Oscilloscope. To repair or not? Fixed.

[neko efecktz]

Hello all,
Just a thought,
Even though the valves are all glowing from the filament check them in the dark to see if there is also a blueish glow.
Sometimes a valve can have a fault but still have a working filament.
even then they don't always have the blue glow.

Just a thought
Bill.

[FlyingHacker]

Be sure to look up the specific tube, as some tubes have a blue glow naturally.

[rf+tech]

Hi Clay,

Now that *I* have taken a few minutes to R(most of)TFM, the CRT is indeed a 5UP1 and the datasheet is attached below. I've also attached the schematic for the Heathkit IO-18 oscilloscope, as a point of reference. This scope uses the 5UP1 CRT - the CRT bias circuitry and bias voltages found in the 555 should be very similar. Other similarites may be noted, possibly with the deflection amplifiers.

Referring to the IO-18 schematic, it shows the CRT beam blanking is capacitively coupled. The 555 may be the same in this respect. With the Intensity control full CW, the cathode to grid voltage difference would be essentially zero. It would be a good first step to check resistances throughout the CRT bias divider in the 555. It would not be unusual to find one or more out-of-tolerance resistors here. An open circuit in one of these resistors would result in an excessively bright and uncontrollable trace. The more likely cause of no trace may be loss of negative 1300 Volts.

Please compare the CRT bias divider in the 555 with this schematic and report the values actually found. From this information, an appliaction of Ohm's law will reveal what voltage to expect at the "bottom" tap in the divider. If this value is found to be less than 600 Volts, and can be measured safely, then by proxy we can determine if the negative 1300 Volts is present.

The following tests can provide insight to the functioning of the deflection amplifiers. A DMM will serve as a visual indication of what is happening to the CRT beam, even though the trace is not visible.

1. Connect DMM across Vertical Deflection Amp 6DJ8 pins 1 and 6, rotate Vertical Position through its full range. Note voltage swings through zero and changes polarity, with the CW and CCW extremes approximately equal magnitude.

2. Adjust Vertical Position for zero volts, set Vertical Volts/CM to 5. Apply 9V battery to Vertical Input (DC coupled) and note DMM reading is no longer zero. Reverse polarity of 9V battery, DMM should now read the same voltage magnitude with opposite polarity.

3. Connect DMM across Horizontal Deflection Amp 6DJ8 pins 1 and 6. Set Time Base to EXT. Rotate Horizontal Position through its full range.  Note voltage swings through zero and changes polarity, with the CW and CCW extremes approximately equal magnitude.

4. Adjust Horizontal Position for zero volts on DMM, apply 9V battery to External Horizontal Input and note DMM reading is no longer zero. Reverse polarity of 9V battery, DMM should now read the same voltage magnitude but with opposite polarity.

5. Set Time Base to 1 second/CM, Triggering to Line +, Level knob to Auto. If the sweep generator is working, the DMM should indicate steadily increasing voltage (polarity unimportant) that resets every 10 seconds.

From the photos, one 6DJ8 visible in RHtube.jpg should be the Horizontal Delfection amplifier, based on the presence of the SWP CAL pot just below. I've made an assumption based on this photo that the second 6DJ8, which serves as the Vertical Deflection amplifier, is on the left side of the CRT board.

While the above measurements reference the 6DJ8 plate pins, the cooresponding deflection plate pins on the 5UP1 may be used just the same.

Let's see how much of this scope is actually functional and what clues can be discovered from real hands-on testing.

RF+ Tech

[clay1905]

First, I'm having difficulty finding a method to compress images that have useful data in them to below about 1Mb each. I'm happy to put pictures here so others may be able to find them in the future, but I will have to continue the hunt for a compression method. Suggestions are welcome, but I don't run Windows.

I see I need a notebook to keep a record of acronyms. They're the steepest part of the curve for me. That was quite an avalanche of information. I spent some time last night reading. Mainly Wikipaedia pages. I didn't know about electrostatic deflection, my knowledge was limited to the basics of picture tubes and magnetic beam steering. There's quite a wealth of information there, but the pages, and especially the one on C.R.O's are a bit terse, and want a few readings.
Between your posts and other reading, I think I've got a reasonable picture of the goings on inside an oscilloscope.

Conventions. I'm seeing both high voltage and high tension used. Is the distinction the same as for cars? That is, high voltage carries current, whereas high tension is like an ignition spark, lots of volts but only a few milliamps or so?

SMPS. I see PS, so probably a power supply, but what are the SM for?

For now I've added three new pictures to the google folder. They are both sides- full shots, and one from the top.
https://drive.google.com/drive/folders/0B3JcpACaWCPlRVEyTzZiR2xxb2c?usp=sharing

My meter is an inherited Dick Smith rebranded Fluke DMM of about 25 years age. It's upper voltage is 500.

The CRT is a Toshiba 5UP1(F) so rf+tech's guesstimate was on the money there. Pin 1 is at the very top of the socket.

So, let me say how I'm seeing things, so you can see if I'm understanding this all correctly.

There are four major blocks, for want of a better word, making up a basic CRO.
There is power conditioning. This can include filtering, transformation and rectification.
There is a circuit that powers the CRT to produce a beam of electrons. There are sub-circuits to control intensity and focus. There is an amplifier associated with this circuit.
There is a circuit that can generate a pulse. The circuit can be made to repeatedly produce pulses, thus making what can be expressed as a sawtooth waveform. This circuit varies an electrostatic charge on two vertical deflector plates in the neck of the CRT. There are sub circuits that control the period of the sawtooth wave, and triggering controls that can determine events that cause one or more sweeps of the electron beam. This circuit controls the CRT's ability to block, or blank off, the electron beam. There is an amplifier associated with this circuit.
There is a circuit that is an amplifier/filter that generates a signal to control the electrostatic charge on the two horizontal deflector plates. The signal is an analogue of variations in voltage detected by the scope probe. There is provision for varying the ratio of input to output in orders of ten.
There is high voltage at the output of every CRT control circuit.
A high -ve charge is put on the cathode and grid, because if you want a stream of electrons, they've got to come from somewhere. Electrons find each other repulsive, so the high -ve grid voltage squeezes the flow into a beam. This charge is likely to be in the low thousands of volts. It is this voltage that  tautech refers to as possibly being not present.
High voltage is applied to all the deflector plates. The voltage difference between opposed plates causes electrons in the beam to change direction, the beam tending towards the more +ve plate. One can expect this voltage to be in the low hundreds.
A strong +ve field may then be applied to the beam to give it a little push along. This accelerator voltage is usually more than enough to burn your toast.
Some voltages required by a given circuit may be obtained by modifying part of the output of another circuit.

So gentlemen, does that seem to be about what's involved in making Lissajous curves?

[alsetalokin4017]

You don't use Windows... neither do I. But you don't say what you _do_ use.... I use Linux, Ubuntu kind.

The image editing package "gimp" is free and available for pretty much all OSs in common use. Here is an example, taking your photo that was previously about 1.2 MB and resizing it to 800 pixels wide and saving as .jpg at 90 percent quality. This brings the file size down by a factor of about 6. Many other image editing programs will allow you to do the same thing.

High voltage and high tension are the same thing. Don't be fooled by any illusions of "low current"... a low current power supply can still charge up a HV capacitor to lethal energy levels which can be discharged at high current if you give it a low resistance pathway -- like your body.

Acronyms do abound. SMPS means "switchmode power supply" which usually will not have a big heavy mains power transformer, but also may suffer from high frequency noise and ripple on the PS output.

If you are going to be working with tube circuits a lot, you might consider investing in a HV probe for your DMM, which will allow you to measure higher voltages safely.

Your verbal "block diagram" description of the oscilloscope is basically correct. However, generating Lissajous figures does not involve a triggered or timed sweep, it simply applies one external signal to the X deflection circuitry and another external signal to the Y deflection circuitry. Some analog scopes have essentially identical X and Y amplifiers, with the X part selectable to either external input, or time-base controlled and triggered sweeps.  Some scopes will use a separate input for the X dimension and others will use the two "normal" input channels, assigning one to X and one to Y when the X-Y mode is selected.

If I were in your shoes I think I'd be more interested in getting the Portascope working, as it will be a much more useful unit than the single-channel 555.

[rf+tech]

Hi Clay,

First-off, my apologies for overwhelming you with too much, too fast. I'll also try to optimize some of the terminology that differs between US technical English and UK technical English - e.g. ground = earth.

There are a few points in your understanding that could use some refinement:

Physically, the direction of effect versus the schematic orientation of deflection plates is 90 degrees. Deflection plates drawn with the broad side vertical deflect the beam at right angles to their broad flat surface, in this case - horizontally. When both plates of a pair have the same voltage/potential (balanced, no differential), the beam feels no pull and remains in the center of that axis.

Electrons from the cathode, having a negative charge, are attracted to positively charged plates. The electron beam will be more strongly attracted toward the plate with the higher positive charge, than the plate with the lower positive charge. It is this differential voltage between the opposing plates, as seen by the beam, that determines which way the beam will go.

The pulse circuit that generates the sawtooth waveform drives the horizontal deflection plates, which have their broad surface oriented vertically inside the CRT and as drawn schematically. This sweeps the beam from left to right. At the right end of the sweep, the beam needs to be returned to the left side. Just as in a television CRT, the beam is blanked during retrace by another pulse of much shorter duration from the sweep generator circuit and applied to the CRT grid. Visualize a sawtooth that rises linearly from zero volts to some positive voltage. The fast return to zero volts corresponds to the blanking pulse.

The signal from the scope probe is attenuated or amplified by the vertical amplifier to produce a much larger analogue of sufficient amplitude to deflect the electron beam in the vertical direction. This signal is applied to the vertical deflection plates and as described above, the beam favors attraction to the plate with the higher positive charge.

Your understanding of the electron beam is basically correct. Since electrons flow from the point of lowest voltage/potential to the point of highest voltage/potential, they don't really care about the absolute voltages as referenced to the chassis/ground/earth. To make deflection amplifier design a bit easier, the highest positive voltage is the nominal plate voltage of the tube/valve deflection amplifiers - something like +300 volts. The cathode is then biased to a much greater negative voltage. The sum of +300 Volts and -1300 Volts combine to create the 1600 Volts accelerating potential mentioned on page 5 of the 555 operating manual.

The typical operating conditions section of the 5UP1 datasheet gives us an approximation of voltages to be expected, and the maximum ratings section an upper limit. Further, the datasheet makes no mention of a Post Deflection Accelerator so there will be no positive high voltage supply present in the 555. This is further substantiated by the 1600 Volts accelerating potential mentioned on page 5 of the operating manual and the -1300 Volts legend silk screened onto the printed circuit board(s).

The high voltage applied to the deflection plates is about one fourth the high voltage applied to the cathode/grid. High is relativistic. As measured against the chassis/ground/earth, the deflection amplifier plate voltages are well within the measurement capability of your DMM.

RF+ Tech

[clay1905]

Easy stuff first. I use Ubuntu as well. I would never have considered simply reducing the image size like that. I would have *assumed* that too much information would be lost. Thanks for the tip.

A high voltage probe for the DMM is looking worthwhile. And a lot better than a wet finger is my guess. The portascope is standing on the workshop floor just now, calling to me. But it will have to wait for a while. First I need to learn more. It does indeed look a much more versatile instrument, and when my skills get to the right level, I'll have at it.

The avalanche of information was not all bad by any means. I'm with Richard Feynman on this, and derive a lot of pleasure from finding things out. When the knowledge is practical too, bonus!

Now, I see I've fallen into a trap for young players. After re-reading my last post and the comments arising, we're on the same page really. It's just that when I was talking about the deflector plates, the terms were 90deg. out. I used vertical to refer to the physical position of the plates, rather than the convention of referring to the beam axis a specific set of plates influence. I shall keep that in mind.

I will go and make the measurements re+tech asked for in a previous post this morning, and post results here. I'll also get cracking on the image reductions and make a post that contains them for future reference.

Thanks,

Clay.

[GK]

Hmm, I've got two of these 555 Kikusui scopes in my dumpster-dive pile. I've been hanging onto these for their CRTs; the model is of such poor overall quality (that's not really wholly conveyed in the photographs) that it's not something I'd personally bother restoring. The styling is an obvious Tektronix rip-off, but that is where the similarities end. Schematics appear to be unobtainable, but I did, surprisingly, find an operation manual (*.pdf) for free download on Kikusui's website a while back.

I remember these being essentially dead, but I just plugged one in and once the heaters warmed up it appeared to be fully operational for several minutes, but the trace slowly dimmed to darkness and eventually couldn't be revived. I'm only interested in getting these operational to the extent of being able to evaluate the condition of the CRTs. The design is pretty simple and it should be a relative doddle to trace the circuit out. Depending on how much time I have I might get around to now.........



 

[tautech]

What's the one top left Glen ?

[clay1905]

Aha!

Three resistors "out of range". They form the part of the voltage divider connected to pin 4 on the CRT.
The focus pot measures infinite between any two legs, and each of the resistors connected to the outside legs of the pot are also open.
I would like have some idea as to why this failure occurred before I cook another set so the parts will be waiting on the bench until I hear back.

Attached is a little sketch of the values of the divider circuit. It doesn't seem to be quite the same as the Osborne though.

Where the diagram is marked 1300V, that isn't strictly true. In the LHhoriz picture, you can see where the board is marked -1300V, and quite close by is a red cotton covered wire connected to the board. The other end of the wire connects to one of the outside legs of the 100K pot in the divider.
The centre leg of the intensity pot is connected to the RH-HTboard. You can see a white/red stripe wire connecting next to a blue PVC covered wire. This board is marked High Voltage, and the hole for the red/white wire is marked 'out'.

Clay.

[alsetalokin4017]

@GK:  Heh... I see an HP180A mainframe in there with the 1801a 2-channel vertical amplifier in the left plugin slot, but I don't recognize the right-hand timebase plugin.
This is a great old scope... although I don't think I've actually turned mine on since I got the Rigol DS1054z....
I'd still like to find a spectrum analyzer plugin for it.

[StillTrying]

It's quite common for high voltage, high resistances to go even higher over many years of use, without there being a fault anywhere else.

[alsetalokin4017]

It's quite common for high voltage, high resistances to go even higher over many years of use, without there being a fault anywhere else.

Yep. I'd just go ahead and replace the pots and resistors (and the capacitor as well.). This could be the whole problem with the scope, right there.

[tautech]

Aha!

Three resistors "out of range". They form the part of the voltage divider connected to pin 4 on the CRT.
The focus pot measures infinite between any two legs, and each of the resistors connected to the outside legs of the pot are also open.
I would like have some idea as to why this failure occurred before I cook another set so the parts will be waiting on the bench until I hear back.

Attached is a little sketch of the values of the divider circuit. It doesn't seem to be quite the same as the Osborne though.

Where the diagram is marked 1300V, that isn't strictly true. In the LHhoriz picture, you can see where the board is marked -1300V, and quite close by is a red cotton covered wire connected to the board. The other end of the wire connects to one of the outside legs of the 100K pot in the divider.
The centre leg of the intensity pot is connected to the RH-HTboard. You can see a white/red stripe wire connecting next to a blue PVC covered wire. This board is marked High Voltage, and the hole for the red/white wire is marked 'out'.

Clay.

AS StillTrying explains this is a common failure in CRO's, more so older ones and related to the quality of the carbon resistors from back then. New components sourced from any of the big suppliers will give most likely better than new service. I have seen them that "baked" that the values are only obtainable from the schematic or BOM. I hope that's not the case for you.
For the pot you could attempt to source a wire wound that will give better service again and some manufacturers used these to ensure long service.

[clay1905]

Hi Gents,

I went down to Jaycar and got baffled with bull. Came away with resistors an order of magnitude small. They don't do a 1meg 5watt resistor, so it was down to building a series, 470K+470K+68 should have done it, but I didn't look and the lad serving gave me 4.7K's instead.
I think part of the problem is I believe I'm coming down with whatever the latest flu release is, and the brain is getting a bit too fuzzy. I was getting all ready to solder in the new bits, and something in the back of the head was saying "something here isn't right". I'll go back in the morning and get some more resistors, but I'm not too keen to be messing about in unfamiliar territory until I feel clear headed enough to be confident. All that high voltage, you know.
Hopefully this will be one of those 24 hour 'flu things and I'll be back on deck soon.

Thanks,
Clay.

[tautech]

Hi Gents,

I went down to Jaycar and got baffled with bull. Came away with resistors an order of magnitude small. They don't do a 1meg 5watt resistor, so it was down to building a series, 470K+470K+68 should have done it, but I didn't look and the lad serving gave me 4.7K's instead.
I think part of the problem is I believe I'm coming down with whatever the latest flu release is, and the brain is getting a bit too fuzzy. I was getting all ready to solder in the new bits, and something in the back of the head was saying "something here isn't right". I'll go back in the morning and get some more resistors, but I'm not too keen to be messing about in unfamiliar territory until I feel clear headed enough to be confident. All that high voltage, you know.
Hopefully this will be one of those 24 hour 'flu things and I'll be back on deck soon.

Thanks,
Clay.

4k7 is Jaycars highest value 5 watt resistor, so good luck getting higher values from them.
IME it is not normal for this resistor chain to be made of 5 watt resistors, 1 watt is normal and should suffice.

Grab one of their catalogues they're handy to have.

[alsetalokin4017]

A little bit of Ohm's Law calculation will show the necessary power ratings of those resistors and potentiometers.

P=I²R and V=IR

[clay1905]

Hi again,

I was pretty fuzzy yesterday, as you can tell from the confusion about power figures etc.
I do like the idea of making me work to learn. It's good to be pushed.

OK, so I set about trying to get the figures to apply Ohm's law. The obvious starting point is the data sheet, which gives all the voltages. However, I don't see anything about current draw. Then I found the data sheet attached. It's a bit of a cheat as all the wattages are included in the "typical circuit" diagram.  I still remember the advice my 1st grade teacher gave the class:   "There's no such thing as cheating in the real world. The only thing that will count is the result."
The result is that 1/2Watt resistors should be good. 1 watt would give a nice margin and Jaycar have these in the right ratings.
Rhubarb! Jaycar aren't open today.
On another note though, they do appear to carry the correct probe and an adapter. Seeing there's the potential for success with replacing these few resistors, I thought it wouldn't hurt to start looking about in this direction.

So, I've made a little shopping list of catalogue numbers, and will splurge tomorrow. Fingers crossed, I'm hoping that replacing the dead resistors won't just be revealing something else wrong further along.

Thanks for all the words of advice so far. This is quite an interesting and fun project, and I'm looking forward to knowing why things like the oscillator attached to the 555 chip isn't doing exactly what the book says it should. The timer is a project to replace a failed circuit in a coffee mill, but that's another story altogether.

I'll report back as soon as I have some sort of result.

Clay.

[clay1905]

The pictures.

Here is the beginning of the pictures. They will be in this and the following three posts.

Clay.

[clay1905]

Four more pictures.

Clay.

[clay1905]

And another four.

Clay.

[clay1905]

The last two pictures.

Clay.

[rf+tech]

Hi Clay,

Good progress, you're learning fast.

I would like to share a few more points and theory to further your understanding.

First, note the 5UP1 datasheet "Examples of Use of Design Ranges" shows two values for maximum G1 voltage for visual cutoff: -45 Volts for an Anode 2 voltage of 1000 and -90 Volts for an Anode 2 voltage of 2000. You are correct in that the -HV at the HV board is higher, however, it will only be higher by something between -45 to -90 Volts. Anything greater would be of no benefit and at worse would serve to limit the useful range of the intensity control by driving the CRT into deeper cutoff. This knowledge implies the -HV from the HV board would be something in the vicinity of -1350 Volts.

With a cathode ray tube (or most any device), the current through the tube is established by the biasing and load resistances/impedances. When given in a datasheet, the values are intended as either absolute maximums or typical, and the latter can be quite broad.

As mentioned by others, a first approximation of the current through the CRT bias network (not the CRT itself) can be obtained from 1350 Volts divided by the sum of the divider resistances to chassis/ground. The current drawn by the CRT from the bias network does have an effect on the actual voltage divisions, and further, the current drawn by the CRT varies with intensity.

I did note in your drawing of the bias network for the 555 CRT that the ground connection is not labeled. How will the current through the bias network return to the -HV power supply? I suspect the ground connection may be the junction of 1M and the 47nF capacitor, or there is another resistor to ground from this junction. Given your illness, it is an understandable oversight. Reverse engineering is an art in itself.

Thanks for sharing the quote from your 1st grade teacher. This reminded me of a quote from one of my grade-school teachers: "Knowledge isn't what you know, knowledge is knowing where to look for information and how to apply it."

Hope you are feeling better quite soon.

RF+ Tech

[clay1905]

Phew,

That's a lot to digest ATM. This bug has a better grip than I first thought. I haven't really attempted to trace the circuit back any further for the while. I just stand there looking while my head swims.

I did find that the second pot, intensity I think, is actually a ganged double. It was dead too, both parts. I have noticed that the ganged pot connects to some whopping capacitors underneath. They even have little insulators sticking up like miniature power station transformers. They're giving me the willies. Any suggestions on how to ensure they're properly discharged before I go sticking my fingers in there? Little ones, up to say 0.2uF I'm not too worried about shorting with an old screwdriver tip, but these have a certain keep away look to them.

Thanks,
Clay.

[tautech]

Purists use something like 100  but most of us just zap them with a driver. 

Do be aware some caps will regain some charge after a short while so either leave a jumper across them of driver them again a time or two.