring 640MHz so accuracy is not great. Total delay has to be multiples of 1/640MHz, cable needs to add to 0.95ns.

Shortest cable delay is:1/640MHz-0.95ns=0.61ns, about 13cm with vf=0.7. HP cable still looks shorter than that. Perhaps delay through the attenuator or my measurement inaccuracy explains the rest. My cable was around 13cm though. If anyone in the group has the cable it would be good to know the exact length.?

I used series of BNC connectors, not locked, and experimented with various delays by sliding them. Picture of the connectors is attached.
Ozan

I measured the delay between ports on my VNA and there is 145 degrees at 640 MHz .. I built a cable for that and it runs at 640.0047,? just +4.7KHz, well within spec? ...? (which is either +- 10 KHz or +-30KHz, depending on which notes you read)?

Physical cable length is 150mm beween phase reference planes of the connectors, plus a standard small SMA attenuator.

AT 640.00470 MHz? into the comb generator, I measure at the mixer input:

.640 GHz? +15.5dBm
1,92G????? -1.0
4.48G???? +2.5
7/04G??? -0.25
9.6G????? -2.25
12.16G ?? -6.7
14.72G???? -9.2
17.28G??? -14.37

No idea how good/bad that is ... I guess I'll have to read the manuel. but at least it seems to run ...

On Wed, Nov 15, 2023 at 09:37 AM, Robin Szemeti wrote:

I measured the delay between ports on my VNA and there is 145 degrees at 640 MHz .. I built a cable for that and it runs at 640.0047,? just +4.7KHz, well within spec? ...? (which is either +- 10 KHz or +-30KHz, depending on which notes you read)?

Physical cable length is 150mm beween phase reference planes of the connectors, plus a standard small SMA attenuator.

AT 640.00470 MHz? into the comb generator, I measure at the mixer input:

.640 GHz? +15.5dBm
1,92G????? -1.0
4.48G???? +2.5
7/04G??? -0.25
9.6G????? -2.25
12.16G ?? -6.7
14.72G???? -9.2
17.28G??? -14.37

No idea how good/bad that is ... I guess I'll have to read the manuel. but at least it seems to run ...

--
Spec is -20dBm minimum at mixer input, looks like your setup worked well.?
Ozan