Date   
Re: Almost finished 468 and parts needed.

Reed Dickinson
 

To: The Infinite Penguin:

Please contact me off line as I have the 468 parts you are seeking.

Reed Dickinson
reed714@...

On 10/11/2018 2:16 PM, lop pol via Groups.Io wrote:
The 468 I have been working on is almost done. https://groups.io/g/TekScopes/album?id=75300
So far it seems to work really well. I ordered a set of EPROMs from the VIntage Tek museum $40. Way cheaper than a programmer which I doubt I would use all too much. I think I'm going to use the Mostek that are in there until they fail. I'll be ready to pop the new ones in if need be.
I am wondering if anyone here can help me out with a few parts? I need the knob for the cursor adjustment and a set of good feet.


Re: Almost finished 468 and parts needed.

Jeff Davis
 

I can probably help you out with feet for your 468. Check eBay or my website www.n0dy.com<http://www.n0dy.com>.

[https://www.n0dy.com/wp-content/uploads/2017/12/B014518-3985-Edit-1.jpg]<http://www.n0dy.com/>

- N0DY Electronics<http://www.n0dy.com/>
www.n0dy.com
Tektronix oscilloscopes and accessories, home automation devices




Regards,
Jeff


________________________________
From: TekScopes@groups.io <TekScopes@groups.io> on behalf of lop pol via Groups.Io <the_infinite_penguin=yahoo.com@groups.io>
Sent: Thursday, October 11, 2018 2:16 PM
To: TekScopes@groups.io
Subject: [TekScopes] Almost finished 468 and parts needed.

The 468 I have been working on is almost done. https://groups.io/g/TekScopes/album?id=75300
So far it seems to work really well. I ordered a set of EPROMs from the VIntage Tek museum $40. Way cheaper than a programmer which I doubt I would use all too much. I think I'm going to use the Mostek that are in there until they fail. I'll be ready to pop the new ones in if need be.
I am wondering if anyone here can help me out with a few parts? I need the knob for the cursor adjustment and a set of good feet.

PG 506 IS IT NECESSARY FOR CALIBRATING?

Jim Olson <v_12eng@...>
 

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

Tek 213

oliver johnson
 

I have put batteries into my tek 213 , i used 8 1200mah aa batteries series parallel to get my 2.4 volts . I was unsure of the capacity of the original d cells but i figured 4800mah should be more than enough , that being said scope worked fine and appeared to charge ok as well .

I charged for about 8 or so hours but when i powered scope nothing happned , i tried with plugging into outlet still nothing , so i opened up scope and checked battery voltage . As it turns out one group of cells were charged while the others were very low , i tried a different battery ,it powers on but not for long . I did also notice that when batteried were left in scope they would die in a day or so just being off .

Is it possible that i could have a bad cap that drained the scope even when powered off???, or is there another reason why this has happened . Now i can only get scope to power with Ac plug working , any thoughts ????

Re: 2445A calibration

maxim.vlasov@...
 

Update:
Measured again the test points, looks like I haven't correctly stated the voltage:
TP2421=-1.2524V
TP2420 = 1.3639V

By looking at the schematics R2422=14.2K and R2421=13.0K we have:

TP2420=-TP2421*R2422/R2421 = 1.368V (1.3639V measured).

Not exactly off target but I'll try replacing U2420A after re-measuring the real R2422 and R2421 resistor values. U2521 is the next candidate for the replacement.

Re: Tek 485 - does anyone recognise these 485 power supply transistor markings?

Lorenzo
 

Thanks Andre. If you had a spare 151-0324-00 lying around it would be good to have one available. I've started to compare how the -9V and -5V supplies are behaving as the first appears to be working fine and the second is not and they're not very dissimilar subcircuits and use the same op-amp. I hope that will shed some more light. Biggest challenge is finding enough time to do the diagnostics.

Re: 2445A calibration

Chuck Harris
 

Hi Maxim,

At no point are you actually measuring the reference. The reference
is really a current that is fed to the DAC. What you are measuring
is the DAC output, and it ranges from -1.25'ish V to +1.25'ish V.

It isn't really important what it is, just that it has 2.500V between
the two ends, and it is roughly centered.

Remember, you didn't get exactly +/-1.25V, for the end points, so you
should not expect the center to be exactly 0.0V. It should be half way
between the two end values you measured at the CCW and CW positions of
the DELTA knob.

-Chuck Harris

maxim.vlasov@... wrote:

Chuck,

Now I understand the DAC calibration routine. The full voltage scale from 0x000 to 0xFFF is 2.500V. My concern is that the DAC reference isn't centered (when I bring the DELTA cursor to the middle graticule division, J119 pin 13 is at 0.006V instead of 0.0V). And it seems that it should be close to the center (the other ppl managed passing CAL02 & CAL03 w/o LIMIT with DAC center around +/-0.001V).
In my 2445a TP2421 measures -0.12424V (very stable in a few hours). TP2420 measures 1.3639V. I wonder what level should be there. In the schematics it's stated -0.125V and +1.36V? At least I see that the negative checkpoint is higher than expected.

I could only assume that the reason of LIMIT is the underperforming DAC and/or demultiplexer and/or sample&hold or repeater...

I'll start from the demultiplexers and opamps, since they could develop some more leakage.

Thank you,

Best regards,

Maxim

155-0004-01 Tektronix Parts

Tom Bowers
 

Hi, I'm a new member of this group. My name is Tom Bowers. I have been a Tektronix fan since the late 1960s. I own an assortment of instruments and run an electronics engineering and fabrication shop from my home. I am looking for a readout decoder for a 576 Curve Tracer, part number 155-0004-01. I see a post from about a year ago, but can't figure out how to get in touch with the person mentioned (Steve Clark). If someone can provide contact info, or knows of a source for this part, I would appreciate the help. Thanks, Tom

Re: 155-0004-01 Tektronix Parts

Craig Sawyers
 

Hi, I'm a new member of this group. My name is Tom Bowers. I have been a Tektronix fan since the
late 1960s. I own an assortment of instruments and run an electronics engineering and fabrication
shop from my home. I am looking for a readout decoder for a 576 Curve Tracer, part number 155-0004-
01. I see a post from about a year ago, but can't figure out how to get in touch with the person
mentioned (Steve Clark). If someone can provide contact info, or knows of a source for this part, I
would appreciate the help. Thanks, Tom
Welcome to the group! Is that device missing entirely, or is the one in your 576 dead?

Craig

Re: 7000-series power-hungry plug-ins

Robert Dixon
 

As I am currently repairing a 7904 (not-A), I'm interested in the PSU limits and also the power needs of various plugins.

Is there a place where both of these are listed, ideally by supply rail?
e.g. 7904 +15V rail supplies 2A, 7S12 +15V needs 3A ?

Thanks,

Robert Dixon

On 12 Oct 2018, at 06:01, Chuck Harris <cfharris@...> wrote:

If I remember correctly, the original 7603 had a
wimpy linear supply, and no fan. Later revisions
used a more robust linear supply, and a quiet little
internal fan.

The 7603 of the later years was the scope intended
for the 7D20, 7L family.... it was the big screen that
made it desirable.

The 7904 (plain) cannot handle the 7L13. It will send
the supply into tick mode. Not the usual tick mode,
but one that was very weird sounding.

The 7904A handles it just fine... but then it has a big
supply and a loud fan.

-Chuck Harris

Dennis Tillman W7PF wrote:
Hi Tim,
I have used a 7D20 in several 7603s I have. All of the plugins could not exceed a well-defined set of power supply voltage, current and power specifications the mainframes had available. Meeting these requirements was sometimes the most difficult aspect of a design.

I suggest you calibrate your 7603 power supplies (and take a look at the power supplies in the 7D20 as well) and then try using the 7D20 again.

I would not disagree with you there were many power hungry plugins. Many of those came along in the mid-1970s and they used power hungry digital logic. The power needs of this type of circuitry was not something the designers of the 7K series could anticipate when they were finalizing the basic requirements of the product line.

The 7D20 is power hungry but it should work if the scope is working properly.

Another really good example would be the 7D01 Hardware Logic Analyzer which is almost entirely made of extremely power hungry ECL logic. It was the only way to reach the 100MHz speed necessary to troubleshoot complex digital circuits.

The 7D02 Microprocessor Logic Analyzer is also a power hungry plugin. It is packed full with TTL which is not as bad as ECL but still needs lots of power. None of the more lower power TTL families were available yet in the numerous logical building blocks that plugin needed. The product engineer on the project told me all the things they tried to lower the power requirements so it would meet the mainframe specs. He was disgusted by the entire process.

The triple wide Spectrum Analyzer plugins are also power hungry.

Dennis Tillman W7PF

Re: 7000-series power-hungry plug-ins

 

On Fri, Oct 12, 2018 at 11:31 AM, Robert Dixon wrote:


Is there a place where both of these are listed, ideally by supply rail?
Part of your request could be found here:
http://www.hakanh.com/dl/docs/hardtofind/7k_PS_currents.pdf
/Håkan

Re: 7000-series power-hungry plug-ins

NigelP
 

I successfully run 7L14s in a 7603 mainframe. I have two of these 7L14 and both are OK. The notable exception of course is the lack of push button illumination on some buttons because the 7603 doesn't have a suitable 5V Lamp supply in-built (you have to add your own circuit into the 7603).

Regards

Nigel G8AYM

Re: 7000-series power-hungry plug-ins

Chuck Harris
 

I ran 7L5's, 7L13's, and the power hog 7L18 in my 7603.

The same 7L18 sank multiple 7904's.... hard.

The 7L18 ran in a 7904A, though.

-Chuck Harris

PS, thanks to Håkan for the 7904 power supply capacity note.

NigelP wrote:

I successfully run 7L14s in a 7603 mainframe. I have two of these 7L14 and both are OK. The notable exception of course is the lack of push button illumination on some buttons because the 7603 doesn't have a suitable 5V Lamp supply in-built (you have to add your own circuit into the 7603).

Regards

Nigel G8AYM






Re: 2445A calibration

Max Vlasov
 

Thank you, Chuck.

Yes, I do understand that it's not necessary for the center point on the screen to have 0.000V, but it's just an observation that persons who had less than +/- 0.002V off center also passed the CAL02/03 w/o a LIMIT error.
I entirely agree that the goal is to calibrate the reference DAC current. However, since the current DAC is utilized to generate the voltage at the multitude of analog controls, then what really counts is the voltage accuracy per channel, IMHO. I.e. after passing the whole signal chain from reference current via DAC to C/V converter (R2620 & R2621) via demultiplexer, sample & hold, buffer etc. the precise programmed current will be translated to the same voltage at the end of the buffer regardless the channel number.
Let's assume that for example, one of the OPAMP TL072 inputs has a high leakage current due to the internal JFET degradation. This will result in the wrong output level on that channel. Even for the DAC being calibrated the voltage offset will be observed on the faulty channel introducing the systematic error in the o-scope operation...
So, my point is - if the circuity is performing well, it's easy to calibrate and the automatic calibration routing is designed to operate only on the good circuit. However, if one of the components underperforming the calibration wouldn't converge.
Also on my scope I've noticed that the pot sampling is a bit weird, like the pots over/under shoot or have a bigger range and then clipping. So, if you go CCW to CW with Intensity (for example), then the center position is shifted to the right. But if you go CW to CCW, the center is shifted to the left. It's like the ADC range is smaller than the voltage swing at its input.

Hence I can only assume that there is something fundamentally wrong with the DAC circuit performance (and the successive approximation ADC as the result). Hoped that I could find an exerciser program for the DAC to feed the same data to all the channels (to match the voltages at the outputs), but there is no such thing, unfortunately...

Have you repaired an o-scope with the similar behavior?

Thank you,

Maxim

Re: Tek 2467B

tekscopegroup@...
 

On Thu, Oct 11, 2018 at 08:33 AM, Chuck Harris wrote:

You might notice that there is a routine to read the calibration
constants, but no routine to write the calibration constants. I seem
to recall that there is a way, but I don't have it in my notes.

I am certain that Tektronix believed, as I do, that the only right
way to write the calibration constants is to do a calibration.
Aside from all the very valid points you point out about the calibration status, I just wanted to point out that if you copied down the calibration constants data with the EXER 02 routine, you can then manually rebuild the information by editing a blank bin file and then write the file into a new NVRAM. I am not sure of the exact step by step procedure, but I understand the general concept and it seems to make sense to me. And I read about more than one instance where someone managed to do it. I think only the operating hours and last scope configuration is all that would be lost but those are irrelevant items when it comes to the calibration constants.

Since I would be only using a scope for hobby use, it would be more than enough to just keep the calibration as-is even after 20-30 years or "wear and tear" and component drift/aging. In my experience top quality second hand equipments like Tek and HP, etc tend to be still pretty much on the button when it comes to accuracy, unless someone went into them and messed around. Sometimes I am even amazed on how little drift the instruments show after going for a long period of time without recalibration. Ok so I don't have a calibration standard to verify, but I have other instruments and when readings closely agree on different instruments that have not been recalibrated I know I'm close enough. Of course, I agree if an instrument needs to go into a high end corporate or research lab, and it would be of course a no-go without a strict up to date calibration. But then again, hobby use? I'm happy if its almost there, within a hair or two of what it was when it was first calibrated way back then. But then, that's just me. Maybe others would not be happy unless their scopes are fully calibrated, even only for hobby bench use. And if I had a calibration service around the corner, which I don't as I am not in the US, then I probably might also consider it.

Re: 2445A calibration

Chuck Harris
 

There is simply no reason the scope's DAC should be
at 0V when the DELTA control positions the cursor on
the center graticule on the screen.

In all the time I have been calibrating 2465's, I have
never, repeat never, had one that was perfectly 1.25000V
and -1.25000V. Never. They have always been offset
a few millivolts one way or the other. I recall one that
was at about 1.35V, and -1.15V that calibrated just fine.

The only requirement is that it not fall out of the range
of the MUX's and OPAMPS in the analog/digital control
circuits in the scope. The A5 board compensates for this
"error" by storing a constant.

You are shooting at the wrong fish.

In my experience calibrating 2465 scopes, LIMIT errors
occur most often by bad vertical preamp/attenuator sections,
and/or bad application of the reference voltage asked for
in the calibration step.

As to the pots behaving a little weird on a 2465B, that
is normal, and the first most common complaint by new users.
The second most common complaint is the random looking
missing chunks from the display... caused by the display
multiplexing.

When tektronix added all of the extra bells and whistles
to the 2465A and B models, they did not increase the A5
processor speed. So, the processor became too busy handling
all of its background tasks to instantly respond to the
changes in pot position. This shows up as a stuttering
when you turn the knob too quickly.

Another complaint is when you are in the AUTO LVL trigger
position, and you have no signal on the screen. If you
look carefully at the trace, you will note it goes through
a little repeating dance as it tries to trigger on the
noise in the vertical preamp. Change from AUTO LVL to
simply AUTO, and it stops. This is due to the scope
crosstalk into the vertical amplifier system as it cranks
through trigger levels searching for a triggerable signal.

Another anomaly most users eventually notice in the 2465A
and B models, is the scope algorithms for reading pots are
faulty.

Try this: set the intensity pot to a normal value, turn
the scope off, set the intensity pot to fully CW, or CCW,
and then turn the scope back on.

Notice that the intensity is the same as it was when you
turned the scope off, and is not represented by the pot's
position.

Now, quickly turn the intensity up fully to the stop, and
then down fully to the stop, and notice that the intensity
control now behaves normally.

Try the same tricks with the horizontal and vertical
position controls, or the trigger level control.

The controls are all simply pots, with +1.38V on one end,
and -1.25V on the other, and the A5 routines use the
DAC/trigger/MUX circuitry to read the voltage coming
out of the pot's wiper... simple and obvious.

On power down, the A5 controller stores the DAC output
values the scope feeds each of the analog circuits that
control intensity/position/level/timing....

But on power up, they read the pot's voltage, and make the
grand assumption that that voltage must be associated with
the intensity value, or position value they stored on
power down.

The programmer never considered that someone might have
fiddled with the knobs while the scope was turned off!

Note: I agree with the idea that they should ignore the
intensity pot's position on power up, and always assume
a safe value. But horizontal/vertical/trigger level???

-Chuck Harris




Max Vlasov via Groups.Io wrote:

Thank you, Chuck.

Yes, I do understand that it's not necessary for the center point on the screen to have 0.000V, but it's just an observation that persons who had less than +/- 0.002V off center also passed the CAL02/03 w/o a LIMIT error.
I entirely agree that the goal is to calibrate the reference DAC current. However, since the current DAC is utilized to generate the voltage at the multitude of analog controls, then what really counts is the voltage accuracy per channel, IMHO. I.e. after passing the whole signal chain from reference current via DAC to C/V converter (R2620 & R2621) via demultiplexer, sample & hold, buffer etc. the precise programmed current will be translated to the same voltage at the end of the buffer regardless the channel number.
Let's assume that for example, one of the OPAMP TL072 inputs has a high leakage current due to the internal JFET degradation. This will result in the wrong output level on that channel. Even for the DAC being calibrated the voltage offset will be observed on the faulty channel introducing the systematic error in the o-scope operation...
So, my point is - if the circuity is performing well, it's easy to calibrate and the automatic calibration routing is designed to operate only on the good circuit. However, if one of the components underperforming the calibration wouldn't converge.
Also on my scope I've noticed that the pot sampling is a bit weird, like the pots over/under shoot or have a bigger range and then clipping. So, if you go CCW to CW with Intensity (for example), then the center position is shifted to the right. But if you go CW to CCW, the center is shifted to the left. It's like the ADC range is smaller than the voltage swing at its input.

Hence I can only assume that there is something fundamentally wrong with the DAC circuit performance (and the successive approximation ADC as the result). Hoped that I could find an exerciser program for the DAC to feed the same data to all the channels (to match the voltages at the outputs), but there is no such thing, unfortunately...

Have you repaired an o-scope with the similar behavior?

Thank you,

Maxim





Re: 2445A calibration

 

Wouldn't it be nice to have a copy of the source for that code.

It would be fun to fix those things.

Regards

----- Original Message -----
From: "Chuck Harris" <cfharris@...>
To: <TekScopes@groups.io>
Sent: Friday, October 12, 2018 10:07 AM
Subject: Re: [TekScopes] 2445A calibration


There is simply no reason the scope's DAC should be
at 0V when the DELTA control positions the cursor on
the center graticule on the screen.

In all the time I have been calibrating 2465's, I have
never, repeat never, had one that was perfectly 1.25000V
and -1.25000V. Never. They have always been offset
a few millivolts one way or the other. I recall one that
was at about 1.35V, and -1.15V that calibrated just fine.

The only requirement is that it not fall out of the range
of the MUX's and OPAMPS in the analog/digital control
circuits in the scope. The A5 board compensates for this
"error" by storing a constant.

You are shooting at the wrong fish.

In my experience calibrating 2465 scopes, LIMIT errors
occur most often by bad vertical preamp/attenuator sections,
and/or bad application of the reference voltage asked for
in the calibration step.

As to the pots behaving a little weird on a 2465B, that
is normal, and the first most common complaint by new users.
The second most common complaint is the random looking
missing chunks from the display... caused by the display
multiplexing.

When tektronix added all of the extra bells and whistles
to the 2465A and B models, they did not increase the A5
processor speed. So, the processor became too busy handling
all of its background tasks to instantly respond to the
changes in pot position. This shows up as a stuttering
when you turn the knob too quickly.

Another complaint is when you are in the AUTO LVL trigger
position, and you have no signal on the screen. If you
look carefully at the trace, you will note it goes through
a little repeating dance as it tries to trigger on the
noise in the vertical preamp. Change from AUTO LVL to
simply AUTO, and it stops. This is due to the scope
crosstalk into the vertical amplifier system as it cranks
through trigger levels searching for a triggerable signal.

Another anomaly most users eventually notice in the 2465A
and B models, is the scope algorithms for reading pots are
faulty.

Try this: set the intensity pot to a normal value, turn
the scope off, set the intensity pot to fully CW, or CCW,
and then turn the scope back on.

Notice that the intensity is the same as it was when you
turned the scope off, and is not represented by the pot's
position.

Now, quickly turn the intensity up fully to the stop, and
then down fully to the stop, and notice that the intensity
control now behaves normally.

Try the same tricks with the horizontal and vertical
position controls, or the trigger level control.

The controls are all simply pots, with +1.38V on one end,
and -1.25V on the other, and the A5 routines use the
DAC/trigger/MUX circuitry to read the voltage coming
out of the pot's wiper... simple and obvious.

On power down, the A5 controller stores the DAC output
values the scope feeds each of the analog circuits that
control intensity/position/level/timing....

But on power up, they read the pot's voltage, and make the
grand assumption that that voltage must be associated with
the intensity value, or position value they stored on
power down.

The programmer never considered that someone might have
fiddled with the knobs while the scope was turned off!

Note: I agree with the idea that they should ignore the
intensity pot's position on power up, and always assume
a safe value. But horizontal/vertical/trigger level???

-Chuck Harris




Max Vlasov via Groups.Io wrote:
Thank you, Chuck.

Yes, I do understand that it's not necessary for the center point on the screen to have 0.000V, but it's just an observation that persons who had less than +/- 0.002V off center also passed the CAL02/03 w/o a LIMIT error.
I entirely agree that the goal is to calibrate the reference DAC current. However, since the current DAC is utilized to generate the voltage at the multitude of analog controls, then what really counts is the voltage accuracy per channel, IMHO. I.e. after passing the whole signal chain from reference current via DAC to C/V converter (R2620 & R2621) via demultiplexer, sample & hold, buffer etc. the precise programmed current will be translated to the same voltage at the end of the buffer regardless the channel number.
Let's assume that for example, one of the OPAMP TL072 inputs has a high leakage current due to the internal JFET degradation. This will result in the wrong output level on that channel. Even for the DAC being calibrated the voltage offset will be observed on the faulty channel introducing the systematic error in the o-scope operation...
So, my point is - if the circuity is performing well, it's easy to calibrate and the automatic calibration routing is designed to operate only on the good circuit. However, if one of the components underperforming the calibration wouldn't converge.
Also on my scope I've noticed that the pot sampling is a bit weird, like the pots over/under shoot or have a bigger range and then clipping. So, if you go CCW to CW with Intensity (for example), then the center position is shifted to the right. But if you go CW to CCW, the center is shifted to the left. It's like the ADC range is smaller than the voltage swing at its input.

Hence I can only assume that there is something fundamentally wrong with the DAC circuit performance (and the successive approximation ADC as the result). Hoped that I could find an exerciser program for the DAC to feed the same data to all the channels (to match the voltages at the outputs), but there is no such thing, unfortunately...

Have you repaired an o-scope with the similar behavior?

Thank you,

Maxim





Re: Tek 2467B

Chuck Harris
 

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.

Doing other insane stuff like loading up a new NVRAM with the
contents of some other scope's NVRAM... well I just don't know
what to say about that...

Even a relatively incompetent embedded system developer can
write a little monitor program for the 6802 microprocessor, in
the 2465 in about a day's time, that would allow reading/writing
and exercising the circuitry in a 2465. That used to be the
first thing I would do when handed a new piece of customer
hardware to verify/validate/program.

99% (hopeless exaggeration) of the NVRAM is not calibration
constants, but rather operating memory for the 6802. That
memory holds stuff like the interrupt pointers, stack, heap,
and data areas for the program. Only about 180 bytes of the
NVRAM contains the calibration constants.

Each constant in the calibration constant block has a parity
bit, and the entire block has a checksum. This is done to
instantly let the CPU know that something is wrong with the
calibration data. Possibilities are:

1) never been calibrated.
2) some constant is beyond its LIMIT
3) NVRAM has expired and data is bad
4) an accident has occurred, and data is bad.

You can simply hand copy the block of data from the screen,
plant it in a blank NVRAM at the appropriate location, and
it will become the new data, no fuss, no muss.

However, if your NVRAM goes bad, you calibration certainly
has been bad for probably 20 years.

-Chuck Harris

Claimer: I am very biased on this subject. I calibrate 2465's.

tekscopegroup@... wrote:

On Thu, Oct 11, 2018 at 08:33 AM, Chuck Harris wrote:

You might notice that there is a routine to read the calibration
constants, but no routine to write the calibration constants. I seem
to recall that there is a way, but I don't have it in my notes.

I am certain that Tektronix believed, as I do, that the only right
way to write the calibration constants is to do a calibration.
Aside from all the very valid points you point out about the calibration status, I just wanted to point out that if you copied down the calibration constants data with the EXER 02 routine, you can then manually rebuild the information by editing a blank bin file and then write the file into a new NVRAM. I am not sure of the exact step by step procedure, but I understand the general concept and it seems to make sense to me. And I read about more than one instance where someone managed to do it. I think only the operating hours and last scope configuration is all that would be lost but those are irrelevant items when it comes to the calibration constants.

Since I would be only using a scope for hobby use, it would be more than enough to just keep the calibration as-is even after 20-30 years or "wear and tear" and component drift/aging. In my experience top quality second hand equipments like Tek and HP, etc tend to be still pretty much on the button when it comes to accuracy, unless someone went into them and messed around. Sometimes I am even amazed on how little drift the instruments show after going for a long period of time without recalibration. Ok so I don't have a calibration standard to verify, but I have other instruments and when readings closely agree on different instruments that have not been recalibrated I know I'm close enough. Of course, I agree if an instrument needs to go into a high end corporate or research lab, and it would be of course a no-go without a strict up to date calibration. But then again, hobby use? I'm happy if its almost there, within a hair or two of what it was when it was first calibrated way back then. But then, that's just me. Maybe others would not be happy unless their scopes are fully calibrated, even only for hobby bench use. And if I had a calibration service around the corner, which I don't as I am not in the US, then I probably might also consider it.

Re: 2445A calibration

Siggi
 

On Thu, 11 Oct 2018 at 20:13 <maxim.vlasov@...> wrote:

Update:
Measured again the test points, looks like I haven't correctly stated the
voltage:
TP2421=-1.2524V
TP2420 = 1.3639V

By looking at the schematics R2422=14.2K and R2421=13.0K we have:

TP2420=-TP2421*R2422/R2421 = 1.368V (1.3639V measured).

Not exactly off target but I'll try replacing U2420A after re-measuring
the real R2422 and R2421 resistor values. U2521 is the next candidate for
the replacement.
The DAC outputs differential currents, which are converted to voltages
across R2520 and R2521. The centerpoint of those two is the 1.36V
reference, which itself is derived from the DAC through U2521, U2420D,
U2420A and Q2320 and misc other components (this is from my 2467
schematics).

I remember figuring out how this works ... time passes ... here's a snippet
from an old email on a different thread:
--- cut here ---
The way these voltages come about is fairly convoluted. Let's walk through
it.

The DAC, U2101, is a current-mode device, and it takes an input reference
current.
The reference current is derived from the 10V REF input, and from R2013 -
which is a 0.1%, low TC resistor. The reference current can then be trimmed
by way of R2010 and the network around that trimmer. If the reference
current is low, the DACs outputs will be low also.

Start by measuring the voltage between pins 14 and 15 on U2101. As these
are the inputs of an opamp in negative feedback, the voltage difference
between those pins should be effectively zero. A visual inspection of the
circuit around R2013 would be in order, and maybe you can measure them
in-circuit, or at least measure the resistance from pins 14/15 to the 10V
rail (<=10k), to ground (<=10k) and to each other (<=5k), just to verify
that the network is connected and sound.

Now, the output of the DAC is a current that's converted to voltage by the
two load resistors R2520/R2521. Unfortunately the service manual I have
doesn't have the waveform you should see across those resistors, but if you
can measure that differentially with a scope, there is a spec somewhere for
the peak-to-peak values you should be seeing. We know the DC offset is
going to be wrong already, because TP2420 sets that.

The way the -1.25 reference voltage is derived, is that the DACs input is
set to a predefined code. The resultant output is sampled through U2521,
held by C2320, and buffered by U2420D. U2420A then creates the +1.36V
reference by ratio across the ground. The ratio of your voltages is right,
so we can assume that U2420A is good, I think.
--- cut here ---

Apparently the DAC's full-scale output is 4mA, and the current-voltage
resistor is 681 Ohm. The code used to derive the reference voltages should
then be (1.36+1.25)/681/0.004*4096 = 3925 (rounded up from 3924.58).
Assuming the DAC's linear, the precision in the output current is 0.025%.

The upshot is, I think, that assuming everything else is healthy, the
difference between TP2420 and TP2421 is a direct reading of the health of
your DAC plus the voltage to current conversion. It looks like you're 0.25%
out, maybe you can trim it to nominal and see what happens? I think you're
likely in spec as-is, though.

Re: Tek 2467B

 

On Fri, Oct 12, 2018 at 04:34 PM, Chuck Harris wrote:
Hi Chuck,
You're writing:

99% (hopeless exaggeration) of the NVRAM is not calibration
constants, but rather operating memory for the 6802. That
memory holds stuff like the interrupt pointers, stack, heap,
and data areas for the program.
Knowing the limited number of write cycles that an FRAM is capable of: Have you ever done any calculations or measurements re. the lifetime to be expected for these devices in a 2445(X) or 2465(X)?
Since these devices do not do any wear leveling, results could be worrying/disappointing.

Raymond