Date   
Re: SC503 storage mode

Mark Wendt
 

On 02/26/2016 04:02 AM, Malcolm Hunter malcolm.r.hunter@... [TekScopes] wrote:
​Here's a video of both issues, which I hope should explain them:

https://drive.google.com/open?id=0B2-hICzIGrsSM0J1VUlOQ0pGdmc
Hi Malcolm,

Get a 404 Not Found clicking on that link.

Mark

Re: Tek 465 intensity problem

Tom Gardner
 

On 15/02/16 19:00, Tom Gardner wrote:
On 14/02/16 11:18, aodiversen@... [TekScopes] wrote:

Hi Tom,
Thinking of Ed's rejuvenation experiments with his 576 CRT, I suggest you might test for connectivity of the grid electrode in a similar way. All you have to do is remove the CRT socket, supply the correct heater voltage to the CRT and check that the "circuit" between grid and kathode pins conducts as a diode. Use some tens of volts of DC supply and a large series resistor to make g positive w.r.t. k, and measure the current. A few uA max is enough to distinguish between connected and disconnected grid terminal.
Albert
Good idea; thanks.

Unfortunately using 50V confirms a disconnected grid. Oh well :(
I pulled the tube out yesterday and it is easy to see the
grid is disconnected. A spot weld between the external
wire and internal metalwork has failed, and they are
now 1mm apart.

The only silver lining is that the diagnostics suggested
and tried were accurate.

Otherwise the tube looks in good condition; no flakes etc.
I notice there are two oval silvered patches about 1.5" along
the neck of the tube, probably associated with pin 8
(a +55V anode?). Are they normal?

Is the tube any more than junk, now? If not, is there any
/safe/ way of destroying the envelope and keeping just the
pretty internal metalwork? Yes, I know I'm somewhat sad.

Re: 7834 horizontal jittery readout

Gordon <tekscopes@...>
 

On 25/02/2016 23:41, Gordon @GordonW [TekScopes] wrote:
Thanks and no, that's a 7633
Duh. While drifting off last night it dawned on me that the 7633 is going to be very similar.

Gordon

Re: SC503 storage mode

Malcolm Hunter
 

​Here's a video of both issues, which I hope should explain them:

https://drive.google.com/open?id=0B2-hICzIGrsSM0J1VUlOQ0pGdmc


On 26 February 2016 at 01:36, David @DWH [TekScopes] <
TekScopes@...> wrote:

I just bought an SC503, which seems to work mostly fine as an analogue
oscilloscope, but when I use the storage mode, the display initially goes
dark but then fills with colour. Any trace that goes across it leaves
virtually no mark on the screen. Pressing erase does trigger the trace and
it briefly leaves an imprint, but this is quickly obliterated. Is this a
sign of a faulty CRT or is there a fault or just a need for adjustment?
What affect does the view time adjustment have?

When the view time is set to maximum (long persistence and fully
clockwise), the background will be dark and when the view time is set
to minimum (fully counterclockwise but not in the off detent), the
background will be bright.

Your description sounds like the view time is set to minimum which may
be because the control is not working or some other fault.
​It's set to infinite. Reducing the view time doesn't help - it just
auto-erases faster and faster.

Also, the trace goes extremely dim with the timebase set to 5ms and above.
I'm hoping this is adjustable.
Do you mean "and slower" or "and faster"?

​5ms and slower.​
At 5ms/div and slower, there is a switch which lowers the beam
intensity to protect the CRT phosphor from being burned. This is a
common feature.

I suspect if you solve the view time problem above that this will be
fixed as well. Test the oscilloscope in non-storage mode.
I thought as much - but the trace is invisible in daylight.

Malcolm​

Re: 468 sweep offset issue

Brian Bloom
 

On another note, I just checked and my new pot for the B intensity/focus on the 7854 will be here tomorrow when I get home.

Re: 468 sweep offset issue

Brian Bloom
 

Well that didn't quite do it, but it definitely helped.

Now B sweep only "lags" A sweep by 0.5 minor divisions. It also ends the same amount before A sweep, where they were both ending at the same time. According to the measurements I took, the transistor-diode combinations are within 200uV of each other. The diode pair for A sweep is also within 200uV, as are Q760 and Q772.

At least I am learning something practical about this type of circuit and how it reacts to matched and unmatched PN junction drops and such. I forgot to check the resistors again, and will do that tomorrow after work.
I guess if what has to be done to get this aligned is to trim the A sweep start circuit, I'll do it and figure out how to make it look clean.. I'll have to buy the trimmer, though, so I might as well buy an entire set of them. For now, maybe there's one that will work from the 475 main board or trigger board that I have laying around.
Probably should check for any out of tolerance resistors and any other unmatched components from A to B sweep start/miller runup, etc.

Re: 468 sweep offset issue

 

On 25 Feb 2016 20:59:01 -0800, you wrote:

At some point all of the sockets were removed from this scope for "reliability reasons", so I have to desolder them regardless. I would go to the extent of picking diode drop values to compensate for any difference between the two sides of the dual transistor as well if I didn't have to go through hell to remove it for testing.
Sockets can be a reliability problem but some are better than others.

I almost always add collet pins wherever I removed a part even if it
is just a resistor or diode:
Where do you get those collet pins at? I'd love to get some to replace those garbage TI sockets that make contact on the side edges of the pins. Those things ... anger me.
It is too expensive to buy just the pins so I either buy machined
collet pin DIP sockets or SIPs and then cut the collet pins out if I
want to use them individually. Usually their outside diameter is just
right to slide into a plated through printed circuit board hole if it
is well cleaned of solder.

http://www.mouser.com/Connectors/IC-Component-Sockets/_/N-211lkZscv7?Keyword=collet

They can hold small signal diode, transistor, and resistor leads just
fine.

Re: 468 sweep offset issue

Bob Macklin <macklinbob@...>
 

If you ant good IC sockets buy either Mil-Max or Auget(SP?).

You can get them from Mouser or DigiKey.

Bob Macklin
K5MYJ
Seattle, Wa.
"Real Radios Glow In The Dark"

----- Original Message -----
From: analogaddict013@... [TekScopes]
To: TekScopes@...
Sent: Thursday, February 25, 2016 8:59 PM
Subject: [TekScopes] Re: 468 sweep offset issue




>>At some point all of the sockets were removed from this scope for "reliability reasons", so I have to desolder them regardless. I would go to the extent of picking diode drop values to compensate for any difference between the two sides of the dual transistor as well if I didn't have to go through hell to remove it for testing.

>Sockets can be a reliability problem but some are better than others.

>I almost always add collet pins wherever I removed a part even if it
>is just a resistor or diode:

Where do you get those collet pins at? I'd love to get some to replace those garbage TI sockets that make contact on the side edges of the pins. Those things ... anger me.

>>I'll just match the diodes for the A sweep start circuit, but for B sweep start I will match the transistors the best I can and then pick diodes to offset any remaining difference in Vbe as best I can.

>At the least it will make an improvement.

>Technically the output transistors should be matched as well as their
>offset also directly adds to the error. The parts list shows that
>Q760 and Q772 are also Tektronix part number 151-0188-00 or 2N3906s so
>that will be easy to do.

>Their operating current is variable so just use the same 5 milliamp
>test.

I'll go ahead and match those as well.

I'm taking a quick break to eat something, but I am going to get this finished up before I go to bed.

On a quick note, the first two transistors I stuck into the socket to test matched down to the 100uV!!
Both read 0.7792V across B-E

I checked the ones that I pulled out of the circuit, and they read:

0.7762V
0.7778V

I haven't checked the diodes yet, but I'm going to guess that they made it worse.

Re: 468 sweep offset issue

Brian Bloom
 

At some point all of the sockets were removed from this scope for "reliability reasons", so I have to desolder them regardless. I would go to the extent of picking diode drop values to compensate for any difference between the two sides of the dual transistor as well if I didn't have to go through hell to remove it for testing.
>Sockets can be a reliability problem but some are better than others.

>I almost always add collet pins wherever I removed a part even if it
>is just a resistor or diode:

Where do you get those collet pins at? I'd love to get some to replace those garbage TI sockets that make contact on the side edges of the pins. Those things ... anger me.

>>I'll just match the diodes for the A sweep start circuit, but for B sweep start I will match the transistors the best I can and then pick diodes to offset any remaining difference in Vbe as best I can.

>At the least it will make an improvement.

>Technically the output transistors should be matched as well as their
>offset also directly adds to the error. The parts list shows that
>Q760 and Q772 are also Tektronix part number 151-0188-00 or 2N3906s so
>that will be easy to do.

>Their operating current is variable so just use the same 5 milliamp
>test.

I'll go ahead and match those as well.

I'm taking a quick break to eat something, but I am going to get this finished up before I go to bed.

On a quick note, the first two transistors I stuck into the socket to test matched down to the 100uV!!
Both read 0.7792V across B-E

I checked the ones that I pulled out of the circuit, and they read:

0.7762V
0.7778V

I haven't checked the diodes yet, but I'm going to guess that they made it worse.

Re: 468 sweep offset issue

 

On 25 Feb 2016 19:35:01 -0800, you wrote:

...

I made a correction post regarding my assessment of the transistor part numbers for Q958 and Q960. They are actually just 151-0188-00's - just plain 2n3906s.
I left it up to you to check the Tektronix part numbers in the service
manual. If they are just plain 2N3906s, then that makes things easy.

At some point all of the sockets were removed from this scope for "reliability reasons", so I have to desolder them regardless. I would go to the extent of picking diode drop values to compensate for any difference between the two sides of the dual transistor as well if I didn't have to go through hell to remove it for testing.
Sockets can be a reliability problem but some are better than others.

I almost always add collet pins wherever I removed a part even if it
is just a resistor or diode:

http://s10.postimg.org/eovyx5y9l/DC505_Lowest_Profile_Collet_Socket_2a.jpg

I'll just match the diodes for the A sweep start circuit, but for B sweep start I will match the transistors the best I can and then pick diodes to offset any remaining difference in Vbe as best I can.
At the least it will make an improvement.

Technically the output transistors should be matched as well as their
offset also directly adds to the error. The parts list shows that
Q760 and Q772 are also Tektronix part number 151-0188-00 or 2N3906s so
that will be easy to do.

Their operating current is variable so just use the same 5 milliamp
test.

Re: 468 sweep offset issue

Brian Bloom
 

---In TekScopes@..., <davidwhess@...> wrote :

On 25 Feb 2016 18:44:31 -0800, you wrote:

>After a little research on transistor matching, I found a site that suggests setting it up as follows:
>
>+15V to 8.2K resistor then to Emitter
>-15V to collector
>uA meter at base to ground..
>(for PNP)
>
>Is it really that simple?
>
>David, is this what you are suggesting I do, but set the current limit to 5mA from the PSU?
>
>Additionally, couldn't I set up a voltage divider at the base to get better resolution? Would that be overkill? I do have a uA capable Fluke 87III.

>There is no need to make it that complicated. I would match the
>forward voltage drop of the base-emitter junction without connecting
>the collector and I would use 5.6 volts through a 1 kilohm resistor.
>This avoids the problem where you accidentally connect the emitter and
>base backwards causing the junction to breakdown which is bad for the
>transistor; the base-emitter breakdown voltage is higher than 5.6
>volts.

>The same test is used for the diodes.

>The absolute value of the current is not important as long as it does
>not drift and is fairly close to the operating current.

>Also read my other post where I suggest deliberately mismatching the
>diodes to compensate for the mismatch in the transistor Vbe values. If
>you do this, then the existing transistors can be left in the circuit.

>When I have done this sort of thing in the past, I have had no trouble
>getting better than a 10 millivolt match and usually I get within 2
>millivolts.

We must have posted at the same time. I saw your post regarding how to do this only after I had posted the method I had found. I'm definitely using the method that you've described.

I made a correction post regarding my assessment of the transistor part numbers for Q958 and Q960. They are actually just 151-0188-00's - just plain 2n3906s.
At some point all of the sockets were removed from this scope for "reliability reasons", so I have to desolder them regardless. I would go to the extent of picking diode drop values to compensate for any difference between the two sides of the dual transistor as well if I didn't have to go through hell to remove it for testing.

I'll just match the diodes for the A sweep start circuit, but for B sweep start I will match the transistors the best I can and then pick diodes to offset any remaining difference in Vbe as best I can.

Re: 468 sweep offset issue

 

On 25 Feb 2016 18:22:27 -0800, you wrote:

With everything set up as per the manual instructions for DC voltage testing:

Q998 Collector: 12.86V
Q967 Collector: 12.72V

Q890 Emitter: -4.655V
Q864 Emitter: -4.726V

There's that 180mV to 200mV or so...

I'm going to follow David's instructions for what to match & replace.
Additionally, I will just replace those 36ohm resistors with new ones while I'm in there with the iron.
The offset between output transistors Q760 and Q772 also adds to it. I
am sorry I did not think to mention that earlier but compensating for
it would just make things more complicated.

Still, reducing the offset in the comparators should help a lot and
mostly only requires matching the diodes.

Re: 468 sweep offset issue

 

On 25 Feb 2016 18:44:31 -0800, you wrote:

After a little research on transistor matching, I found a site that suggests setting it up as follows:

+15V to 8.2K resistor then to Emitter
-15V to collector
uA meter at base to ground..
(for PNP)

Is it really that simple?

David, is this what you are suggesting I do, but set the current limit to 5mA from the PSU?

Additionally, couldn't I set up a voltage divider at the base to get better resolution? Would that be overkill? I do have a uA capable Fluke 87III.
There is no need to make it that complicated. I would match the
forward voltage drop of the base-emitter junction without connecting
the collector and I would use 5.6 volts through a 1 kilohm resistor.
This avoids the problem where you accidentally connect the emitter and
base backwards causing the junction to breakdown which is bad for the
transistor; the base-emitter breakdown voltage is higher than 5.6
volts.

The same test is used for the diodes.

The absolute value of the current is not important as long as it does
not drift and is fairly close to the operating current.

Also read my other post where I suggest deliberately mismatching the
diodes to compensate for the mismatch in the transistor Vbe values. If
you do this, then the existing transistors can be left in the circuit.

When I have done this sort of thing in the past, I have had no trouble
getting better than a 10 millivolt match and usually I get within 2
millivolts.

Re: 7834 horizontal jittery readout

 

R749 on the 7904 and R1749 on the 7834 are part of the vertical
amplifier. Gordon says the problem is horizontal jitter.

On 25 Feb 2016 17:46:51 -0800, you wrote:

Hi Gordon,

My 7904 manual says to adjust LF Comp R749 for minimum jitter. Your scope has this adjustment as R1749, but the adjustment is not stated in the 7834 manual.

Dallas

---In TekScopes@..., <tekscopes@...> wrote :

Hi All. Just trying to sort out a jittery readout on my 7834. The ramp
outputs of the horizontal board look very stable and the trace itself
doesn't appear to be jittery, at least nothing like the readout. In
inches I'd say it's jittering about 0.040" maybe a bit less but only
horizontally. The supplies don't appear to have significant ripple and
putting an extra 4.7u on the 130V doesn't make any difference.

Checking TP12 and 13 on the readout board (x and y out) reveals a rather
odd waveform http://www.mgcsoft.com/tek/7834_tp12_readout_bd.jpg. http://www.mgcsoft.com/tek/7834_tp12_readout_bd.jpg. The
waveforms in the manual aren't really clear enough. Could someone
confirm whether or not the waveform is something like normal please.

Cheers

Gordon

Re: 468 sweep offset issue

Brian Bloom
 

->Hey Brian,

>nice - that'll do it. David knows his stuff :).
As a tip, you get more resolution from your DMM if you simply measure
across the two points. That way your DMM does the math-ing, and it can do
it at the 200mV or 2V scale. Doing two measurements on the 20V scale and
subtracting is extra effort plus the loss of one or two orders of magnitude
of precision - not that it matters here.


I see what you're saying... just look at the difference between the two points - say Q998-C and Q967-C and just measure the difference between the two.. I didn't think of that.



After a little research on transistor matching, I found a site that suggests setting it up as follows:

+15V to 8.2K resistor then to Emitter
-15V to collector
uA meter at base to ground..
(for PNP)

Is it really that simple?

David, is this what you are suggesting I do, but set the current limit to 5mA from the PSU?

Additionally, couldn't I set up a voltage divider at the base to get better resolution? Would that be overkill? I do have a uA capable Fluke 87III.

Re: 468 sweep offset issue

 

On 25 Feb 2016 17:45:29 -0800, you wrote:

---In TekScopes@..., <davidwhess@...> wrote :

The diodes only need to be matched for forward voltage drop in pairs;
you do not need 4 matched diodes. The 2 discrete transistors in the B
sweep comparator should also be matched for Vbe. Ignore Q987 in the A
sweep comparator for now since as a dual transistor, it should be well
matched anyway.
Use a 5 milliamp test current for matching. See below.
----- Matching the diodes will be easy enough since I have plenty. The transistors are supposed to be 151-0220-03, which are selected 2n3906. Do modern 2n3906's work without being selected ones, or is there some kind of parameter set I should use to pick out the ones I should use? I do have a few 151-0220-03's from a 465 parts scope, but there's only 5 good ones total.
Based on the Tektronix parts book, it looks like they selected them
for a high Ft of 600 MHz minimum or rb'Cc < 50 picoseconds. I wonder
how they did that. rb'Cc is the collector-base time constant.

This Agilent technical note has a good description of it:

https://web.fe.up.pt/~hmiranda/etele/trans_primer1.pdf

Unless you or someone knows a simple way to measure the collector-base
time constant, I think there are three good options. In order of
cost:

1. Ignore the transistors for now and just match the pairs of diodes.
2. Match some 2N3906s for Vbe at 5 milliamps and try them out.
3. Match some PN4917s for Vbe at 5 milliamps. The PN4917 has the
specified collector-base time constant and other characteristics:

http://www.mouser.com/search/ProductDetail.aspx?R=0virtualkey0virtualkeyPN4917
https://www.fairchildsemi.com/datasheets/PN/PN4917.pdf

If you want to be clever, you could match each diode and Vbe series
pair to its opposite diode and Vbe series pair so that the diodes
compensate for the mismatched Vbe of the transistors. I guess the
easy way to do that would be to measure the mismatch in the transistor
Vbe and then find a pair of diodes that have the opposite mismatch.

Never checked a transistors specs before, actually... only JFETs. I made rigs for those. I'll need to look up how to do it.

Also, it's apparent that the four 36ohm resistors in that circuit have been replaced. I checked what I could in circuit, and none that I was able to measure were out of tolerance. Then again, spec tolerance being 5% on many of these resistors could easily cause a voltage difference of tens or hundreds of mV.
Operating current for the comparators is about 10 milliamps total so
the voltage drop across each resistor is 5 mA * 36 Ohms = 180
millivolts. Worst case tolerances would make that 18 millivolts of
difference which is pretty small compared to the typical mismatch in
the diodes and B sweep transistors. It would not hurt to replace them
with 1% metal film resistors. The values are not critical.
--- How convenient.. I just picked up a couple of huge 135 value variety packs from resistorland. I figured I might as well since even if I need a value I don't have, I'll certainly be able to make a temporary series or parallel replacement. I'll go ahead and replace them if they aren't already within 1% - If I remember correctly, they were in the 1% margin even though they are 5% ones.
The values are not critical as long as they match. Anywhere from 33
to 39 ohms should work fine.

When I need a 1% resistor which I do not have, I buy 100 and add
another marked drawer to my parts cabinet.

Since the changes I suggest above are pretty easy to do, I would try
them.
I'll go ahead and do them as soon as I learn how to match a pair of transistors by their Vbe. :-)
I usually just use the diode mode on my multimeter but in this case we
know the exact operating current. Set your variable power supply to
5.6 volts and put a 1 kilohm resistor in series with it. That will
produce 5 milliamps through the base emitter junction which you can
measure with your voltmeter.

The low voltage will prevent damage if you connect the emitter-base
junction backwards. Breakdown will usually be about 7 volts.

---- No, I bet you are thinking of the topic "465B Degraded Bandwidth". I read it all, and it appears that topic abruptly ended without the OP putting out any test results or anything. Never even said if they managed to fix it or not. Someone did reply to them and rattle off a few items to check, which I wrote down and will try.

The signal source is a 191 leveled sine generator, which is the manual recommended gear. I also have the 65Mhz to 500Mhz 067-0532-01, but that's just another leveled sine gen. Other than my 2 not-properly-calibrated 106 square wave gens (rated <1ns), the best I've got puts out approx 10ns risetime.

I read about the 3.5ns test briefly, but forget how to perform it. I could try it with both of my poorly cal'd 106s.
Your 106's will probably work fine for this. Just apply the fast
rise/fall signal to the oscilloscope through a 50 ohm terminated
cable, set the sweep speed as fast as necessary, and visually measure
the rise/fall time between 10% and 90%. For a 468 and most other 100
MHz oscilloscopes, it should be 3.5 nanoseconds:

bandwidth = 0.35 / (transition time)

VGA connector as in old school PC monitor type VGA connector?? I could pull out one of my old desktops and stick a video card in it if I have one with VGA out, if that's even what you mean..
That is the one. If you have a DVI-I output which is more modern, VGA
vertical sync is pin 8 and horizontal sync is C4:

https://en.wikipedia.org/wiki/Digital_Visual_Interface

Re: Need Help Fixing A Couple of 2445 Scopes

Siggi
 

On Thu, 25 Feb 2016 at 18:25 gohim@... [TekScopes] <
TekScopes@...> wrote:



I spent some more time looking on the Web for a Service Manual for the
@$$% GPIB Option, and all I coiuld find was the Service Manual for 24X5A
Options, so I went ahead and dowloaded,it figuring that I could compare it
to what my later 2445 contained, and maybe it would be the same.

It may be worth your while to check out Artek Manuals <
http://artekmanuals.com/manuals/tektronix-manuals/>. Dave generally has
excellent scans, and it looks like the price of the 2445 bundle is below
the price of each manual individually - Dave?

Siggi

Re: 468 sweep offset issue

Brian Bloom
 

With everything set up as per the manual instructions for DC voltage testing:

Q998 Collector: 12.86V
Q967 Collector: 12.72V

Q890 Emitter: -4.655V
Q864 Emitter: -4.726V


There's that 180mV to 200mV or so...

I'm going to follow David's instructions for what to match & replace.
Additionally, I will just replace those 36ohm resistors with new ones while I'm in there with the iron.

Re: 468 sweep offset issue

Siggi
 

On Thu, 25 Feb 2016 at 21:06 analogaddict013@... [TekScopes] <
TekScopes@...> wrote:

With everything set up as per the manual instructions for DC voltage
testing:


Q998 Collector: 12.86V
Q967 Collector: 12.72V

Q890 Base: -4.655V
Q864 Base: -4.726V


There's that 180mV to 200mV or so...

I'm going to follow David's instructions for what to match & replace.
Additionally, I will just replace those 36ohm resistors with new ones
while I'm in there with the iron.
Hey Brian,

nice - that'll do it. David knows his stuff :).
As a tip, you get more resolution from your DMM if you simply measure
across the two points. That way your DMM does the math-ing, and it can do
it at the 200mV or 2V scale. Doing two measurements on the 20V scale and
subtracting is extra effort plus the loss of one or two orders of magnitude
of precision - not that it matters here.

Siggi

Re: 468 sweep offset issue

Brian Bloom
 

Maybe I should have my eyes checked. Q958 and Q960 are just 2n3906s.

I still wonder about the specs for 151-0220-03's.

Parts matching and after pics to come.....


With everything set up as per the manual instructions for DC voltage testing:

Q998 Collector: 12.86V
Q967 Collector: 12.72V

Q890 Emitter: -4.655V
Q864 Emitter: -4.726V


There's that 180mV to 200mV or so...

I'm going to follow David's instructions for what to match & replace.
Additionally, I will just replace those 36ohm resistors with new ones while I'm in there with the iron.