Topics

Tek 576 noise and looping

peter bunge
 

The performance check/calibration that starts on page 5-4 was perfect up to step 16. Check Horizontal and Vertical Displayed Noise on page 5-15. The noise is twice what it should be for the horizontal and four times for the vertical. I 'scoped the power supplies and there is no ripple. Any suggestions?
I then switched to step 3. Locating Malfunctioning Circuits on page 4-5. Step D revealed that the rectified sine wave is flattened at the top and distorted on the right. This may be a cause of looping but not the noise problem. Any suggestions?
I have a parts unit that I decided to repair so I could compare signals. I had to remove the Collector switch box to repair shaft couplings so got to find out where the collector bridge rectifiers are and how to set up the shafts of the switches. I was surprised to find that the floating 50 volt supply is only 5 volts checked on both units with a 'scope and DVM. There is no ripple. It is plainly marked 50 volts in several places. Any comments?
I am going through the calibration on the second 576 but it is not looking good either.
Too bad I cannot show the photos here but if anyone wants my notes to get involved please ask off line.
PeterB

Kevin Oconnor
 

Is the looping truly visible in every day transistor measurement?
Mine is a virgin. It has some looping under some conditions but others the sweep lines are fine. Knobs & switches are somewhat noisy but not worth putting unit into the ER.
There is a line of thought that says sometimes you should leave well enough alone. Are you there or is major surgery called for?

Kjo KO3Y

Ed Breya
 

There is a looping compensation variable capacitor in the collector supply, that normally should help. However, the capacitor shaft may be seized up, and the driving shaft extension or knob may be too loose to actually rotate the cap or show that it's stuck. So, while the knob may turn, the cap may be stuck (for many years) in one spot, and have no apparent effect. Both of my 576s have this problem, and before I finally figured out what was going on years ago, I thought it just plain didn't work - a useless feature. Because it's such a PITA to work on, I let it slide, and adapted myself to the looping and ignore it.

If you've got it apart anyway, you may want to check for proper mechanical operation of the cap. As I recall, the extension shaft is plastic for electrical isolation, making it even harder to keep a tight grip.

Don't worry about distortion of the collector sinewave - line voltage looks pretty ugly anyway, due to all the nonlinear loads on it, regardless of what the 576 is doing.

Ed

peter bunge
 

Kevin; It is useable but has big loops at low current and as I said it
fails the noise test but the manual does not suggest why.
2N3568 at low current. I'm not sure if pictures show up, let me know.
[image: 2N3568 low current.jpg]

On Sun, Nov 10, 2019 at 9:14 PM Kevin Oconnor <@KO3Y> wrote:

Is the looping truly visible in every day transistor measurement?
Mine is a virgin. It has some looping under some conditions but others the
sweep lines are fine. Knobs & switches are somewhat noisy but not worth
putting unit into the ER.
There is a line of thought that says sometimes you should leave well
enough alone. Are you there or is major surgery called for?

Kjo KO3Y



Mlynch001
 

"I'm not sure if pictures show up, let me know.
[image: 2N3568 low current.jpg]"

Peter,

Attachments are not allowed within our posts. However, you can post your picture(s) in the "PHOTOS" section. Make a File and share the filename in a post. Then anyone can see the photos. Please edit and reduce the size of the files to save our limited storage space on the forum.

--
Michael Lynch
Dardanelle, AR

peter bunge
 

I suspected that I could not attach photos.
Any one interested please see the file I uploaded with examples of looping
on pages 5 and 6.
File is: Tektronix 576 Curve Tracer Repair Log 1.pdf
It also covers several repairs and my current attempt to find out about the
noise.
PeterB

On Mon, Nov 11, 2019 at 12:19 PM Mlynch001 <@mlynch001> wrote:

"I'm not sure if pictures show up, let me know.
[image: 2N3568 low current.jpg]"

Peter,

Attachments are not allowed within our posts. However, you can post your
picture(s) in the "PHOTOS" section. Make a File and share the filename in
a post. Then anyone can see the photos. Please edit and reduce the size
of the files to save our limited storage space on the forum.

--
Michael Lynch
Dardanelle, AR



Kevin Oconnor
 

Peter,
I had forgot about that looping compensation capacitor Ed mentioned. Do look into it since it sounds like you are digging inside. Let us know what you fine.
KJO KO3Y

peter bunge
 

Kevin; I found the looping capacitor broken probably from dropping the 576
during shipping. I repaired it with epoxy and used a different shaft
coupling that can be seen in a photo in my notes. Have you looked at them?
Why do you ask? It does nothing to the noise or looping I'm seeing. I have
the parts unit working, with the help of epoxy, so I can compare both
units. I will try to get thru the adjustments to the noise check on the
parts unit and see if there is a difference.
Peter.

On Mon, Nov 11, 2019 at 5:53 PM Kevin Oconnor <@KO3Y> wrote:

Peter,
I had forgot about that looping compensation capacitor Ed mentioned. Do
look into it since it sounds like you are digging inside. Let us know what
you fine.
KJO KO3Y



 

Hi Peter,
TekScopes couldn't find your photos when I asked it to search on the file name you gave. It took me 8 or 9 search queries before TekScopes finally found it.
In the future it would be very helpful to include the actual URL in addition to the name of the file so people could immediately see what you need help with.
Dennis Tillman W7PF

-----Original Message-----
From: peter bunge
Sent: Monday, November 11, 2019 12:58 PM

I suspected that I could not attach photos.
Any one interested please see the file I uploaded with examples of looping on pages 5 and 6.
File is: Tektronix 576 Curve Tracer Repair Log 1.pdf
It also covers several repairs and my current attempt to find out about the noise.
PeterB

On Mon, Nov 11, 2019 at 12:19 PM Mlynch001 <@mlynch001> wrote:

"I'm not sure if pictures show up, let me know.
[image: 2N3568 low current.jpg]"

Peter,

Attachments are not allowed within our posts. However, you can post
your
picture(s) in the "PHOTOS" section. Make a File and share the
filename in a post. Then anyone can see the photos. Please edit and
reduce the size of the files to save our limited storage space on the forum.

--
Michael Lynch
Dardanelle, AR







--
Dennis Tillman W7PF
TekScopes Moderator

 

Hi Peter,
The big loops in one photo, which is showing a worst case scenario, are due to the Miller Effect which takes the very small capacitance (Ccb) between the collector and base and multiplies it by the gain of the transistor (Beta).
As a result the sweeping collector voltage (dV/dt) generates current loops (Ib) because a tiny capacitance is multiplied by a factor of 100 and shows up as loops in the base steps. Ib = Beta x Ccb x (dV/dt).

The solution is to shut off the collector sweep and switch to the DC position which supplies a fixed voltage to the collector. Now by manually sweeping the DC slowly from 0V to your maximum voltage you will eliminate the loops. Of course on a 576 all you see is a dot which, as you slowly turn the collector voltage up, the dot will move. With a grease pencil (or equivalent) mark the progress of the trace as you increase the collector voltage and you will get a sort of useful set of curves.

A MUCH SIMPLER SOLUTION is to get a 577-D1 (Storage version). All you have to do is turn the storage on and slowly increase the DC collector voltage and the screen stores the resulting curves without loops as you increase the DC collector current slowly. This is just one of the many things you can do with the 577-D1 storage that you can't do with any other curve tracer.

Dennis Tillman W7PF

-----Original Message-----
From: peter bunge
Sent: Monday, November 11, 2019 12:58 PM

I suspected that I could not attach photos.
Any one interested please see the file I uploaded with examples of looping on pages 5 and 6.
File is: Tektronix 576 Curve Tracer Repair Log 1.pdf
It also covers several repairs and my current attempt to find out about the noise.
PeterB

On Mon, Nov 11, 2019 at 12:19 PM Mlynch001 <@mlynch001> wrote:

"I'm not sure if pictures show up, let me know.
[image: 2N3568 low current.jpg]"

Peter,

Attachments are not allowed within our posts. However, you can post
your
picture(s) in the "PHOTOS" section. Make a File and share the
filename in a post. Then anyone can see the photos. Please edit and
reduce the size of the files to save our limited storage space on the forum.

Michael Lynch
Dardanelle, AR




--
Dennis Tillman W7PF
TekScopes Moderator

 

Hi Peter,
The looping compensation will help somewhat. See my previous email for what causes the loops and how to remove them.

Dennis Tillman W7PF

-----Original Message-----
From: peter bunge
Sent: Monday, November 11, 2019 6:52 PM

Kevin; I found the looping capacitor broken probably from dropping the 576 during shipping. I repaired it with epoxy and used a different shaft coupling that can be seen in a photo in my notes. Have you looked at them?
Why do you ask? It does nothing to the noise or looping I'm seeing. I have the parts unit working, with the help of epoxy, so I can compare both units. I will try to get thru the adjustments to the noise check on the parts unit and see if there is a difference.
Peter.

On Mon, Nov 11, 2019 at 5:53 PM Kevin Oconnor <@KO3Y> wrote:

Peter,
I had forgot about that looping compensation capacitor Ed mentioned.
Do look into it since it sounds like you are digging inside. Let us
know what you fine.
KJO KO3Y







--
Dennis Tillman W7PF
TekScopes Moderator

peter bunge
 

Yes, I have trouble even finding my own postings. I searched on "tek 576
repair log" and got it right away.
https://groups.io/g/TekScopes/files/Tek%20576%20Curve%20Tracer%20Repair/Tek%20576%20Curve%20Tracer%20repair%20Log%201.pdf

I have uploaded several repair logs packed with useful information but
obviously no one has been able to find them. I include the name of the
instrument (HP 8566A) in the title so they should not be too difficult to
find.
I just searched on my "HP 8566A Repair Log.pdf" and realized I did not
upload it because it has photos and ideas from other people that are not
credited. These are my notes that include any source I can find and were
never intended for general distribution. What is the policy? I can send
them privately.
That's a good idea to include the URL when referencing them but it does not
help next month when someone is having the same problem.
Peter.

On Tue, Nov 12, 2019 at 9:23 AM Dennis Tillman W7PF <@Dennis_Tillman_W7pF>
wrote:

Hi Peter,
TekScopes couldn't find your photos when I asked it to search on the file
name you gave. It took me 8 or 9 search queries before TekScopes finally
found it.
In the future it would be very helpful to include the actual URL in
addition to the name of the file so people could immediately see what you
need help with.
Dennis Tillman W7PF

-----Original Message-----
From: peter bunge
Sent: Monday, November 11, 2019 12:58 PM

I suspected that I could not attach photos.
Any one interested please see the file I uploaded with examples of looping
on pages 5 and 6.
File is: Tektronix 576 Curve Tracer Repair Log 1.pdf
It also covers several repairs and my current attempt to find out about
the noise.
PeterB

On Mon, Nov 11, 2019 at 12:19 PM Mlynch001 <@mlynch001> wrote:

"I'm not sure if pictures show up, let me know.
[image: 2N3568 low current.jpg]"

Peter,

Attachments are not allowed within our posts. However, you can post
your
picture(s) in the "PHOTOS" section. Make a File and share the
filename in a post. Then anyone can see the photos. Please edit and
reduce the size of the files to save our limited storage space on the
forum.

--
Michael Lynch
Dardanelle, AR







--
Dennis Tillman W7PF
TekScopes Moderator



Kevin Oconnor
 

Hey Peter,
Have you really been working on this for 8 years? That’s dedication.

I apologize, after looking at your pics I realized you are working on 576, and I have a 577. The 577 is much simpler in some regards. But the analog display information is essentially identical. However, as Dennis indicated, the looping is in all likelihood accurate. Miller effects and hysteresis along with wire loops can be the cause. Your 36” leads are going to be a nightmare at low current. Your modest current curves are the best I think you can expect.

Kevin

peter bunge
 

No, I have not been working on the 576 curve tracer for 8 years but i have
been curious about the looping for at least that long.
Thanks for your explanation which I tested by comparing curves with slow
and fast transistors. I had some trouble trying to find a couple with
matching gain but the high freq (900 MHz) transistor performed perfectly
with nice curves at low current. On the other hand the slow transistor had
huge loops. I would like to concentrate on the noise problem next. It mat
be finger trouble.
I also tried the DC method but find the dots turn into lines in a "noisy
dot, small line, big line" repeating sequence as I rotate the VERTICAL
control CW. At least that is what I remember. It was totally unexpected and
I intend to go back and investigate it. I was trying to follow one line of
thought at the time.
I am stuck at Step 16, Performance Check/Calibration, Section 5, in the
manual. This step is the "Check Horizontal and Vertical Displayed Noise".
The manual goes all the way from Step 8, where the Initial Settings are
used, with numerous setting changes to Step 16 and it is easy to get them
wrong. I have printed a list of all the setting changes to step 16 and find
that some are messed with but never reset to a specific value (Var
Collector Supply and Step Gen Amplitude for example.
Tektronix really should have listed all settings prior to Step 16 and I
suspect following the manual may be wrong and causing my inability to get
within specs. It does not help that the manual does not tell me what they
are trying to accomplish, or where to look for the problem either.
Does anyone know what the intention of Step 16 is and the correct settings
prior to the tests?
I will send my Setting Notes separately to you and anyone else interested
(contact me off line).
Regards, Peter

On Tue, Nov 12, 2019 at 7:38 PM Kevin Oconnor <@KO3Y> wrote:

Hey Peter,
Have you really been working on this for 8 years? That’s dedication.

I apologize, after looking at your pics I realized you are working on 576,
and I have a 577. The 577 is much simpler in some regards. But the analog
display information is essentially identical. However, as Dennis indicated,
the looping is in all likelihood accurate. Miller effects and hysteresis
along with wire loops can be the cause. Your 36” leads are going to be a
nightmare at low current. Your modest current curves are the best I think
you can expect.

Kevin



Brad Thompson
 

peter bunge wrote on 11/15/2019 3:27 PM:

No, I have not been working on the 576 curve tracer for 8 years but i have
been curious about the looping for at least that long.
Thanks for your explanation which I tested by comparing curves with slow
and fast transistors. I had some trouble trying to find a couple with
matching gain but the high freq (900 MHz) transistor performed perfectly
with nice curves at low current. On the other hand the slow transistor had
huge loops. I would like to concentrate on the noise problem next. It mat
be finger trouble.
<good info snipped>

Hello--

This is a long shot, but I wonder whether there's a thermal issue involved? If the
device's die is poorly attached to its header or if there's a poor thermal connection between
internal lead wires and the die attachment(s) that might cause an offset between the
increasing and decreasing base current.

If it's a leakage effect, I'd expect that germanium transistors (remember those<g>)
would show fat loops.

73--

 Brad  AA1IP

peter bunge
 

The post by Dennis talking about Miller capacity prompted me to run a test
with slow transistors and RF transistors which showed no looping with a 900
MHz transistor. You won't see loping at higher currents (10 mA) with either.
The base current is constant for each curve. The collector voltage is a
full wave rectified sine wave and rises to a maximum then returns along the
same path to 0 V. The Collector current should shoot up to a set level and
return along the same path but it does not at very low currents. I think
the miller capacity feeds the decreasing voltage as a bucking charge into
the base turning off the transistor and reducing the current creating the
loop. A fast transistor has lower capacity and does not do this. Try it
with two transistors and see.
The curve is collector current in the vertical plotted against collector
voltage on the horizontal. Ideally the line should jump from zero
instantaneously to a current and stay there as a horizontal line as long as
there is any collector voltage. But nothing is ideal.
I don't believe either one of my Curve Tracers has a fault, I just want to
complete the adjustments which are stuck at the noise test.
Peter.

On Fri, Nov 15, 2019 at 4:15 PM Brad Thompson <brad.thompsonaa1ip@...>
wrote:

peter bunge wrote on 11/15/2019 3:27 PM:

No, I have not been working on the 576 curve tracer for 8 years but i
have
been curious about the looping for at least that long.
Thanks for your explanation which I tested by comparing curves with slow
and fast transistors. I had some trouble trying to find a couple with
matching gain but the high freq (900 MHz) transistor performed perfectly
with nice curves at low current. On the other hand the slow transistor
had
huge loops. I would like to concentrate on the noise problem next. It mat
be finger trouble.
<good info snipped>

Hello--

This is a long shot, but I wonder whether there's a thermal issue
involved? If the
device's die is poorly attached to its header or if there's a poor
thermal connection between
internal lead wires and the die attachment(s) that might cause an offset
between the
increasing and decreasing base current.

If it's a leakage effect, I'd expect that germanium transistors
(remember those<g>)
would show fat loops.

73--

Brad AA1IP



 

Hi Peter,
That is a clever idea to compare high frequency transistors (low Miller Effect capacitance results in small loops) to low frequency transistors (greater capacitance means larger loops).

Even bigger loops will be found in power transistors because their E, B, and C have to be larger to handle the currents. Their loops will be large provided their current gain (Beta) is good. But as a rule of thumb the gain of a power transistor is optimized to peak at high currents since it will be controlling high power. At very low currents a power transistor's current gain may not be very much resulting in a relatively small loop even though it has a large capacitance between base and collector.

The size of the loop is directly proportional to the capacitance between base and collector and to the Beta (current gain) of the transistor. So a transistor with low gain will have a smaller loop than a transistor with very high Beta all things being equal.

Dennis Tillman W7PF

-----Original Message-----
From: peter bunge
Sent: Friday, November 15, 2019 12:27 PM
Subject: Re: [TekScopes] Tek 576 noise and looping

<SNIP>
Thanks for your explanation which I tested by comparing curves with slow and fast transistors. I had some trouble trying to find a couple with matching gain but the high freq (900 MHz) transistor performed perfectly with nice curves at low current. On the other hand the slow transistor had huge loops.




--
Dennis Tillman W7PF
TekScopes Moderator