开云体育

Ozan,

Sorry, I’m running on a bit of a sleep deficit, and I got the sender names confused in the thread.

Can I just use a multimeter to check then impedance of the scope input?

— Jeff Dutky

My measurements put it at about 290 MHz.? I find the -3 dB display frequency, repetitive signal.
Bob

On Tue, Mar 22, 2022 at 07:48 PM, James55 wrote:

This electronics is addictive stuff!

Even though I am falling asleep, couldn't resist one last 'hit' of voltage
readings before passing out.

You'll like this...

With CR1732 removed, R1757 replaced and all pins of U1724 connected, we
finally have over +60v on the +65v test point!

It was 61.8 volts, without any R1736 adjusttment

Great progress. At this stage please do not remove the opamp. TP+65V will fly to 80+ volts.

CR1732 is nothing special and given you have replaced (or will replace) many components with not-exact replacements you can go with 1N4148. Its job is to protect BE junction, most of the time it is off, reverse biased by 0.7V and having an easy life. Same with CR1733, used exactly the same way.

Just to make sure: the diode checker should show ~ -0.7V (for silicon diodes) in forward direction and show open circuit (or high impedance) in the reverse direction. When you get your new diodes you may want to confirm your diode checker works properly.

Ozan

This electronics is addictive stuff!

Even though I am falling asleep, couldn't resist one last 'hit' of voltage readings before passing out.

You'll like this...

With CR1732 removed, R1757 replaced and all pins of U1724 connected, we finally have over +60v on the +65v test point!

It was 61.8 volts, without any R1736 adjusttment

I'll get some diodes tomorrow and let's see what happens next?


Night night



James


zzzzzzzzzzzzzzz

Interleaved:

On 3/22/2022 10:01 PM, James55 wrote:

Hi Ozan and Harvey.
I do believe we are closing in on the issue.

One certainly hopes <grin>

Following the results of lifting the #5, 6, and 7 pins, and realising that the dual op-amps were in fact not as 'isolated' as I had imagined, I thought I would go and have a look at the +15v rail.
This rail has had a short for a while but I hadn't bothered to investigate as the +65v was the priority.

Now that could do nasty things.........

Anyways......

After having a quick probe on the voltages, it was clear that there was nothing on the Q1756E, so I went to have a closer look at the schematic and saw that R1757 (0.6Ω) was the next component after the emitter. To cut a long story short, R1757 was open. Not only that but as I lifted a leg, it turned out to be in two pieces. That would certainly explain the lack of voltage on the 15v TP.

Resistors are not meant to be two piece assemblies.

Look at U1724B. Pin 7 (output) can both source and sink current. In U1724A, it can only sink current. (more on that later)

This has been a revelation that certain op amp outputs can flow both ways. Definitely helps to understand circuitry

Like the output of a hi-fi amplifier, there's a transistor from the op amp output to the + rail, and one from the - rail to the output.

If there's no load, and the output ought to be at, say, 10 volts, then no current flows into the output.? If you're driving a resistive load, then current flows from the output to the load. If, however, you connect to some voltage source (through a resistor), the op amp still wants to keep the output at 10 volts, even though the output pin wants to go to 11.? The op amp will have to sink current to keep the output at the desired level.

Normally, all the current to the +15 volts comes from the +15 volt supply, the 22.5 volts at the + end of C1751. No problem.

Yes, we have that 22.5v

I assume that CR1757 and C1757 are good and there's no load on the +15 at all.

C1757 is a new tantalum and all the jumper connectors as well as the A5 and A7 boards are disconnected.
As for CR1757, after looking into CR1732 (which I'll explain in a minute), I am now thinking that this could be a source of one of the problems. Is a forward voltage of 0.585v too low?

Not necessarily so.? generally about .6 to .7, but it varies with current as well.? Go find a forward voltage curve vs current on a diode.? The knee is not necessarily all that sharp.

I'd still like to see what happens if you disconnect U1724B pin 7, which would keep the op-amp from sourcing current to the 15 volt supply.

At the moment pins 5, 6, and 7 are disconnected

Then what I'd expect is that the voltage at pin 8 on the op amp would be close to 22 volts, especially if the junction of CR1724 and VR1724 is grounded.

Another alternative would be to connect the 15 volt bulk supply o CR1751, which would convert the two transistors to transistors rather than BE diodes.

Not quite sure that I follow, however I'm happy to give it a whirl if todays developments don't make a significant headway.

In a transistor, the BE junction acts like a forward biased diode, and the BC junction acts like a reverse diode.? Putting in current to the base lead effectively makes the composite CE junction leaky (one way to think of it) allowing collector current to flow.? If the collector of the transistor is open, you get a diode formed by the BE junction.? Note that the current in the transistor flows *through* the base region in normal operation. So with a transistor of a beta of 100, for each 1 ma flowing in the collector, you need 10 ua flowing into the base.? The emitter current therefore becomes (Ib + Ic) or 1.01 ma.? WIth no collector current source, you just have a base-emitter junction, which is? a forward biased diode.

@Ozan

It leaves CR1732 as suspect.

Funnily enough, I had measured and re-measured the voltage on that diode several times as It didn't seem right for some reason. I just wasn't really sure what I was looking out for.

When you measure the voltage drop on a diode, the forward voltage drop on silicon ought to be about .6 to .7.? There should be no reverse voltage drop, it should effectively be open (unless a zener, zeners are reverse biased).? Looking at the Q1732 and CR1732, though, they are effectively a diode (CR1732) and a BE junction on a transistor.? They're connected back to back with opposite polarity.? Measured one way, you get the forward drop of CR1732 and the BE junction of the transistor is reverse biased, measure the other way, CR1732 is reverse biased and you see the forward voltage drop of the BE junction.

On to this evening... CR1732 and CR1733 are identical 30v 175mA diodes, so when in diode mode on the DMM, CR1732 measured 0,315v one way and slightly higher in reverse, whilst CR1733 measured 0.679-ish both ways, I pulled CR1732 and sure enough it still has the low readings. Even though I am fully aware that necessary voltage across a silicon junction is around 0.7v,
I realise now that I have occasionally overlooked lower diode readings as long as the device only give a reading in one direction, and that is wrong.

Same situation as above.? Vbe one direction, Vfwd the other. Always measure diodes both ways, then understand why you get what you get.

Last one is CR1732. Not only it should register a diode in forward direction but it should work well in reverse direction. When you do diode check it is important to check both directions.

My case in point.
I do always check diodes in both directions however, it is clear to me that I need to review the different diode types.


So, tomorrow I shall be out again searching for a 1N4152 equivalent as they don't exist here. Any suitable suggestions would be appreciated.
The R1757 has already been replaced.

Work on seeing why you get the readings you do.

Harvey

Once again it is late here but progress is being made.


Thanks to all for the support and input.


James

Looking online for a 40v 150mA 1N4152 replacement, the 1N4148 keeps popping up along with Schottky options.

Hi Ozan and Harvey.

I do believe we are closing in on the issue.

Following the results of lifting the #5, 6, and 7 pins, and realising that the dual op-amps were in fact not as 'isolated' as I had imagined, I thought I would go and have a look at the +15v rail.
This rail has had a short for a while but I hadn't bothered to investigate as the +65v was the priority.

Anyways......

After having a quick probe on the voltages, it was clear that there was nothing on the Q1756E, so I went to have a closer look at the schematic and saw that R1757 (0.6Ω) was the next component after the emitter. To cut a long story short, R1757 was open. Not only that but as I lifted a leg, it turned out to be in two pieces. That would certainly explain the lack of voltage on the 15v TP.

Look at U1724B. Pin 7 (output) can both source and sink current. In U1724A, it can only sink current. (more on that later)

This has been a revelation that certain op amp outputs can flow both ways. Definitely helps to understand circuitry

Normally, all the current to the +15 volts comes from the +15 volt supply, the 22.5 volts at the + end of C1751. No problem.

Yes, we have that 22.5v

I assume that CR1757 and C1757 are good and there's no load on the +15 at all.

C1757 is a new tantalum and all the jumper connectors as well as the A5 and A7 boards are disconnected.
As for CR1757, after looking into CR1732 (which I'll explain in a minute), I am now thinking that this could be a source of one of the problems. Is a forward voltage of 0.585v too low?

I'd still like to see what happens if you disconnect U1724B pin 7, which would keep the op-amp from sourcing current to the 15 volt supply.

At the moment pins 5, 6, and 7 are disconnected

Another alternative would be to connect the 15 volt bulk supply o CR1751, which would convert the two transistors to transistors rather than BE diodes.

Not quite sure that I follow, however I'm happy to give it a whirl if todays developments don't make a significant headway.


@Ozan

It leaves CR1732 as suspect.

Funnily enough, I had measured and re-measured the voltage on that diode several times as It didn't seem right for some reason. I just wasn't really sure what I was looking out for.

On to this evening... CR1732 and CR1733 are identical 30v 175mA diodes, so when in diode mode on the DMM, CR1732 measured 0,315v one way and slightly higher in reverse, whilst CR1733 measured 0.679-ish both ways, I pulled CR1732 and sure enough it still has the low readings. Even though I am fully aware that necessary voltage across a silicon junction is around 0.7v,
I realise now that I have occasionally overlooked lower diode readings as long as the device only give a reading in one direction, and that is wrong.

Last one is CR1732. Not only it should register a diode in forward direction but it should work well in reverse direction. When you do diode check it is important to check both directions.

My case in point.
I do always check diodes in both directions however, it is clear to me that I need to review the different diode types.


So, tomorrow I shall be out again searching for a 1N4152 equivalent as they don't exist here. Any suitable suggestions would be appreciated.
The R1757 has already been replaced.


Once again it is late here but progress is being made.


Thanks to all for the support and input.


James

The 2440 X-Y bandwidth was 300MHz it says at probe tip, also for the 2439 and 2432/A however the maximum single-event useful storage bandwidth is
200 MHz(2440) 100MHz (2432A) 40MHz (2430A) The 2440 service manual says adjust transient responce for greater than or equal to 310 MHz each channel.
Jeff
--
Jeffeelcr

On Tue, Mar 22, 2022 at 05:42 PM, Jeff Dutky wrote:

I do seem to see correct DC attenuation at 50 mV, 500 mV, and 1 V settings
(I've now checked that multiple times). I was expecting to see the wrong
attenuation at DC, was surprised when I did not see that, and am now wondering
if the stuck switches are tarnished or carbonized so that they act like a tiny
capacitor (maybe with high EPR leakage?)

Perhaps one of the 10x paths stays connected to ground (perhaps with a tiny cap as you wrote) even when it shouldn't. Any additional cap will kill the BW as impedance is pretty high in that network. Does the input resistance stay ~ 1M in all modes?

I see what you mean about there being a path through the attenuators even with
all switches open: it would act like ÷100 has been selected, which could
explain what I'm seeing, if the attenuation of a DC signal were as attenuated
as the AC signals.

I had been following that discussion ("Tek 2465 stuck in horizontal sweep")
and had seen your posts and pictures. Your mention of having to repair or
replace the tabs between the attenuator and the damping network has me
worried.

I believe Siggi posted the comments and pictures. I don't have any direct experience on these attenuators. Only attenuators I fixed were on a Tek 485, and replacing all 16 relays on a TDS7104 attenuator. It was interesting to see BW beyond 1GHz was possible with regular looking relays.

Ozan

The 2200 ohm resistors go to the screen, same as the 470pF capacitor. The suppressor is internally connected to cathode.

Dave Wise

From: [email protected] <[email protected]> On Behalf Of Mark Vincent via groups.io
Sent: Tuesday, March 22, 2022 6:09 PM
To: [email protected]
Subject: Re: [TekScopes] Type 106 Saga (Again)

Notice the screens are floating with a 470mmfd to ground. The suppressor has the 2200 ohm resistors to ground. With a floating screen, the plate will barely get a signal. I am going by the first schematic link on the W140 site. I have disconnected the screen on pentodes, signal and output, and the signal was greatly attenuated and distorted. The screens need to be at some negative voltage potential reference to ground and be less positive than the plate. If I am wrong, what did I miss?

Mark

Notice the screens are floating with a 470mmfd to ground. The suppressor has the 2200 ohm resistors to ground. With a floating screen, the plate will barely get a signal. I am going by the first schematic link on the W140 site. I have disconnected the screen on pentodes, signal and output, and the signal was greatly attenuated and distorted. The screens need to be at some negative voltage potential reference to ground and be less positive than the plate. If I am wrong, what did I miss?

Mark

Ozan,

I do seem to see correct DC attenuation at 50 mV, 500 mV, and 1 V settings (I've now checked that multiple times). I was expecting to see the wrong attenuation at DC, was surprised when I did not see that, and am now wondering if the stuck switches are tarnished or carbonized so that they act like a tiny capacitor (maybe with high EPR leakage?)

I see what you mean about there being a path through the attenuators even with all switches open: it would act like ÷100 has been selected, which could explain what I'm seeing, if the attenuation of a DC signal were as attenuated as the AC signals.

I had been following that discussion ("Tek 2465 stuck in horizontal sweep") and had seen your posts and pictures. Your mention of having to repair or replace the tabs between the attenuator and the damping network has me worried.

-- Jeff Dutky

Interleaved, and I think you're found something.? (Watch this, nothing up my sleeve...........)


On 3/22/2022 6:12 PM, James55 wrote:

Just a quick one as out the door but will give more details in a bit.

Harvey,
I didn't yet connect the junction to ground, however you asked to disconnect pins #5, 6 and 7 form U1724B.

The following voltages were then taken;
Pin # 1 - 21.2v
Pin # 2 - 4.12v
Pin # 3 - 9.23v
Pin # 8 - 21.84v

Look at U1724B.? Pin 7 (output) can both source and sink current.? In U1724A, it can only sink current.? (more on that later)

Let's assume that for some reason there's a fault in the +15 volt supply, such as perhaps a short to ground, or a high current drain.

Current from pin 7 of the op amp can go to the base of Q1752, which emitter drives the base of Q1756, which drives the +15 volts.

Normally, all the current to the +15 volts comes from the +15 volt supply, the 22.5 volts at the + end of C1751.? No problem.

Suppose it got disconnected, and that supply is now floating, the +15 volts unregulated.? No current can come from that.? However, we have pin 7 on the op amp conneced through the BE junction of Q1752, which then goes through the BE junction of Q1756 (remember that without a collector voltage they're just diodes), and then to the +15 volt output.? It looks as if the op amp is trying to source all the current needed to make the +15 volts run

I assume that CR1757 and C1757 are good and there's no load on the +15 at all.

(it would make a difference)

So maybe the problem isn't in the +65 volt circuit, but in the +15.

Grounding the anode of CR1724 removes the ability of U1724A to control the voltage of the +65 supply.? IF U1724A were bad, then the +65 would go to the lowest voltage capable of being selected.

I'd still like to see what happens if you disconnect U1724B pin 7, which would keep the op-amp from sourcing current to the 15 volt supply.

Another alternative would be to connect the 15 volt bulk supply o CR1751, which would convert the two transistors to transistors rather than BE diodes.

Interesting thought to consider the *other* half of the chip....

Harvey

So U1724B and circuitry is definitely not helping our cause.

R1735 measured 57.8K and C1735 didn't register as a capacitor on the DMM but is not shorted.

Ozan,
you asked for measurements too. The following are with the above op-amp pins disconnected...

VR1724 - 4.462v
C1724 - 4.468v
R1724 - 0.00v

Removing Q1736 had no effect on the relevant voltages and R1723 measured 15.6K with a leg lifted.

Must dash. Back soon.


James

On Tue, Mar 22, 2022 at 05:18 AM, James55 wrote:

I don’t like testing components without having a good guess of which one failed but this may be a case where
testing Q1732, Q1736, and CR1732 may identify a bad component.

Q1732 did fail before but was replaced. It currently tests ok using diode mode
on the DMM and when replaced just now with a 2N5551 the Base reads 25.90v with
the Emitter at 25.48v.
Q1734 is a new TIP31C following my carelessness and Q1736 also tests fine in
diode mode.

Hi James,
From my list of devices to check Q1732 was new so probably fine, your other e-mail said you removed Q1736 and it didn't make any difference. Last one is CR1732. Not only it should register a diode in forward direction but it should work well in reverse direction. When you do diode check it is important to check both directions.

Why focus on this part of the circuit: Even with opamp out, and Q1736 out base of Q1732 is not above ~ 26V or so. There are only two paths for the current of R1723 to flow in this condition. Base of Q1732 or CR1732. We know that Q1732 is newly replaced. It leaves CR1732 as suspect.

Ozan

On Tue, Mar 22, 2022 at 02:35 PM, Jeff Dutky wrote:

Yes, I saw that, but it doesn't explain how the three attenuation modes (÷1,
÷10, and ÷100) are combined with three gain modes (2, 4, and 10) to get
eleven input scales.

2445 SM I looked at by accident mentions two more attenuation settings 1 and 2.5 for the pre-amp. I didn't cross check whether 2445 and 2445A use the same pre-amp.

On 2467B 2mV/5mV/10mV/20mV/50mV settings have no relay click. It seems there are 5 possible gain settings outside the attenuator. Then 100mV/200mV/500mV use one relay setting (10x) and 1V/2V/5V use the other (100x, two 10x in series) relay setting.

...
The ARE actuating where I thought, but I
noticed that the relays for Ch 1 sounded different from the relays for Ch 2:
the Ch 1 relay clicks sound like single clicks, while the Ch 2 relay clicks
sound like a pair of clicks very close together. This is true for the clicks
produced as you change input scale, as well as for the clicks produced when
you change input coupling.

The schematic clearly shows that each relay is a pair of switches in parallel.
Maybe one half of the switches in the Ch 1 attenuator are stuck closed?

(the switches would have to be stuck closed, or I would be seeing input scales
that produced no signal at all, and I'm not seeing that)

This is a very likely cause. Even when all the switches open there is still a path through the 10x attenuators. I would think that a stuck relay should give wrong volts/div as well even at DC. Do you see correct DC attenuation at 50mV, 500mV, and 1V settings? Does the input resistance stay ~ 1M-ohm in all settings?

Message
/g/TekScopes/message/191878
mentions repair of a similar attenuator.


Ozan

Hi David,

I had considered LED's, after being unable to locate any kind of bulb even remotely similar to the originals, here in Brazil, but then discovered that the old-skool Xmas decoration bulbs (the two-wire ones which just plug in), make a perfect replacement.
Very bright at full intensity and easy to find in junk markets. I now have about 100 spares. :)


James

On Tue, Mar 22, 2022 at 02:59 PM, <toby@...> wrote:

I have a 466 here but it died while I was testing it out. I haven't
opened it up to investigate yet, but it needs to happen.

Don't mind helping out other 466 owners to the extent I can... Have
never repaired one before, though. I follow threads like this one for
hints. :)

--Toby

Toby,

Go for it!

I have repaired a few of these older type Tek scopes and this one is only difficult because of a mistake which I made after already having repaired it. Doh!

If yours is currently not working, the worst that can happen is that it will remain that way, but you will definitely learn in the process.


James

Just a quick one as out the door but will give more details in a bit.

Harvey,
I didn't yet connect the junction to ground, however you asked to disconnect pins #5, 6 and 7 form U1724B.

The following voltages were then taken;
Pin # 1 - 21.2v
Pin # 2 - 4.12v
Pin # 3 - 9.23v
Pin # 8 - 21.84v

So U1724B and circuitry is definitely not helping our cause.

R1735 measured 57.8K and C1735 didn't register as a capacitor on the DMM but is not shorted.

Ozan,
you asked for measurements too. The following are with the above op-amp pins disconnected...

VR1724 - 4.462v
C1724 - 4.468v
R1724 - 0.00v

Removing Q1736 had no effect on the relevant voltages and R1723 measured 15.6K with a leg lifted.

Must dash. Back soon.


James

Ozan,

Yes, I saw that, but it doesn't explain how the three attenuation modes (÷1, ÷10, and ÷100) are combined with three gain modes (2, 4, and 10) to get eleven input scales.

I have some more observations about the fault that may be enlightening:

While trying to work out how the input scales were constructed from the attenuation and gain modes I went back to verify that the relays were actuating where I thought they were. The ARE actuating where I thought, but I noticed that the relays for Ch 1 sounded different from the relays for Ch 2: the Ch 1 relay clicks sound like single clicks, while the Ch 2 relay clicks sound like a pair of clicks very close together. This is true for the clicks produced as you change input scale, as well as for the clicks produced when you change input coupling.

The schematic clearly shows that each relay is a pair of switches in parallel. Maybe one half of the switches in the Ch 1 attenuator are stuck closed?

(the switches would have to be stuck closed, or I would be seeing input scales that produced no signal at all, and I'm not seeing that)

-- Jeff Dutky

On Tue, Mar 22, 2022 at 05:18 AM, James55 wrote:

Having a quick look, R1724 is 324mV

I have a suspicion that your DVM may have ~ 300mV offset. Some diode and Vbe voltages are about 300mV off. In your earlier measurement:

C1724 - 23.3v
VR1724 - 23.3v

Voltage across R1724 should simply be the difference of them, which should be zero but you are measuring 324mV.
Could you re-measure all three voltages across the components again to make sure DVM is telling the truth: C1724, VR1724, R1724

Q1732 did fail before but was replaced. It currently tests ok using diode mode
on the DMM and when replaced just now with a 2N5551 the Base reads 25.90v with
the Emitter at 25.48v.

This is one such measurement where there seems to be ~ 300mV error: Vbe is 0.42V, expected voltage is 0.6-0.7V.

Q1734 is a new TIP31C following my carelessness and Q1736 also tests fine in
diode mode.

Diode mode won't catch all faults. If you remove Q1736 does voltage recover? Q1736 is there for short circuit protection. I assume at this point power supply is not connected to rest of the scope so short circuit is not a concern.

One more component can cause all this: If R1723 went high resistance. If you measure it lifting one leg up, do you measure close to correct value (15k)?

Ozan