Thoughts on buying "non-working" or "for-parts" Oscilloscopes?

[alm]

In my opinion it doesn't make sense to spend $$$ on spare parts you don't currently need. I don't think the hybrid is known to have a high failure rate (check TekScopes to be sure), so who knows if that's the first to fail? Maybe some other part will die first, or maybe you'll replace/sell it before it dies. In the mean time, you've saved $40 that you can spend on something more useful.

You bought this scope and immediately found a spare hybrid for sale, so it's not likely a rare item. Might as well buy it when you actually need it.

[FenderBender]

After considering it myself, it does seem silly to buy the chip. Solid state stuff usually doesn't die too often, but who knows since this was a unique Tek chip. I made a post on TekScopes. Hopefully the guys can help me.

I have my fingers crossed that it works. If not, I'm going to be quite sad...

[djsb]

I'm looking for a Tektronix digital scope to supplement my 2225. I've read on the yahoo group that the 2440 is the last model that is really repairable down to component level. This is because (according to the posts on there) the schematics on models after this one are no longer available in the service manuals. It is expected that if faulty the relevant PCB is replaced, impossible if they are no longer available.
The 2440 has a 300MHz bandwidth and is usable up to 500Mhz. I'd like a digital scope so I can see non-repetitive signals and also observe RS232 comms etc.
I'd love an Agilent MSO model but I can't stretch to £7K.
Any thoughts.

David.

[don.r]

I'm looking for a Tektronix digital scope to supplement my 2225. I've read on the yahhoo group that the 2440 is the last model that is really repairable down to component level. This is because (according to the posts on there) the schematics on models after this one are no longer available in the service manuals. It is expected that if faulty the relevant PCB is replaced, impossible if they are no longer available.
The 2440 has a 300MHz bandwidth and is usable up to 500Mhz. I'd like a digital scope so I can see non-repetitive signals and also observe RS232 comms etc.
I'd love an Agilent MSO model but I can't stretch to £7K.
Any thoughts.

David.

Do you need the bandwidth? Anything above 100MHz is not too useful for most situations but a storage scope is really handy many times. Why not plump for the Rigol 100MHz which is probably £300 now. With a new scope with warranty you don't have to worry too much about repairability (and for £300, no worries after warranty either  ).

[alm]

The 2440 has a 300MHz bandwidth and is usable up to 500Mhz. I'd like a digital scope so I can see non-repetitive signals and also observe RS232 comms etc.

Single shot bandwidth will be much less than 300 MHz. In my opinion buying old analog scopes makes less sense than buying old digital scopes. The best analog scopes are no longer made and essentially the only analog scopes built today are cheaply made and much less powerful than eg. the Tek 2465 or 7000 series. Today's digital scopes are better than the old digital scopes in all aspects except may build quality. I would only buy an old digital scopes if I really needed one of the features (eg. a brand new DSO with 300 MHz will be fairly expensive) and could live with the limitiations (low sample rate/memory depth, sometimes slower, some of the early models had complex controls). Models from late 1990s or early 2000s have less limitations compared to the first few generations.

[tom66]

I know what you mean and I would have scrapped it for parts but at some point this scope made it personal.
at first it was totally dead but now there's the trace and so on but it's like the focus is drastically out so it's about half an inch thick.

Check this: http://www.cromwell-intl.com/radio/tek2445a.html

Failed resistor in focus supply.

[Gridstop]

I was just gifted a Tek 7401 (1Ghz) with two 1Ghz vertical modules and 1Ghz timebase & 500Mhz timebase units. Pretty exciting! It's in really good shape, no broken knobs/etc, and everything seems to work great, with no burnout on the screen or anything. Unfortunately the 1Ghz inputs are 50 ohm, and no probes were included.

So my plan is...
Build one or two of the DIY 1Ghz resistive probes for probing CMOS/digital stuff with the 1Ghz vertical amps:
http://koti.mbnet.fi/jahonen/Electronics/DIY%201k%20probe/

Pick up one of the dual 200Mhz (7A26), 120Mhz (7A12), or 75Mhz (7A18) vertical modules on ebay. ($20-40) (1M 20~24pF input on most of them)
Pick up two cheaper replacement probes from Rigol/Instek/Velleman/TPI ($25-50 ea, depending) Not sure what would be best.

And then figure out how to calibrate the thing at least reasonably close. Vertical amplitude I can do because my DMM gives extremely accurate AC measurements and the internal calibrator looks spot on voltage-wise, but timebase I don't see how I can verify.

One day I'd like to explore using a HD webcam or camcorder streaming to a PC to blow up the tiny ~4" screen and allow freeze frame, accumulation, and hopefully capture single-shot images.

Of course, a good working 7401 with Ghz modules could fetch a few hundred on ebay, and buy me something like the GW-Instek 1072-AU... but I'm hesitant to give up this kind of speed and the ability to really image square waves in the 20-100Mhz range, which most affordable digital scopes ($<500) can't come close to doing.

[tekfan]

...but timebase I don't see how I can verify.

Now that's a nice scope! Be careful with the beam intensity. You can easily burn the phosphor on the screen.

There are two trimpots on the calibrator board. One for the frequency (R375) and one for amplitude. If you've got a frequency counter you can first try to tweak the frequency trimpot to bring the calibrator as close to 1kHz as possible. Then feed the signal to one of the channels and adjust the SWP CAL control on the time base (set at 0.1ms/div) to get exactly one cycle across the screen.  The other ranges on the time base should still be pretty close to within calibration.

[Gridstop]

Yes, I did the trick with the intensity turned way down, move a flat trace below the screen, and then in a dark room turn the intensity up so you can see reflected 'glow'. The image is very smooth, with only some slight burn-in where the readouts are. When we first powered it up, I had no idea what a microchannel CRT was, and I kept trying to turn the intensity up, and it kept switching off on me, so I thought something was wrong with it and almost left it to be recycled! At least I know the limiter works properly. My understanding is that in addition to the CRT burning in due to the incredible electron output, the microchannel detector itself wears out cumulatively over time? I can't imagine using it more than a couple hours a week, and often not for months at a time, I'm just a hobbyist, so I don't think I'll have any problems.

My main long-term concern is do I want to invest into this scope very much, and then have one of those irreplaceable hybrid IC things die on me... Not that $100 for a 2-ch vertical amp and some probes is a big 'investment'.

As for the timing, the Hz counter on my DMM is accurate to 0.03% + 1 LSD so I should be able to dial it in to 1khz exactly. (or plenty close). But calibrating at 1khz really is close enough for say, 10Mhz, 100Mhz, or even 1Ghz? I figured I'd need some insane oscillator & long-period frequency counter to dial in those ranges...

[tekfan]

I don't think the hybrids are much of a problem in these scopes. The Tek 2400 series seem to be the ones with the problematic hybrids. If you're in the US then you're lucky. You can get plugins and probes for next to nothing. It's not like this in Europe. People seem to realise the power of older well built instruments. A Tek 7104 goes for about 500 euros here.

The time bases have air variable capacitors for the fastest ranges and trimpots for the slower ranges. If nobody touched them since the thing was last calibrated the sweep speeds should be fine.

All that's really needed to adjust all the sweep speeds is a time mark generator:
https://www.eevblog.com/forum/index.php?topic=5574.0

[FenderBender]

Hey I saw that Time Mark Gen a while ago. Cool piece of kit.

My scope should be here either Monday or Tuesday. I have a little digital function generator. Can I do the calibration solely with a Function Generator?

[tekfan]

As long as you know the frequency, you can easily calibrate the slower ranges. If for example you have a 10MHz function generator, then the fastest sweep speed you can calibrate is 10ns/div. Since you need to see the whole cycle to accurately calibrate the sweep speed.

Most oscilloscopes have several sweep speeds grouped to a single trimpot. You are relying on the accuracy of the resistors in the time base for the other time/div settings.

[FenderBender]

You would use Square wave I presume?

My function generator can only do 5Mhz. Pretty pointless?

[tekfan]

You can use any waveshape but the square wave has the most defined edges making it easy to align it to the graticule markings.

It's pretty pointless to try to calibrate a scope with a 5MHz function gen. Although you can still try and see if the various lower sweep speeds are still relatively accurate.

Remember, a scope isn't made for precise frequency measurments. That's more the realm of frequency counters.

I'm assuming you've got the Tek 465M. First check if there even is the need to calibrate it. If there isn't then just leave it alone.
Dave made a nice video on how he calibrated his Tek 2225. That's really all you need to do for general use. Again using square waves for both frequency and amplitude adjustments is much easier than using sine waves.

[FenderBender]

Right right. I guess a scope is better for seeing how a waveform is affected by different factors. I did watch that video but I can't say I remember anything. To youtube...

And this is my beloved function generator: http://www.ebay.com/itm/5MHz-DDS-Digital-Signal-Generator-Module-Sweep-Function-CPLD-STM32F103-/150725450410?pt=LH_DefaultDomain_0&hash=item2317efdaaa

=/

I didn't even realize at the time I bought it that I couldn't adjust the amplitude with this function generator...WTF is that? Grrr

[alm]

Note that square waves don't necessarily have a duty cycle of 50%, so using anything less than a full period (eg. half a period) may introduce an error.

[FenderBender]

I don't quite understand what you mean..