2235 Horizontal Calibration Issue


Stephen
 

Hi,
Since I’ve had this scope (2235), it’s one that I enjoy using because it’s fairly light in comparison to my other ones; the form factor is convenient. And it gets the job done for my needs.
However, for some reason(s), it’s the only scope I haven’t been able to properly calibrate horizontally; the center is always a bit off. It’s probably not such a huge deal overall, in the grand scheme of things, but it bugs me.
I’ve tried everything I know, or can think of: Geometry, etc... I regret that there is no Y-Axis settings like on most of my other scopes.
Any ideas are welcome.

The 2 Picts are both precisely 10Khz signals.
One from a function generator, and the other from my Tek 184.

The calibration procedure says to align the 8 middle pips. And they are as best as I can do it on this scope, but the central one is always off...

See picts:
https://groups.io/g/TekScopes/album?id=259498


Dave Peterson
 

I don't know about the 22xx series calibration procedures, but recently familiarized myself with the 465. There were several adjustments to the horizontal sweep that used/addressed the Time Delay Position potentiometer, B DLY'D sweep, as well as the 10x Mag to precisely set the horizontal sweep timing and centering.
I have a 2236 coming in the mail right now, and so will soon be familiarizing myself with the 22xx series calibrations. But I'd be very surprised if it does not include several adjustment potentiometers to set these parameters. There are several links (under the 2236) on the Tek Wiki site to the service manual(s). I'd recommend going there, if you don't already have a copy. Look up the calibration section. Having brought this up, I'll be looking it up myself as soon as I can.
https://w140.com/tekwiki/wiki/2235

Hope this helps.Dave

On Sunday, January 17, 2021, 04:40:32 AM PST, Stephen <stephen.nabet@gmail.com> wrote:

Hi,
Since I’ve had this scope (2235), it’s one that I enjoy using because it’s fairly light in comparison to my other ones; the form factor is convenient.  And it gets the job done for my needs.
However, for some reason(s), it’s the only scope I haven’t been able to properly calibrate horizontally; the center is always a bit off.  It’s probably not such a huge deal overall, in the grand scheme of things, but it bugs me.
I’ve tried everything I know, or can think of:  Geometry, etc... I regret that there is no Y-Axis settings like on most of my other scopes.
Any ideas are welcome.

The 2 Picts are both precisely 10Khz signals.
One from a function generator, and the other from my Tek 184.

The calibration procedure says to align the 8 middle pips.  And they are as best as I can do it on this scope, but the central one is always off...

See picts:
https://groups.io/g/TekScopes/album?id=259498


 

On Sun, Jan 17, 2021 at 01:40 PM, Stephen wrote:


The calibration procedure says to align the 8 middle pips. And they are as
best as I can do it on this scope, but the central one is always off...
Spec says +/- 5%, i.e. 0.1 division, so it's within spec. See SM, "HORIZONTAL", "PROCEDURE STEPS", "1. Check Timing Accuracy and Linearity", point "e.".

Raymond


Ozan
 

As Raymond wrote this could be within the accuracy of the oscilloscope. Do you see similar accuracy if you use B sweep?

Ozan


Dave Peterson
 

Page 5-14 of the 2235 Service Manual:
https://w140.com/tekwiki/images/3/30/070-4206-00_2235_Service_Manual.pdf

1. Adjust Horizontal Amplifier Gain (R740 and R730)...c. ADJUST - A Sweep Gain (R740) for 1 time marker per division over the center 8 divisions....e. ADJUST - B Sweep Gain (R730) for 1 time marker per division.
Don't know if R740/R730 are easily accessible, but that's what seems to be called for here. Should be accessible and easy since this is fairly routine calibration. But I'm an armature, so take my advise with an appropriate level of caution.
Dave

On Sunday, January 17, 2021, 11:14:09 AM PST, Ozan <ozan_g@erdogan.us> wrote:

As Raymond wrote this could be within the accuracy of the oscilloscope. Do you see similar accuracy if you use B sweep?

Ozan


Stephen
 

Hi Dave,

I also have several 400 series scopes. Some things are similar, and others are lacking (for me) on the 2000 series. Some adjustments are not on the 2000 series.
I have all the manuals. But thanks.


Stephen
 

On Sun, Jan 17, 2021 at 06:09 AM, Raymond Domp Frank wrote:


On Sun, Jan 17, 2021 at 01:40 PM, Stephen wrote:


The calibration procedure says to align the 8 middle pips. And they are as
best as I can do it on this scope, but the central one is always off...
Spec says +/- 5%, i.e. 0.1 division, so it's within spec. See SM,
"HORIZONTAL", "PROCEDURE STEPS", "1. Check Timing Accuracy and Linearity",
point "e.".

Raymond
Hi Raymond,
You’re right. But is bothers me because it’s the ONLY scope that has that issue. All the others are dead on.


Stephen
 

On Sun, Jan 17, 2021 at 08:13 AM, Ozan wrote:


As Raymond wrote this could be within the accuracy of the oscilloscope. Do you
see similar accuracy if you use B sweep?

Ozan
I haven’t checked sweep B. Good idea. Maybe they interact?


Stephen
 

On Sun, Jan 17, 2021 at 08:39 AM, Dave Peterson wrote:


Page 5-14 of the 2235 Service Manual:
https://w140.com/tekwiki/images/3/30/070-4206-00_2235_Service_Manual.pdf

1. Adjust Horizontal Amplifier Gain (R740 and R730)...c. ADJUST - A Sweep Gain
(R740) for 1 time marker per division over the center 8 divisions....e. ADJUST
- B Sweep Gain (R730) for 1 time marker per division.
Don't know if R740/R730 are easily accessible, but that's what seems to be
called for here. Should be accessible and easy since this is fairly routine
calibration. But I'm an armature, so take my advise with an appropriate level
of caution.
Dave
I have the manual, thanks.
R740 and R730 are just next to each other on the horizontal board.


 

On Sun, Jan 17, 2021 at 08:39 PM, Dave Peterson wrote:


Don't know if R740/R730 are easily accessible, but that's what seems to be
called for here.
Don't think so. It's not a problem of the amplitude (A sweep gain), because beginning-to end is ok, but a linearity problem. Amplitude corrections won't help. I guess there are some linearity adjustments for the highest sweep speeds but these (around 100 us/div) aren't.

Raymond


Dave Peterson
 

Ah, yes, I see now.
I've wondered this about the "tuning" of these scopes. When they were new I'm sure they were very linear - kind of the point of the careful design. But as they age and drift it's not a surprise that linearity can suffer. In familiarizing myself with the calibration I was wondering if it would go so far as recommending component replacement when a non-adjustable parameter goes out of spec. The procedures appear not to provide such guidance. And there's a practical limit to how tunable Tek could make such things. I presume the point of providing the schematics and circuit descriptions. Facilitating such debug and repair.
But herein lies the fun for me. Sorry Stephan, I don't have a recommendation for you at this point, but this is the sort of issue I'd enjoy debugging. I misunderstood the issue at first, but now that I do I'll have to study it some. My 2236 (parts - non working) is a day or two away. I'll need a manual too. Might be known or apparent to some, but this is where the learning is.
Dave

On Sunday, January 17, 2021, 02:07:25 PM PST, Raymond Domp Frank <hewpatek@gmail.com> wrote:

On Sun, Jan 17, 2021 at 08:39 PM, Dave Peterson wrote:


Don't know if R740/R730 are easily accessible, but that's what seems to be
called for here.
Don't think so. It's not a problem of the amplitude (A sweep gain), because beginning-to end is ok, but a linearity problem. Amplitude corrections won't help. I guess there are some linearity adjustments for the highest sweep speeds but these (around 100 us/div) aren't.

Raymond


Ozan
 

It is unlikely A and B sweep interaction is causing the sweep nonlinearity but if B looks more linear then you can focus on the sweep generator. A and B sweeps have different paths until they hit U760. Again, this could just be just be within spec and nothing to fix. I don't own this particular scope so I don't know what is normal.

If both A and B sweeps behave the same I would focus on the horizontal amplifier (sheet 6 in the SM), assuming it is not normal and it bothers you enough to debug it. In your pictures sweep speed is slightly faster (horizontal gain is slightly higher) at mid point (zero volts differential) compared to end points. This is common for most differential amplifiers, some kind of local or global feedback is used to improve linearity. I would check DC operating points marked on the schematic, especially after U760 to see if any voltages drifted. For example if voltage at bases of Q780/Q770 are off and/or voltage at top of VR764 zener is off diodes CR770/CR780 may be slightly conducting and compressing the signal. Note that 100V is present at this stage so safety is important.

Q770/Q780 diff amp is an interesting one. It is made of NPN and PNP with diodes at the emitters. Does anyone know the source of this circuit (patent, paper, etc)?

Ozan


Tom Lee
 

Hi Ozan,

I don't know of a specific patent, or even of a paper to cite, but it, and versions of it, have been around for a while. It's a handy way to make a diff pair out of a PNP-NPN pair, while keeping the base voltages nominally equal, as they would be in a conventional unisex diff pair. Coupling the emitters through a series connection of diodes gives you a level shift that cancels out the base-emitter voltage drops. I think I first saw a version of it in an early Wavetek function generator, and it is unlikely that I happened to have seen the first incarnation of it.

-- 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 1/17/2021 17:21, Ozan wrote:
It is unlikely A and B sweep interaction is causing the sweep nonlinearity but if B looks more linear then you can focus on the sweep generator. A and B sweeps have different paths until they hit U760. Again, this could just be just be within spec and nothing to fix. I don't own this particular scope so I don't know what is normal.

If both A and B sweeps behave the same I would focus on the horizontal amplifier (sheet 6 in the SM), assuming it is not normal and it bothers you enough to debug it. In your pictures sweep speed is slightly faster (horizontal gain is slightly higher) at mid point (zero volts differential) compared to end points. This is common for most differential amplifiers, some kind of local or global feedback is used to improve linearity. I would check DC operating points marked on the schematic, especially after U760 to see if any voltages drifted. For example if voltage at bases of Q780/Q770 are off and/or voltage at top of VR764 zener is off diodes CR770/CR780 may be slightly conducting and compressing the signal. Note that 100V is present at this stage so safety is important.

Q770/Q780 diff amp is an interesting one. It is made of NPN and PNP with diodes at the emitters. Does anyone know the source of this circuit (patent, paper, etc)?

Ozan





keantoken
 

Hi Ozan, on DIYAudio.com this arrangement is known as a "Rush cascode", because it was used by Christopher Rush in connection with THX. But it was always one potential way to make a differential pair so it probably goes back far before THX. It seems more common in audio because audio is less focused on absolute precision (not unlike a CRT).
I wasn't so bright in 2009, but here is a thread I started that shows some different ways it was used.
https://www.diyaudio.com/forums/solid-state/137509-rush-cascode-wiki-page.html


It is hard for me to see why the Q770 circuit couldn't just be a mirror of the Q780 circuit. But I will say that the diodes keep the transfer function of the circuit symmetrical on transients so maybe it is good enough for visual display. Perhaps there is some intentional obfuscation at play. It's also possible that an asymmetrical circuit was used so that the two feedback loops would be dissimilar enough to avoid common mode oscillation.
The diodes remove the offset between the differential inputs, but by this point engineers usually would rather just use a conventional LTP with better performance in most areas. The benefit of this circuit is mainly in having a positive and negative rail output rather than both outputs being on one rail for the LTP. And it also has virtually unlimited slew rate in one direction. If you combine two of them you can achieve extremely high slew rates. None of these things are taken advantage of in this application, so I don't think performance was the chief concern for this circuit. On Sunday, January 17, 2021, 07:44:51 PM CST, Tom Lee <tomlee@ee.stanford.edu> wrote:

Hi Ozan,

I don't know of a specific patent, or even of a paper to cite, but it,
and versions of it, have been around for a while. It's a handy way to
make a diff pair out of a PNP-NPN pair, while keeping the base voltages
nominally equal, as they would be in a conventional unisex diff pair.
Coupling the emitters through a series connection of diodes gives you a
level shift that cancels out the base-emitter voltage drops. I think I
first saw a version of it in an early Wavetek function generator, and it
is unlikely that I happened to have seen the first incarnation of it.

-- 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 1/17/2021 17:21, Ozan wrote:
It is unlikely A and B sweep interaction is causing the sweep nonlinearity but if B looks more linear then you can focus on the sweep generator. A and B sweeps have different paths until they hit U760. Again, this could just be just be within spec and nothing to fix. I don't own this particular scope so I don't know what is normal.

If both A and B sweeps behave the same I would focus on the horizontal amplifier (sheet 6 in the SM), assuming it is not normal and it bothers you enough to debug it. In your pictures sweep speed is slightly faster (horizontal gain is slightly higher) at mid point (zero volts differential) compared to end points. This is common for most differential amplifiers, some kind of local or global feedback is used to improve linearity. I would check DC operating points marked on the schematic, especially after U760 to see if any voltages drifted. For example if voltage at bases of Q780/Q770 are off and/or voltage at top of VR764 zener is off diodes CR770/CR780 may be slightly conducting and compressing the signal. Note that 100V is present at this stage so safety is important.

Q770/Q780 diff amp is an interesting one. It is made of NPN and PNP with diodes at the emitters. Does anyone know the source of this circuit (patent, paper, etc)?

Ozan






Tom Lee
 

A horizontal amplifier has very different requirements from those for a vertical amplifier. A vertical amplifier is supposed to be reproduce arbitrary waveforms within a large specified bandwidth. It drives the most sensitive pair of deflection plates to help lighten the constraints on choice of output transistors. A horizontal amplifier has to preserve a clean linear ramp, and a wicked-fast retrace, period; it doesn't have to handle arbitrary waveforms. That's one difference. Another is that the swings have to be substantially larger for the horizontal circuitry because the corresponding deflection plates are farther away, so the horizontal amp has to make use of transistors designed for higher breakdown voltage. That, in turn, implies transistors with worse ft. These differing requirements and available devices drive the design of these amplifiers in quite different directions, so much so that to say "engineers usually would rather just use a conventional LTP with better performance in most areas" makes me wonder just who these engineers might be. Tek engineers didn't choose different architectures (and different output transistor types) for horizontal and vertical amplifiers just to express their individuality. They were /compelled/ to make them different because of the different requirements.

As to "It is hard for me to see why the Q770 circuit couldn't just be a mirror of the Q780 circuit", the answer is that one could in fact do the obvious and implement the circuits exactly the way you describe. But if you look at the horizontal circuitry of high-performance scopes, they almost never do it that way. So again, the self-aware student should be asking "What could be the reason for doing the non-obvious?" The answer is that the output stage runs largely Class A at low speeds and transitions to Class AB only for fast edges, where you need strong push and strong pull to provide the necessary slew rates without burning huge static power when it's not needed. It has nothing to do with worries about CM oscillation. There is negligible CM feedback here, so CM oscillation is a non-issue.

-- 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 1/17/2021 19:49, keantoken via groups.io wrote:
Hi Ozan, on DIYAudio.com this arrangement is known as a "Rush cascode", because it was used by Christopher Rush in connection with THX. But it was always one potential way to make a differential pair so it probably goes back far before THX. It seems more common in audio because audio is less focused on absolute precision (not unlike a CRT).
I wasn't so bright in 2009, but here is a thread I started that shows some different ways it was used.
https://www.diyaudio.com/forums/solid-state/137509-rush-cascode-wiki-page.html


It is hard for me to see why the Q770 circuit couldn't just be a mirror of the Q780 circuit. But I will say that the diodes keep the transfer function of the circuit symmetrical on transients so maybe it is good enough for visual display. Perhaps there is some intentional obfuscation at play. It's also possible that an asymmetrical circuit was used so that the two feedback loops would be dissimilar enough to avoid common mode oscillation.
The diodes remove the offset between the differential inputs, but by this point engineers usually would rather just use a conventional LTP with better performance in most areas. The benefit of this circuit is mainly in having a positive and negative rail output rather than both outputs being on one rail for the LTP. And it also has virtually unlimited slew rate in one direction. If you combine two of them you can achieve extremely high slew rates. None of these things are taken advantage of in this application, so I don't think performance was the chief concern for this circuit. On Sunday, January 17, 2021, 07:44:51 PM CST, Tom Lee <tomlee@ee.stanford.edu> wrote:
Hi Ozan,

I don't know of a specific patent, or even of a paper to cite, but it,
and versions of it, have been around for a while. It's a handy way to
make a diff pair out of a PNP-NPN pair, while keeping the base voltages
nominally equal, as they would be in a conventional unisex diff pair.
Coupling the emitters through a series connection of diodes gives you a
level shift that cancels out the base-emitter voltage drops. I think I
first saw a version of it in an early Wavetek function generator, and it
is unlikely that I happened to have seen the first incarnation of it.

-- Cheers,
Tom


Ozan
 

Hi Tom and keantoken,
Thank you for the information.

I don't want to hijack this thread, maybe we should start a new one if we have more ideas.

I have seen series NPN-PNP pairs before, LM 741 input uses a similar structure
https://www.researchgate.net/figure/Circuit-diagrams-of-two-amplifiers-at-the-top-of-a-widely-used-IC-LM741-type_fig1_237625509
and
MOS class-AB input stage with NMOS-PMOS version has been around for some time:
http://www.cse.psu.edu/~chip/course/analog/lecture/01OpampTut.pdf (Gray & Meyer paper Fig 20).

However, I haven't seen a discrete implementation with differential input and differential output before. It turns out in reality I have but not noticed: My Tektronix 485's horizontal output is NPN-PNP type.

As Tom mentioned it looks like this choice has more to do with the characteristic of the sweep signal, where ramping up and down has different requirements and class-AB operation is desirable.

Ozan


Stephen
 

Although I don’t understand 90% of all this, it seems very interesting.

Maybe you guys have an idea where I should start looking, besides the voltages mentioned earlier.

Worth noting that when I got this scope over a year ago, I went thru all the calibration procedures, and the voltages were within specs. Although the 100V rail was a tiny bit high, I remember it was within specs too.
However, the horizontal sweep gain has always been non linear. And as I mentioned, it’s the only scope in my « collection » that has that particular issue.


Stephen
 

Thanks Raymond,

That’s exactly it « a linearity problem « . That’s the term I was looking for and which best describes the issue.

Thank you


keantoken
 

Check the voltages on the horizontal amp schematic on page 193. If those are okay, then move to whatever is before the horizontal amp and check those voltages. If you have another scope you can check the waveforms at the input and output of the sweep amplifier to see if the distortion occurs in the amplifier or before it.


teamlarryohio
 

Don't rule out the possibility of a slightly deformed plate. The test is of course to swap out the jug and see if it changes.
-ls-