Date   
Re: 2445A calibration

maxim.vlasov@...
 

Chuck,

Regarding the ROMs it's certain that the thing is hand-written in ASM (but not macro ASM). Looks like the work was spliced between the team and they have put almost the double effort into creating the infrastructure 2 or even more times. They could simply create the common low level drivers and put all of them together with the interrupt service routines into the lower RAM working area or at least use the macro ASM and use the same code modules for the time critical applications. But they have prefered another way of doing things. Or maybe the current state of the project is the result of a long and painful evolution (usually when things are getting pretty ugly at the end). Also the code hierarchy and the depth of the nesting is limited by the stack heap size, which is also sitting in the bottom 2K area.
I'm nobody to judge the mighty Tek decisions, but I would make the system more simple. One of the complexities in the current FW are due to the fact that there is no common always "visible" ROM area with the interbank calls and routines. Also it would be good to keep the interrupt vectors in RAM (just give the ROM under reset) especially since there is plenty of free space at the end of each 32K ROM partition.
But at the end looks like apart from jumping all over the place the thing is well debugged (the most important) and does its function.

Also thank you for pointing to the screen adjustment, which I have always skipped. I'll start from that and check out again whether I can put the cursor beyond the bound of the graticule after that. Regarding the ripple, what was measured on J119 is well in the spec (almost no ripple at all).

Thanks again,

Re: 2445A calibration

maxim.vlasov@...
 

Hello Siggi,

Thank you for your suggestions,

I'll try checking the leakage in the store capacitors (10MOhm probe is ok for that IMHO).

A1 board was repaired ("butchered" by me) from above without unscreweing and removing is at all. I.e. the opamps desoldering in the CHANNEL 1 and the hybrid replacement was done by keeping the A1 board in place (my fault).
It seems like I would have to do more substantial time investment, remove A1 and recap the board and finally verify/replace C851-C854 caps.
Also I follow Chuck's recommendation about the raster adjustment since I've never done that before.
On top of that I've measured the old removed electrolytic caps from A5 and they seem to be fine. Low ESR, good capacitance, very low self-discharge current. But in the power supply the most of the caps were really bad. Also A1 runs in much tougher temperature mode, so likely the recapping is necessary also.

looking at the schematics just noticed, that the -1.25V and +1.36 references from A5 are used on A1 all over the place.

Re: What is the meaning of ULU? Topic was: 7000-series power-hungry plug-ins

 

Hi Håkan,

After your post I re-read the two pages you provided but I still do not see the headline on the first page that says Universal Load Unit. I think ULU must be defined on a different page from the two pages you scanned.

I re-read this 5 times and I still don't see it. Maybe this is a case of RTFM but I don't think so.

Dennis

-----Original Message-----
From: TekScopes@groups.io [mailto:TekScopes@groups.io] On Behalf Of
zenith5106
Sent: Sunday, October 14, 2018 2:50 AM

On Sun, Oct 14, 2018 at 02:49 AM, Dennis Tillman W7PF wrote:


So far no one has offered an explanation for ULU. I can't think of
anything that fits those letters and there are very few ULU acronyms.
How about Universal Load Unit referring to the 067-fixtures mentioned in
the headline of the first page.
/Håkan



--
Dennis Tillman W7PF
TekScopes Moderator

Re: 468 horizontal jitter

Brendan
 

On Sat, Oct 13, 2018 at 08:22 PM, lop pol wrote:


On Sat, Oct 13, 2018 at 10:38 AM, lop pol wrote:

I'm working on my second 468 and I noticed some horizontal jitter.

On Sat, Oct 13, 2018 at 09:51 AM, Jeff Urban wrote:


In case you haven't, it is important to check if it is a centering problem
or
a triggering problem. Even in auto or free run it is still triggered, so
if
it
is the problem could be all the way back at the ramp generator.

If the trace itself is moving and not the sweep (I hope you got that...)
then
I would more look for hash on a PS line somewhere. Of course that applies
to
the ramp generator as well but different PS legs of course.
Thanks Jeff.
There is jitter on the A gate out with no input just a free running trace. I
had to put my 465 in 10x to see it but it does match the jitter on the 468.
So... This looks like its going to be a fun one to figure out.
Q181 on the sweep and z axis board was not seated properly. I have no idea how
it happened. I have never messed around inside but the base was bent instead
of seated. Looking at the schematic I can't tell how that could cause the
trace jitter though. The base is connected to ground. So far it seems to be
working correctly though. I also installed the EPROMs and images from the
Vintage Tek museum. Boards fit correctly and so far so good.
Well I spoke too soon. Still the same issue. I hate erratic occurring issues. It has to be something in the horizontal amplifier. XY mode the dot will sometimes just slightly wiggle horizontally other times it will quickly shoot 2 to 3 divisions horizontally. Its almost unnoticeable until you push the 10x button but is still occurring. I think I have about fully curve traced every transistor in a 468. I need to get some freeze spray.

Re: PG 506 IS IT NECESSARY FOR CALIBRATING?

Jim Olson <v_12eng@...>
 

Albert,
Thanks for the reply you covered part of the reason I put the query up to see if what I had was OK to cal the scopes I have or if I should seriously consider putting the other plug ins up on the ebay and figure a way to get a PG506. Problem there is they are all in Europe and expensive.
The SG503 is not so hard to get and more affordable watching some on the bay. I don't have the 191 on hand at this time as the person I got all the stuff from forgot to include it and haven't been able to drive there and pick it up it is a 110 mile round trip.
I am assuming that the other modules are OK for my purpose?

Jim O

On October 14, 2018 at 2:05 AM Albert Otten <aodiversen@... mailto:aodiversen@... > wrote:


On Fri, Oct 12, 2018 at 12:54 AM, Jim Olson wrote:

>

> > This is for the folks here that are a lot more knowledgeable and experienced
than me with these 4xx series scopes with the testing and calibrating of them.
I have a number of the TM series plugins plus case for them that I got really
lucky and got for a great deal. Here's a list of the modules: DM501, DM502A,
PG501, PG502, PS503A, FG504, Type 187 Time mark Generator, Type 191 Signal
Generator and a CT3, need to get the BNC adapters for it I see there two types
one straight through and one with a 50ohm terminator should i get both? I
still need the 50ohm terminators and coax cables. I have 100' of RG58U just
need to pick up some cable connectors for it.
I also have a good 2215 scope with a P6208 probe.
So my question is do I need the PG506 to really calibrate the scopes and also
should i get a SG503 to get the higher Mhz output?
Once i get the 466 together I will really need to calibrate it so do i have
enough cal instruments? and I will be asking for much help too!

Jim O

> Jim,
What you miss in your series of plugins is the possibility of quick DUT V/div checks, i.e. you miss the Standard Amplitude mode of the PG506.
The PG502 is fast enough (rise time) and flat (pulse top flatness) to replace the PG506 for waveform checks.

The Type 191 is sufficient to check the frequency response up to 100 MHz, so good enough for a 466. You might be curious how much higher the 466 goes; then of course you need a SG503 or similar.

I use the 50R feed through terminators 011-0049-01. But for a 100 MHz scope a T-connector with an arbitrary 50R end stop in one leg will do. I gathered some end stops when the local computer coax network at my work was dismantled.

Albert



Re: What is the meaning of ULU? Topic was: 7000-series power-hungry plug-ins

 

On Sun, Oct 14, 2018 at 05:47 PM, Dennis Tillman W7PF wrote:


After your post I re-read the two pages you provided but I still do not see
the headline on the first page that says Universal Load Unit. I think ULU must
be defined on a different page from the two pages you scanned.
It's not defined on any of the pages. What I meant was that it referred to the
067-0883-99 / 067-0919-99 which are on the top of the page.
067-0883-99 is a Universal Load Unit and IIRC the 067-0919-99 was used to program it.
/Håkan

Re: 2445A calibration

Chuck Harris
 

One of the guys I know replaced the 68B02 CPU
in an HP frequency counter that is highly coveted
by timenuts, with a board that hosted a Beagle
Bone Black single board computer running linux,
that in turn hosts a 6800 emulator.

Because the BBB SBC runs linux, it also gives the
counter USB, ethernet, an onboard web server, and
its own internal webpage. The extra compute power
allowed a whole slew of enhancements to the original
code... including self contained Allen Variance
calculations of time reference/standard quality.

The emulator easily runs the original ROM code as
well as the original CPU, faster actually until they
throttled it back, and is much faster in its ability
to ship data from the counter to the outside world.

I routinely run my counter half a building away
from my office computer.

Now, you can get even smaller SBC versions of
the BBB that are equal in capability.

So, why would I mention this? Well, because of the
SMD electrolytic capacitors leaking all over the
2445/65/67B A5 boards, we are starting to run out
of serviceably good A5 cards.

Might be a fun way of, fixing the SMD capacitor problem,
and, uhmmm?, replacing the NVRAM.

-Chuck Harris


maxim.vlasov@... wrote:

Chuck,

Regarding the ROMs it's certain that the thing is hand-written in ASM (but not macro ASM). Looks like the work was spliced between the team and they have put almost the double effort into creating the infrastructure 2 or even more times. They could simply create the common low level drivers and put all of them together with the interrupt service routines into the lower RAM working area or at least use the macro ASM and use the same code modules for the time critical applications. But they have prefered another way of doing things. Or maybe the current state of the project is the result of a long and painful evolution (usually when things are getting pretty ugly at the end). Also the code hierarchy and the depth of the nesting is limited by the stack heap size, which is also sitting in the bottom 2K area.
I'm nobody to judge the mighty Tek decisions, but I would make the system more simple. One of the complexities in the current FW are due to the fact that there is no common always "visible" ROM area with the interbank calls and routines. Also it would be good to keep the interrupt vectors in RAM (just give the ROM under reset) especially since there is plenty of free space at the end of each 32K ROM partition.
But at the end looks like apart from jumping all over the place the thing is well debugged (the most important) and does its function.

Also thank you for pointing to the screen adjustment, which I have always skipped. I'll start from that and check out again whether I can put the cursor beyond the bound of the graticule after that. Regarding the ripple, what was measured on J119 is well in the spec (almost no ripple at all).

Thanks again,

Re: What is the meaning of ULU? Topic was: 7000-series power-hungry plug-ins

 

Hi Håkan,
I have never come across the 067-0883-99 Universal Load Unit before now. There was only one hit on google but that had no picture or information.
It is not listed in the Calibration Fixtures Catalogs so I have no idea even what it looks like.
Ditto for the 067-0919-99 Programmer. No hits at all on google

Is there any chance you have photos of these units or a manual?

Dennis

-----Original Message-----
From: TekScopes@groups.io [mailto:TekScopes@groups.io] On Behalf Of
zenith5106
Sent: Sunday, October 14, 2018 11:27 AM
On Sun, Oct 14, 2018 at 05:47 PM, Dennis Tillman W7PF wrote:

After your post I re-read the two pages you provided but I still do
not see the headline on the first page that says Universal Load Unit.
I think ULU must be defined on a different page from the two pages you
scanned.
It's not defined on any of the pages. What I meant was that it referred
to the
067-0883-99 / 067-0919-99 which are on the top of the page.
067-0883-99 is a Universal Load Unit and IIRC the 067-0919-99 was used
to program it.
/Håkan



--
Dennis Tillman W7PF
TekScopes Moderator

Re: What is the meaning of ULU? Topic was: 7000-series power-hungry plug-ins

 

On Sun, Oct 14, 2018 at 09:05 PM, Dennis Tillman W7PF wrote:


I have never come across the 067-0883-99 Universal Load Unit before now.

Is there any chance you have photos of these units or a manual?
Dennis, I may be totally out in the blue here ...... but, I have a vague memory that we had one in our field office.
If it is the one I think of I don't think it was ever used or at least not often. Again from a fading memory, it was a
multi channel electronic load and it looked like being built into a 7603 chassi and used a 7k style plugin
for programming using various resitors. I have no idea what happened to it. It may have gone to another field office or
just been scrapped.
Not many of the -99 fixtures, if any, were listed in the Calibration Fixtures Catalog.
I have tried to find info about it in the microfiche but so far not found any.
Maybe some of the guys in the group who worked in the factory could confirm or prove me wrong in what I just said.
/Håkan

Re: 2445A calibration

maxim.vlasov@...
 

Chuck,

Running the emulated machine in the sandbox under Linux is a good approach if there is no notion of the real-time. Linux is not real-time OS with the determinism of nanoseconds.
This brings us to the rhethorical question: HW emulation or SW.

In the gaming industry, while the SW emulators dominate the market, the hard core gamers only choose the HW emulation. For 24xx A5 board the 6802 is very much involved in the control path to the point that the loops are counted and trimmed in the instruction cycles. For example, the whole ADC and DAC multiplexing is delay based and employs the CPU loops in the 2445/55/65 FW.

I would rather use the FPGA with the "cycle accurate" clone 6802 model and the digital logic replica. For example, Xilinx Spartan 6 or Artix 7 (or even Spartan 3) microboards could fit the entire digital subsystem, memory included. But I would keep the analog circuit untouched, since the FW is designed to run with exactly the original HW and uP as its central timing element. Knowing the operating principle of the AD/DA interface the latter could be re-designed using modern and more precise components, or even left unchanged for the reason of full compatibility with the existing service manual calibration/performance checking routines.
IMHO, going HW emulation gives a chance preserving the original timing+functionality and keep all the original controls at its HW/FW/System optimal performance.

But HW emulation doesn't exclude the goodies usually coming with the SW emulation, where the user can also remember/store/load/manipulate the emulated system state to/from SDCard/USB etc. However, adding this kind of services into the HW emulator requires a bit more involvement.

All the above is feasible, but requires a motivation and a time. But it's absolutely feasible, IMHO.

Re: 468 horizontal jitter

Brendan
 

On Sun, Oct 14, 2018 at 10:05 AM, lop pol wrote:


On Sat, Oct 13, 2018 at 08:22 PM, lop pol wrote:


On Sat, Oct 13, 2018 at 10:38 AM, lop pol wrote:

I'm working on my second 468 and I noticed some horizontal jitter.

On Sat, Oct 13, 2018 at 09:51 AM, Jeff Urban wrote:


In case you haven't, it is important to check if it is a centering
problem
or
a triggering problem. Even in auto or free run it is still triggered, so
if
it
is the problem could be all the way back at the ramp generator.

If the trace itself is moving and not the sweep (I hope you got that...)
then
I would more look for hash on a PS line somewhere. Of course that
applies
to
the ramp generator as well but different PS legs of course.
Thanks Jeff.
There is jitter on the A gate out with no input just a free running trace.
I
had to put my 465 in 10x to see it but it does match the jitter on the
468.
So... This looks like its going to be a fun one to figure out.
Q181 on the sweep and z axis board was not seated properly. I have no idea
how
it happened. I have never messed around inside but the base was bent instead
of seated. Looking at the schematic I can't tell how that could cause the
trace jitter though. The base is connected to ground. So far it seems to be
working correctly though. I also installed the EPROMs and images from the
Vintage Tek museum. Boards fit correctly and so far so good.
Well I spoke too soon. Still the same issue. I hate erratic occurring issues.
It has to be something in the horizontal amplifier. XY mode the dot will
sometimes just slightly wiggle horizontally other times it will quickly shoot
2 to 3 divisions horizontally. Its almost unnoticeable until you push the
10x button but is still occurring. I think I have about fully curve traced
every transistor in a 468. I need to get some freeze spray.
I made a video of what is happening the 468 is in XY mode and 10X and the 465 is hooked to the deflection plates and is in 10X and 1V division. The video is kind of junky but you can see what is going on.

http://tinypic.com/player.php?v=2i0885y&s=9

After I made the video I rolled my digital scope over to take a look at it. I let it run for at least an hour and NOTHING the XY dot sat in the middle of the screen and didn't flinch an inch. No trace, wont power on? Thats fine but I hate intermittent stuff.

Re: PG 506 IS IT NECESSARY FOR CALIBRATING?

Albert Otten
 

Jim,

Frequency response is only checked afterwards, adjustment is done using step response. So the 191 or SG503 is not that urgent.
I think the PG502 can be used op to 5V peak-peak (when terminated in 50R) so that can also replace the High Amplitude of the PG506 when frequency compensation is adjusted up to 5 V/div. BTW the PG504 Square wave rise and fall times can be set to very small values so this plugin probably is also very useful.
The FG504 is fine for low frequency trigger adjustments.
Note that for the 191 and CT3 you also need some GR to BNC adapters.
For the checks at small amplitude you will also need some feed through attenuators; the minimum output amplitude of the PG502 is 0.5 V I think.
Some T connectors and some equal length short coax cables are also handy, f.i. in comparing CH1 and CH2 and checking common mode suppression.

I have no experience with a 466, other group members might have additional advice (or corrections...)

Albert

On Sun, Oct 14, 2018 at 08:07 PM, Jim Olson wrote:


Albert,
Thanks for the reply you covered part of the reason I put the query up to see
if what I had was OK to cal the scopes I have or if I should seriously
consider putting the other plug ins up on the ebay and figure a way to get a
PG506. Problem there is they are all in Europe and expensive.
The SG503 is not so hard to get and more affordable watching some on the bay.
I don't have the 191 on hand at this time as the person I got all the stuff
from forgot to include it and haven't been able to drive there and pick it up
it is a 110 mile round trip.
I am assuming that the other modules are OK for my purpose?

Jim O

On October 14, 2018 at 2:05 AM Albert Otten <aodiversen@...
mailto:aodiversen@... > wrote:


On Fri, Oct 12, 2018 at 12:54 AM, Jim Olson wrote:

>

> > This is for the folks here that are a lot more knowledgeable and
experienced
than me with these 4xx series scopes with the testing and
calibrating of them.
I have a number of the TM series plugins plus case for them that I
got really
lucky and got for a great deal. Here's a list of the modules:
DM501, DM502A,
PG501, PG502, PS503A, FG504, Type 187 Time mark Generator, Type
191 Signal
Generator and a CT3, need to get the BNC adapters for it I see
there two types
one straight through and one with a 50ohm terminator should i get
both? I
still need the 50ohm terminators and coax cables. I have 100' of
RG58U just
need to pick up some cable connectors for it.
I also have a good 2215 scope with a P6208 probe.
So my question is do I need the PG506 to really calibrate the
scopes and also
should i get a SG503 to get the higher Mhz output?
Once i get the 466 together I will really need to calibrate it so
do i have
enough cal instruments? and I will be asking for much help too!

Jim O

> Jim,
What you miss in your series of plugins is the possibility of quick DUT
V/div checks, i.e. you miss the Standard Amplitude mode of the PG506.
The PG502 is fast enough (rise time) and flat (pulse top flatness) to
replace the PG506 for waveform checks.

The Type 191 is sufficient to check the frequency response up to 100
MHz, so good enough for a 466. You might be curious how much higher the 466
goes; then of course you need a SG503 or similar.

I use the 50R feed through terminators 011-0049-01. But for a 100 MHz
scope a T-connector with an arbitrary 50R end stop in one leg will do. I
gathered some end stops when the local computer coax network at my work was
dismantled.

Albert



Re: 2445A calibration

Chuck Harris
 

You are mistaken about linux not being a real-time OS. The
usual desktop doesn't need RT, but there are RT extensions
that make it amazingly useful in situations like instrumentation,
and motion.

I use LinuxCNC to drive my CNC milling machine. It simultaneously
drives the X, Y, and Z axes, while calculating paths, acceleration
and deceleration, curves and arcs, all RT.

It can do this while displaying videos, web browsing, playing
MP3's, burning a DVD, and downloading files over the network.

That is the jitter and latency test I use. I am using a puny
little dual core DELL Optiplex 780 that was built in 2005.

Any variance in the timing of the stepper pulses will result
in surface irregularities, and tool wear or damage.

I also use LinuxCNC on a BBB board to run my tiny 3 axis CNC
circuit board milling machine.

There are RT extensions to linux, and, the BBB has an I/O
processor for areas where timing is super critical.... the
DELL Optiplex doesn't.

The counter I mentioned was a HP5370A/B, which is a picosecond
resolution recriprocal frequency counter, and counter timer.

To the extent that a 68B02 can do accurate timing loops, linux
on a BBB can do it better. The emulator is cycle accurate to
the 6800.

The DAC muxing loops would be trivial for a BBB running under
linux.

-Chuck Harris

maxim.vlasov@... wrote:

Chuck,

Running the emulated machine in the sandbox under Linux is a good approach if there is no notion of the real-time. Linux is not real-time OS with the determinism of nanoseconds.
This brings us to the rhethorical question: HW emulation or SW.

In the gaming industry, while the SW emulators dominate the market, the hard core gamers only choose the HW emulation. For 24xx A5 board the 6802 is very much involved in the control path to the point that the loops are counted and trimmed in the instruction cycles. For example, the whole ADC and DAC multiplexing is delay based and employs the CPU loops in the 2445/55/65 FW.

I would rather use the FPGA with the "cycle accurate" clone 6802 model and the digital logic replica. For example, Xilinx Spartan 6 or Artix 7 (or even Spartan 3) microboards could fit the entire digital subsystem, memory included. But I would keep the analog circuit untouched, since the FW is designed to run with exactly the original HW and uP as its central timing element. Knowing the operating principle of the AD/DA interface the latter could be re-designed using modern and more precise components, or even left unchanged for the reason of full compatibility with the existing service manual calibration/performance checking routines.
IMHO, going HW emulation gives a chance preserving the original timing+functionality and keep all the original controls at its HW/FW/System optimal performance.

But HW emulation doesn't exclude the goodies usually coming with the SW emulation, where the user can also remember/store/load/manipulate the emulated system state to/from SDCard/USB etc. However, adding this kind of services into the HW emulator requires a bit more involvement.

All the above is feasible, but requires a motivation and a time. But it's absolutely feasible, IMHO.

Re: 2445A calibration

 

On Sat, Oct 13, 2018 at 03:08 PM, <maxim.vlasov@...> wrote:

Hi Maxim,


Hello Raymond,

I see the comment ;) Well we are not jailbreaking the new iPhone herein with
all the possible and impossible protection switched on ;))

Well, we can start it in a few ways.
I see that you're making progress. I have been using my Logic Analyzers (!) (Tek TLA715, HP16702B) recently in an effort to follow the instruction flow in an 80C186-based low-frequency spectrum analyzer (Stanford Scientific SR760). The out-of sequence/pipeline architecture makes that task a lot more difficult than when I was having fun using my HP 1615A logic analyzer while developing 6802 systems 30 years ago.
Good luck and have fun with this project!

Raymond

Re: Tek 2467B

Robert Calk Jr.
 

On Fri, Oct 12, 2018 at 09:34 AM, Chuck Harris wrote:


Consider what you guys routinely do to your scopes in the name
of repair/restoration/polishing, and then think on the value of
your NVRAM's contents:

1) replace burned out attenuators, and vertical hybrids.
2) replace malfunctioning trigger hybrids.
3) replace power supply capacitors.
4) replace A5 SMD capacitors and DAC reference parts.
5) replace bad CRT's, and CRT anode triplers.
6) swap about A5 boards.
7) replace 4051 MUX's, TL07x opamps, and S-and-H capacitors.
8) twiddle the DAC reference adjustment and the 10V adjustment.

Any one of the above list will likely instantly invalidate some
or all of your calibration constants.
Oh, come on!! So if I change the caps in my Tek 2465A DV, my calibration will be toast??!! You just broke my heart!!

I changed my memory back-up battery with no problems. Anyone can google it and find my article. Right now I am working on writing articles on rebuilding the Siemen's fan motor in my scope to show others that there is nothing to be scared about. It isn't that difficult or scary of a job to do, even for someone that doesn't have my experience in repairing electronics devices.

There has to be a way that I can replace those caps without needing to have my scope calibrated! Please tell me that I'm right, Chuck!!

Re: Tek 2467B

 

Hi Robert,
Chuck is a very smart guy and a very prolific contributor to TekScopes for 15 years. But I always encourage people to challenge our assumptions and the ways we learned to do things. More power to you if you can find a solution to replacing any of these components while still preserving the calibration constants.

Technology evolves. The concept of calibration constants that were stored in PROMs was an improvement over trim pots, variable trimmer caps, tunable cores, etc.

But soon even settable calibration constants were made obsolete by Self Calibrating plugins and mainframes. It takes them anywhere from seconds to a minute to calibrate themselves any time you want to 0.1% accuracy. Compare that to the late 1960s when scopes and plugins required time consuming manual calibration every 6 months to maintain their accuracy to 3%.

So by all means take as much time as you need to find workarounds to the necessity of recalibrating whenever you change any or the things in Chuck's list. Keep us posted on your progress so we can make suggestions and give you encouragement. We may all benefit from what you discover along the way.

Dennis Tillman W7PF

-----Original Message-----
From: Robert Calk Jr.
Sent: Sunday, October 14, 2018 12:12 PM

On Fri, Oct 12, 2018 at 09:34 AM, Chuck Harris wrote:


Consider what you guys routinely do to your scopes in the name of
repair/restoration/polishing, and then think on the value of your
NVRAM's contents:

1) replace burned out attenuators, and vertical hybrids.
2) replace malfunctioning trigger hybrids.
3) replace power supply capacitors.
4) replace A5 SMD capacitors and DAC reference parts.
5) replace bad CRT's, and CRT anode triplers.
6) swap about A5 boards.
7) replace 4051 MUX's, TL07x opamps, and S-and-H capacitors.
8) twiddle the DAC reference adjustment and the 10V adjustment.

Any one of the above list will likely instantly invalidate some or all
of your calibration constants.
Oh, come on!! So if I change the caps in my Tek 2465A DV, my calibration
will be toast??!! You just broke my heart!!

I changed my memory back-up battery with no problems. Anyone can google
it and find my article. Right now I am working on writing articles on
rebuilding the Siemen's fan motor in my scope to show others that there
is nothing to be scared about. It isn't that difficult or scary of a job
to do, even for someone that doesn't have my experience in repairing
electronics devices.

There has to be a way that I can replace those caps without needing to
have my scope calibrated! Please tell me that I'm right, Chuck!!


--
Dennis Tillman W7PF
TekScopes Moderator

Re: Tek 2467B

Robert Calk Jr.
 

Hi Dennis,

I was hoping that I misread it... I'm just a hobbyist. If yall can't find a workaround, then I'm not even going to try. But the caps have to be replaced. I guess I'll just have to buy the equipment and learn how to calibrate my scope myself, or use it for target practice and chalk it up as a loss. It seems like the scope could make allowances or self correct about something as simple as changing a few caps.

I guess I should not have listened to Dave and just went digital from the beginning. I don't understand why changing a few caps would mess up the calibration - it doesn't make any sense to me. But I don't know much about working on scopes.

Re: Tek 2467B

Harvey White
 

On Sun, 14 Oct 2018 19:35:14 -0700, you wrote:

Hi Dennis,

I was hoping that I misread it... I'm just a hobbyist. If yall can't find a workaround, then I'm not even going to try. But the caps have to be replaced. I guess I'll just have to buy the equipment and learn how to calibrate my scope myself, or use it for target practice and chalk it up as a loss. It seems like the scope could make allowances or self correct about something as simple as changing a few caps.
Caps in a bulk supply, once changed, can change the output voltage.
Not necessarily by a lot, but some.

Very few regulators (bulk supply down to regulated supply) are immune
from supply variations.

A very rough example might be (numbers are fictitious) 14.5 volt bulk
supply for a 12 volt regulator changing to 15.0. When adjusted to 12
volts output (if you can) at 14.5 volts, the output may change to 12.2
at 15.0 volts.

Now if the 15 volts is taken to be a standard and a reference, then
your reference is now 0.2 volts off.

Even if it is further regulated by another circuit, *that* circuit may
behave similarly.

So ultimately, it depends on how close you need to be.



I guess I should not have listened to Dave and just went digital from the beginning. I don't understand why changing a few caps would mess up the calibration - it doesn't make any sense to me. But I don't know much about working on scopes.

analog scopes are quite useful, less complicated, and have a very
different set of advantages.

In a digital scope, you sample, store, then wait, then sample, then
store, then wait. Doing this enough reconstructs the waveform.

Now what you don't know is what happens to the waveform when you do
that store, and wait. all you know is what happens when you store.
Even that is a bit of an average.

Now because of something called the Nyquist theorem, a storage scope
can lie to you. If a frequency is the same as the rate you sample,
then that's one thing. Now take the same frequency, and multiply it
by two. When you sample, and then reconstruct the waveform (give it a
try with graph paper), you will get the exact same waveform
reconstructed, even though it is twice the frequency and the scope
CANNOT tell the difference.

This ignores bandwidth limits, but it's an illustration. To keep this
from being a problem, digital scopes have very stiff bandwidth limits,
which distorts the waveforms to an extent.

Analog scopes don't do this. What you see is what you get. If you
feed an analog scope rated for 100 Mhz something above it, you'll get
less sensitivity, but no frequency aliasing (which happens in the
digital scope).

Now, where are digital scopes better?

(imho)

1) where you need to extract data for a computer
2) where you need to do math on the data
3) where the repetition rate is slow, you can get a very nice and
stable waveform that's bright. (yes, storage scopes get around this,
but only so much)
4) where you need a memory of what you did last to compare with
5) where you don't have a problem with what happens between samples.
knowing this stuff is often half the problem

I have both, I use both.

Harvey




Re: Tek 2467B

Robert Calk Jr.
 

Hi Harvey,

I have read some of that stuff before, but this is the first time that I heard changing a few caps could mess up the calibration. My pulse rate is back to normal now. lol

I really love this scope and it has the Television/Video option that might come in handy some day, among other options. Plus it looks brand new inside; like it has spent 90% of it's life boxed up in a non-smokers closet. I'm going to study the schematics and see if I can figure something out. If there is a way, I'll find it.

I was planning on getting a digital scope anyway. I'll just have to get one sooner than I had planned because I believe it's better to have both also. Thanks for the info.

Re: Tek 2467B

Chuck Harris
 

Hi Robert,

Some of the items I mention are pretty drastic in their
affect, and others not so much.

Consider a scope that was a little long in the tooth when
it was last calibrated. If a couple of the capacitors in
one of the inverter's filters were old, and had high ESR's,
the ripple will get through, and averaging the ripple waveform
will present a different average DC voltage than it would
if the capacitors were in prime shape. Your DVM responds
to averages.

Now, being a good tekscopes citizen, you replace out the
capacitors, you will now have shifted the DC outputs by
whatever amount the original ripple was fooling the DVM used
in the last calibration.

Triplers? Well, the only regulated voltages in the HV section
are the cathode voltages. It is assumed that the tripler will
produce the correct CRT anode voltage when presented with the
correct regulated cathode voltage from the HV transformer.

But suppose that it doesn't. Triplers are capable of only
sourcing very tiny currents, and variations in internal leakage
resistance of capacitors, bleeders, and diodes are normal.

Variations in the ultimate CRT anode voltage will cause the gain
of the vertical and horizontal deflection plates to change
somewhat... an effect that is corrected by calibration...

Burn out a filament in your CRT, and put in a new one? Each
CRT has subtle differences in internal structure resulting in
different deflection points, gains, geometries, speeds...etc.

Calibration accounts for these differences. Your new CRT won't
be the same as your old CRT prior to burning out its filament.

OOPSIE! I burned out my CH1 attenuator. Let's get a new one
from ebay. When you replace it, it will seem to work, but each
attenuator section is going to be slightly different from the
original CH1 attenuator, and the calibration constants reflect
the original attenuator, not the new one.

There is a reason the manual suggests that calibration should
be done after major repairs. Each of the items I mentioned is
a major repair that is often treated like it wasn't.

The caps are probably the least critical of the items I listed,
but they can affect calibration. Simply adjusting the 10V ref
pot is not going to save the old calibration because you don't
know what the voltages were like when it was last calibrated.

Errors in the last guy's DVM, coupled with errors in your DVM
could easily be enough to push the scope out of calibration.

I am not trying to cause panic, I just want to shake things up
a bit and make people think.

-Chuck Harris

Robert Calk Jr. wrote:

On Fri, Oct 12, 2018 at 09:34 AM, Chuck Harris wrote:


Consider what you guys routinely do to your scopes in the name
of repair/restoration/polishing, and then think on the value of
your NVRAM's contents:

1) replace burned out attenuators, and vertical hybrids.
2) replace malfunctioning trigger hybrids.
3) replace power supply capacitors.
4) replace A5 SMD capacitors and DAC reference parts.
5) replace bad CRT's, and CRT anode triplers.
6) swap about A5 boards.
7) replace 4051 MUX's, TL07x opamps, and S-and-H capacitors.
8) twiddle the DAC reference adjustment and the 10V adjustment.

Any one of the above list will likely instantly invalidate some
or all of your calibration constants.
Oh, come on!! So if I change the caps in my Tek 2465A DV, my calibration will be toast??!! You just broke my heart!!

I changed my memory back-up battery with no problems. Anyone can google it and find my article. Right now I am working on writing articles on rebuilding the Siemen's fan motor in my scope to show others that there is nothing to be scared about. It isn't that difficult or scary of a job to do, even for someone that doesn't have my experience in repairing electronics devices.

There has to be a way that I can replace those caps without needing to have my scope calibrated! Please tell me that I'm right, Chuck!!