Date   
Re: Tek and Darlingtons...

Ed Breya
 

Gary wrote: "Please remember that integrated pairs track real good with temperature as
long as you keep the same initial current in each one."

Yes, the same current in equal devices is the best condition for straight-up mirroring. In my application I'm trying to divide the current by a large factor, using an array with N "identical" transistors, with the unknown emitter (all paralleled) current going through them (paralleled B&C), except for one with its collector free, carrying 1/N of the current, which is the desired signal. So, each transistor will carry about 1/N of the current. In an IC design, this would be easy, by using a whole bunch of equal transistors, or by scaling their sizes in proportion, but I'm trying to implement it with off-shelf arrays.

So far the biggest array I've found with suitable arrangement is the CA3081 - 7 NPNs with common-emitter, so I can divide the input emitter current by 7. The output will be reflected in a PNP mirror, made from an MPQ3906. (It's even harder to find PNP arrays, and I think a quad is the biggest available.) The PNP is likewise set up to divide by 4, and turn the direction, but because of this, its output collector current departs and isn't included in the net, so the end result is divide by 3. That's OK - the NPN emitter current is what counts - but it reduces the available divide ratio. Overall, whatever current is pulled from the CA3081 emitters, 1/21th of that will be sourced from one of the MPQ3906 collectors. All of the transistors will operate at more or less constant voltage, so Early voltage compensation isn't needed, keeping it simple. The NPN output will load into the PNP mirror, while the PNP output will go to a P-channel MOSFET in cascode.

I would prefer a higher divide ratio, like 30-50, but the 21 is as high as I can get it with readily available parts. Ideally, arrays with more transistors would do the trick. If anyone is wondering what's the big deal about the divide ratio, it's about power. The mirror circuit will float at about +350V, and be self-biased with a small (few V) voltage drop from there. Current drawn from the NPN emitters needs to be accurately divided and reflected out of the PNP, and delivered via the cascode MOSFET to ground-referenced circuits. The exact divide ratio doesn't matter, since everything will be scaled later, so it just needs to be consistent over a wide current range (0-100 mA or so), and reasonably high to conserve power in the source, and limit dissipation in the MOSFET when it's near full scale.

This all may seem rather complicated, but topologically, it's actually quite simple, if I can get good enough tracking between the devices in the arrays, without needing degeneration resistors. It would just be the two arrays, MOSFET, and a few support parts for biasing and protection, which also support another section of amplifier circuitry. I started out planning to use a MAX471 or 472 type part as the scaling device, but then gravitated to current mirrors of various sorts, with the appeal of relative simplicity. If the array type setup I envision doesn't work, there are of course a number of other current mirror topologies or opamp circuits that will do, so I'm not too worried about making something work. If I could just find some arrays with a couple more transistors, I think it will work out nicely. With what I have so far, it should work, but not as nicely as I had hoped.

If anyone is wondering why power and simplicity are so important here, it's because this is for my 576 vacuum tube curve tracer plug-in, which has very limited power and space to provide all the stuff I want it to do. This circuit is to provide the option of measuring screen current vs plate voltage, by first getting the current signal, then performing some trickery in the way the 576 measures stuff with the signals that pass through the test fixture. I'm not yet sure if it's possible, but my gut instinct says yes - I have to study the 576 a bunch more, but I'm confident enough that I'm putting the current measurement circuit in.

So, after all these gyrations, since it's about the 576 and tubes, this is on-topic after all. Started with Darlingtons, then led to the 576 - which could also check them.

Ed

Re: Tek and Darlingtons...

Gary Robert Bosworth
 

Please remember that integrated pairs track real good with temperature as
long as you keep the same initial current in each one.

Gary

On May 17, 2017 5:12 AM, "David @DWH [TekScopes]" <
TekScopes@...> wrote:



Oh, they have their place and especially in integrated form and in
power devices. I am just dubious about small signal darlingtons and
for whatever reason, they seem to have a higher failure rate than
other transistors.

The question of matched integrated transistors is completely different
and there is a dearth of suitable parts. There are a lot of dual
transistors and transistors arrays which basically are not matched at
all

Diodes Incorporated makes some matched pairs intended for current
mirror applications. NXP makes matched pairs. But they are all only
available in surface mount.

National (TI) still makes the LM3046 and it is inexpensive. THAT
Corporation makes NPN and PNP quads in DIP packages for audio
applications but they are expensive.

For current mirrors I usually end up matching the Vbe of 2N3904s and
2N3906s and gluing them together.

On 16 May 2017 18:15:00 +0000, you wrote:

Actually, Darlingtons are alive and well, and all over the place - but
mostly in IC form. Many peripheral driver ICs use the configuration,
including integrated resistors and clamp diodes that you would normally
have to add for discrete devices. Look at the ULN2000 family, for example.

I recently looked at a number of transistor arrays and drivers, trying to
find a bunch of single transistors on a common die, for a current mirror
application - I had a hard time finding anything that was NOT Darlington in
the modern driver IC class. I had to settle for good old CA-series arrays.
Another common topology is a PNP driving an NPN, or vice versa.

Ed

Re: Starting point suggestion: 475 with a Jittery Trace

 

I would use a new radial lead cap. There is a very high probability you will pull track and the plated through holes out when you remove the bad cap. Install jumpers where the can made the circuit on the back side of the board.

Regards

----- Original Message -----
From: mark2382@... [TekScopes]
To: TekScopes@...
Sent: Wednesday, May 17, 2017 2:37 AM
Subject: [TekScopes] Re: Starting point suggestion: 475 with a Jittery Trace



Poked around, and C1412 is a goner. 48k ohm and <100nf... Should be 350uf....

It bridges between the 50+ unreg and the +105/+160 that is later used as an input for the +110V.

Not sure about the -15 noise. The e-cap there checks out.


I have apparently a good C1412 in my doner machine. Generally, I would go for a new one, but the 3 points of contacts makes me think twice... Recommendations?


Mark

Re: Starting point suggestion: 475 with a Jittery Trace

Mark Hatch
 

Tack another capacitor in parallel worked c1412 worked! All the ghosts are gone.

Still need to go back and check the voltage ripple on my scope. But definite progress!


Thanks!


Mark

Re: Starting point suggestion: 475 with a Jittery Trace

Tom Gardner
 

On 17/05/17 15:26, mark2382@... [TekScopes] wrote:


What do people typically use to replace those e-caps? The cans are going to be difficult to match with their three mounting point structure. Can I just put in a modern 2 point radial? Will not be as secure as the original. (I am assuming that the two outer points are just redundant and that I only really need to connect one of them...)
People either bodge it, or make a standoff "converter" PCB.

Sometimes the capacitor can is used to connect two PCB tracks, so make sure those are duplicated.

I recommend taking photos before starting, so as to have something to refer back to.

A temporary "fix" can be to simply tack another capacitor in parallel, and see if that solves any problem.

Re: Starting point suggestion: 475 with a Jittery Trace

Mark Hatch
 

David,

Thanks for the advice and confirmation. The +110 tp that I am measuring is on the bottom board of the scope, near the +50 reg, +15, etc. and is on schematic 11.


What do people typically use to replace those e-caps? The cans are going to be difficult to match with their three mounting point structure. Can I just put in a modern 2 point radial? Will not be as secure as the original. (I am assuming that the two outer points are just redundant and that I only really need to connect one of them...)


Do I need to go for special low noise types or just a good quality?




Thanks again!


Mark

Re: Tek and Darlingtons...

 

Oh, they have their place and especially in integrated form and in
power devices. I am just dubious about small signal darlingtons and
for whatever reason, they seem to have a higher failure rate than
other transistors.

The question of matched integrated transistors is completely different
and there is a dearth of suitable parts. There are a lot of dual
transistors and transistors arrays which basically are not matched at
all

Diodes Incorporated makes some matched pairs intended for current
mirror applications. NXP makes matched pairs. But they are all only
available in surface mount.

National (TI) still makes the LM3046 and it is inexpensive. THAT
Corporation makes NPN and PNP quads in DIP packages for audio
applications but they are expensive.

For current mirrors I usually end up matching the Vbe of 2N3904s and
2N3906s and gluing them together.

On 16 May 2017 18:15:00 +0000, you wrote:

Actually, Darlingtons are alive and well, and all over the place - but mostly in IC form. Many peripheral driver ICs use the configuration, including integrated resistors and clamp diodes that you would normally have to add for discrete devices. Look at the ULN2000 family, for example.

I recently looked at a number of transistor arrays and drivers, trying to find a bunch of single transistors on a common die, for a current mirror application - I had a hard time finding anything that was NOT Darlington in the modern driver IC class. I had to settle for good old CA-series arrays. Another common topology is a PNP driving an NPN, or vice versa.

Ed

Re: Starting point suggestion: 475 with a Jittery Trace

 

On 17 May 2017 05:52:38 +0000, you wrote:

As suggested by David.... Measured voltage ripple, should have done that first...

Results pretty telling...

Supply Tolerance Tolerance Actual Ripple (p-p) Actual 110 3% 3.3v 102.4 1V 40V (photo) 50 0.50% 250mv% 49.88 2mv 2mv 15 1.50% 225mv 14.95 2mv 2mv 5 1.50% 75mv 4.98 2mv 2mv -15 1.50% 225mv -14.9 2mv 5mv (photo) -8 1.50% 120mv -7.97 2mv 2mv 50(unreg) 67 1.5v

Couldn't find a standard for 50V unregulated. But that seemed high both DC and ripple.
The +50V unregulated is just that, unregulated. What is important is
that it never drop low enough to affect the +50V regulated output

Link to +100 photo of 40V P-P https://groups.yahoo.com/neo/groups/TekScopes/photos/photostream/lightbox/963857539?orderBy=mtime&amp;sortOrder=desc&amp;photoFilter=ALL#zax/963857539
Is that the +110V test point at the bottom of schematic 11?

I assume so and obviously that is a problem. You say C1412 is
basically open and that explains it.

Link to -15 movie with strange 5mv p-p ripple https://groups.yahoo.com/neo/groups/TekScopes/photos/photostream/lightbox/1985584798?orderBy=mtime&amp;sortOrder=desc&amp;photoFilter=ALL#zax/1985584798
That sure looks odd. My guess is that it is an interaction with the
spike in the +110V regulated output; some other circuit is breaking
into oscillation.

Any chance that this is just capacitors? The 40V p-p seems to be likely more and neither the -15 nor 50V unregulated look good....
On 17 May 2017 06:37:46 +0000, you wrote:

Poked around, and C1412 is a goner. 48k ohm and <100nf... Should be 350uf....

It bridges between the 50+ unreg and the +105/+160 that is later used as an input for the +110V.
It is kind of an odd way to do it but this allows C1412 to have the
same voltage rating as C1414 instead of double.

Not sure about the -15 noise. The e-cap there checks out.

I have apparently a good C1412 in my doner machine. Generally, I would go for a new one, but the 3 points of contacts makes me think twice... Recommendations?

Mark
I would not bother with a donor part since it will be at the end of
its life anyway.

Once I gained access, I would change all of the aluminum electrolytic
capacitors but that is up to you. Replacing C1412 will likely be good
enough. Of you could just temporarily put a new capacitor in parallel
with the old C1412 for now if you do not want to risk removing parts.

Re: Starting point suggestion: 475 with a Jittery Trace

Mark Hatch
 

Poked around, and C1412 is a goner. 48k ohm and <100nf... Should be 350uf....

It bridges between the 50+ unreg and the +105/+160 that is later used as an input for the +110V.

Not sure about the -15 noise. The e-cap there checks out.


I have apparently a good C1412 in my doner machine. Generally, I would go for a new one, but the 3 points of contacts makes me think twice... Recommendations?


Mark

Re: 43 year old 7S14 Dual Trace Dual Timebase Sampling Plugin still works

Dave Brown <tractorb@...>
 

Even before that there were button bypass caps- as used in a number of
VHF/UHF WW2 radios. Small and round with I think a mica dielectric. They
usually had a threaded hole in the base 'plate' to screw mount them to the
chassis or other grounded part of the set, the top terminal typically was a
small solder tag.
DaveB, NZ

-----Original Message-----
From: TekScopes@... [mailto:TekScopes@...]
Sent: Wednesday, May 17, 2017 2:48 PM
To: TekScopes@...
Subject: [TekScopes] Re: 43 year old 7S14 Dual Trace Dual Timebase Sampling
Plugin still works

Those caps I think you're referring to are called "leadless" caps - an apt
name - before surface mount devices became common. Picture the manufacturing
steps to make a radial lead cap, but instead of attaching leads and
overcoating, you just have a ceramic disk with metallized pads on each side.
No leads = no lead inductance, so they were very good for RF and high speed
stuff. Tek used these a lot in the early days. They could be mounted flat on
a ground plane or transmission line surface, or mounted vertically in a slot
in the board or on air-suspension wiring. I would imagine they are still
made (in the same form - not the same as SMT) and used for the same reasons
now as then.

I'd guess that the lifetime of the Hg cells depended mostly on the leakage
of the particular sampler diodes installed, and on storage conditions.
Clean, cool, and dry is good. BTW I noticed back when I studied the 7S14
extensively, that somewhere along the line the Hg cell size was changed from
something smaller to something around 225 mAh, as I recall. It could have
been due to part availability issues, but I'm guessing the small ones didn't
last long enough, so were upped significantly in size to cover the diode
leakage and environment conditions. I once did a rough estimate of how long
a 225 mAh cell should last, based on a SWAG of what the diode leakage was at
normal storage temperatures, and came up with about thirty years, as I
recall, so I could picture your numbers being real.

Also, do you know for sure that the cells are the original ones?

Ed







------------------------------------
Posted by: edbreya@...
------------------------------------


------------------------------------

Yahoo Groups Links

Re: Starting point suggestion: 475 with a Jittery Trace

Mark Hatch
 

As suggested by David.... Measured voltage ripple, should have done that first...

Results pretty telling...


Supply Tolerance Tolerance Actual Ripple (p-p) Actual 110 3% 3.3v 102.4 1V 40V (photo) 50 0.50% 250mv% 49.88 2mv 2mv 15 1.50% 225mv 14.95 2mv 2mv 5 1.50% 75mv 4.98 2mv 2mv -15 1.50% 225mv -14.9 2mv 5mv (photo) -8 1.50% 120mv -7.97 2mv 2mv 50(unreg) 67 1.5v


Couldn't find a standard for 50V unregulated. But that seemed high both DC and ripple.


Link to +100 photo of 40V P-P http://https://groups.yahoo.com/neo/groups/TekScopes/photos/photostream/lightbox/963857539?orderBy=mtime&amp;sortOrder=desc&amp;photoFilter=ALL#zax/963857539


Link to -15 movie with strange 5mv p-p ripple https://groups.yahoo.com/neo/groups/TekScopes/photos/photostream/lightbox/1985584798?orderBy=mtime&amp;sortOrder=desc&amp;photoFilter=ALL#zax/1985584798 http://https://groups.yahoo.com/neo/groups/TekScopes/photos/photostream/lightbox/963857539?orderBy=mtime&amp;sortOrder=desc&amp;photoFilter=ALL#zax/963857539

Any chance that this is just capacitors? The 40V p-p seems to be likely more and neither the -15 nor 50V unregulated look good....


Thanks


Mark



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

To me that looks like a problem with ripple on the low voltage power
supply outputs.

And checking the low voltage power supply outputs for level, ripple,
and noise is where to start anyway.

On 16 May 2017 05:11:12 +0000, you wrote:

>Just got a 475 with a trace problem. Looks like every other refresh is a little off the prior one.
>
> See photo:
>
> https://groups.yahoo.com/neo/groups/TekScopes/photos/albums/1766650730/lightbox/1894781410?orderBy=ordinal&sortOrder=asc&photoFilter=ALL#zax/1894781410 https://groups.yahoo.com/neo/groups/TekScopes/photos/albums/1766650730/lightbox/1894781410?orderBy=ordinal&sortOrder=asc&photoFilter=ALL#zax/1894781410
>
> Any suggestions on where to start isolating this? I do have a donor set of boards, so I could do some swapping to isolate things.
>
> FYI: I have not checked the power supply values....
>
> Thanks
>
> Mark

Re: 43 year old 7S14 Dual Trace Dual Timebase Sampling Plugin still works

Ed Breya
 

Those caps I think you're referring to are called "leadless" caps - an apt name - before surface mount devices became common. Picture the manufacturing steps to make a radial lead cap, but instead of attaching leads and overcoating, you just have a ceramic disk with metallized pads on each side. No leads = no lead inductance, so they were very good for RF and high speed stuff. Tek used these a lot in the early days. They could be mounted flat on a ground plane or transmission line surface, or mounted vertically in a slot in the board or on air-suspension wiring. I would imagine they are still made (in the same form - not the same as SMT) and used for the same reasons now as then.

I'd guess that the lifetime of the Hg cells depended mostly on the leakage of the particular sampler diodes installed, and on storage conditions. Clean, cool, and dry is good. BTW I noticed back when I studied the 7S14 extensively, that somewhere along the line the Hg cell size was changed from something smaller to something around 225 mAh, as I recall. It could have been due to part availability issues, but I'm guessing the small ones didn't last long enough, so were upped significantly in size to cover the diode leakage and environment conditions. I once did a rough estimate of how long a 225 mAh cell should last, based on a SWAG of what the diode leakage was at normal storage temperatures, and came up with about thirty years, as I recall, so I could picture your numbers being real.

Also, do you know for sure that the cells are the original ones?

Ed

Re: 43 year old 7S14 Dual Trace Dual Timebase Sampling Plugin still

Kevin Oconnor
 

It does not surprise me that some of these mercury cells could last
that long in an essentially no load application. Some batches were
probably manufactured "just right".

Regarding true mercury cells, I think they can live a long time on the shelf. I have several pocket radiation monitors from the late 40's likely. They look like cigar cases. Recap some have micro tubes other transistors. They use some line of relaxation oscillator like a camera flash for the detector. All are powered by 4.5v mercury battery. I can't remember the code, but like a bloated AA battery. The box these came in contained a NOS battery, and the last time I measured it it was still 4.5v. No idea when it was made though.
Oh, and those monitors still work if powered and exposed to ionizing radiation!

Kjo

Sent from kjo iPhone

Re: 2235 Trigger on one channel

Tom Jobe <tomjobe@...>
 

There are some new CA3102E arrays on eBay.
I believe Kevin is in the UK.
Here's one from Switzerland:
http://www.ebay.com/itm/CA3102E-IC-RCA-2x-5962-01-093-4610-/302278280934?hash=item46613086e6:g:yi0AAOSwnF9Y6mS7
Two of several from the US:
http://www.ebay.com/itm/INTERSIL-HARRIS-CA3102E-DIP-14-Dual-High-Frequency-/361788292835?hash=item543c433ee3:g:lXEAAOSwcLxYDh5H
http://www.ebay.com/itm/CA3102E-Original-New-Harris-Integrated-Circuit-/121052570211?hash=item1c2f4ba263:m:m-2FOnEaLRPmCo6nwF1MBzg
There are many available from China, some of which look like they might
not be fake.
http://www.ebay.com/itm/1PCS-CA3102E-Dual-High-Frequency-Differential-Amplifier-/221115321652?hash=item337b800d34:m:mdO4W1kfigq6oNKBsBWxfMw





On 5/16/2017 9:02 AM, David @DWH [TekScopes] wrote:

On 16 May 2017 14:03:52 +0000, you wrote:

Hi David

Been a little busy so took a while to get back to this. So today
lunch time I started to measure some voltages with my DMM as
suggested, results as follows, something a little amiss (I think?)

U555B Output pin 6 ch1 = 0.139V (0.2V) (0.06V error)
U555B Output pin 6 ch2 = 1.476V (2.5V) (1.02V error)

U310 Output pin 1 ch1 = -3.26V (-3.4V) (0.14V error)
U310 Output pin 1 ch2 = -2.443V (-2.1V) (0.34V error)
You can check the values of R314, R315, R317, R318, and R319 around
U310 but I think the answer is going to be that U310 is bad.

The signal into pin 1/11 is suppose to be low when channel 1 is
selected. When that happens, the output of U555B is disconnected
(pins 1/11 of U310 should not draw any current) and should reach all
the way to ground which is what we are seeing and U310 should be
working but it is not.

When the signal into pin 1/11 is high, channel 2 is selected but pin
1/11 draws more current from U555B. In this case, it is drawing too
much current indicating that either U310 is damaged or R314 and/or
R315 are open.

The fact that U555B is not reaching its full output is irrelevant
because under that condition, U310 should not be forwarding the
trigger signal anyway.

I also looked at the feed from the A3 Board Connection as follows:-

Pin 18 (6.2V) = 6.5V (0.3V error), this also shows (0V) for XY but
can't see how to check this on front panel.
Pin 32 (0.0V) = 0V
Pin 33 (5.0V) = 4.996V
Pin 34 (5.0V) = 4.996V

Some of the values are slightly higher / lower than they should be,
is this the issue?

No, nothing is wrong there.

I wonder if there are any good sources for CA3102 transistor arrays. I
do not see any so if you cannot find one, blast me an email with your
mailing address and I should be able to mail one to you. I will have
to pull it from a junked board but will test it somewhat.



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

Re: Deane Kidd's Estate

Tom Gardner
 

On 16/05/17 20:11, Jochen Feldhaar jochen_feldhaar@... [TekScopes] wrote:

Another case, I was given a big lot of Ham rigs by the relatives of a
deceased Ham, and they were astonished when I gave them the money (make
that thunderstruck) I got when I sold most of the stuff (at reasonable
prices) at a Ham fest.
The local hackspace was recently given two scopes
(2245, DM63) in similar circumstances. They didn't want
money but did want to know that they wouldn't be
flogged off.

My parents have a stipulation that some of their
stuff which we don't want has to be sold "in a
reputable auction".

Re: Deane Kidd's Estate

Jokken Feldhaar
 

Hi group,

it gave me a fit of sorts when a very active HAM friend died of cancer,
he was single most of his life and had accumulated a lot of Ham
equipment, especially magnetic antennas. When he passed away, his
relaitves cleared out his flat, and some Hams of our local club were
notified.

So, the situation arose where the relatives thanked the Hams for
carrying out stuff, lots of stuff, all very well maintained. They were
happy that the dumpster was not filled with his electronics.

I still have a case of bad conscience, thinking of the relatives
thanking us for carrying off his stash of fine electronice.

Another case, I was given a big lot of Ham rigs by the relatives of a
deceased Ham, and they were astonished when I gave them the money (make
that thunderstruck) I got when I sold most of the stuff (at reasonable
prices) at a Ham fest.

Makes me wonder.....

Regards, Jochen DH6FAZ (age 49)


Am 16.05.2017 um 01:30 schrieb AncelB mosaicmerc@... [TekScopes]:


Perhaps the group should maintain an 'estate' database of member
submitted eqpt and useful component lists with a contact #.
Then each member can give his family a contact for the moderator of the
group in case of 'unexpected departure' as can occur by getting snuggly
with a live CRT flyback.

Then the group can assist by having an internal auction and the
moderator collects the $$ and pays the family for the haul.
So at least most of the gear and gizmos still go on working for others
and the family realizes something as well.

Given that few of us here are young with another 2 or 3 decades to
spare, it is a real issue.

Ancel

Re: Tek and Darlingtons...

Ed Breya
 

Actually, Darlingtons are alive and well, and all over the place - but mostly in IC form. Many peripheral driver ICs use the configuration, including integrated resistors and clamp diodes that you would normally have to add for discrete devices. Look at the ULN2000 family, for example.

I recently looked at a number of transistor arrays and drivers, trying to find a bunch of single transistors on a common die, for a current mirror application - I had a hard time finding anything that was NOT Darlington in the modern driver IC class. I had to settle for good old CA-series arrays. Another common topology is a PNP driving an NPN, or vice versa.

Ed

Re: Tek and Darlingtons...

 

Ah, I have seen that also.

1. Where low base current is needed like in precision controlled
current sources (7CT1N). These days we might use FETs for this.
2. High impedance error amplifiers. We might use FETs for this now
although super beta operational amplifiers (LMN308) are often better.
3. Where a high gain switch is needed like in some of the plug-in
readout circuits.
4. Scale illumination circuits.

In almost all cases, there are better ways.

On Tue, 16 May 2017 02:48:39 -0400, you wrote:

I have seen them used to drive the scale illumination lamps in several models. The 465 and 475 come immediately to mind.

Regards

Re: 2235 Trigger on one channel

 

On 16 May 2017 14:03:52 +0000, you wrote:

Hi David

Been a little busy so took a while to get back to this. So today lunch time I started to measure some voltages with my DMM as suggested, results as follows, something a little amiss (I think?)

U555B Output pin 6 ch1 = 0.139V (0.2V) (0.06V error)
U555B Output pin 6 ch2 = 1.476V (2.5V) (1.02V error)

U310 Output pin 1 ch1 = -3.26V (-3.4V) (0.14V error)
U310 Output pin 1 ch2 = -2.443V (-2.1V) (0.34V error)
You can check the values of R314, R315, R317, R318, and R319 around
U310 but I think the answer is going to be that U310 is bad.

The signal into pin 1/11 is suppose to be low when channel 1 is
selected. When that happens, the output of U555B is disconnected
(pins 1/11 of U310 should not draw any current) and should reach all
the way to ground which is what we are seeing and U310 should be
working but it is not.

When the signal into pin 1/11 is high, channel 2 is selected but pin
1/11 draws more current from U555B. In this case, it is drawing too
much current indicating that either U310 is damaged or R314 and/or
R315 are open.

The fact that U555B is not reaching its full output is irrelevant
because under that condition, U310 should not be forwarding the
trigger signal anyway.

I also looked at the feed from the A3 Board Connection as follows:-

Pin 18 (6.2V) = 6.5V (0.3V error), this also shows (0V) for XY but can't see how to check this on front panel.
Pin 32 (0.0V) = 0V
Pin 33 (5.0V) = 4.996V
Pin 34 (5.0V) = 4.996V

Some of the values are slightly higher / lower than they should be, is this the issue?
No, nothing is wrong there.

I wonder if there are any good sources for CA3102 transistor arrays. I
do not see any so if you cannot find one, blast me an email with your
mailing address and I should be able to mail one to you. I will have
to pull it from a junked board but will test it somewhat.

Slightly OT: caps again - what a difference a few thousand uF makes

Ed Breya
 

I recently got a Boonton 1120 audio analyzer, which seemed to power up and run just fine - until I set it up to look at its own source. The distortion reading was way out of spec. Looking at the source on a scope, it clearly showed line modulation on the envelope. I didn't even bother looking at PS ripple - went right to the big filter caps and measured in-circuit, then out. Both caps for the +/-15V supplies were equally shot, reading about 50 uF/150 ohms ESR. The nominal value is supposed to be 4,500 uF and way less ESR of course. Fortunately I had some equivalents in stock, tested them, popped them in (easy since they're screw-mount type), and that took care of the problem.



I'm surprised that it even appeared to work at all with the bad caps. The 5V was OK, and overall it made approximately correct measurements.


Ed