开云体育

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of si_emi_01
Sent: Saturday, April 30, 2022 10:21 PM
To: [email protected]
Subject: Re: [TekScopes] 2465 Calibration

Some responses.

Do you have a lot of Overshoot or Undershoot on the waveforms you are supplying it? On the other scope, what is the Rise and Fall Time? If it is too fast, since the Impedance is High, it could cause reflections. It says the Rise/Fall time is less than 10ns.

You are likely been bit by a Signal Integrity issue. The Signal Generator edge is too fast. With an edge as fast as that is, it needs to be terminated into the cable Impedance. This isn't possible in your case because the Oscilloscope Calibration routine expects certain values to be met.

As a test you could use a 50 Ohm in-line Terminator, connect that to the Oscilloscope Input. Set the Generator for twice the Voltage level (only below 2V to protect your load and Generator), start the test and see if the problem goes away (as a test only). If it works it is likely caused by Undershoot or Undershoot. If not, the problem might be somewhere else in the Signal Chain.

You might try a real short cable too reducing the cable Propagation Time and hence the ringing that could occur.

If the Rise/Fall Time is programmable make it slower per the information I provided. It may use a current source for the output, which makes the following option possible. It would give you symmetrical edge rates. If they are not, the Rise Time or Fall Time may not be the same. I'm not sure how close you need to be to my values.

Another option is that the 1kHz Frequency is pretty slow - lots of settling time possible. You could slow the edge down without affecting the Square Wave output level. I have provided information below that will help you get near the Rise/Fall Times my PG506 produces.

If you are able to slow it down, it might fix it.



I remeasured my two PG506s. The numbers are almost identical, so I will give the numbers just one:

Amplitude Output (STD and High) Connector into my Tektronix 5104 Oscilloscope
-------------------------------------------------------------------------------------------------------------
Termination - 1 MegOhm Internally Terminated input Frequency - 1.024kHz Amplitude - Follows the Amplitude Knob setting Rise Time - 12.86us (microseconds) Fall Time - 12.56 us (microseconds)


Trigger Output Connector into my Tektronix 5104 Oscilloscope
------------------------------------------------------------------------------------
Termination - 50 Ohm Internally Terminated input Frequency - Follows the Period knob setting Amplitude - Didn't record it - Crap!
Rise Time - 8.66ns (nanoseconds)
Fall Time - 4.86ns (nanoseconds)


Fast Rise (+) Going Output Connector into 50 Ohm Internally Terminated input of Tektronix 5104 Oscilloscope
-----------------------------------------------------------------------------------------------------------------------------------
Termination - 50 Ohm Internally Terminated input Frequency - Follows the Period knob setting Amplitude - 300mV Rise Time - 504.0ps (picoseconds) Fall Time - 664.0ps (picoseconds)


Fast Rise (-) Going Output Connector into 50 Ohm Internally Terminated input of Tektronix 5104 Oscilloscope
-----------------------------------------------------------------------------------------------------------------------------------
Termination - 50 Ohm Internally Terminated input Frequency - Follows the Period knob setting Amplitude - 300mV Rise Time - 504.0ps (picoseconds) Fall Time - 504.0ps (picoseconds)


Ross

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of jimbert4 via groups.io
Sent: Saturday, April 30, 2022 4:44 PM
To: [email protected]
Subject: Re: [TekScopes] 2465 Calibration

That makes sense. Oh, and I discovered that in Cal 02, vertical calibration, the 2465 is NOT put into 50 Ohm DC input mode. It is in high impedance mode. So my voltages are not being divided by 2. So now I don't know what the heck is causing the Limit errors. Sigh. Could it be that my square waves low side is not exactly at 0V? It's hard to get it exactly right on my PSG9080 signal generator. I'm using another scope to set the high and low levels as close as I can, but they interact and I can rarely get them both exactly right. Close though, within 10s of mV.

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of jimbert4 via groups.io
Sent: Saturday, April 30, 2022 4:44 PM
To: [email protected]
Subject: Re: [TekScopes] 2465 Calibration

That makes sense. Oh, and I discovered that in Cal 02, vertical calibration, the 2465 is NOT put into 50 Ohm DC input mode. It is in high impedance mode. So my voltages are not being divided by 2. So now I don't know what the heck is causing the Limit errors. Sigh. Could it be that my square waves low side is not exactly at 0V? It's hard to get it exactly right on my PSG9080 signal generator. I'm using another scope to set the high and low levels as close as I can, but they interact and I can rarely get them both exactly right. Close though, within 10s of mV.

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of toby@...
Sent: Saturday, April 30, 2022 12:57 PM
To: [email protected]
Subject: OT+DC was Re: [TekScopes] 2465 Calibration

On 2022-04-30 2:25 p.m., Jeff Dutky wrote:

On Sat, Apr 30, 2022 at 09:27 AM, <toby@...> wrote:

On 2022-04-30 9:18 a.m., Raymond Domp Frank wrote:

On Sat, Apr 30, 2022 at 12:49 AM, si_emi_01 wrote:

No one tells you that when they say "signal" they mean the Square Wave not the DC Level Output.

In a cocky mood I'd say a DC level is not a signal; it contains no information. Switching a DC level on or off however is.

In a quibbling mood I might point out that DC is frequently defined
as "current flowing in one direction" without implying a fixed
potential :)

but if there is any variation in potential then you can select an
arbitrary ground potential that will result in some portion of the
signal have a reversed flow of current, so the only unquestionably
direct current signal is also a signal with unvarying potential. (I
can quibble with the best of them!)

I didn't make it up. Take it up with the textbooks, or, indeed, wikipedia :)

(Quibble^2: There is no such thing as an unvarying potential in the real
world.)

(I admit for many years I too assumed a more or less fixed potential, probably because that's the most common subset of DC that we encounter.
It was only more recently that I learned the more general definition.)

--T

-- Jeff Dutky

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of si_emi_01
Sent: Saturday, April 30, 2022 1:16 PM
To: [email protected]
Subject: Re: [TekScopes] 2465 Calibration

Correct.

If the Source Impedance (PG506), Destination Impedance (2465), and the Transmission Path (Cable and Connectors), are matched ie. 50 Ohms for all of them, you will have the maximum power delivered from the source to the load.

The Voltage drop at the node may be accounted for in the internal Calibration Routine it goes through to calibrate the Vertical Amplifier system from input connector to CRT Plates.


All of this Impedance match stuff is frequency limited. The weakest link in this use case is the BNC Connectors. The Cable and BNC Connectors introduce loss dominated by the BNC Connectors. At the 1kHz frequency (it has around 10ns Rise Time on my PG506, Fall time is much greater), it doesn't matter too much.

The RG-58 or RG-223 cable has minimum loss up to a 2GHz (my value), it is specified in dB loss per unit of length (ie. 10dB for 1000ft).

The BNC Connectors are lossy because the dielectric isn't homogenous (there are gaps where it changes during the connection length). The Dielectric Constant of the plastic/Delrin/whatever material is around 2.5, the Dielectric Constant of Air is 1.0. The Cable Shield is stripped off in most BNC Connectors before the transition areas from Cable to Connector. Any change in distance or dielectric media between the center conductor and metallic outside conductor (Shield or Connector metal), changes that distance and hence the Impedance during that transition.

Before the shield was stripped off, it was tightly controlled (optical coverage considered), around the cable dielectric material maintaining a constant Impedance. Impedance change incurs loss due to transfer Impedance mismatch, reflections, VSWR, etc. The connector mate also causes this due to losses of the Pin/Socket and Shell/Shell mates.

As I said before, this is all frequency dependent. You wouldn't use a BNC Connector system to transfer 10GHz without incurring expected losses. You can do it and accept the losses if you want to. You certainly wouldn't get a good launch into the connector, transition across the connector, exiting the connector, etc. You certainly wouldn't want to run any higher RF power through it either.

No, you would choose an SMA, K or 3.5mm, 2.92mm, 2.4mm, 1.85mm connector system depending on frequency. You might even change the cable material to one with lower losses like Hard-Line if needed. The design is more coaxial, has a better (more controlled) Pin/Socket arrangement, better Shield attachment (more controlled again), connector shell mating and connector torque (control again), arrangement. They introduce a better match, more homogenous dielectric for a longer length of the transition, lower transfer Impedance, less loss, lower VSWR, etc.

This applies to Digital/Analog/RF applications. For Digital, you base it on the edge rate (Rise/Fall), of the signal (for example a 1PPS signal has a 1 second period, variable pulse width - like 2us, but a <1ns screaming edge - 200ps in some cases). The <1ns is to reduce uncertainty in Phase Detectors of PLLs. There is a TIA specification for 1PPS.


If this is TMI, let me know.

Ross


-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of jimbert4 via groups.io
Sent: Saturday, April 30, 2022 7:29 AM
To: [email protected]
Subject: Re: [TekScopes] 2465 Calibration

Okay, I looked up some stuff in my long forgotten transmission lines book from college. To match impedances, the source resistance must also be 50 Ohms, in series, to match the cable and the 2465 input, hence the divide-by-2. How does the Tektronix test equipment like the PG506 handle this? Would you see double its voltage setting if you terminated its output at the 1 M Ohm oscilloscope input?

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of jimbert4 via groups.io
Sent: Saturday, April 30, 2022 7:29 AM
To: [email protected]
Subject: Re: [TekScopes] 2465 Calibration

Okay, I looked up some stuff in my long forgotten transmission lines book from college. To match impedances, the source resistance must also be 50 Ohms, in series, to match the cable and the 2465 input, hence the divide-by-2. How does the Tektronix test equipment like the PG506 handle this? Would you see double its voltage setting if you terminated its output at the 1 M Ohm oscilloscope input?