开云体育

Firstly, thank you for taking the time for a lengthy explanation.

Secondly, I didn't take any offence in any way. I'm only too aware that I
have a lot to learn, especially when I ask a question!

I've eventually managed to make progress, after real-life intervened. I
made my own normaliser and, while it significantly improved matters, the
result was still "sub-optimal". Eventually I noticed that the shape of
transitions depended on the level. I was, I thought, using a C0G capacitor
- but either I wasn't or they aren't as voltage insensitive as I believed.
Removing that and using only a trimcap solved that issue, and I made
progress.

Currently there is a good waveshape on both channels and at all
sensitivities, with a 50ohm terminator and a conventional *10 probe. The
notable exception is that one channel's /2 attenuator is very sensitive to
movement; the AC response varies significantly. I've ordered a replacement.

So, thank you very much indeed.



On 21/02/18 14:57, Fabio Trevisan wrote:

Hello Tom,
I own a 466 and, from your description, it seems your 475's vertical
attenuator works exactly as the one in mine (I`m not looking at the 475
manual right now).

Again, from your description, it seems that either you're getting the
frequency compensation adjustment / input capacitance adjustment procedure
wrong, or the input capacitance trimmer of *all* your attenuator modules
are bad or flaky.

Since it's unlikely that *all* them are bad, it seems more likely that
you're getting the procedure wrong (forgive me if I`m assuming too much
here).
It's easy to overlook the input capacitance and have them wrong if you're
always sourcing the signal through a low impedance generator (or low
impedance terminated cable), because whatever they're right or wrong,
these
tiny capacitances, won't trouble a 50 Ohm sourced signal...

Every attenuator module has a freq-compensation trimmer (which is in
series
with the signal, paralleled to the attenuator resistor in question), and
an
input capacitance trimmer, that's between the input of the attenuator and
ground.

The compensation capacitor (the one in series with the signal, paralleled
to the attenuator resistor) works against the input capacitance of the
scope *or* against the input capacitance trimmer of any other attenuator
module that is in series, down in the attenuation chain...so, it's
paramount that each attenuator module's input capacitance is adjusted to
*exactly* the same as the oscilloscope's bare front end capacitance.

So in order to get it right, first of all, you need an input capacitance
normalizer (or feed the signal through a x10 oscilloscope probe, but then
your generator needs to be able to generate some hefty 30Vpp for a 6 div
waveform).

Then you have to follow the sequence of adjustments...

1. Use the most sensitive setting of the scope (5mV/div, no atten.
inserted), source the signal through the input normalizer and adjust the
its trimmer to get a good square edge (or adjust your x10 probe
compensation).
This will assure that the normalizer's capacitance equals the input
capacitance of the oscilloscope front end's bare capacitance.
Adjust the frequency of the generator so that it's low enough for you to
see the square wave become round or pointy... don't set the frequency too
high, or it will only seem that the square wave is changing amplitude.

2. Double the output voltage of your generator, set the scope to 10mV/div
(x2 attenuator inserted), and then...

2.1. Source the signal to the scope through a 50Ohm pass-thru terminator,
and adjust the freq. compensation trimmer of the x2 attenuator module...
This will get the x2 freq. compensation right and will eventually change
by
some amount the input capacitance of the scope... which won't trouble your
50Ohm source.

2.2.Now, remove the 50Ohm pass-thru terminator, insert the input
normalizer
and adjust the input capacitance trimmer from the x2 attenuator module.
Since the input normalizer is adjusted to match the scope's bare input
capacitance (from step 1), when you now adjust the input capacitance
trimmer of the x2 module, you're
making the x2 input capacitance the same as the bare input, meaning that,
as from now, the x2 input capacitance is "normalized" and therefore, will
be seen by any preceding attenuator
that may be inserted in the chain, as having the same input capacitance as
the oscilloscope's bare front end.
Note:
This step is assuming that your connecting cable is short or that you're
hooking up your generator directly to the input of the
normalizer...(because the cable will be not properly terminated)
Anyway, at the frequency that you will be using, in the 10s of kHz, it
shouldn't be troubled too much by the cable's reflections.
If you're a purist, you can keep the 50Ohm pass through terminator before
the input normalizer, but then your square wave will be halved and it may
require you to double the output of the generator.
No big deal at this step, but it can be more difficult to get the
appropriate test level when you're setting the x100 attenuator.

3. Walk you way up through the next attenuation levels, x4, x10 and x100,
repeating the same procedure of step 2, 2.1 and 2.2, as you increase the
output of the generator.

At the end, all input attenuator blocks will present the same input
capacitance to the input BNC of your scope, or to any precedent attenuator
block, as the scope's front end bare capacitance.

Hope I guessed it right, because other than that, to me it simply doesn't
make sense (in considering the way how the blocks are interconnected and
how they're performing correctly when each one is inserted alone).

rgrds,

Fabio




2018-02-21 6:16 GMT-03:00 Tom Gardner <tggzzz@...>:

I'm having a problem setting up the attenuators in my 475. It appears that

each section is working individually, but not when there are two sections
in series.


The background...

The 475 was very cheap because the timebase was unreliable. It was a
little mucky on the outside, but there's no visible problem inside.
There
are signs that a previous owner had been inside before. The power
supplies
are all good, and so far the only components I've replaced has been the
motor controller quad transistors (with ZTX630s that just happened to be
available).

I've got the timebase working by a combination of using:
- IPA to clean the pin in the timebase knob that prods the A/B
microswitch
- IPA on smooth paper to clean the timing board finger contacts
- since that wasn't completely successful, using a jeweller's screwdriver
to add a tiny drop of Caig DeOxit to each switch contact
I'd have preferred not to use DeOxit since it might attract crap over
time, but it wasn't good enough without it. Conclusion: those switch
contacts were imperfect.


The problem...

So now I've moved onto the input channels. I've used a multimeter to
check
the continuity of the topside attenuator finger contacts, and they are
all
fine (<<1ohm). I've reseated all the attenuators and "sel"
resistors/jumpers in the attenuator and added a tiny drop of DeOxit,
without effect.

But I'm getting ambiguous behaviour when 2 attenuator sections are in
series:

1) on 5mV/div the displayed waveform is perfect from DC to 1.8ns
risetime.
I interpret this to everything after the attenuator is working well, and
finger contacts underneath the attenuator board are working.

2) on 10mV/div, I tweaked the *2 attenuator capacitors to get a perfect
display. Therefore that attenuator and those contacts are working.

3) on 20mV/div, ditto with *4 attenuator

4) on 50mV/div, ditto with *10 attenuator

5) on 100mV/div, the displayed waveform shows pronounced HF compensation
error. An attenuator can be re-tweaked to get a good display, but it is
very sensitive to position. Of course that invalidates steps 2 or 4

6) ditto 200mV/div, and invalidating steps 3 or 4

Summary: it appears the *2, *4, *10 attenuator/switch are working
individually, but not in combination.


Future...

So, what's my next step?
I could use IPA and/or DeOxit on the finger contacts. But that might
attract crap, and if a film bridged the gaps it might subtly affect HF
performance.
I believe that sometimes it helps to resolder the contacts inside the
attenuator. But that feels like kill-or-cure, particularly since I don't
know the temperature and the plastic looks like it might melt easily.

Are there any other diagnostics I can do?
What's the best course of action?
What iron temperature and re-soldering techniques should I use?
Would 465 attenuator sections be substitutes for some of the 475's
attenuator?

Thanks

Hello Tom,
I own a 466 and, from your description, it seems your 475's vertical
attenuator works exactly as the one in mine (I`m not looking at the 475
manual right now).

Again, from your description, it seems that either you're getting the
frequency compensation adjustment / input capacitance adjustment procedure
wrong, or the input capacitance trimmer of *all* your attenuator modules
are bad or flaky.

Since it's unlikely that *all* them are bad, it seems more likely that
you're getting the procedure wrong (forgive me if I`m assuming too much
here).
It's easy to overlook the input capacitance and have them wrong if you're
always sourcing the signal through a low impedance generator (or low
impedance terminated cable), because whatever they're right or wrong, these
tiny capacitances, won't trouble a 50 Ohm sourced signal...

Every attenuator module has a freq-compensation trimmer (which is in series
with the signal, paralleled to the attenuator resistor in question), and an
input capacitance trimmer, that's between the input of the attenuator and
ground.

The compensation capacitor (the one in series with the signal, paralleled
to the attenuator resistor) works against the input capacitance of the
scope *or* against the input capacitance trimmer of any other attenuator
module that is in series, down in the attenuation chain...so, it's
paramount that each attenuator module's input capacitance is adjusted to
*exactly* the same as the oscilloscope's bare front end capacitance.

So in order to get it right, first of all, you need an input capacitance
normalizer (or feed the signal through a x10 oscilloscope probe, but then
your generator needs to be able to generate some hefty 30Vpp for a 6 div
waveform).

Then you have to follow the sequence of adjustments...

1. Use the most sensitive setting of the scope (5mV/div, no atten.
inserted), source the signal through the input normalizer and adjust the
its trimmer to get a good square edge (or adjust your x10 probe
compensation).
This will assure that the normalizer's capacitance equals the input
capacitance of the oscilloscope front end's bare capacitance.
Adjust the frequency of the generator so that it's low enough for you to
see the square wave become round or pointy... don't set the frequency too
high, or it will only seem that the square wave is changing amplitude.

2. Double the output voltage of your generator, set the scope to 10mV/div
(x2 attenuator inserted), and then...

2.1. Source the signal to the scope through a 50Ohm pass-thru terminator,
and adjust the freq. compensation trimmer of the x2 attenuator module...
This will get the x2 freq. compensation right and will eventually change by
some amount the input capacitance of the scope... which won't trouble your
50Ohm source.

2.2.Now, remove the 50Ohm pass-thru terminator, insert the input normalizer
and adjust the input capacitance trimmer from the x2 attenuator module.
Since the input normalizer is adjusted to match the scope's bare input
capacitance (from step 1), when you now adjust the input capacitance
trimmer of the x2 module, you're
making the x2 input capacitance the same as the bare input, meaning that,
as from now, the x2 input capacitance is "normalized" and therefore, will
be seen by any preceding attenuator
that may be inserted in the chain, as having the same input capacitance as
the oscilloscope's bare front end.
Note:
This step is assuming that your connecting cable is short or that you're
hooking up your generator directly to the input of the
normalizer...(because the cable will be not properly terminated)
Anyway, at the frequency that you will be using, in the 10s of kHz, it
shouldn't be troubled too much by the cable's reflections.
If you're a purist, you can keep the 50Ohm pass through terminator before
the input normalizer, but then your square wave will be halved and it may
require you to double the output of the generator.
No big deal at this step, but it can be more difficult to get the
appropriate test level when you're setting the x100 attenuator.

3. Walk you way up through the next attenuation levels, x4, x10 and x100,
repeating the same procedure of step 2, 2.1 and 2.2, as you increase the
output of the generator.

At the end, all input attenuator blocks will present the same input
capacitance to the input BNC of your scope, or to any precedent attenuator
block, as the scope's front end bare capacitance.

Hope I guessed it right, because other than that, to me it simply doesn't
make sense (in considering the way how the blocks are interconnected and
how they're performing correctly when each one is inserted alone).

rgrds,

Fabio




2018-02-21 6:16 GMT-03:00 Tom Gardner <tggzzz@...>:

I'm having a problem setting up the attenuators in my 475. It appears that
each section is working individually, but not when there are two sections
in series.


The background...

The 475 was very cheap because the timebase was unreliable. It was a
little mucky on the outside, but there's no visible problem inside. There
are signs that a previous owner had been inside before. The power supplies
are all good, and so far the only components I've replaced has been the
motor controller quad transistors (with ZTX630s that just happened to be
available).

I've got the timebase working by a combination of using:
- IPA to clean the pin in the timebase knob that prods the A/B microswitch
- IPA on smooth paper to clean the timing board finger contacts
- since that wasn't completely successful, using a jeweller's screwdriver
to add a tiny drop of Caig DeOxit to each switch contact
I'd have preferred not to use DeOxit since it might attract crap over
time, but it wasn't good enough without it. Conclusion: those switch
contacts were imperfect.


The problem...

So now I've moved onto the input channels. I've used a multimeter to check
the continuity of the topside attenuator finger contacts, and they are all
fine (<<1ohm). I've reseated all the attenuators and "sel"
resistors/jumpers in the attenuator and added a tiny drop of DeOxit,
without effect.

But I'm getting ambiguous behaviour when 2 attenuator sections are in
series:

1) on 5mV/div the displayed waveform is perfect from DC to 1.8ns risetime.
I interpret this to everything after the attenuator is working well, and
finger contacts underneath the attenuator board are working.

2) on 10mV/div, I tweaked the *2 attenuator capacitors to get a perfect
display. Therefore that attenuator and those contacts are working.

3) on 20mV/div, ditto with *4 attenuator

4) on 50mV/div, ditto with *10 attenuator

5) on 100mV/div, the displayed waveform shows pronounced HF compensation
error. An attenuator can be re-tweaked to get a good display, but it is
very sensitive to position. Of course that invalidates steps 2 or 4

6) ditto 200mV/div, and invalidating steps 3 or 4

Summary: it appears the *2, *4, *10 attenuator/switch are working
individually, but not in combination.


Future...

So, what's my next step?
I could use IPA and/or DeOxit on the finger contacts. But that might
attract crap, and if a film bridged the gaps it might subtly affect HF
performance.
I believe that sometimes it helps to resolder the contacts inside the
attenuator. But that feels like kill-or-cure, particularly since I don't
know the temperature and the plastic looks like it might melt easily.

Are there any other diagnostics I can do?
What's the best course of action?
What iron temperature and re-soldering techniques should I use?
Would 465 attenuator sections be substitutes for some of the 475's
attenuator?

Thanks

Both the grid bias pot and U1890A shunt current away from Q1980. By your
description it sounds like extra current is sneaking into Q1980s base -
maybe C1991 has gone leaky?

Page 3-37 has this description of this part of the circuitry:

"Transistor Q1980 is configured as a shunt-feedback amplifier, with C1991
and R1994 as teh feedback elements. The feedback current through R1994
develops a voltage across the resistor that is positive with respect to
the +42.6 V on the base of the transistor. The value of this additive
voltage plus the diode drop across CR1950 sets the upper clamping
threshold. Grid Bias potentiometer R1878 sinks varying amounts of current
away from the base node of the transistor and thus sets the feedback
current through R1994. The adjustment range of the pot can set the nominal
clamping level between +71 V and +133 V."

So the nominal adjustment range of the grid bias pot should result in 71V
to 133V on Q1980s collector. What are you seeing?

Siggi

I have no lights but I know they work. Im assuming because p10 is empty. I have looked at the schematic for hours but I cant figure out where the input to p10 comes from.