Old analog or modern digital oscilloscope for low budget. Need advice.

[oldway]

It is not easy to compare CRT's about sharpness and brightness because you are limited by the risk of burning the phosphor and there is no way to measure it.
Then, it is a bit subjective.
But I think the best CRT I know is the one of my Hameg HM 1505 (150Mhz)
The CRT of HP1740A is also very good.

A great feature of the HM 605 and other HAMEG analog oscilloscopes is the component tester.
It works like the HUNTRON tracker and is very useful for repair.

The 465 / 465B are very reliable, far more reliable than the 2235.
I recomend to install a fan in your 2235, like the 2236, because SMPS works very hot.
475 has 3 time bases, 2 mains (slow and fast) and 1 delay and is more complicated than 465 for this reason.

Probes of higher bandwith than 100 Mhz are expensive ...It is a nonsense to buy an high bandwith oscilloscope (200 or more Mhz) and use it with 100Mhz probes.

The 465/475 was designed to fit under aircraft seats.

So far I know, it was the 2335/2336/2337 not the 465/475.

HP\Agilent 546xxx are digital and not analog oscilloscopes...Not as simple as an analog one....their weaker point is the display : it is a little monochrome TV crt that has not a very long lifetime and hard to find.

[David Hess]

At one time, it was sought to sophisticate it and to make it perform functions for which it was not originally conceived as measures of tensions and time with high precision.   These are the analog oscilloscopes with cursors.

Not all analog oscilloscopes use cursors for increased accuracy.  Differential comparator inputs allow slideback voltage measurements and delta delayed sweep allows slideback time measurements.

Besides usually being more accurate, cursors are also easier to use so there is human factors engineering to consider.  Automatic measurements are easier yet except when they automatically measure the wrong thing without notifying you.  There is a big convenience factor here.

We also wanted to combine the advantages of the analog oscilloscope and digital, which produced "ugly little lame ducks" called combiscopes.

Do not waste your time and money with these lame hybrids.

I usually disapprove of unneeded complexity if only because of a loss in reliability but I make an exception for some of the Tektronix combination analog and digital storage oscilloscopes.  The only reason I do not recommend them for beginners is that their extra complexity does make maintenance more difficult.  In the case of the OP's question, they are also likely outside the budget limitation of $100 although I bought each of my 2230s for less than that.

The various 22xx combination analog and digital storage oscilloscopes use a separate "storage" board for their digital storage functions.  If the board is removed or disabled, then they operate identically to the basic analog 2235 or 2225 model.  Instead of a storage board, the 2236 and 2236A have a universal time/counter board which gives them almost uniquely useful capabilities shared with the 4 channel 2247A and 2252 models.

The 22xx oscilloscopes which include DSO functionality also have a readout in analog mode which is nice.

The DSO displays on these oscilloscopes are very high resolution making them clearer than you would expect.  100 points per linear centimeter or one division is 254 dpi which is considerably higher than most if not all modern DSOs.  HP had some raster CRT based DSOs with unusually high resolution as well.

If you want the storage function or if you want to make accurate measurements, buy a digital oscilloscope, there is no match.

The analog oscilloscope is still very useful in various fields, in repair, in low precision measurements on complex periodic analog signals.

Slideback measurements may be more accurate than a DSO or cursor measurement but general purpose analog and digital storage oscilloscopes are not the instrument to use for precision.  The old Tektronix 2440 series DSOs support dual delta delayed sweep which can be used for slideback time measurements but I have no idea why Tektronix included that.

Good DSOs are however much more convenient to use for measurements especially because of their automatic continuous measurement capability.

[David Hess]

The 475, 475A, and 485 may have a brighter and sharper CRT than the 22xx oscilloscopes.

Why is that ? Better or different CRT?

These CRTs are all basically the same design but the faster CRTs, especially the 485, have higher total acceleration voltages making them brighter and sharper.

The 465B and 485 have alternate dual delayed sweep.  Most of the other 4xx oscilloscopes have mixed sweep which is not as useful.

So in 4xx series, these two are considered to be top models ?

The 465B (includes the 468) was the last 465 series design.  The 485 is actually a year older than the original 465 but could be considered a performance and feature at any price design.  I actually consider the 465M which is based on but very different than the 465 to be best of them except for the lack of alternate sweep.

[oldway]

Besides usually being more accurate, cursors are also easier to use so there is human factors engineering to consider.  Automatic measurements are easier yet except when they automatically measure the wrong thing without notifying you.  There is a big convenience factor here.

This may be a concern on a production line where you have to make hundreds of measurements every hour, but not for a beginner.
If you need such convenience, buy a new DSO, not an old analog scope.
These analog oscilloscopes are at least 20 years old or more and reliability and maintenance are the greatest concern.
Other thing to remember: SMPS and SMD's electrolytic capacitors have well known reliability issues...try to avoid oscilloscopes with such problems.

High bandwith CRT's (like 2465) have distributed vertical deflection plates and this reduce the sharpness of the beam.

Oscilloscopes that where the best choice when they where new are no more the best choice 20 years later because they are to difficult to maintain.

How to measure the frequency of the displayed signal without cursors ?
My Hameg HM 605 has an Y output on the rear panel. I connected a low cost frequency counter to this output so I can measure the frequency of the A channel.

[David Hess]

It is not easy to compare CRT's about sharpness and brightness because you are limited by the risk of burning the phosphor and there is no way to measure it.
Then, it is a bit subjective.
But I think the best CRT I know is the one of my Hameg HM 1505 (150Mhz)
The CRT of HP1740A is also very good.

Tektronix used a scan expansion mesh in most of their CRTs sometime after the 50 MHz 547 which is great for increased deflection sensitivity which is important for high bandwidth and short CRT necks but has a side effect of making the traces thicker and causing ghosts.  Some competing oscilloscopes did not allowing them to offer a sharper display.  I think the one contemporary I have seen which looked obviously sharper than a 2235 was made by Hitachi.

One story I have heard is that TV stations were loath to give up their Tektronix 547s because the 7000 replacements with their scan expansion meshes were not able to resolve the NTSC horizontal sync pulses nearly as well.

Sharpness can be measured using a raster.  Brightness can be measured by using a low repetition rate signal at a fast sweep.

The 465 / 465B are very reliable, far more reliable than the 2235.
I recomend to install a fan in your 2235, like the 2236, because SMPS works very hot.

Most of the problems in the 2235 series seem to be from the switching power supply which suffers from at least one mysterious failure mode (transformer damage from overheating?) but the original 2213/2215 power supply was just as bad if not worse.  I do not really consider this a serious problem for someone with electronics experience and I suspect most of the failures are do to not replacing the aluminum electrolytic capacitors in a timely manner.

The 4xx series oscilloscopes definitely have an advantage here with their linear power supplies but that comes with the cost of increased weight.

475 has 3 time bases, 2 mains (slow and fast) and 1 delay and is more complicated than 465 for this reason.

I am not sure what you are referring to here.  The delay function is part of the main (A or delaying) timebase.  The 485 had an innovation shared with the 7B92 dual delaying timebase where the A and B timebases are swapped at high sweep speeds simplifying the sweep integrators.

Probes of higher bandwith than 100 Mhz are expensive ...It is a nonsense to buy an high bandwith oscilloscope (200 or more Mhz) and use it with 100Mhz probes.

I am not sure who is making them, but Texas (Hong Kong) sells the same inexpensive x10 probes up to 300 MHz which HP/Agilent/Keysight and some other oscilloscope manufacturers sell and include with their oscilloscopes and they seem pretty good.

HP\Agilent 546xxx are digital and not analog oscilloscopes...Not as simple as an analog one....their weaker point is the display : it is a little monochrome TV crt that has not a very long lifetime and hard to find.

I think these were the ones where HP made a special effort to almost double the horizontal resolution of the monochrome CRT which earned a patent.

The Tektronix went with using an LCD shutter to deliver color from a monochrome CRT because color CRTs were not good enough.

[tggzzz]

Probes of higher bandwith than 100 Mhz are expensive ...It is a nonsense to buy an high bandwith oscilloscope (200 or more Mhz) and use it with 100Mhz probes.

And, as I'm sure you know, probing technique is at least as important as the probe.

The 465/475 was designed to fit under aircraft seats.

So far I know, it was the 2335/2336/2337 not the 465/475.

A little googling makes me think we are both wrong: it was the 453. http://readingjimwilliams.blogspot.co.uk/2011/07/scope-sunday-3.html (My 475 also says "Property of IBM")

[tggzzz]

The 475, 475A, and 485 may have a brighter and sharper CRT than the 22xx oscilloscopes.

Why is that ? Better or different CRT?

These CRTs are all basically the same design but the faster CRTs, especially the 485, have higher total acceleration voltages making them brighter and sharper.

My favourite CRTs are the 485 and the HP1740A. The advantage of the hp1740a is that the screen is larger - but having a small screen didn't hurt the original Macintosh computers

[tggzzz]

Besides usually being more accurate, cursors are also easier to use so there is human factors engineering to consider.  Automatic measurements are easier yet except when they automatically measure the wrong thing without notifying you.  There is a big convenience factor here.

This may be a concern on a production line where you have to make hundreds of measurements every hour, but not for a beginner.
If you need such convenience, buy a new DSO, not an old analog scope.

Most beginners think, possibly correctly, that automated measurements are a big advantage of DSOs. But, as we know, David's point about silently measuring the wrong thing is only too true.

[tautech]

On CRO CRT's the best sharpest and brightest/crisp that I've had were:
1. Telequipment D83, a 50 MHz dual channel unit with a mainframe construction (w plugins) made by Tek after they acquired UK based Telequipment. It had the largest display of any CRO I've owned and the blue filter along with the sharp raster made it a joy to use. 15 KV acceleration voltage.

2. An Aussie made BWD, #452 IIRC, a nice and reasonably compact 50 MHz CRO, nothing special but the exceptionally sharp raster.

3. HP1740A, nice to work on but the trace pales in comparison to the above.

A few others, not worth mentioning.

[David Hess]

High bandwidth CRT's (like 2465) have distributed vertical deflection plates and this reduce the sharpness of the beam.

I do not think the distributed deflection plates have much to do with it.  Distributed deflection plates are found on CRTs which also have scan expansion meshes simply because both help achieve high bandwidth.  The increase in deflection sensitivity provided by scan expansion reduces the output voltage requirements of the vertical output amplifier.  Scan expansion also allows for a shorter CRT neck

Oscilloscopes that where the best choice when they where new are no more the best choice 20 years later because they are to difficult to maintain.

Analog oscilloscopes are the best choice if their low used price is the difference between having an oscilloscope and not having one.  They are also a good choice if high bandwidth is required at the lowest cost.

How to measure the frequency of the displayed signal without cursors ?
My Hameg HM 605 has an Y output on the rear panel. I connected a low cost frequency counter to this output so I can measure the frequency of the A channel.

Or count divisions.  Or use delayed sweep.  Or make a slideback measurement with delta delayed sweep.  Or use a DSO to make an automatic measurement.  None of these will be as accurate as a frequency counter but oscilloscopes are also not as accurate as voltmeters.  Use the right tool for the job.

My favorite method is to connect the vertical output to the frequency counter and the B gate output to the frequency counter's arm input.  Now a gated measurement of frequency and time can be made which is more accurate than a common DSO can achieve.  Or a 2236, 2247A, or 2252 analog oscilloscope which have this capability built in can be used.

[tggzzz]

Analog oscilloscopes are the best choice if their low used price is the difference between having an oscilloscope and not having one.  They are also a good choice if high bandwidth is required at the lowest cost.

That succinct statement should be posted in any of the threads about "which scope should I buy". The earlier that posting, the shorter the thread will be (As if!)

Or count divisions.  Or use delayed sweep.  Or make a slideback measurement with delta delayed sweep.  Or use a DSO to make an automatic measurement.  None of these will be as accurate as a frequency counter but oscilloscopes are also not as accurate as voltmeters.  Use the right tool for the job.

Yes Too many people have hammers and not screwdrivers - which is OK and is the inspiration for interesting engineering.
Too many people argue that you only need a hammer - which isn't OK.

[David Hess]

These CRTs are all basically the same design but the faster CRTs, especially the 485, have higher total acceleration voltages making them brighter and sharper.

My favorite CRTs are the 485 and the HP1740A. The advantage of the hp1740a is that the screen is larger - but having a small screen didn't hurt the original Macintosh computers

Over years I gotten used to CRTs in the the 22xx series (14kV) but then I brought my 7603 (15kV) out of retirement and despite its dimmer CRT, its large size is very nice.  Then I picked up a 7834 (10kV or 12kV) which is bright for a storage CRT but roughly equivalent to the 7603.  The first time I really noticed a difference was with the 400 and 500 MHz 7000 mainframes which have much higher acceleration voltages.  My 7904 (24kV) makes every other oscilloscope I have look dim.

Most of the 465 series are 18kV and the 485 is 21kV.

There is more to brightness and sharpness than acceleration voltage like age, operating hours, and design, but acceleration voltage is a good place to start.

[oldway]

475 has 3 time bases, 2 mains (slow and fast) and 1 delay and is more complicated than 465 for this reason.

I am not sure what you are referring to here.  The delay function is part of the main (A or delaying) timebase.  The 485 had an innovation shared with the 7B92 dual delaying timebase where the A and B timebases are swapped at high sweep speeds simplifying the sweep integrators.

From 475A service manual: two non delaying sweep generators, one slow and one fast, and one delaying sweep generator.

How to measure the frequency of the displayed signal without cursors ?
My Hameg HM 605 has an Y output on the rear panel. I connected a low cost frequency counter to this output so I can measure the frequency of the A channel.

Or count divisions.  Or use delayed sweep.  Or make a slideback measurement with delta delayed sweep.  Or use a DSO to make an automatic measurement.  None of these will be as accurate as a frequency counter but oscilloscopes are also not as accurate as voltmeters.  Use the right tool for the job.

My favorite method is to connect the vertical output to the frequency counter and the B gate output to the frequency counter's arm input.  Now a gated measurement of frequency and time can be made which is more accurate than a common DSO can achieve.  Or a 2236, 2247A, or 2252 analog oscilloscope which have this capability built in can be used.

HM 605 has no delayed sweep, nor B gate output.
The first 2236 (serial number below more or less 10.000 24.000) have a very bad problem (very bad and fragile potentiometer D time position R1617 on counter/timer/multimeter board) who make it useless.

[David Hess]

475 has 3 time bases, 2 mains (slow and fast) and 1 delay and is more complicated than 465 for this reason.

I am not sure what you are referring to here.  The delay function is part of the main (A or delaying) timebase.  The 485 had an innovation shared with the 7B92 dual delaying timebase where the A and B timebases are swapped at high sweep speeds simplifying the sweep integrators.

From 475A service manual: two non delaying sweep generators, one slow and one fast, and one delaying sweep generator.

I see what you mean.  I might have known this at one time but certainly forgot.

This may figure into the improvements Tektronix made in the 485 and 7B92 which only require one integrator for each sweep generator.  For them, the delaying sweep generator is slow and the delayed sweep generator is fast.  When the delaying sweep is used for the fastest sweeps, the delaying and delayed sweep generators are transparently swapped.  This threw a for a loop a couple years ago on the TekScopes@yahoogroups.com email list when we were trying figure out how the delaying timebase generated the fastest sweeps..

Note that the release data are reversed here with the 475 and 475A coming after the 485 and 7B92 but I suspect they were handled by different design teams.

The first 2236 (serial number below more or less 10.000) have a very bad problem (very bad and fragile potentiometer D time position R1617 on counter/timer/multimeter board) who make it useless.

This would not be the first Tektronix oscilloscope design with trimmer potentiometer problems.  Luckily trimmer potentiometers are easy to replace once the problem is found.

[oldway]

This would not be the first Tektronix oscilloscope design with trimmer potentiometer problems.  Luckily trimmer potentiometers are easy to replace once the problem is found.

It would be easy to replace if you could find a spare part....But this is a special trimmer...hard to find.
Look at the service manual and you will understand.
It is a trimmer used for Delta time position, that's a needed control potentiometer.

[siggi]

The old Tektronix 2440 series DSOs support dual delta delayed sweep which can be used for slideback time measurements but I have no idea why Tektronix included that.

I just discovered this feature on my old 2430, it's well hidden. Beyond the basic horizontal/vertical controls, it's hard to love the UI on that scope .

In any case it looks like it has quite accurate delays. The max delay time decreases with increased sample rate, suggesting that the base delay is counting off sample clocks, but the time resolution seems to well exceed the sampling clock.
I'm not sure how they'd achieve that sort of delay resolution, as clearly each of the dual delay captures have to be done on separate trigger events. Maybe the delayed signal is simply offset digitally post capture?

[David Hess]

This would not be the first Tektronix oscilloscope design with trimmer potentiometer problems.  Luckily trimmer potentiometers are easy to replace once the problem is found.

It would be easy to replace if you could find a spare part....But this is a special trimmer...hard to find.
Look at the service manual and you will understand.
It is a trimmer used for Delta time position, that's a needed control potentiometer.

Oh, it is a control potentiometer and not a trimmer potentiometer.  Based on the parts list which shows 4 different parts over time, it does look like Tektronix had problems with it.

The schematic shows a configuration which I would normally assume is a coarse and fine control potentiometer but the operating instructions describe and the photograph show a single control.  I will have to ask over on TekScopes@yahoogroups.com what the deal is.

The old Tektronix 2440 series DSOs support dual delta delayed sweep which can be used for slideback time measurements but I have no idea why Tektronix included that.

I just discovered this feature on my old 2430, it's well hidden. Beyond the basic horizontal/vertical controls, it's hard to love the UI on that scope .

In any case it looks like it has quite accurate delays. The max delay time decreases with increased sample rate, suggesting that the base delay is counting off sample clocks, but the time resolution seems to well exceed the sampling clock.
I'm not sure how they'd achieve that sort of delay resolution, as clearly each of the dual delay captures have to be done on separate trigger events. Maybe the delayed signal is simply offset digitally post capture?

The the delta delayed acquisition is a completely separate acquisition because of the short record length.  It looks to me like the processor just alters the delay clock counter between acquisitions so two separate delayed acquisitions are captured.  I could not find anything in the theory section about the frequencies used by the delay counter but I assume it is a fixed fraction of the sample rate.

I do not know why a cheap modern DSO could not do the same thing trivially but my guess is that the use case is not there.

On the 2247A series and the 2465 series, dual delta delayed sweep is used *instead* of cursors.  You can use one or the other but not both at the same time and on these oscilloscopes, dual delta delayed sweep is activated by selecting the time measurement cursors so the front panel controls serve double duty.

[oldway]

Oh, it is a control potentiometer and not a trimmer potentiometer.  Based on the parts list which shows 4 different parts over time, it does look like Tektronix had problems with it.

The schematic shows a configuration which I would normally assume is a coarse and fine control potentiometer but the operating instructions describe and the photograph show a single control.  I will have to ask over on TekScopes@yahoogroups.com what the deal is.

Yes, in serial number under 24.000, they used a trimmer as control potentiometer and that is bullshit because a trimmer's lifetime has a (very) limited number of operations.
Yes, it is only a single control, it does not match the schematic...and yes, Tektronix had problems with it...

This multimeter/counter board is very sophisticated, that's why I recomend not to buy 2236, but 2235 instead.

[tautech]

The 465B (includes the 468) was the last 465 series design.  The 485 is actually a year older than the original 465 but could be considered a performance and feature at any price design.  I actually consider the 465M which is based on but very different than the 465 to be best of them except for the lack of alternate sweep.

So this one might be worth grabbing if you knew what you were doing:
http://www.trademe.co.nz/electronics-photography/other-electronics/other/auction-1230850237.htm

[David Hess]

The 465B (includes the 468) was the last 465 series design.  The 485 is actually a year older than the original 465 but could be considered a performance and feature at any price design.  I actually consider the 465M which is based on but very different than the 465 to be best of them except for the lack of alternate sweep.

So this one might be worth grabbing if you knew what you were doing:
http://www.trademe.co.nz/electronics-photography/other-electronics/other/auction-1230850237.htm

For the right price, absolutely it would be worth it but include the cost of buying a duplicate with the hope of combining the two to make one good oscilloscope.  That auction does not really say anything which the photographs do not reveal.  Maybe major portions of the instrument are missing inside.

Also consider how many of a specific instrument are available in your market which may take time.  If it is rare like a 2212, 2214, or 2216 and requires repairs, then parts availability will likely be more of a problem.

However there is a lot of uncertainly in many auction descriptions.  The seller may not be prepared to test the equipment at all or there could be a trivial fault which prevents any operation.  The two 2230s I bought with the fallback intention of combining them were "as is" but except for an odd problem with one of them, they both worked fine.  The only thing I knew from the photographs was that they powered on, presented an apparently working display, and showed no physical damage.

[tautech]

From the Q&A the sellers replies:
Hi! I believe it is related to the HV power supply but I didn't trouble shoot it. It does not come on at all. The service manual should help if your are into this type of thing. As the title indicates it is a 100 MHZ scope.

No probes. I bought it faulty and I am selling it as is without attempting to repair it. I have a later model that I am working on.

I'd guess it's beyond him but I won't be grabbing it, if it was another 2465 or a B then I'd be interested.

[David Hess]

From the Q&A the sellers replies:
Hi! I believe it is related to the HV power supply but I didn't trouble shoot it. It does not come on at all. The service manual should help if your are into this type of thing. As the title indicates it is a 100 MHZ scope.

No probes. I bought it faulty and I am selling it as is without attempting to repair it. I have a later model that I am working on.

I'd guess it's beyond him but I won't be grabbing it, if it was another 2465 or a B then I'd be interested.

High voltage problems may require trivial repairs but are often difficult to diagnose.

[T@ller]

Happy New Year to all !


Thanks for tips.

[T@ller]

Finally, I have got Tek 2235!

First impression - it is smaller and lighter than I thought.

Second impression - yes, it is old man.

After 5 min working something has blown inside.  It was C904 cap (EMI). Was changed on new one, also X2 rated but .22 mkF instead 0.068 mkF (I guess it is not so important there).
Also I have checked (by ESR meter) all electrolytic caps in power supply area - no suspicious caps were found, while I had been thinking otherwise..

Scope seem to work right but I need manual to check all options. Is there some short manual ?

[Basileus]

You sir made a good choice.

Just look for tek 2235 service manual.

Getting a vintage tek scope is a great idea. The manuals tek e are amazing, there are plenty of infos on the net and if something fails you can ask for help at the tekscopes yahoo group (those guys are amazing, they could have lead my grandma through the repair i did on my tek 2215, faulty dtb).

They are beatiful machines, and imo its a ton of fun repair them. A good project for a beginner (if its not a crt fault, dont want to touch that).