Evaluating old CRO probes

[David Hess]

The second probe shows a faster response because it is peaked.

Test the oscilloscope itself without the probes by attaching the pulse generator directly to a vertical input.  If you have to use a cable or adapters, then connect it through a 50 ohm feed-through adapter directly at the oscilloscope input.

If you are going to be messing around with probes and testing oscilloscopes, then a feed-through adapter is a good investment anyway.

https://www.amazon.com/Rigol-ADP0150BNC-Ohm-Impedance-Adapter/dp/B01C3EFUF0/
https://www.amazon.com/Zyyini-Adapter-Through-Terminator-Device/dp/B07SD3TC51/

[jdutky]

The second probe shows a faster response because it is peaked.

Test the oscilloscope itself without the probes by attaching the pulse generator directly to a vertical input.  If you have to use a cable or adapters, then connect it through a 50 ohm feed-through adapter directly at the oscilloscope input.

If you are going to be messing around with probes and testing oscilloscopes, then a feed-through adapter is a good investment anyway.

https://www.amazon.com/Rigol-ADP0150BNC-Ohm-Impedance-Adapter/dp/B01C3EFUF0/
https://www.amazon.com/Zyyini-Adapter-Through-Terminator-Device/dp/B07SD3TC51/

It took a while for the feed-through adapter to arrive, but now it has and I tested the scope with the pulse generator directly attached to the scope inputs. Both channels appear to show a rise time of something like 1.4-1.5 ns, which, if my math is correct, comes to something like 470 MHz. I also tested again with the various probes, with similar results (in the range of 1.5-1.6 ns).

Attached is an image from one of the channels with the pulse generator directly attached through the terminator-adapter, with the timebase set at 0.01 us/div (10 ns/div) and using the 10x mag switch to give 1 ns/div. The other channel, and the images using the probes, look substantially similar (there's slightly more ringing with the probes in place, and the cheap P2200s actually seem to perform slightly better than the original Tek P6075As, but the rise times are basically the same to within 0.2 ns. At least I'll have no further hesitation about ordering cheap probes off Amazon).

I've been using the scope for a micro-processor project for a couple weeks, and I'm getting more confident in its capabilities, as well as its shortcomings. The channel 2 selector switch needs a cleaning pretty badly, as do both vertical position POTs and the beam intensity POT, and there's some kind of crosstalk between the channels when one of them is set to the highest sensitivity. Also, I'm pretty sure that the timebase has drifted a bit since it was last calibrated (in Dec. 1979!), which only adds to my uncertainty about the actual rise time, but otherwise this old thing is a real trooper and a pleasure to use.

I've ordered a "for parts" 475 off eBay, and it will arrive in a couple weeks. My intention is to keep that around for spare parts, but I'm going to open it up, try to diagnose it, and just get a feel for what it will be like refurbishing my father's old scope. I have hesitated to open up his scope without a pressing reason because I don't want to make any stupid mistakes with it, but a stranger's non-working scope is completely fair game.

[srb1954]

The second probe shows a faster response because it is peaked.

Test the oscilloscope itself without the probes by attaching the pulse generator directly to a vertical input.  If you have to use a cable or adapters, then connect it through a 50 ohm feed-through adapter directly at the oscilloscope input.

If you are going to be messing around with probes and testing oscilloscopes, then a feed-through adapter is a good investment anyway.

https://www.amazon.com/Rigol-ADP0150BNC-Ohm-Impedance-Adapter/dp/B01C3EFUF0/
https://www.amazon.com/Zyyini-Adapter-Through-Terminator-Device/dp/B07SD3TC51/

It took a while for the feed-through adapter to arrive, but now it has and I tested the scope with the pulse generator directly attached to the scope inputs. Both channels appear to show a rise time of something like 1.4-1.5 ns, which, if my math is correct, comes to something like 470 MHz. I also tested again with the various probes, with similar results (in the range of 1.5-1.6 ns).

Attached is an image from one of the channels with the pulse generator directly attached through the terminator-adapter, with the timebase set at 0.01 us/div (10 ns/div) and using the 10x mag switch to give 1 ns/div. The other channel, and the images using the probes, look substantially similar (there's slightly more ringing with the probes in place, and the cheap P2200s actually seem to perform slightly better than the original Tek P6075As, but the rise times are basically the same to within 0.2 ns. At least I'll have no further hesitation about ordering cheap probes off Amazon).

I've been using the scope for a micro-processor project for a couple weeks, and I'm getting more confident in its capabilities, as well as its shortcomings. The channel 2 selector switch needs a cleaning pretty badly, as do both vertical position POTs and the beam intensity POT, and there's some kind of crosstalk between the channels when one of them is set to the highest sensitivity. Also, I'm pretty sure that the timebase has drifted a bit since it was last calibrated (in Dec. 1979!), which only adds to my uncertainty about the actual rise time, but otherwise this old thing is a real trooper and a pleasure to use.

I've ordered a "for parts" 475 off eBay, and it will arrive in a couple weeks. My intention is to keep that around for spare parts, but I'm going to open it up, try to diagnose it, and just get a feel for what it will be like refurbishing my father's old scope. I have hesitated to open up his scope without a pressing reason because I don't want to make any stupid mistakes with it, but a stranger's non-working scope is completely fair game.

Your maths doesn't seem right. If your are measuring risetimes in the range 1.4-1.5ns this corresponds to bandwidths in the range 233-250MHz, which is about what you would expect for a 475.

The formula for calculating bandwidth is BW = 0.35/Tr

This formula applies to waveforms that have a smooth rise and don't overshoot too much.

[jdutky]

Your maths doesn't seem right. If your are measuring risetimes in the range 1.4-1.5ns this corresponds to bandwidths in the range 233-250MHz, which is about what you would expect for a 475.

The formula for calculating bandwidth is BW = 0.35/Tr

This formula applies to waveforms that have a smooth rise and don't overshoot too much.

and that's what I get for doing cargo-cult math: I mangle the constant without noticing.

So I looked up a definition of the bandwidth formula (https://www.tek.com/support/faqs/how-bandwidth-related-rise-time-oscilloscopes) and I really don't understand what they're talking about (because I'm bailed to computer science before hitting the real engineering courses in college). I don't really know what "This corresponds to a 1- or 2-pole filter roll-off in the frequency domain" means.

Still, it's good to know that even with the correct math my scope seems to be performing as expected.

[David Hess]

The first picture below shows the transient response of a 465 (100 MHz) that I am working on; obviously it needs adjustment.

The second picture shows the transient response of the 2232 (100 MHz) that I use regularly which in within specification.

All aberrations are from the oscilloscopes; the PG506 used as a source and the RG-400 patch cable have been tested on a sampling oscilloscope and are completely clean with no aberration at all.

[srb1954]

and that's what I get for doing cargo-cult math: I mangle the constant without noticing.

So I looked up a definition of the bandwidth formula (https://www.tek.com/support/faqs/how-bandwidth-related-rise-time-oscilloscopes) and I really don't understand what they're talking about (because I'm bailed to computer science before hitting the real engineering courses in college). I don't really know what "This corresponds to a 1- or 2-pole filter roll-off in the frequency domain" means.

Still, it's good to know that even with the correct math my scope seems to be performing as expected.

The reference to 1- or 2- pole filter roll-off comes from the mathmatics for representing the frequency response for a low-pass filter and implies a certain circuit configuration to produce such a response. The number of poles generally corresponds to the number of reactive elements i.e. capacitors or inductors, in the circuit.

A 1-pole filter is the type of response produced by simple single-stage RC low-pass filter. This type This type of filter starts rolling off gradually above its cut-off frequency and eventually reaches a ultimate roll-off rate of -6dB/octave i.e. for every doubling of frequency the filter output reduces 6dB. The transient step response of a 1-pole filter is smooth and monotonic and there is no overshoot of the output waveform.
 
A 2-pole filter contains 2 reactive elements and can be implemented with two cascaded RC filters or by an LC filter or by an active filter with two capacitors. This frequency response of a 2-pole filter rolls off more rapidly at -12dB/octave and the transient step response will generally have little (<5%) or no overshoot.

[jdutky]

The first picture below shows the transient response of a 465 (100 MHz) that I am working on; obviously it needs adjustment.

The second picture shows the transient response of the 2232 (100 MHz) that I use regularly which in within specification.

All aberrations are from the oscilloscopes; the PG506 used as a source and the RG-400 patch cable have been tested on a sampling oscilloscope and are completely clean with no aberration at all.

Nice, I think my scope trace looked more like the 2232, before I engaged the 10x mag. I'll have to set up the test again and verify that memory. I will also have to work harder to get good camera images of the screen: your images are nice and sharp. Are you doing that by hand or do you have a rig to hold the camera?

(I've considered buying one of the Tek camera rigs on eBay and using it to mount a digital camera, but that seems like the wrong way to be spending money when I could save it for a nice, modern digital scope)

[jdutky]

The reference to 1- or 2- pole filter roll-off comes from the mathmatics for representing the frequency response for a low-pass filter and implies a certain circuit configuration to produce such a response. The number of poles generally corresponds to the number of reactive elements i.e. capacitors or inductors, in the circuit.

A 1-pole filter is the type of response produced by simple single-stage RC low-pass filter. This type This type of filter starts rolling off gradually above its cut-off frequency and eventually reaches a ultimate roll-off rate of -6dB/octave i.e. for every doubling of frequency the filter output reduces 6dB. The transient step response of a 1-pole filter is smooth and monotonic and there is no overshoot of the output waveform.
 
A 2-pole filter contains 2 reactive elements and can be implemented with two cascaded RC filters or by an LC filter or by an active filter with two capacitors. This frequency response of a 2-pole filter rolls off more rapidly at -12dB/octave and the transient step response will generally have little (<5%) or no overshoot.

Wow. That is a wonderfully clear and helpful explanation of a topic that I was despairing of understanding any time in the next few weeks.

Thank you.

[jdutky]

Sad news on the Tek 475. When I went to use it just now the horizontal sweep seems to have died, as has the second vertical channel. Also, when I went to pack it up, it looks like the bail has pulled out of the right side of the case.

None of this is really surprising, the instrument is over 50 years old, and it spent its first 10 years being hauled back and forth across the country, so the mount points for the bail were probably badly fatigued all along. The 475 is pretty heavy, after all.

Still, I'm sad about this. It really looked like it was in good repair and mostly functional, and I was just getting familiar with its features. Now I have had to fall back to my father's other scope, an only slightly newer 2213. This might the straw that forces me to join the 21st century and buy a nice new DSO (especially if the 2213 fails as well: then I'll need a new scope, even if it's just to diagnose and fix the old ones).

 

Thank you, everyone, for all the help on this. I really appreciate it. I certainly learned a lot, and I'll be using it both to evaluate the 2213, and on whatever DSO I end up buying.

[David Hess]

I will also have to work harder to get good camera images of the screen: your images are nice and sharp. Are you doing that by hand or do you have a rig to hold the camera?

I bought a Canon Powershot SX150 IS which uses 2 x AA batteries and has optical image stabilization for taking oscilloscope photographs and documenting other work.  It works well enough freehand that I just do it that way and then clean up the photograph in Photoshop.

(I've considered buying one of the Tek camera rigs on eBay and using it to mount a digital camera, but that seems like the wrong way to be spending money when I could save it for a nice, modern digital scope)

A camera mount could also be made from copper clad board to fit in the mounting rails on the bezel.

The reference to 1- or 2- pole filter roll-off comes from the mathmatics for representing the frequency response for a low-pass filter and implies a certain circuit configuration to produce such a response. The number of poles generally corresponds to the number of reactive elements i.e. capacitors or inductors, in the circuit.

A 1-pole filter is the type of response produced by simple single-stage RC low-pass filter. This type This type of filter starts rolling off gradually above its cut-off frequency and eventually reaches a ultimate roll-off rate of -6dB/octave i.e. for every doubling of frequency the filter output reduces 6dB. The transient step response of a 1-pole filter is smooth and monotonic and there is no overshoot of the output waveform.
 
A 2-pole filter contains 2 reactive elements and can be implemented with two cascaded RC filters or by an LC filter or by an active filter with two capacitors. This frequency response of a 2-pole filter rolls off more rapidly at -12dB/octave and the transient step response will generally have little (<5%) or no overshoot.

That applies when the oscilloscope's bandwidth limiting is engaged however you might noticed from my screen shots that the response is *not* a common 1 or 2 pole response; it is actually closer to a Gaussian response from a number of cascaded amplifiers which gives it that nearly constant slope instead of an exponential curve which is what would be shown if the 20 MHz bandwidth limit was active.