Topics

Difference between a 475 and a 475A


 

I have a 475 that belonged to my father, manufactured in late 1974, that I have a sentimental attachment to and would like to keep running. To this end I have purchased several "for parts" scopes on eBay, both 475s and 475As, because many of the mechanical parts appear to be completely interchangeable. I had terrible luck with the first parts scope: while it is in terrible cosmetic condition (busted knobs, spattered with something that does not wash off with soap and water or with isopropyl alcohol), it appears to be operating flawlessly. Next I received a 475A which appears to be a much better donor candidate as it appears to be quite sick.

I have done thorough inspections of both scopes when I received them, and I noticed something odd: the sweep board of the first parts scope (475) does NOT look like the sweep board on my father's 475, but it DOES look almost identical to the sweep board on the 475A, which makes me wonder if it might not actually be a 475A with the faceplate of a 475.

I pulled down the schematics for both the 475 and the 475A and compared their vertical amplifier sections, on the basis that a higher bandwidth scope would require upgrades to the vertical amplifiers. I also made a thorough assay of the part numbers for the PCBs and custom ICs in all three scopes, hoping to find some definitive evidence for which scopes were 475s and which were 475As. Sadly, aside from the layout of the sweep board, I have not found anything conclusive.

The collected information on the three scopes follows:

475, vintage 1974
vertical amp board PN 670-224-01 or GH-2780-01
- switch IC PN 155-0091
- 1st pre-amp IC PN 155-0085-01
- 2nd pre-amp IC PN 434-0078-02 / 437-0078-02
main board PN 670-2239-04 or GE-2779-01
trigger board PN 670-2241-03 or GH-2781-01
- sweep IC PN 155-0059-01
- trigger ICs PN 155-0032-01
sweep board PN 670-2244 or GI-2784-05

475/475A? vintage 1976
vertical amp board PN 670-2240-08 or GH-2780-01
- switch IC PN 155-0091-00
- 1st pre-amp IC PN 155-0085-10
- 2nd pre-amp IC PN 019-0078-10 / 936-0078-10
main board PN 670-2239 or GE-2779-01
trigger board PN 670-2241-03 or GE-2781-02
- sweep IC PN 155-0049-02
- trigger ICs PN 155-0217-00
sweep board PN ? or GD-3980-02

475A, vintage 1978
vertical amp board PN 670-2240 or GH-2780-01
- switch IC PN 155-0091-00
- 1st pre-amp IC PN 155-0085-01
- 2nd pre-amp IC PN 813-0098-03
main board PN 670-2239 or GF-2779-01
trigger board PN 670-2241-02 or GA-2781-02
- sweep IC PN 155-0049-02
- trigger IC PN 155-0032-01
sweep board PN ? GD-3990-02

There are also differences between the parts in the 3rd pre-amp, but I'm not sure what to make of them because the physical circuit on the boards differs significantly from the schematics and parts lists.

Are the 475 and 475A really just the same scope with a marketing upgrade? I know that Tek over-engineered their scopes, and that a 475 may well be able to handle signals up to (and beyond) 250 MHz, but I expected to find noticeable upgrades to the vertical amplifiers. I suppose that it's possible that my father's scope was upgraded in the field (that is, by Tek field service personnel, not by my father), but other such modifications have been accompanied by option indicators, and there are not such indicators in my father's scope.

Have I been looking in the wrong place for the modifications that would distinguish a 475 from a 475A?

Are there, in fact, no such modifications, and the 475A was merely a marketing upgrade?

How would I be able to tell if a scope had had its front panel swapped with another scope? Is the scope serial number recorded anywhere other than on the plate at the bottom of the front panel?

Thanks,

Jeff Dutky


John Gord
 

Jeff,
My unresearched recollection is that the 475A was basically the same as the 475, but adjusted for higher bandwidth but perhaps slightly inferior pulse response. The 475A lacked the 2mV/div vertical setting.
--John Gord


 

John Gord wrote:

My unresearched recollection is that the 475A was basically the same as the 475, but adjusted for higher bandwidth but
perhaps slightly inferior pulse response. The 475A lacked the 2mV/div vertical setting.
Cool. That's something that I can verify! The scope whose faceplate reads "475A" does indeed lack a 2mV range (and it has a 100V range, which the 475s lack). That means that I should be able to see a difference in the cams for the vertical range switches (same number of cam positions, but they should engage different sets of attenuator blocks, or the blocks should have different values).

If the 475A were working better I could also check the pulse response, as I acquired the tools to do that measurement while evaluating my father's scope a few months ago. I'll be checking the pulse response of the both of these scopes under any circumstance, once I've repaired whatever is wrong with the 475A, just so I know what parts are good candidates as spares.

I suppose that this would also mean that, if you had swapped the front panel and knobs from a 475 onto the innards of a 475A, the V/div would be wrong (everything would be off by one position). That should be easy to verify on the suspect 475 (except that the V/div knob skirt is basically illegible from whatever is spattered all over the poor scope).

Any advice on what kinds of solvents are reasonable to use on the front panel and knobs of a 475?

I've tried both of the suggested cleaning agents -- detergent with water and isopropyl alcohol -- to no effect on this scope. I'd like to clean up some of the surviving knobs and skirts, but don't want to use anything that will damage the plastic or paint.


 

John Gord wrote:

the 475A was basically the same as the 475, but adjusted for higher bandwidth but perhaps slightly inferior pulse response
I'm not sure I understand what you are saying here: wouldn't "inferior pulse response" mean a longer rise time? Wouldn't a longer rise time necessarily mean a lower bandwidth?

I don't really understand all the math, but my impression, from evaluating my father's 475 and probes a couple months ago, was that you calculated bandwidth as the inverse of rise time multiplied by a constant (0.3-something?). Am I misunderstanding what is meant by "pulse response"?

-- Jeff Dutky


Tom Lee
 

You're narrowly interpreting "inferior" to mean "slow", but there are other qualities to consider, such as overshoot, ringing, "dribble" and so on. If you allow a little extra overshoot, it's possible to have a higher bandwidth, for example, but if you're fussy about the time response, then that tradeoff would be considered inferior.

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 11/19/2020 17:58, Jeff Dutky wrote:
John Gord wrote:
the 475A was basically the same as the 475, but adjusted for higher bandwidth but perhaps slightly inferior pulse response
I'm not sure I understand what you are saying here: wouldn't "inferior pulse response" mean a longer rise time? Wouldn't a longer rise time necessarily mean a lower bandwidth?

I don't really understand all the math, but my impression, from evaluating my father's 475 and probes a couple months ago, was that you calculated bandwidth as the inverse of rise time multiplied by a constant (0.3-something?). Am I misunderstanding what is meant by "pulse response"?

-- Jeff Dutky




Harvey White
 

There's a subtle "gotcha" in here.  It may be best thought of as the problem when designing a filter (lowpass, highpass, bandpass, doesn't matter, but something designed to cut off a bunch of frequencies - or allow them to go though)

In general, as I remember it, the sharper the response (the better the filter is at rejecting things really close to the edge of what it thinks is ok), the more ripple in the filter passband.

To state it another way, the filters that have the flattest response across the frequencies that are "OK" are the ones with the most shallow slope on the cutoff.  That is, they're the ones that aren't sharp.  The sharper the filter (the better it is for rejecting the "out of band" frequencies), the worse it is for flatness.  You can fix that, but only at the expense of adding more parts and other compromises.  Certain filter designs are optimized for flatness, certain for edge response.

Oscilloscope wise (and I can be wrong here), you optimize an oscilloscope response (because it has a LOT of filters for adjusting response and compensating for it, so think of the entire vertical amplifier as a big filter with gain) for either flatness (good for bandwidth measurements) OR you optimize it for pulse response (which means that the gain across the bandwidth is not even).

You get one or the other, depending on what you are measuring. Since the two adjustments (pulse fidelity and frequency fidelity (so to say) are complementary, you don't get both.

Typically, pulse response accentuates the high end response of the scope (and makes the pulse edges look good) at the expense of the frequency response (which may make the flat part of the pulse look bad, depending on frequency).

Hope that I explained it well enough, and I got it right.

Harvey

On 11/19/2020 8:58 PM, Jeff Dutky wrote:
John Gord wrote:
the 475A was basically the same as the 475, but adjusted for higher bandwidth but perhaps slightly inferior pulse response
I'm not sure I understand what you are saying here: wouldn't "inferior pulse response" mean a longer rise time? Wouldn't a longer rise time necessarily mean a lower bandwidth?

I don't really understand all the math, but my impression, from evaluating my father's 475 and probes a couple months ago, was that you calculated bandwidth as the inverse of rise time multiplied by a constant (0.3-something?). Am I misunderstanding what is meant by "pulse response"?t

-- Jeff Dutky





 

Dr. Lee, and Harvey White,

Thanks for the informative responses, that does make more sense, especially in light of an impulse input, which I understand implies a broad spectrum of underlying frequencies including very high frequencies (in order to reconstruct the impulse from the elements of the Fourier decomposition). Those high frequencies, which are required to get a good steep rise time, will also show up as high frequency ringing after the rise and fall, which messes up the fidelity to the overall shape of the pulse.

This also makes sense to me from a mechanical analog to the electrical signal: if you have a damped spring system you can either get fast changes in the displacement, but have lots of wiggling afterwards, or you can have very little wiggling but the displacement is slowed down a lot. A loose piston attached to the spring lets the spring change length very quickly, but doesn't do much to damp overshoot and undershoot or to dissipate the energy in the spring after the change in length, but a stiff piston which will rapidly stop the under/overshoot will also make it much harder and slower the change the length of the spring.


Tom Lee
 

Yes, that's basically it. An impulse and a step both have spectra that extend to infinity, so both will exercise the channel over an infinite bandwidth. The requirements for a perfect time-domain response are a flat gain to all frequencies, and also a phase response that varies linearly with frequency. That latter criterion is just a high-falutin' way of saying that you want all Fourier components to experience the same time delay.

Real amplifiers roll off and shift phase nonlinearly, of course, so you want to approximate the above two criteria to the best possible extent. If you focus on only one, the other tends to suffer. Scope designers face a tough challenge, and Tek's engineers mastered the art of designing amplifier chains with remarkably well-behaved response over a broader band than others thought possible with the technologies of the day. The tricks (oops, I mean methods) they came up with remain relevant to this day.

-- Cheers,
Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 11/19/2020 18:55, Jeff Dutky wrote:
Dr. Lee, and Harvey White,

Thanks for the informative responses, that does make more sense, especially in light of an impulse input, which I understand implies a broad spectrum of underlying frequencies including very high frequencies (in order to reconstruct the impulse from the elements of the Fourier decomposition). Those high frequencies, which are required to get a good steep rise time, will also show up as high frequency ringing after the rise and fall, which messes up the fidelity to the overall shape of the pulse.

This also makes sense to me from a mechanical analog to the electrical signal: if you have a damped spring system you can either get fast changes in the displacement, but have lots of wiggling afterwards, or you can have very little wiggling but the displacement is slowed down a lot. A loose piston attached to the spring lets the spring change length very quickly, but doesn't do much to damp overshoot and undershoot or to dissipate the energy in the spring after the change in length, but a stiff piston which will rapidly stop the under/overshoot will also make it much harder and slower the change the length of the spring.