开云体育

I think I figured out why the calibrator rise time measurement doesn't read to spec. I warmed up the 11810 for about 40 minutes to do the Enhanced Accuracy adjustments. That seemed to go fine except afterwards when I measured the rise time I was getting >400 ps. The 11801 will remember the incorrect settings through a power cycle.

If you do the Loop Gain adjustment in the Enhanced Accuracy menu it resets something which results in incorrect values. If you go to the Utility menu and press Initialize you should now be able to set the Trigger to Internal, press Autoset, adjust the time base, turn on Hardware measurement and get <35 ps rise time.

Unfortunately, it does not appear to be possible to measure the rise time on more than one one channel.

I think a bunch of the Chinese DSO designers must have spent a lot of time using an 11801 and thought that was what a scope UI should look like. Even after reading all the way through the User manual the UI is still confusing because of the strange locations of various settings.

The tip off came from this line on p 82 of the User manual:

"Whenever you begin a new task using the 11801. you should initialize the system so that all the settings are at "factory default" . That way you do not get unexpected results because of settings remaining from the last use of the 11801."

BTW I have a 2 port divider feeding the upper channel of my SD-22s and the waveforms look to be exactly the same. The pieces of hardline from the divider to the heads are different lengths and I have not been able to shift one relative to the other to see how closely they overlay as there is a 340 ps delay between the two. The divider increases the rise time to 42 ps. It's an MBC Technology unit. No frequency rating specified.

I just discovered that it reset the time base readings when I removed the splitter and turned off one channel to take a look at how much of the 7% overshoot was the splitter which it turns out is about 5% of the total.

There's a little bit of ringing on the calibrator step at 12.2 GHz which is probably the result of reflections where the semi-rigid SMA cable connects to the 3.5 mm calibrator and sampling head connectors.

I am quite agog at how precise this thing is. Getting the full performance out of it will take some serious skill.

Reg,

If you display the two channels simultaneously then they share the same time base speed and position. I don't see how you tried to align the rising edges. On my CSA's I do this: display both channels CH1 and CH2, then Store CH1 en turn CH1 off, then Recall the stored trace of CH1. Now the displayed stored trace is "frozen" and remains in position regardless how you change the time base of the CH2 trace. (Note you can select one of the displayed traces by touching the trace somewhere.)

Albert

That was my conclusion. But the UI is so bad that I thought there might be a variable delay line in there somewhere that would adjust channel skew. I've also discovered it is buggy as all hell. You pretty much have to reinitialize the scope for a bunch of operations. I tried switching from vector to dot mode. Menu said vector was off, but the display said otherwise. Once I initialized the scope I was able to set it to dot mode.

I've had the 11801 running for 4-5 hours with the calibrator feeding an SD-22 with cursors set on the peaks of the reflection in the cable. I've noticed that the internal timebase varies by a little over 1 ps over the course of 5-10 minutes. I have a postit on the screen and the cursors are not moving. The waveform is moving. I've got the internal clock output connected to my 5386A with the OCXO option and that confirms the internal time base period is changing.

From looking at the block diagram in the service manual, it appears to me that the OCXO is not as stable as the one in my 5386A which I have compared to a GPSDO over long periods.

This has me considering whether installing one of Leo Bodnar's single output GPSDOs in place of the 200 MHz OXCO would be worthwhile. In particular in improving the jitter spec from the factory 1.1 ps.

Watching the clock period vary by 100 femtoseconds over the course of 15-30 seconds with a 1 s gate time is rather interesting.


Reg

--------------------------------------------