Topics

2445A calibration

maxim.vlasov@...
 

Dear forum members,

I need your help in calibrating 2445A o-scope after it was repaired
(re-capped). Currently it says FAIL04 (bad SRAM content). I have the
TG501 time mark generator (2 ns option is broken however) and HP 8175A
AW programmable generator. I'd like to perform the calibration in steps
and therefore here is my first question.

When having FAIL04 is it possible to do the CALxx steps one by one? I
would like to start from CAL01 and then store its settings and then
switch off the scope and bring it into another place for the vertical
calibration. Is it possible?

The service manual is sometimes very confusing. When I was doing the
first round of steps on CALxx I still don't know if I did it right. In
the first step of CAL01 the manual tells to do the superimpose and
matching of the B-sweep delayed marks over the 6th graticule (screen
center). However for all the other steps nothing is mentioned. I have
assumed that when aligning the B-sweep (zoomed) marks I have to join
them on the vertical axis in the screen center with the precision of 1.2
or 0.2 graticule. Is this correct or not at all?

Also initially since the gain x1 and x10 were not set correctly the time
marks would not align with the graticules. So, I had to run the CAL01
calibration a few times.

I'll try to upload the images, but somehow whatever I try to do online
on groups.io doesn't seem to work.

Thank you for your help,

All the Best,

Maxim

Max Vlasov
 

Hello,

I'm still struggling with 2445A calibration. Good news - I have completely repaired and calibrated TG501. Bad news - I don't have neither the tunnel diode pulser nor the PG506 and I don't really know whether I can do w/o these units.
For the vertical calibration I pretty much only need a some sort of AWG, IMHO. I do possess HP8175A with the option 002 (analog 50 MS/s AWG) which can drive up to 16V into 50 Ohm load.

I assume that the standard amplitude signal required for the 2445a is 1KHz positive polarity (DC coupled) with selectable amplitude.

Yet one more problem to be solved. I had to change the channel 1 hybrid and its from a different scope : 2465a. The seller told me that the hybrid will certainly work after recalibration. However, the trigger AUTO LVL function doesn't find the signal level on the channel 1 (there is a large offset in the auto level measurement). Otherwise the channel 1 works fine.
Does the CAL03 TRIGGERING procedure takes care of AUTO LVL trigger function and adjusts the offset in the ADC algorithm?

If someone could also explain for CAL01 where to overlay the B delay sweep signals (in the center of the screen over the 6th vertical line or anywhere on the screen for all the steps but step 1), this would be awesome.

Thank you,

Best regards,

Maxim

Max Vlasov
 

Hi,

Also have anyone transferred the 2445B NVRAM content into the 2445A SRAM?
There is a 2445B NVRAM file kindly uploaded by Mitchell Kobierowski in this forum. Technically it's possible to upload it into the 2445A SRAM (connect the voltage source via the current limiter while being in the ROM programmer tester and then bring it back to the A5 board). I hope it should fix the FAIL04 prompt and restore the calibration data integrity. Since my 2445a has option 05 (video triggering), I would have to recalibrate it later on together with CAL01, CAL02, CAL03.... but one at a time.

I wonder if doing CAL one at a time is possible at all..

Thank you,

All the Best,

Maxim

joe price
 

I have a scope that fails 04 my need calibration and can't I find the 2445B NVRAM in files section can someone point me in the correct direction?

Joe

Max Vlasov
 

Hello Joe,

Here is the file I've mentioned in my previous post:

https://groups.io/g/TekScopes/files/2445B%20B060XXX%203-15-2015.zip

kindly uploaded by Mitchell Kobierowski,

However in my case I'm much further from the full working o-scope... Hope that someone would help me on this forum,

Good luck,

Maxim

Chuck Harris
 

You can replace the CMOS RAM and battery with a NVRAM
(or FRAM) with a little work.

You have to remove the battery, cover it over with a label
describing the modification (because you are nice).

You also have to provide a voltage to the CPU's battery
detect circuit to fool it into thinking it has a good battery.

After doing that, you program the new NVRAM with the old
NVRAM's contents, and plug it in.

However, your calibration will be wrong for your scope, as
the NVRAM's files are unique to the scope that created it.

I wouldn't do the above, as the battery/cmos RAM combination
lasts a lot longer than an NVRAM, and cribbed calibration data
is worse than no calibration data at all.

When you plug a new NVRAM into your scope, and turn the power
on for the first time, the firmware installs a default set of
calibration constants into the constant area, and sets the
uncalibrated flag so that the screen has a row of dots along
the bottom, to announce to the world the scope is uncalibrated.

The only logical reason to program the NVRAM with someone else's
data is if you want to fool a buyer into thinking the scope is
calibrated.

-Chuck Harris



Max Vlasov via Groups.Io wrote:

Hi,

Also have anyone transferred the 2445B NVRAM content into the 2445A SRAM?
There is a 2445B NVRAM file kindly uploaded by Mitchell Kobierowski in this forum. Technically it's possible to upload it into the 2445A SRAM (connect the voltage source via the current limiter while being in the ROM programmer tester and then bring it back to the A5 board). I hope it should fix the FAIL04 prompt and restore the calibration data integrity. Since my 2445a has option 05 (video triggering), I would have to recalibrate it later on together with CAL01, CAL02, CAL03.... but one at a time.

I wonder if doing CAL one at a time is possible at all..

Thank you,

All the Best,

Maxim

Max Vlasov
 

Hello Chuck,

Thank you for answering.

My 2445a repair/calibration project is the life long (almost 10 years, almost no time). I should of given up on it long before, but it's just not an approach of an engineer to give up on something which almost already works (put lots of hours already, but 2445a in my setup will be more of an artifact). I just wanted to understand how serious is the data integrity failure (FAIL04) and whether I can do the calibration in steps powering up the o-scope every time for the new CALxx routine?

I simply assumed that if the data integrity is OK then I can do the CAL routines selectively (therefore the idea was to get the calibration data w/o errors and then do one calibration routine by another). Otherwise I have to go from the beginning though the end non-stop otherwise nothing will be stored... So far whatever I have calibrated wasn't taken into account by o-scope since I didn't have the right set of equipment to perform the CAL ALL. For example I did perform CAL01 and CAL02 separately, then powered off the scope then I'm back to the same problems. Like it was not simply stored.

Have you managed to calibrate the 24xx scope in separate CAL sessions by having initially FAIL04?

Also do you know which CAL routing will contribute to the AUTO LVL trigger function. In my case channel 1 with the replaced hybrid has the offset too high on AUTO LVL, so trigger level isn't found automatically.

Thank you again and all the Best,

Maxim

Chuck Harris
 

Hi Maxim,

When you complete any calibration routine (eg. CALXX), the scope
will have stored the calibration constants, and when you leave the
calibration menu, it will recalculate the checksum for the block
of all calibration data. It will *not* recalculate the parity for
the individual calibration constants... only those that it
affected by calibration.

Do not waste your time with calibration until you have fixed the
cause of the Fail04. That means replacing the NVRAM or battery.

If you decide that you really want it calibrated, contact me off
group.

-Chuck Harris

Max Vlasov via Groups.Io wrote:

Hello Chuck,

Thank you for answering.

My 2445a repair/calibration project is the life long (almost 10 years, almost no time). I should of given up on it long before, but it's just not an approach of an engineer to give up on something which almost already works (put lots of hours already, but 2445a in my setup will be more of an artifact). I just wanted to understand how serious is the data integrity failure (FAIL04) and whether I can do the calibration in steps powering up the o-scope every time for the new CALxx routine?

I simply assumed that if the data integrity is OK then I can do the CAL routines selectively (therefore the idea was to get the calibration data w/o errors and then do one calibration routine by another). Otherwise I have to go from the beginning though the end non-stop otherwise nothing will be stored... So far whatever I have calibrated wasn't taken into account by o-scope since I didn't have the right set of equipment to perform the CAL ALL. For example I did perform CAL01 and CAL02 separately, then powered off the scope then I'm back to the same problems. Like it was not simply stored.

Have you managed to calibrate the 24xx scope in separate CAL sessions by having initially FAIL04?

Also do you know which CAL routing will contribute to the AUTO LVL trigger function. In my case channel 1 with the replaced hybrid has the offset too high on AUTO LVL, so trigger level isn't found automatically.

Thank you again and all the Best,

Maxim

maxim.vlasov@...
 

Thank you very much for the detailed answer, Chuck,

So tihs means that actually I can do the calibration in separate steps by running the CALxx procedure from the beginning to the end. This will fix then the FAIL04 problem once all the calibration constants are updated and the checksum is OK for all of them.

Sending the scope from Europe to the US and back is more than 500 USD (including the taxes, customs etc) so its beyond the economical repairs. The battery is good, it's just our lab technician pulled out all the chips in the failed attempt to quickly repair the scope by reseating the ICs and clearly lost all the settings once the memory has lost its power. Then the scope was thrown to the company auction and I scored it for 150$. After that I had to replace the U200 hybrid, opamps around, etc. Looks like someone decided to probe HV with this scope. After recapping the power supply I decided to calibrate it, but I didn't have by that time the time mark generator. So now, I'm better prepared and can go slowly running the CAL routines one by one.

I'm going to re-run tonight the CAL01. Could you, please, tell me whether I always have to join the zoomed time marks around the Y axis in the middle of the screen. Sometimes two zoomed traces disagree more in the middle of the screen, The service manual only precises that the traces must be joined around the division 6 (screen center) only for the step 1. I assume that in order to align the signal timing they have to be precisely aligned at x10 zoom around the screen center (vertical axis). So I'll try re-running the CAL01 with the signal alignment around the screen center.

Thank you again and all the Best,

Maxim

Chuck Harris
 

Yes, you can calibrate only the part you need, without
having to complete the entire routine. The only restrictions
are that some calibrations rely on other calibrations having
been done, or being good already.

For example, power supply, and DAC accuracy adjustments affect
everything. The CRT adjustments are important pretty early on,
as changing the geometry will change the CRT's sensitivity.

It is not essential that the overlaid waveforms be in the center
of the screen. The important part is they be overlaid, and they
be the specified pulses (eg, 2 and 10, 4 and 28, 3 and 19...) pulse.

Also, you need a bench scope to do the "B" timebase part of the
horizontal calibration. You can also eyeball them, ala the 2465's
instructions. The extra scope allows an enhancement that is somewhat
important to the measurement routines.

-Chuck Harris

maxim.vlasov@... wrote:

Thank you very much for the detailed answer, Chuck,

So tihs means that actually I can do the calibration in separate steps by running the CALxx procedure from the beginning to the end. This will fix then the FAIL04 problem once all the calibration constants are updated and the checksum is OK for all of them.

Sending the scope from Europe to the US and back is more than 500 USD (including the taxes, customs etc) so its beyond the economical repairs. The battery is good, it's just our lab technician pulled out all the chips in the failed attempt to quickly repair the scope by reseating the ICs and clearly lost all the settings once the memory has lost its power. Then the scope was thrown to the company auction and I scored it for 150$. After that I had to replace the U200 hybrid, opamps around, etc. Looks like someone decided to probe HV with this scope. After recapping the power supply I decided to calibrate it, but I didn't have by that time the time mark generator. So now, I'm better prepared and can go slowly running the CAL routines one by one.

I'm going to re-run tonight the CAL01. Could you, please, tell me whether I always have to join the zoomed time marks around the Y axis in the middle of the screen. Sometimes two zoomed traces disagree more in the middle of the screen, The service manual only precises that the traces must be joined around the division 6 (screen center) only for the step 1. I assume that in order to align the signal timing they have to be precisely aligned at x10 zoom around the screen center (vertical axis). So I'll try re-running the CAL01 with the signal alignment around the screen center.

Thank you again and all the Best,

Maxim

maxim.vlasov@...
 

Chuck,

Thank you for your help!

First I did the power supply 10V and DAC adjustment (performed the checks as per the manual). Then did the time marks (I'll upload tomorrow the photos). But then between step 17 and 23 I have failed miserably with all my digital scopes (Tek DSO 3054, LeCroy 9410, Agilent) which couldn't perform the stroboscopic display (the holdoff is not short enough), so I could only superimpose the time marks on the bench scope for the slower frequencies (had to switch off all the fancy features but still can't achieve fast re-triggering). Tomorrow I'll take my very old Tek 7834 eating dust in the celler and will repeat the steps 17 to 23 again in CAL01. Otherwise it is much clear now. Even without correcly overlaying time marks on the bench scope this time the CAL01 have completed and when 2445 is booted I can see that the horizontal timing is accurate.
Also it seems like even the semi-automated calibration procedure is very labor intensive.

Thanks for clarifying how to deal with the overlays. Looks like I can even superimpose the time marks on the center of the screen within the specified accuracy (0.2 square for 2/10 or 1.2 square for 4/28). Seems like I shouldn't get rid of my 7000 series scopes (I just was about to do that) since till this evening I couldn't find the good use for them ;)

Maxim

Max Vlasov
 

Hello Chuck,

I have posted the photos from the CAL01 to the forum:
https://groups.io/g/TekScopes/album?id=74577

Could you, please, have a look and tell me if you see something not right?

I'll have to re-do the CAL01 with 7603 or 7834 scope as a bench scope this evening.

Thank you again,

Maxim

Chuck Harris
 

Basically, it looks as it should, but I think there is a problem
with the amount being displayed. You should have a marker on
each of the 11 divisions on the screen. I don't see them all
the way to the right... Perhaps you need to adjust the horizontal
position control so that the first marker is aligned with the
left most graticule line?

If that is the case, you may have the wrong markers illuminated.

-Chuck Harris

Max Vlasov via Groups.Io wrote:

Hello Chuck,

I have posted the photos from the CAL01 to the forum:
https://groups.io/g/TekScopes/album?id=74577

Could you, please, have a look and tell me if you see something not right?

I'll have to re-do the CAL01 with 7603 or 7834 scope as a bench scope this evening.

Thank you again,

Maxim



maxim.vlasov@...
 

Hello Chuck,

Thank you for reviewing the traces. I'll try to count the markers in B-sweep display to make sure that the markers to be superimposed are the right ones.

Will go through the calibration again. 7834 yesterday showed its age. Had to work out the rotary switches on 7A18 and 7B53 to make it working again... Hope to finish the calibration tonight.

Looks like in order to calibrate an analog scope you need another analog scope ;)

All the Best,

Maxim

Max Vlasov
 

Hi Chuck,

Regarding the CAL01, I've followed your suggestion and counted the markers on the zoomed screen. The numbers do perfectly match. IMHO what looks like the rate problem is the long fall time of the time mark signal. Therefore the illuminated markers are not all the way down to the bottom.

Also I've run the calibration from step 17 to 23. I used both the Tektronix 7834 and LeCroy 9410 as the bench scope.

https://groups.io/g/TekScopes/photo/74577/14?p=Name,,,20,1,0,0

On LeCroy finally it was possible to reconstruct two signals to be superimposed. What was a bit shocking is that even on Tektronix 7834 using 7B53 timebase the 2445a stroboscopic display setup wasn't triggering that well (in some modes perfectly fine, in some not). So finally I've resorted to LeCroy, since its trigger was very stable.
Also since 7A18 amplifier has no 50 Ohm input termination I had to put the 50 Ohm terminators on a Tee connector just on the face of 7A18 (had no flow through terminated plugs). I looked and couldn't see any measurable reflections generated by such a setup...

So I wonder what's wrong and why I can't always see two signals to be superimposed (have to manipulate with the trigger level on 7B53 when changing the time mark rate).

Do you have a stable trigger with two displayed signals in all the steps from 17 to 23? What time mark generator are you using?

Thank you,

Maxim

Chuck Harris
 

The reason I mentioned the number of markers is your photos
did not show the full count.

The display of the two visually overlaid pulses requires that
the sweep is capable of being triggered on successive B-Gate
signal edges. It sometimes requires an adjustment of the
sweep "hold-off" control.

In answer to your question, I do get a stable display with
some plugins on the 7000 series, but not others. I suspect
that it is mostly a problem with the plugin's needing repair
or alignment. I always get a stable display when I use the
2465... but then, my 2465's are always well calibrated, given
that I calibrate 2465's for hire.

If you can align two successive traces from triggering on the
B-Gate signal on the LaCroy, it should be fine. The analog
scope is just more automatic, allowing you to use the visible
trace as an immediate form of feedback.

There is nothing wrong with using a "T" connector and a 50 ohm
load as a thru terminator.

-Chuck Harris

Max Vlasov via Groups.Io wrote:

Hi Chuck,

Regarding the CAL01, I've followed your suggestion and counted the markers on the zoomed screen. The numbers do perfectly match. IMHO what looks like the rate problem is the long fall time of the time mark signal. Therefore the illuminated markers are not all the way down to the bottom.

Also I've run the calibration from step 17 to 23. I used both the Tektronix 7834 and LeCroy 9410 as the bench scope.

https://groups.io/g/TekScopes/photo/74577/14?p=Name,,,20,1,0,0

On LeCroy finally it was possible to reconstruct two signals to be superimposed. What was a bit shocking is that even on Tektronix 7834 using 7B53 timebase the 2445a stroboscopic display setup wasn't triggering that well (in some modes perfectly fine, in some not). So finally I've resorted to LeCroy, since its trigger was very stable.
Also since 7A18 amplifier has no 50 Ohm input termination I had to put the 50 Ohm terminators on a Tee connector just on the face of 7A18 (had no flow through terminated plugs). I looked and couldn't see any measurable reflections generated by such a setup...

So I wonder what's wrong and why I can't always see two signals to be superimposed (have to manipulate with the trigger level on 7B53 when changing the time mark rate).

Do you have a stable trigger with two displayed signals in all the steps from 17 to 23? What time mark generator are you using?

Thank you,

Maxim

Max Vlasov
 

Thank you Chuck,

Yesterday I've tried to get through CAL02, but I had LIMIT error problem. The DC balance walked automatically without generating any error message.
Then I've aligned two points for 10 horizontal divisions like this:
https://groups.io/g/TekScopes/photo/74577/16?p=Name,,,20,1,0,0
But then when the scope walked through the step 113 (I could see the beam jumping diagonally from the left bottom to the right top corner) it spit out LIMIT 113. I have pressed again on TRIGGER COUPLING button and it jumped immediately to the step 114. Then again, the story repeated with LIMIT error.

I wonder I the problem is with the hardware. Scope passes all the tests and generates FAIL04 3 (bad checksum). Otherwise the vertical amplifiers on all channels measure the input signal correctly (uncalibrated). The only problem I see is with AUTO LVL function (doesn't work for the channel 1), but I assume that it's addressed in the CAL03.

If I have the LIMIT error what do I have to do in the CAL02? The 2445a manual doesn't state what to do if the LIMIT ERROR is encountered (here is a nice re-written calibration procedure for 2465B):
https://wiki.epfl.ch/carplat/documents/PDF/Tektronix_2465B_calibration.pdf
If I advance the step after having LIMIT error is it assumed that the current step has failed and no calibration data is registered?

Thanks again,

Best regards,

Maxim

Chuck Harris
 

Your Dots as shown in the picture seem to be awfully
messy. It may just be the photo, but they look to me
as thought the astigmatism is not as well adjusted as
it could be.

Limit in CAL02 usually means that your voltage source
isn't accurate, or is on the wrong setting. It can also
mean that your vertical preamp, or the attenuator section
has a failure.

Also, the instructions are not clear in this area. When
you set the two dots to 10 divisions, you are using the
vertical amplifier in the UNCAL position. In subsequent
steps, you need it to be back in the CAL position. The
instructions don't tell you when to switch it back to CAL.

I put a note in my instruction cheat sheet to remind me
to do that before I acknowledge the step that needed the
UNCAL position, and go to the next step.

If you do the steps over again, and they were correctly
set previously, you will find that the dots are exactly
positioned without using the UNCAL position...

Limit errors get registered, and will show up as UNCAL
when you restart the scope.

-Chuck Harris

Max Vlasov via Groups.Io wrote:

Thank you Chuck,

Yesterday I've tried to get through CAL02, but I had LIMIT error problem. The DC balance walked automatically without generating any error message.
Then I've aligned two points for 10 horizontal divisions like this:
https://groups.io/g/TekScopes/photo/74577/16?p=Name,,,20,1,0,0
But then when the scope walked through the step 113 (I could see the beam jumping diagonally from the left bottom to the right top corner) it spit out LIMIT 113. I have pressed again on TRIGGER COUPLING button and it jumped immediately to the step 114. Then again, the story repeated with LIMIT error.

I wonder I the problem is with the hardware. Scope passes all the tests and generates FAIL04 3 (bad checksum). Otherwise the vertical amplifiers on all channels measure the input signal correctly (uncalibrated). The only problem I see is with AUTO LVL function (doesn't work for the channel 1), but I assume that it's addressed in the CAL03.

If I have the LIMIT error what do I have to do in the CAL02? The 2445a manual doesn't state what to do if the LIMIT ERROR is encountered (here is a nice re-written calibration procedure for 2465B):
https://wiki.epfl.ch/carplat/documents/PDF/Tektronix_2465B_calibration.pdf
If I advance the step after having LIMIT error is it assumed that the current step has failed and no calibration data is registered?

Thanks again,

Best regards,

Maxim



Max Vlasov
 

Chuck,

Thank you for checking the photo. There were 2 small problems: the photo is overexposed but also as you noticed the dots are not round, but rather oblong. The generator output doesn't seem to be noisy, but I'll try replacing it by another AWG and run the test again.

The main repair was made to the channel 1. But I don't really see any difference in terms of noise, offset etc when comparing the old "good" channels 2/3/4 and the repaired channel 1. After the repairs the channel 1 I received a hybrid U100 not from 2445a but from 2465a. Would it be the problem I don't really know.
I'll try advancing the test to the CHANNEL 2 calibration and see whether I'll have the same LIMIT error.

Regarding CAL/UNCAL state - this is all new to me. Are you referring to the R639 state, meaning that on the first go R638/R639 aren't set properly, so the state is UNCAL? But after running the first time through CAL02 on the second run the state is CAL, since the resistors are supposed to be set?

Thank you and all the Best,

Maxim

Chuck Harris
 

Hi Maxim,

The UNCAL/ CAL refer to the CH1 VAR pot on the CH1
Attenuator. I have seen them called various names,
but in the step where you are to make the two dots
line up with the outer graticule lines, you do so
by adjusting the pot on the CH1 attenuator out
of the click and into the VAR position.

Page 5-14, step g tells you to adjust the VOLTS/DIV VAR
control to position the dots...

After finishing that adjustment, you press the upper
trigger coupling switch to advance.

You must then put the VOLTS/DIV VAR control back into
the CAL position! The instructions do not mention this.

The Hybrid could be the problem, but input overload usually
destroys one or more thin film resistors in the attenuator
section of the vertical preamp... It all depends on what
position the ATTEN switch was in when the overload occurred.

To test for that failure, you have to check all positions
of the attenuator, and AC/DC/50/GND.... the calibration
routine does.

Most times, however, the out of range failure happens
because you switched the calibrator's voltage/div switch
wrongly... or in the case of the PG506, the switch contacts
didn't switch correctly, and the PG506 is putting out the
wrong voltage. When I use the PG506, I always wiggle the
switch CCW and CW a little bit, after changing v/div settings,
to settle the switch cams.

I don't have to do that with the CG5011.

-Chuck Harris



Max Vlasov via Groups.Io wrote:

Chuck,

Thank you for checking the photo. There were 2 small problems: the photo is overexposed but also as you noticed the dots are not round, but rather oblong. The generator output doesn't seem to be noisy, but I'll try replacing it by another AWG and run the test again.

The main repair was made to the channel 1. But I don't really see any difference in terms of noise, offset etc when comparing the old "good" channels 2/3/4 and the repaired channel 1. After the repairs the channel 1 I received a hybrid U100 not from 2445a but from 2465a. Would it be the problem I don't really know.
I'll try advancing the test to the CHANNEL 2 calibration and see whether I'll have the same LIMIT error.

Regarding CAL/UNCAL state - this is all new to me. Are you referring to the R639 state, meaning that on the first go R638/R639 aren't set properly, so the state is UNCAL? But after running the first time through CAL02 on the second run the state is CAL, since the resistors are supposed to be set?

Thank you and all the Best,

Maxim