¿ªÔÆÌåÓý

Take a look at
Connect the S-6 to the pulser.
Set time/Div to 50ps, distance to zero, vertical to 100 mV, Scan to REP, locate off (button in), scan speed high (knob to the right), high resolution off (button out).
Locate the trace vertically with the offset knob. If everything works, you should be able to bring the step into view with the FINE (ZERO SET) control.

"IDENRIFY" is a typical symptom of a mis-alignment in the readout system.
I don't remember off the top of my head which of the trimpots you need to tweak, but the procedure is described in the manual.

Thank you Dave.
Hopefully he¡¯ll see the post¡­
I¡¯m about 12h away from the guy showing up with the scope.

The 7S12 with S-6 adn S-52, if working, are about worth that, so in my book, yes, seems a good deal. Others might feel differently.

Richard Steedman is in your time zone, and got me started with the 7S12, and S-52. But I shouldn't speak for others - he may not be available. If you can stand the time difference (I'm in San Francisco) I'm glad to share what I've learned. But I am still in a learning phase myself.

I can be reached via my Gmail account with the same username as well.

Good luck!
Dave

Not that I need another scope, but I am offered what I believe to be a good deal, and an opportunity to buy this 7704A that comes with a 7A26, 7B70, and a 7S12 (S-6 & S-52).
I¡¯ve never ever used a sampling plugin before. I wouldn¡¯t even know how to see if it even works properly, nor where to start¡­
Could someone tell me how to check whether it¡¯s ok. The guy¡¯s bringing the scope over this evening. Asking price is 280€ (Approx $315) for the whole package. Good deal?

Thanks

Searching on "paraphase amplifier" turns up lots of hits for me, including a dictionary definition (that sounds surprisingly technical), a definition on element14, a definition at Newark.com which includes the term "push-pull", and a patent titled "Balanced Paraphase Ampllifier Including a Feed Forward Path", so I'd say that this is a pretty commonly used term. I do also get a lot of hits for "paraphRase amplifier" a number of which appear to be typos.

-- Jeff Dutky

Seems to be a Tektronix term. Looking up the term "paraphase amplifier" doesn't yield much. I did find a wiki article that used the term "paraphrase". Maybe a typo?

In any case, commonly used in 465, and likely other, circuit descriptions for the front-end of vertical amplifiers. Used in the referenced video on push-pull amplifiers.

Tek Museum: have "Part 1" of the video?

Dave

Paraphase, eh?? That's a new one on me.? Thanks for the new word.? ? ? ? Jim Ford?Sent from my Verizon, Samsung Galaxy smartphone

Found it:

And it's a push-pull amplifier, not paraphase. Paraphase is single ended to differential. I knew that. Duh.

On Wed, Dec 22, 2021 at 07:46 AM, Albert Otten wrote:

Stephen,
I think the oscillator frequency was too low because your tube is too good.
That is, has high emission. Then the anodes switch between +100 V and a rather
negative voltage, perhaps -55 V or - 60 V. The manual suggest -50 V. Then the
capacitors have a relatively longer discharge traject. ( 50/150 vs 55/155 or
60/160 if the switching grid voltage were -100 V; that voltage in reality is
more positive because of the common cathode resistor). In my unit the
negative voltage is about 35 V, frequency 1.08 kHz.
BTW the tube operates near or even above the maximum allowable cathode-grid
voltage, 50 V.
Albert

Very interesting Albert,
I¡¯ll check the voltages on the tube. Thank you.

Well, thus the mystery. How does an amp fail in this way? I'm at a loss.

Along the same lines of thinking: if I'm seeing the roll off at the interface pins 7A/7B then it must be something between the 5A18N pins and Q821/Q931.

To address a couple of things you brought up:

I was just down in the garage checking for differences between 7A and 7B. Long story short, I haven't been able to discern anything revealing. As you noted, there's no ground reference in the emitter side of Q821/Q931. I did watch an old Tek video some time ago about amplifier design, and the presenter showed how paraphase amplifiers can have a virtual ground. And that they will take a offset input and produce a equal differential output despite the input offset. I'm not finding it right now, but I'll keep looking.

I don't know if Prof. Hajmiri's next set of lectures have gotten into the capacitive feedback parasitics of BJTs. I know he covered it somewhat in small signal model development in the Device Physics portions. But I do recall this being called "Miller" capacitance in my course work, and it's a consideration in CMOS performance. Also this amplifier design seems that it could be considered an op-amp, but I'm not quite that up to speed on up scaling single transistor amps to op-amps. I recall Prof. Lee mentioning it in regards to the Z-axis amp you were trying to debug. If this horizontal front-end is an op-amp configuration, is there capacitive feedback? Even if purely parasitic? Can parasitic feedback capacitance cause this sort of BW short fall?

Dave,

So, I'll ask again, what kind of failure in Q821/Q931 (or around them) would result in lower bandwidth observed at their bases? Pardon my ignorant floundering, but it seems like the obvious failure (or change in component parameters) would be an increase in capacitance at or near the base of these transistors. As you said, your instrument does not have C921, which would be the obvious suspect, but suppose that the base capacitance of one of the two transistors had increased. That should decrease the bandwidth of one half of the differential amplifier, which should affect the entire amplifier (right?). And the increase in capacitance at the base one of the transistors would be visible at the corresponding connector pin as a change in the bandwidth at that pin.

At least, that's what my inexpert speculation would be.

Maybe I'm wrong. My feeling is that Q821 and Q931 should insulate the signal pins from seeing effects from components further to the right (in the schematic). Maybe there are other ways that one of the transistors could fail that could reduce the bandwidth? My feeling is that other failure modes (e.g. shorting from B to E, or failing open) would cause the amplifier to fail in much more obvious ways (e.g. not working at all), but I'm not doing any kind of rigorous, quantitative analysis, just making qualitative guesses.

-- Jeff Dutky

Cleaned the Delay Time switch, not fun. Used paper and MG Alcohol. See what Strange behavior I¡¯m getting now, still can¡¯t move pointer to left with position control.

Dave,

Now that I'm thinking about it, is it even sensible for me to ask if the rolloff happens on only one of the two differential signals to the amplifier? Is there an easy way to measure just half of the differential signal? Are the + and - signals referenced to ground? I don't see ground in the amplifier circuit (except through a C721 to the collectors of Q820 and Q930, does that count?) anywhere, so I'm not convinced that measuring between + and ground would tell you anything.

-- Jeff Dutky

Dave,

I agree that the problem is likely to be in the horizontal amplifier, but, as I said, given that you are seeing the rolloff at the connector pins, it can't be after Q821/Q931. The only components between the connector pins and the transistors are a pair of resistors, and I doubt that they are the culprits here (but what do I know? I have considered that you should check all the resistors around Q821 and Q931 to see if any have drifted out of spec).

Have you verified that the rolloff occurs on both + and - signals? It is much easier to believe that one of the two transistors is bad, and is causing the rolloff, than that both are bad (though, I guess they are probably from the same batch, so they might be failing in a synchronized fashion).

There are so few components involved here that I'm sure you could find the failed components quickly with an exhaustive search. My 5111A is currently inaccessible, or I would try to reproduce your work.

-- Jeff Dutky

Do you have a copy of the correct service manual? A&B Sweep Generators are suspect. Horizontal amp appears out of balance, since the spot will only move slightly left of center and off the screen in the other direction. Look for dislodged components as there are many sockets in these scopes, including individual transistor sockets. . Start by looking for the wave forms indicated at the various test points of the service manual.

--
Michael Lynch
Dardanelle, AR

I tend to ramble, so I might have been too brief in an effort to be concise.

I have the 5111A in question and a 5110 (5103/D10), so 6 slots. I have one 5A18N that I feel pretty good about - I've run through the adjustment procedure on it and checked it's gain at the 7A/7B outputs (taken from the rear of the interface connectors). From several of the available slots. I think I didn't measure off the back of the 5110 vertical interface. That would have called for removing the left cabinet. I know I did measure the horizontal interface on the 5110 because I had the right cabinet panel off. Otherwise I have very good consistency of both the gain at the output of the 5A18N and set the vertical and horizontal gains of both scopes based on it.

I have a PG506 and TG501 which I have good confidence in. I also have a SG503 that appears to work well. It seems to do the things it's supposed to do, and I get reasonable BW response out of the equipment I've applied it to. I won't say more than that about it because I have yet to try to calibrate it. The PG506 and TG501 I have run through check and adjustment procedures. For time reference I have a GPSDO, and for level reference I have a AD584. I agree that the level reference is of limited precision, but it's something. I have room for improvement on my level calibration, but between a DM501, the PG506, and the scopes I've calibrated with them, everything is in very good agreement. I have good confidence in the 5A18N and do not feel it is the source of the issue.

I ran through the check procedures for both scopes and the BW on the 5111A horizontal as measured with the SG503 has rolled off significantly lower than all other 5 slots available. I've checked both via the scope displays and with 10X probes on the interface connectors 7A/7B (differentially on a 465 - i.e. CH1, CH2 INV, ADD). As noted, I have not measured the at vertical interface connectors on the 5110. I've tried my best to eliminate dumb mistakes in measurement. The BW on the horizontal slot of the 5111A consistently rolls off at about 1.5MHz, where all other slots have at least 2MHz BW. The 5110's are all measuring about 2.1MHz and the 5111As vertical slots go up to 2.5MHz when not probing the interface pins.

The BW numbers quoted here are based on a 6 division output from the SG503 set to 50kHz, then the BW measurement is the point at which the display reduces to 4.2 divisions while raising the frequency on the SG503. This is per the procedures in the corresponding service manuals for the 5110 and 5111A. (6 x 0.7 = 4.2).

After measuring BW on screen, I have measured the 5111A horizontal BW at interface connector pins 7A/7B in the hopes of proving the roll-off was happening later in the horizontal amplifier. I had expected the issue to be later in the Horizontal Deflection, but I was surprised to see the roll-off on the output of the 5A18N. Just to note, with probes the BW rolls off at about 1.5MHz, and without probes connected is seems to roll off at 1.6MHz. The probes do reduce the BW, but the issue is there without them as well. The vertical slots show a BW reduction with probing: from 2.5MHz without, down to 2.1MHz with. So contribution of probe loading (P6105 10X) is something to consider, but it is not the source of the 1.5MHz limit.

To me this seems to indicate the first stage of the interface amplifier is creating a large capacitive load at Q821 and Q931? Noting that R922 and R932 are out of circuit when the plug-in is removed I checked those. They are both just at the high end of tolerance at 345 ohm.

I've noted that the horizontal amp front end is the same between the 5111A and the 5110. Both use an "SPS8801" NPN, which is MPS6521 according to the semiconductor cross reference. It seems innocuous enough - a high Beta/ft version of a 2N3904. I suspect I'll try pulling those and checking them out on the curve tracer once Santa drops off the de-soldering station I asked for. But there seems nothing esoteric about the interface amplifier. This also suggests to me that there is an issue in the 5111A horizontal amplifier and not the 5A18N or my measurement technique. If there were issues there I'd see this issue on the 5110, but instead I'm getting a consistent 2MHz BW limit in all slots on that scope.

I'm still wary of concluding the issue is Q821/Q931. I'm getting better at understanding these discrete component amplifiers, but have a ways to go before I feel like I know what to look for. As you are the one who introduced me to Prof. Ali Hajimiri's Analog Circuit Design course: I've just finished the Device Physics portion (nice refresher!) and am about to embark on the Basic Amplifier Stages. The revisit of small signal models of BJT/MOSFETs does have me considering the small signal BW aspects of these amplifiers. It seems this low BW issue indicates a capacitive issue in the amplifier? Thinking beyond the input Qs, could there be feedback capacitance playing a roll here? Might these Qs check out fine on a curve tracer, but still have an issue with capacitance? How do I check that?

While eyeing the interface board on the 5111A - it does appear relatively easy to remove - I'm trying to consider how there might be something physical to look for here. Perhaps damage to the connectors? Solder blobs? Etc. I'll likely start plucking parts and trying to measure them. But I would like to have some idea of what I should be looking for.

That's enough rambling. Hope this clarifies things.
Dave

Stefano,

There is vertical deflection from the calibration signal, it's just that in the early part of the video the vertical scale is set too high to see the deflection. Later in the video, when he changes the VOLTS/DIV settings, you can see vertical deflection.

The lopsided horizontal deflection suggests that either one half of the horizontal amplifier is malfunctioning, or that one of the deflection lines has come loose from the CRT terminal. I would first check that the horizontal deflection terminals on the CRT neck are both connected.

As to why there is no sweep, that is more difficult to determine. I had a similar failure in a 475 that was due to a decoupling cap on the sweep board leaking it's guts out. It was a simple diagnosis and a simple fix. I would open the scope and have a close look at the boards in the horizontal section to see if anything looks obviously damaged. In my case, the cap in question, an axial electrolytic about half and inch long, had one leg turned all green and crusty, even an idiot (like me) could spot it.

If nothing obviously damaged turns up, then you will need to do a more systematic investigation.

-- Jeff Dutky

Here is a list I made for someone that wanted to get his going with new parts:

The parts I selected are the best in quality I could find. I want parts to
be in a piece for a long time without replacing because of a cheap part
that barely works. The part numbers are Mouser numbers. Ones that do not
have a part number after them are in the amount above with the same value.
No prices since I do not know the conversion for different countries. I did
not list the axial tantalums on the A10 and A11 boards as they rarely go
bad. They could be changed if the owner wants. R420 and R421 on the A16
board were lowered to bleed the B+ off faster. There is no reason other
than it as a random selection that is a common to find. What I called for
is a 1% type although a 5% will be fine in this case. The battery is a
larger capacity one than originally used to give a longer life. Early
models have a battery in one place and the chip to the rear while newer
ones use the Dallas type with the internal battery. This is a comprehensive
list. Check the s/n of your unit to make sure it matches parts. Higher
capacitance values are in some cases. It will filter better. The board
numbers, Axx:, are listed in order they appear in the manual. The power
supply diodes are not listed. Replace as necessary with heavier duty ones
when originals fail.

A10: C110, C111, C112, C122, C140, C141, C142, C169, C190, C202, C211,
C213, C215, C223, C263, C265, C340, C465, C468, C471, C480, C511C C523,
C524, C652, C768, C851- 47MFD 25V (40) 647-ULD1E470MDD1TD
C530- 22MFD 10V 581-TAP-226K010SRW
A11: C130, C131, C700, C701, C702- 22MFD 25V to 47MFD
A12: BT800 battery 667-TL5903P
C590, C882, C884, C886, C904- 47MFD 25V
C938- 10MFD 25V 647-ULD1H1100MDD1TD
A13: C702, C731- 22MFD 25V (2) 581-TAP226K025SRW
C881 1MFD 35V (film) (10) 80-R82DC4100DQ60J
A14: C903, C904- 47MFD 25V
A16: C105, C305- 680MFD 200V (2) 647-LGR2D681MELA40 or (2)
647-LGR2D821MELC30 for higher filtering if it will physically fit.
C128, C138, C184, C829, C900- 1MFD 35V (film)
C244- 180MFD 40V to 35V 647-UHE1V221MPD1TD
C262, C553- 840MFD 12v TO 1200MFD 16V (2) 647-UHE1C122MPD
C455- 1200MFD 6,3V TO 1500MFD 16V 647-UHE1C152MHD6
C460, C856, C956- 250MFD 20V to 330MFD 16V (3) 647-UHE1C331MPD
C461, C487, C494, C550, C585, C594, C595, CC650, C695, C947- 100MFD 25V
(11) 647-UHE1E101MED1TD
C664, C764- 4,7MFD 35V (2) 80-R82CC4470Z330J
C706, C816- ,068MFD film (2) 594-F339X136848MFP2B
C625- ,01MFD 80-F872BB103M480R
C219, C225- ,0022MFD 594-F339X122248MDA2B
C223- .001MFD 871-B32911A5102M
C750, C756- 180MFD 25V to 330MFD (2) 647-UHE1E331MPD6
R420, R421- 470,000 to 150,000 1W (2) 594-MBE04140C1503FC1
R223- 100,000 1/2W to 2W 71-CPF3100K00FHE14
R627- 150,000 1/2W 660-MF1/2CC1503F OR (100) 273-150K Xicon only sells
1/2W in packs of 100
A17: C218- 4,7MFD 100V 647-ULD2A4R7MDD1TD
C317- 10MFD 100V 647-ULD-2A100MED1TD
C613- 100MFD 25V 647-UHE1E101MED1TA
A10: R220, R512, 360 1/2W to 1W (2) 279-H4P360RFZA
A17: R245- 357,000 1/2W 594-HVR3700003573FR5
R246- 442,000 1/2W 603-MFR50SFTE52-442K
R247- 121,000 1W 71-CCF60-121K-E3
R248- 169,000 1W (2) 71-CMF5584K500FHEK (put these two in series)
R689- 357,000 1W MFR100FTE52-330K in series with 756-MFR4-27KFI
Optional- change the two NTC from the mains input from 5 ohms to 10 ohms to
reduce inrush current. (2) 527-CL60

Use a high deg. C m/k (4 or more) heatsink compound from heatsink to inside
base to help transfer heat and on any devices on a heatsink. Something like
MX-4 compound. Using this compound is advised for the 2465/7 series. Put
the compound on the CCD chips with the black heatsinks.The RIFA capacitors
have been speced and changed manufacturer to X1 values which will easily
handle 230/240V mains. The voltage rating of these is higher than the X2
types. Be sure to oil the fan with a heavier oil. That has not been
lubricated since it was new. Using a shielded mains cord will help keep
noise from being radiated from the cord by the noise coming in and noise
generated from the supply. Any FAIL modes might change to PASS after
replacing the battery and power cycling it a couple of times. My 2440 did
this. I do not know how it went from FAIL to PASS without calibration
unless the internal resistance of the new battery is the reason even though
the original checked good with a VTVM.

That is strange, you get a vertical movement with your finger, but nothing with the calibrator that a working instrument confirms there is a signal.
The horizontal contro looks like isn't properly working, since you can't move the "point" end to end.
While restoring my 7104 I had huge and unexplainable problems with bad contacts, I would start from complete cleaning connectors, switches, etc. with contact cleaner after removing dust with a brush and dirt with IPA. Basing on my (small) experience I wouldn't stop to IPA... I had to use contact cleaner.
Then check voltages out of the PSU... when you get all voltages in spec you can start having fun with the service manual.