Tek 2230 PSU


Thomas Drage
 

Hi All,


I was hoping for some advice on the power supply of a 2230 I'm working on.


I received the unit with mains fuse blown. Checked and found shorted rectifiers in the main bridge. I also found C904 cracked and with evidence of arcing. I found many of the LV electrolytics on both sides of the PSU having very poor ESR so I replaced pretty well all the electrolytics.


I then powered the unit up; no fuse blowing. However, the 43V supply to the inverter was not existent and C925 not charging up to power the preregulator. I did some more checking and found Q908 shorted, R909 open and C944 shorted. The FET Q9070 isn't shorted but I've not checked that it actually works.


Further, measuring resistance across C925 gives only 25ohms (having lifted R907, CR920). Is this normal or is it likely U930 is cooked too? C944 being shorted is also a bit interesting and leaves me wondering how much else of this power supply I should be suspicious of.

At present, replacing the components I know to be bad and powering up doesn't seem to to be the best idea given I don't know the cause of the issue.


http://members.iinet.net.au/~drage/other/2230PS.png http://members.iinet.net.au/~drage/other/2230PS.png <------ Schematic



http://members.iinet.net.au/~drage/other/2230PS.png

http://members.iinet.net.au/~drage/other/2230PS.png http://members.iinet.net.au/~drage/other/2230PS.png


View on members.iinet.net.au http://members.iinet.net.au/~drage/other/2230PS.png
Preview by Yahoo




Any thoughts?


Many thanks,


Thomas


 

On 06 Nov 2015 03:57:32 -0800, you wrote:

Hi All,

I was hoping for some advice on the power supply of a 2230 I'm working on.

I received the unit with mains fuse blown. Checked and found shorted rectifiers in the main bridge. I also found C904 cracked and with evidence of arcing. I found many of the LV electrolytics on both sides of the PSU having very poor ESR so I replaced pretty well all the electrolytics.

I then powered the unit up; no fuse blowing. However, the 43V supply to the inverter was not existent and C925 not charging up to power the preregulator. I did some more checking and found Q908 shorted, R909 open and C944 shorted. The FET Q9070 isn't shorted but I've not checked that it actually works.

Further, measuring resistance across C925 gives only 25ohms (having lifted R907, CR920). Is this normal or is it likely U930 is cooked too?
I do not see how Q908 and R909 could be bad without Q9070 and U930
also being bad.

C944 being shorted is also a bit interesting and leaves me wondering how much else of this power supply I should be suspicious of.
I would be suspicious of everything except high value resistors at
this point.

At present, replacing the components I know to be bad and powering up doesn't seem to to be the best idea given I don't know the cause of the issue.

...

Any thoughts?
These power supplies tend to go big when they fail so I would plan on
testing every part and replacing every semiconductor and aluminum
electrolytic capacitor. There are modern replacements for every part
except the transformers.

If you have a suitable power supply, the inverter can be disconnected
from the preregulator and tested separately. The preregulator can be
tested using a 42.8 volt load.


Thomas Drage
 

Thanks David - good idea on testing the inverter.

I replaced C944 and powered it up....I ramped slowly up to about 25V and it seemed to work fine. The inverter drive works (viewed pulse train on CRO). The LED lit, the fan spun and "beam find" worked.


At just over 25V, the current increases dramatically and my supply limited at 4A. I'm not sure if this is due to the digital section powering up as the CRT display also changed at that point...but also could have been because voltage would have been all over the place.


The first time I did this I got a small puff of smoke - couldn't see where it had come from. I did it again, no smoke, worked okay until I pushed it back over 25V and then there was another puff of smoke and current increase. I can't quite see what component released the smoke! It was somewhere in the region of the pre-reg: around CR920 (currently disconnected). Nothing looks burned at this point but I think I need to do some more digging.


 

I would disconnect the preregulator from the inverter when doing this
test. At the very least, the body diode of Q9070 will conduct in
reverse which should not cause a problem but C906 may be shorted from
previous abuse. CR907 is also a suspect and usually fails if Q9070
fails.

Of course the failure could be on the secondary side. Check the
secondary rectifiers for shorts.

On 07 Nov 2015 23:39:52 -0800, you wrote:

Thanks David - good idea on testing the inverter.

I replaced C944 and powered it up....I ramped slowly up to about 25V and it seemed to work fine. The inverter drive works (viewed pulse train on CRO). The LED lit, the fan spun and "beam find" worked.

At just over 25V, the current increases dramatically and my supply limited at 4A. I'm not sure if this is due to the digital section powering up as the CRT display also changed at that point...but also could have been because voltage would have been all over the place.

The first time I did this I got a small puff of smoke - couldn't see where it had come from. I did it again, no smoke, worked okay until I pushed it back over 25V and then there was another puff of smoke and current increase. I can't quite see what component released the smoke! It was somewhere in the region of the pre-reg: around CR920 (currently disconnected). Nothing looks burned at this point but I think I need to do some more digging.


Thomas Drage
 

Hi David and all,

I've looked into this in some more detail and found the behaviour I was seeing is due to Q935 clamping. However, I removed and checked the 51V zener - zener okay, still clamps when inverter supply exceeds 27V. I removed the SCR and it seems okay in that it doesn't seem to break down at 27V.


Therefore, the SCR must be triggered into conducting by the pulse train from CR/R948. What is the purpose of this connection? The other thing is that looking at the waveform at CR948 anode/R949 I don't see what waveform 47 in the manual should have me seeing; there is a lot more of a "spike". I did change C944 for a 1.5uF tantalum but I can't see that having made much difference...


I will try to upload some waveform images in the next few days.


Many thanks,




Thomas


 

On 29 Nov 2015 05:49:12 -0800, you wrote:

Hi David and all,

I've looked into this in some more detail and found the behaviour I was seeing is due to Q935 clamping. However, I removed and checked the 51V zener - zener okay, still clamps when inverter supply exceeds 27V. I removed the SCR and it seems okay in that it doesn't seem to break down at 27V.

Therefore, the SCR must be triggered into conducting by the pulse train from CR/R948. What is the purpose of this connection? The other thing is that looking at the waveform at CR948 anode/R949 I don't see what waveform 47 in the manual should have me seeing; there is a lot more of a "spike". I did change C944 for a 1.5uF tantalum but I can't see that having made much difference...

I will try to upload some waveform images in the next few days.

Many thanks,

Thomas
Q935 will crowbar the 42.8 volt supply under either of two conditions;
if the 42.8 volt supply rises enough for 51 volt zener diode VR935 to
conduct, or if the voltage across R949 exceeds about 1.2 volts (the
combined voltage drop of CR948 and the SCR gate turn-on voltage)
indicating a current of 2.4 amps or greater.

The inverter made up of Q946 and Q947 operates continuously with a 50%
duty cycle controlled by T944. Based on waveform 47, it appears that
switching takes quite a bit of time.

It sounds like there is a problem on the secondary side of the
transformer which is causing excessive current draw. I would start by
taking a close look at the secondary side rectifiers and capacitors.
Have you done any work on the secondary side yet?


Tom Jobe <tomjobe@...>
 

I'm really enjoying this discussion about the 2230 PSU repair, as you
use excellent logic and reason to work your way through the problem.
The 2230 PS differs from the simpler small 22xx's such as the 2235 in
that it has this extra bit of circuitry CR948 / R948 which adds another
level of complication to possible failures involving the SCR Q935.
David's analysis of how some extra current across R949 could fire the
SCR Q935 through CR948 / R948 is very enlightening to this amateur.

One low tech thought to add to this discussion is the general experience
the Tek world has had with the original SCR Q935 which had C106 in its
part number. They are very prone to failure and I believe Tektronix
later replaced it with the higher rated MCR72-4 SCR in their 22xx power
supply update kit. The MCR72-4 part is probably not available new, but
higher rated MCR72's such as the -8 are still available from Mouser and
others. Q935 is a Sensitive Gate SCR so you might want to pick your
replacement carefully if you can't use an MCR72-4 or higher dash number.
I have used these MCR72's on the small 22xx scopes and they work perfectly.

Another simple thought for your investigation might be to provide the
42.8 VDC again from an external source but where you can see the amount
of current you are providing, I don't know the exact number, but I would
guess a 2230 with no problems would need 2 to 2.5 amps of the 42.8
volts. It has been awhile... but I think the simple 22xx's use something
like 1.5 to 2 amps of 42.8 VDC. This test might tell you if the
secondary is the part of the power supply that is loading things down?
tom jobe...



On 11/29/2015 11:27 AM, David davidwhess@gmail.com [TekScopes] wrote:

On 29 Nov 2015 05:49:12 -0800, you wrote:

Hi David and all,

I've looked into this in some more detail and found the behaviour I
was seeing is due to Q935 clamping. However, I removed and checked the
51V zener - zener okay, still clamps when inverter supply exceeds 27V.
I removed the SCR and it seems okay in that it doesn't seem to break
down at 27V.

Therefore, the SCR must be triggered into conducting by the pulse
train from CR/R948. What is the purpose of this connection? The other
thing is that looking at the waveform at CR948 anode/R949 I don't see
what waveform 47 in the manual should have me seeing; there is a lot
more of a "spike". I did change C944 for a 1.5uF tantalum but I can't
see that having made much difference...

I will try to upload some waveform images in the next few days.

Many thanks,

Thomas
Q935 will crowbar the 42.8 volt supply under either of two conditions;
if the 42.8 volt supply rises enough for 51 volt zener diode VR935 to
conduct, or if the voltage across R949 exceeds about 1.2 volts (the
combined voltage drop of CR948 and the SCR gate turn-on voltage)
indicating a current of 2.4 amps or greater.

The inverter made up of Q946 and Q947 operates continuously with a 50%
duty cycle controlled by T944. Based on waveform 47, it appears that
switching takes quite a bit of time.

It sounds like there is a problem on the secondary side of the
transformer which is causing excessive current draw. I would start by
taking a close look at the secondary side rectifiers and capacitors.
Have you done any work on the secondary side yet?



[Non-text portions of this message have been removed]


 

On Sun, 29 Nov 2015 17:52:59 -0800, you wrote:

I'm really enjoying this discussion about the 2230 PSU repair, as you
use excellent logic and reason to work your way through the problem.
The 2230 PS differs from the simpler small 22xx's such as the 2235 in
that it has this extra bit of circuitry CR948 / R948 which adds another
level of complication to possible failures involving the SCR Q935.
David's analysis of how some extra current across R949 could fire the
SCR Q935 through CR948 / R948 is very enlightening to this amateur.
I never even noticed this difference between the 2235 and 2230 but the
2235 schematic that I have is pretty poor. The 2235A added the
current trip to the SCR crowbar.

The 2230 also has the thermal shutdown board which the 2235 is
lacking.

One low tech thought to add to this discussion is the general experience
the Tek world has had with the original SCR Q935 which had C106 in its
part number. They are very prone to failure and I believe Tektronix
later replaced it with the higher rated MCR72-4 SCR in their 22xx power
supply update kit. The MCR72-4 part is probably not available new, but
higher rated MCR72's such as the -8 are still available from Mouser and
others. Q935 is a Sensitive Gate SCR so you might want to pick your
replacement carefully if you can't use an MCR72-4 or higher dash number.
I have used these MCR72's on the small 22xx scopes and they work perfectly.
My notes on the 22xx power supplies have not gotten as far as the
crowbar circuit so I do not know what changes they made over time.
Looking at it now, I see that they changed it from a 2A SCR to an 8A
SCR in the 2232 and 2235A but the Tektronix part number, 151-0565-00,
is not in the catalog. The same part was used in the update kit.

The MCR72-4 looks like it could have been the one Tektronix used; it
matches the abbreviated specifications given in the parts list.
Motorola/ON still lists the MCR72-6 and MCR72-8 which should be ideal
replacements although I see lots of options on Mouser.

Hehe. They still make the C106 but now it is available in the
TO-126/TO-225 package instead of the TO-202.

I wonder if a sensitive gate SCR is really needed. Maybe they have a
more consistent gate voltage which would be important in this case.

Another simple thought for your investigation might be to provide the
42.8 VDC again from an external source but where you can see the amount
of current you are providing, I don't know the exact number, but I would
guess a 2230 with no problems would need 2 to 2.5 amps of the 42.8
volts. It has been awhile... but I think the simple 22xx's use something
like 1.5 to 2 amps of 42.8 VDC. This test might tell you if the
secondary is the part of the power supply that is loading things down?
tom jobe...
I usually estimate this based on the nameplate power and efficiency.
For a 2230 this is 85 watts. If the first stage efficiency is 85% then
the current should be 72 watts divided by 42.8 volts which is 1.7
amps. 1.7 amps across the current sense resistor of 0.51 ohms is 0.86
volts which is comfortably low enough not to trigger the SCR through
CR948.


Thomas Drage
 

Hi David,


I replaced the LV side caps and checked the rectifiers at the start. Also tried disconnecting the upper 5V section to no avail.


I powered it up today sans SCR to test. I was able to ramp is up to almost 40V with the current around 1.5A.

The analogue section appears to work and front panel controls respond as expected. Unfortunately, my bench PSU has re-developed an old issue and prevents more testing for a little bit. I'm wondering if the trigger voltage of the SCR has drifted somewhat and it triggers too early. Once I figure that out I'll be back to repairing the pre-regulator...


 

The SCR trigger voltage is just a bipolar transistor Vbe of about 0.6
volts. Variation there is not enough to explain what is going on.

A more likely cause is that the SCR is damaged and conducting at 20
volts without triggering. It may have become so during an earlier
episode when it was triggered do to an overvoltage or overcurrent
condition.

The power supply will operate without the SCR in place. Its purpose
is to protect the rest of the oscilloscope should the power supply
output voltages rise above safe levels or if too much current is
drawn.

My recommendation is to check the other parts in the SCR circuit and
replace both the SCR and VR935.

On 06 Dec 2015 03:56:36 -0800, you wrote:

Hi David,

I replaced the LV side caps and checked the rectifiers at the start. Also tried disconnecting the upper 5V section to no avail.

I powered it up today sans SCR to test. I was able to ramp is up to almost 40V with the current around 1.5A.

The analogue section appears to work and front panel controls respond as expected. Unfortunately, my bench PSU has re-developed an old issue and prevents more testing for a little bit. I'm wondering if the trigger voltage of the SCR has drifted somewhat and it triggers too early. Once I figure that out I'll be back to repairing the pre-regulator...


Thomas Drage
 

I've since tested the SCR on a curve tracer - it doesn't conduct without triggering, so seems fine.


Also repaired my test supply and got it running again. The 2230 seems to work - displays a waveform just fine. But back to the power supply:

This is the waveform at R949 (waveform 47 in the manual):
http://members.iinet.net.au/~drage/other/IMG_5264.JPG http://members.iinet.net.au/~drage/other/IMG_5264.JPG



Note the large "spike" prior to switching: this is evidently what is *just* tripping the crowbar but looking at waveform 47, it shouldn't be there at all...


 

On 29 Dec 2015 23:02:16 -0800, you wrote:

I've since tested the SCR on a curve tracer - it doesn't conduct without triggering, so seems fine.
Crowbar duty is pretty hard on SCRs so I would have planned on
replacing it whether it tested good or not especially if I had to
remove it to test it.

Also repaired my test supply and got it running again. The 2230 seems to work - displays a waveform just fine. But back to the power supply:
This is with the test supply just powering the inverter without the
preregulator operating, right?

This is the waveform at R949 (waveform 47 in the manual):
http://members.iinet.net.au/~drage/other/IMG_5264.JPG

Note the large "spike" prior to switching: this is evidently what is *just* tripping the crowbar but looking at waveform 47, it shouldn't be there at all...
It is sometimes helpful to check the waveform examples in the other
2235 series oscilloscopes which use the same power supply design but
they show the same thing in this case. The level is about 800
millivolts and they lack that current spike just before switching
which I would normally associate with transformer saturation.

I suspect we have seen this before on 2235 series power supplies but
never identified a specific cause.

How exactly are you measuring this? What kind of isolation is being
used?


Thomas Drage
 

Hi David,

Noted, I will replace the SCR when I put it back together.


Yep - isolated test supply is powering the inverter, the preregulator is disconnected. 40V to the point where R907 was connected and -ve to TP950.


Measurement is with a 1:1 probe 200mV/div, zero is about 3 minor divs from the bottom of the screen. The probe was to the end of R949, ground clip to the test supply -ve.


Thanks,


Thomas


 

On 30 Dec 2015 20:38:28 -0800, you wrote:

Noted, I will replace the SCR when I put it back together.
When a part is old, possibly damaged, and replacements are
inexpensive, it just make sense to change it for a new one if it has
to be removed for testing.

Yep - isolated test supply is powering the inverter, the preregulator is disconnected. 40V to the point where R907 was connected and -ve to TP950.

Measurement is with a 1:1 probe 200mV/div, zero is about 3 minor divs from the bottom of the screen. The probe was to the end of R949, ground clip to the test supply -ve.
That is how I would do it. I just want to make sure that there is no
ground loop between the probe ground and the power supply ground
corrupting the measurement.

So what is going on here . . .

The oscillation frequency is controlled by saturation of T944 and
looks to be correct. Your DSO indicates that this is 42.5 kHz but it
is seeing both cycles so divide by 2 to get 21.25 kHz which is very
close to what the photograph in the service manual shows. I actually
like to see markers or cursors for this type of measurement which the
service manual photograph shows so what is being measured is
unequivocal.

The current peak shown at the end of every cycle by your DSO
definitely looks like transformer saturation in T948 and I have no
other explanation for it. T944 saturates like that to control
switching of Q946 and Q947 but its current level is so much smaller
that it is not even visible in the service manual photographs. We had
a recent post here discussing this.

Maybe someone else on the list has a better idea but my conclusion is
that T948 is damaged causing it to have low inductance. I do not
think it has a shorted turn which would cause high primary current
through the entire waveform but instead it either has a cracked or
separated core creating an air gap or maybe the core material was
damaged by high temperature.

Does physical inspection of T948 reveal anything?

You might try doing the same test while physically pressing on T948 to
see how that affects the measured current waveform.

As far as replacing T948, I have not checked but I suspect many of the
2235 series oscilloscopes used the same transformer. Perhaps an
inexpensive parts donor can be found.


Dave Hills
 

Hi,

I just picked up on this thread and curiously looked at your R949 waveform. The schematic you linked does not show the DC transformer, T948, so I am assuming that it (the transformer connection) is the same as the one on K4OBB's 2235 page. The key point is that the transformer's primary center tap is connected to a low impedance voltage source, (C940, 1000uF in this case). If this is the case your R949 waveform is indicating that Q946 and Q947 are overlapping, that is, both ON at the same time during the switching interval.

Have you scoped the Q946 and Q947 collectors at the same time to see what, if any, the 'dead time' is? There should be a short interval when both transistors are OFF. If there is no dead time, then, either Q946 and Q947 are being over driven, or one or both of them are bad. What is the DC voltage at the primary center tap of T944?

Dave


 

Do not take my following comments personally. I agree that what you
suggest in measuring the collector voltages is worth doing as a
confirmation.

On Thu, 31 Dec 2015 09:16:29 -0600, you wrote:

Hi,

I just picked up on this thread and curiously looked at your R949
waveform. The schematic you linked does not show the DC transformer,
T948, so I am assuming that it (the transformer connection) is the same
as the one on K4OBB's 2235 page. The key point is that the
The 2230 and 2235 power supplies are practically identical. There are
layout differences in the schematics so the 2230 has all of the power
supply control circuitry on a separate schematic from the CRT
circuitry while the 2235 everything on the same schematic.

The 2213A, 2215A, 2221, 2221A, 2230, 2224, 2232 (early), 2235, 2236,
2232 (late), 2235A, and 2236A all use the same power supply design
with minor additions like the thermal shutoff and changes to
accommodate greater output power. I think the 2220 also uses this
power supply but I have never seen its service manual.

transformer's primary center tap is connected to a low impedance voltage
source, (C940, 1000uF in this case). If this is the case your R949
waveform is indicating that Q946 and Q947 are overlapping, that is, both
ON at the same time during the switching interval.
If the on times of Q946 and Q947 are overlapping, then why is the
current dropping to normal between switching cycles? Are they
overlapping at the end of each half cycle, both turning off, then one
turning back on?

Have you scoped the Q946 and Q947 collectors at the same time to see
what, if any, the 'dead time' is? There should be a short interval when
both transistors are OFF. If there is no dead time, then, either Q946
and Q947 are being over driven, or one or both of them are bad. What is
the DC voltage at the primary center tap of T944?
Both transistors would have to be bad. The measured current waveform
shows the current through each transistor on alternate cycles and they
look identical which would be the case with the transformer in balance
and saturating on every half cycle.

Look carefully at the measured waveform and you can see something else
in comparison to the reference waveform in the 2230 and 2232 service
manual; after the start of each half cycle, the measured current
during each half cycle slowly grows before rising rapidly while in the
reference waveform it slowly falls. That slow rise is caused by low
inductance in the output transformer allowing its magnetizing current
to rise enough during each half cycle to become visible. The 2235
service manual does not show this as well because it has a drawing
rather than a photograph of the actual waveforms on an oscilloscope.


Tom Jobe <tomjobe@...>
 

Hi David and Dave,
Thank you for this informative exchange that applies to the power supply
of so many of these nicer 22xx oscilloscopes!
David Hess always tries to analyze the problems and the related circuits
which helps us amateurs understand a little more of the circuit's
'theory of operation', and having another person join the conversation
who understands these circuits at a higher level just makes it better
for us all.
Thanks again... and happy New Year to all!
tom jobe...



On 12/31/2015 8:05 AM, David davidwhess@gmail.com [TekScopes] wrote:

Do not take my following comments personally. I agree that what you
suggest in measuring the collector voltages is worth doing as a
confirmation.

On Thu, 31 Dec 2015 09:16:29 -0600, you wrote:

Hi,

I just picked up on this thread and curiously looked at your R949
waveform. The schematic you linked does not show the DC transformer,
T948, so I am assuming that it (the transformer connection) is the same
as the one on K4OBB's 2235 page. The key point is that the
The 2230 and 2235 power supplies are practically identical. There are
layout differences in the schematics so the 2230 has all of the power
supply control circuitry on a separate schematic from the CRT
circuitry while the 2235 everything on the same schematic.

The 2213A, 2215A, 2221, 2221A, 2230, 2224, 2232 (early), 2235, 2236,
2232 (late), 2235A, and 2236A all use the same power supply design
with minor additions like the thermal shutoff and changes to
accommodate greater output power. I think the 2220 also uses this
power supply but I have never seen its service manual.

transformer's primary center tap is connected to a low impedance voltage
source, (C940, 1000uF in this case). If this is the case your R949
waveform is indicating that Q946 and Q947 are overlapping, that is, both
ON at the same time during the switching interval.
If the on times of Q946 and Q947 are overlapping, then why is the
current dropping to normal between switching cycles? Are they
overlapping at the end of each half cycle, both turning off, then one
turning back on?

Have you scoped the Q946 and Q947 collectors at the same time to see
what, if any, the 'dead time' is? There should be a short interval when
both transistors are OFF. If there is no dead time, then, either Q946
and Q947 are being over driven, or one or both of them are bad. What is
the DC voltage at the primary center tap of T944?
Both transistors would have to be bad. The measured current waveform
shows the current through each transistor on alternate cycles and they
look identical which would be the case with the transformer in balance
and saturating on every half cycle.

Look carefully at the measured waveform and you can see something else
in comparison to the reference waveform in the 2230 and 2232 service
manual; after the start of each half cycle, the measured current
during each half cycle slowly grows before rising rapidly while in the
reference waveform it slowly falls. That slow rise is caused by low
inductance in the output transformer allowing its magnetizing current
to rise enough during each half cycle to become visible. The 2235
service manual does not show this as well because it has a drawing
rather than a photograph of the actual waveforms on an oscilloscope.



[Non-text portions of this message have been removed]


 

We may have seen this particular failure before and it might be the or
one of the "mysterious 22xx failures" which I posted about a while
ago. The difference is that in this case, we also have excellent
measurements by Thomas showing the inverter's current waveform.

I remember that the output transformer suffering from some kind of odd
failure has been suggested as a problem before but we never had any
measurements to back it up. The measurement is tricky to do without a
DC supply to power the inverter without the preregulator.

On Thu, 31 Dec 2015 15:31:31 -0800, you wrote:

Hi David and Dave,
Thank you for this informative exchange that applies to the power supply
of so many of these nicer 22xx oscilloscopes!
David Hess always tries to analyze the problems and the related circuits
which helps us amateurs understand a little more of the circuit's
'theory of operation', and having another person join the conversation
who understands these circuits at a higher level just makes it better
for us all.
Thanks again... and happy New Year to all!
tom jobe...


Tom Jobe <tomjobe@...>
 

Hi David,
This is all going to be fun to read and learn about, but as a parallel
practical check on it all I would like to see the person with the
problem 2230 power the inverter with about 43 VDC (as it does in real
life) and report how much current it actually uses... and also tell us
very clearly whether the 2230 functions properly when powered by an
external 43 volt source using 'X' amps.
tom jobe...





On 12/31/2015 3:40 PM, David davidwhess@gmail.com [TekScopes] wrote:

We may have seen this particular failure before and it might be the or
one of the "mysterious 22xx failures" which I posted about a while
ago. The difference is that in this case, we also have excellent
measurements by Thomas showing the inverter's current waveform.

I remember that the output transformer suffering from some kind of odd
failure has been suggested as a problem before but we never had any
measurements to back it up. The measurement is tricky to do without a
DC supply to power the inverter without the preregulator.

On Thu, 31 Dec 2015 15:31:31 -0800, you wrote:

Hi David and Dave,
Thank you for this informative exchange that applies to the power supply
of so many of these nicer 22xx oscilloscopes!
David Hess always tries to analyze the problems and the related circuits
which helps us amateurs understand a little more of the circuit's
'theory of operation', and having another person join the conversation
who understands these circuits at a higher level just makes it better
for us all.
Thanks again... and happy New Year to all!
tom jobe...


[Non-text portions of this message have been removed]


 

I think he reported doing just that but without measuring the DC input
current. My understanding at this point, and Thomas should correct me
if I get something wrong, is that the 2230 operated normally under the
following conditions:

1. The preregulator was bypassed and the inverter was powered with an
isolated 42.8 volt DC supply. Note that the absolute voltage is not
critical as long as it is high enough since the linear regulator built
around differential pair Q938 and Q939 controls the bias to the
inverter such that the output transistors also operate as pass
elements to control the output voltage.

2. The crowbar circuit built around Q935 was disabled by removing the
SCR because excessive current into the inverter detected by R949 was
triggering it. The current spikes shown in the measured waveform
presumably are enough to trigger it.

A short term solution I would consider is adding a capacitor across
R935 to filter out the detected current spikes. Somewhere between
0.12 and 0.68 microfarads should work. This runs the risk of allowing
the output transistors to run hot and fail.

On Thu, 31 Dec 2015 16:21:56 -0800, you wrote:

Hi David,

This is all going to be fun to read and learn about, but as a parallel
practical check on it all I would like to see the person with the
problem 2230 power the inverter with about 43 VDC (as it does in real
life) and report how much current it actually uses... and also tell us
very clearly whether the 2230 functions properly when powered by an
external 43 volt source using 'X' amps.

tom jobe...