It's OK now! was->Help - it's broken! was->Tektronix 2467b - ok for beginner?

[ModemJunki]

So I've got the opportunity to pick up a Tektronix 2467b with some probes for free. It's been "obsoleted" but supposedly is in working order (and it's on an OEM Tek cart, nice!).

Would this unit be OK for a beginner working on simple low power (low speed) microcontroller projects based on Arduino and the like?

The first thing I will do with it is learn how to to see if my free function generator generates. 
https://www.eevblog.com/forum/index.php?action=profile;area=showposts;u=63453

[Todd]

Those are terrific scopes.  If it is in good shape you may never outgrow it.

[edavid]

It's an awesome scope for anyone.  One minor issue is that it has a fairly aggressive screen blanker to keep you from burning the somewhat delicate MCP CRT.  You end up tweaking the intensity control frequently to keep it from blanking while you're looking at a waveform.

[ModemJunki]

Good news: it powers up, and the self-calibration point displays a waveform.

Bad news: Attempting to bring up any menu results in a thin line across the middle of the display. So it looks like whatever drives the menu is bad.

Looking through the service manual (found online) it looks like the output of the readout is broken. There is a section on calibrating the readout but all it mentions is centering and gain. I was able to figure out how to run the self-test for the readout board and neither of the trigger LEDs lit - which is a good sign as having either light up would have meant a RAM or shift register failure.

But the schematic of the readout board is of little help to me, I'm a complete beginner and don't even know where to start looking. Perhaps someone here has seen something like this before?

Attached is what I see - flickr is down at the moment so I can't post photos.

[c4757p]

Just a quick beginning for you: Start from pin 26 of J411, the vertical output from the readout board, and work backward. Make sure the vertical signal is being generated. It should come out somewhere near a TL074 on the readout board. This then goes through some analog stuff on the main board - look at TP800, and also check U485 (appears to be a quad matched transistor).

You'll need another scope to diagnose this. See if you can borrow one.

It's almost 2 AM here, so I haven't dug through diagrams much, but a lot of this on the main board should be in the vicinity of U400 (one of the big hybrids). The readout board is near the power supply.

[KJDS]

Before doing anything, work out where the dangerous voltages are within the scope and how to avoid them. I don't want to come over all elf'n'safety but a little caution now is wise.

[c4757p]

They are quite well contained in this one, as long as you don't go taking off any covers. Worst exposed voltage is 87V.

[ModemJunki]

Thanks, all.

Borrowing a second scope may not be possible at the moment, my source for the scope hasn't got another that can leave the building except to go to a cal lab.

But there is the possibility for me to budget for a cheap used scope, say ~$100. Checking Craigslist for my local area shows some in the range, e.g., BK Precision model 2120 (kind of far away), EICO 460S (possibly really inexpensive), Heathkit 5mhz with a Heathkit multimeter of the same vintage, all documentation, and extras for $125 (looks very clean and is and conveniently nearby).

So if that was the case, what would be the minimum spec scope I would need to troubleshoot this problem? Or should I just buy one of those cheap disposable single-channel MP3 player sized DSOs and sell that later after making the repair (or just keep it and use as a fancy meter..)? A plus for this concept is future storage space - my electronics workspace is a bit on the small side.

I will take the advice to avoid the high-voltage areas seriously, I have a healthy respect for electricity.

Note that this will be a real challenge for me - I know almost nothing of electronics, this will be a true learning experience for me and I will have to budget time for it as well. But I look forward to it, and will certainly be asking for more guidance after I follow the 'scope tutorial.

Thanks to all for the advice.

[c4757p]

Minimum spec? Turns on. You'll be fine with a cheap 20 MHz for this. If a TL074 can handle that signal anything can. But if you buy one of the MP3 player turds you'll be screaming at it before you even have a clue what's going on in your real scope. Don't do that, for the sake of your own sanity.

Note that this will be a real challenge for me - I know almost nothing of electronics, this will be a true learning experience for me and I will have to budget time for it as well. But I look forward to it, and will certainly be asking for more guidance after I follow the 'scope tutorial.

Luckily, looking at the signal path from the vertical circuit in the readout board through to the channel switch (which we know works because the analog channels work), this looks like a fairly simple one. I'm suspecting the TL074 in the output from the readout board, or that matched transistor array on the main board.

Hint: Non-matched transistors will work, just not well. You can test by temporarily replacing it with PN2222s - just make sure you check pinouts. You'll at least get something on the screen, even if it's not quite perfectly linear like the matched pairs would give. And unlike matched transistor arrays (a bit hard to find these days), TL074 is still very common and very cheap.

[ModemJunki]

Would a Tek 2205 be able to use the leads that came with my 2467? In general, can leads from the Tek be used with other manufacturers?

A lot of used units seem to be missing leads.

[PA0PBZ]

The leads can normally be used on all scopes, they all have BNC inputs if you are not looking at 1950´s  or very special GHz models.

I wonder, if the scope measures ok and it is only the readout that doesn´t work, should it be possible to measure the readout signals with the scope itself? Both the ground of the probe and the ground on the readout board are the same... 

[ModemJunki]

I wonder, if the scope measures ok and it is only the readout that doesn´t work, should it be possible to measure the readout signals with the scope itself? Both the ground of the probe and the ground on the readout board are the same... 

That is brilliant - and would be easy to try (if I knew what I was doing).

I shall make a go of it. 

[PA0PBZ]

That is brilliant - and would be easy to try (if I knew what I was doing).
I shall make a go of it. 

Let's wait to see what others have to say first, it doesn't feel right for some reason but I can't see a flaw in it (yet...)

[c4757p]

  Yep, that'll do just fine. Wish I could say I thought of that. 

Only problem I can think of is if you probe something in the signal path for the channel itself. Even then, you're just likely to create a garbage signal, you shouldn't break anything. The readout signal path might as well be a separate device.

[nctnico]

There problably is some multiplexer circuit which shows the readout signals when no trace is written. So basically the readout is only shown when there is no trace on the scope. OTOH this should be simple to fix. Usually there is an X and Y DAC and it seems there is an interruption somewhere between the Y DAC and the place where the readout signal and the trace signal are muxed. All in all I strongly doubt the problem can be found by using the scope on itself.

[PA0PBZ]

Ah, yes, when the readout is only active when there is no trace I can see a problem there...

[c4757p]

So basically the readout is only shown when there is no trace on the scope. OTOH this should be simple to fix. ...  All in all I strongly doubt the problem can be found by using the scope on itself.

Wow. How did I miss that?  Yeah, that signal isn't going to do anything while the trace is visible. You'll need another one.

[ModemJunki]

OK, then I will do the following:

1) Open the scope and look for any obvious burnt component
2) Work on getting a second working scope, minimum 20MHz, borrowed or bought.

I think even if 1) yields bad components, 2) might still be needed to do the repair properly.

I'll post back if I find anything burnt, otherwise I must wait until I source a 'scope.

[ModemJunki]

Well, I found a ribbon cable which had been damaged/chafed so that one lead was open. It was rubbing against the chassis. I repaired the open spot with a dab of liquid tape and wrapped it with electrical tape to protect it from further rubbing against the chassis.

Now the scope shows the menus, but the grid illumination backlight has stopped working. 

I can't figure out how to remove the knobs so I can pull the faceplate to check the lamps and see which one is out. Do they just pull off, or is there a secret trick to it?

[c4757p]

They indeed pull off, if they don't crumble first. They're quite brittle. Be very careful.

[ModemJunki]

Ok, I have two open bulbs. They look exactly like the little wheat bulbs used by hobbyists when I was a kid. They were working before, so I think when I was re-assembling the unit I gave them a jolt, I recall struggling a bit with the leading edge of the case when I was fitting it, it was a gentle jolt but apparently enough to do them in.

It also looks like you can't really replace the bulbs with an LED+resistor.

Are these just 5v wheat bulbs?

[c4757p]

The exact part number is actually still in production. The manual lists them as "aged and selected", though... Leave it to Tek to have special requirements for light bulbs.  Probably just for longevity, since they're such a pain in the ass to replace, so I'm sure any old one will be just fine.

(Sorry about the TinyURL link - it's Mouser. The forum software was mangling the URL.)

[c4757p]

By the way, pretty much the whole thing disassembles with ease, unless you have to take out the main board. If you're struggling I suggest you have a look at the disassembly instructions in the service manual before something else gets broken. Some of the cosmetic parts especially are fragile and hard to find replacements for.

[ModemJunki]

Even as you were posting, I found the exact same bulb on the Mouser site.

I had no problems with dis-assembly once I got past pulling the knobs off. Interesting how the "knob" is actually just a skin on the pot handle. I won't put the knobs back until after I have the lamps as they seem kind of thin.

For the moment I will put it back with some 10-ohm resistors underneath in place of the bulbs.

[c4757p]

10 ohm? They take 115mA at 5V, that's 43 ohms. It's relying on their resistance to divide 15V across them, so it matters - you'll blow the last good one if you do that.