Conversion of 500MHz TDS744A to 1GHz TDS784A

[Wuerstchenhund]

The problem with CCFLs is they have a very finite life, while for the most part CRTs don't.

Of course CRTs have a finite life. The average 'half-time' of a CRT tube is around 20k hrs, after which the brightness will have fallen by 50%, and after which it will decline further. In addition, the lifetime is limited by some of its parts, i.e. the heater filament.

CCFL tubes have a life span of somewhere in the range of 10k to 100k hrs.

(Though, to be fair, I hear some of the NuColor shutters also have a finite life due to bonding issues.)

This is just another issue on top.

Just because LCDs are a cheaper technology doesn't make them better.

LCD is not better because it's cheaper, it is better because it is the better performing technology.

The fact that it's cheaper (these days, it wasn't back in the mid-90s) is just the bonus.

I personally think plasma TVs have much better pictures compared to LCD sets. Unfortunately it's an expensive technology that has heat and burnin issues, that's why it's basically dead.

Yes, and also because modern LCDs are much better than they were ten years ago.

Plasma's strenght has been its superior reproduction of black, but that was back then. The price one paid for plasma, as you said, was fragility, heat/power consumption, and a very limited service life.

The NuColor display on the TDS series is capable of reproducing all 256 colors of the VGA palette pretty clearly, so I don't know why you're saying it's not capable of it?

What makes you think the TDS uses all 256 colors? From what I remember (OK, it's been a while) the UI itself uses just 16 out of 256 colors, and color grading I believe was limited to 16 or 32 colors.

I'd like to see that press release from Tek talking about NuColor being cheaper than a color CRT. For some reason I just can't imagine it being cheaper.

I don't have the original release but I managed to find this:

http://www.electronicproducts.com/Test_and_Measurement/Oscilloscope_makers_ride_the_color_bandwagon.aspx

...The substantially lower cost of the NuColor display when compared with traditional shadow-mask displays played a big part in Tektronix being able to price the scopes so aggressively. Shadow-mask displays used in color DSOs (including their additional memory, power, and circuit needs) from Tektronix, Hewlett-Packard, and LeCroy in the past have  added about $2,000 to $3,000 to the cost of a color scope. The estimated added costs with the NuColor display is only about $500 or so, according to Martinez.

It shows that at least Tek believed it's cheaper to do NuColor than use a color CRT.

Hell, HP had been using color CRTs since the 80's on their logic analyzers. LeCroy had some of those big ass 1GHz scopes with color CRTs, mid to late 90's vintage, right? (The ones with the thermal printers; they replaced the models with Amber CRTs.)

Yes, but that doesn't show that color CRTs were cheaper. It shows that, at that time, LeCroy and HP thought that this is the appropriate technology for their digital scopes (and other kit in case of HP). Don't forget that in the digital scope market Tek suddenly found itself surrounded by much stronger competition than in the analog scope market (where Tektronix was pretty much the technology leader).

Keep in mind that NuColor wasn't originally designed for use on a raster scan display. It was designed to produce a color electrostatic CRT for analog oscilloscopes. (It also commanded a $1500 premium compared to the B&W only model when it was originally released in 1984!)

I can't remember about NuColor for analog scopes. What scope models was it used in, and what for?

For use on a scope, faced with the choice between:

A) NuColor Display from 1995
B) Color CRT from 1995
C) Color LCD from 1995

I'll take A every single time, with B coming in second. In fact I'd take a monochrome CRT before I took option C!

In 1995 there was no big brand scope with color LCD I remember of. Which means for a 1995 scope I'd probably vote B over A, mainly because I would get a larger screen (the scopes with color CRT like the HP 54700 Series had a larger screens than the TDS series). NuColor would have given me more contrast but for a scope display that is pretty irrelevant, although NuColor would also give less flickering.

Going forward to say 1998, I'd go for C every time. Even back then the readability and sharpness of a LCD like the ones used in the HP 545xxC, 546xxC or the LeCroy WR LT is *a lot* better than even a mono CRT or NuColor could ever provide. Colors might not be so overly saturated as on NuColor (but then, color saturation is was still great at least when the CCFL tube was new, it's only after many hours when the lamp has faded and shifted its color temperature that colors start looking washed out), but frankly for a scope I couldn't care less. As long as the display is clear, crisp and easily readable then it's fine.

I guess my point is, up until the early 2000's, LCDs were not a realistic option for oscilloscopes (yes, a few had them, but they were either absolute budget models with blurry monochrome LCDs or very high end models with clear yet expensive color displays).

Yes, but then the TDS700 wasn't exactly a bargain bin scope, too. Price-wise it very much did compete with the HP 545xx/546xx and 547xx, or the LeCroy LC 600 and WR LTs. In fact, some of the TFT equipped scopes were notabily cheaper than the TDS700 with NuColor. It may have saved Tek some money, but they certainly didn't pass that on to the customer.

[dxl]

My TDS754 has a TFT, but only because i had to replace the NuColor Display. The Picture was almost invisible after the tube had being overdriven for 15 Years or so. That's one of the downsides on NuColor: The Shutter display 'eats' a certain amount of light, so Tek decided to crank up the brightness of the Tube which decreases it's life span. Another thing i never liked on the TDS700 series is the glossy shutter. Put the TDS700 in a Room where there's a window on the other side of the room. You have a reason to be happy every day when you have the Sun as Scope background

I don't have much devices from the 90s with color display. But i can say for sure that even the HP70004A display from my Spectrum Analyzer doesn't look bad compared to NuColor. But that display has a Sony Trinitron tube, which was probably way more expensive then a normal Color tube.

[timb]

I don't have the original release but I managed to find this:

http://www.electronicproducts.com/Test_and_Measurement/Oscilloscope_makers_ride_the_color_bandwagon.aspx

...The substantially lower cost of the NuColor display when compared with traditional shadow-mask displays played a big part in Tektronix being able to price the scopes so aggressively. Shadow-mask displays used in color DSOs (including their additional memory, power, and circuit needs) from Tektronix, Hewlett-Packard, and LeCroy in the past have  added about $2,000 to $3,000 to the cost of a color scope. The estimated added costs with the NuColor display is only about $500 or so, according to Martinez.

It shows that at least Tek believed it's cheaper to do NuColor than use a color CRT.

Keep in mind that NuColor wasn't originally designed for use on a raster scan display. It was designed to produce a color electrostatic CRT for analog oscilloscopes. (It also commanded a $1500 premium compared to the B&W only model when it was originally released in 1984!)

I can't remember about NuColor for analog scopes. What scope models was it used in, and what for?

The 5116 was the first use of Tek's NuColor technology.

The purpose was to allow different channels to be displayed in different colors. (And for the readout to be color coded as well.)

As for that article you linked, I imagine a lot of that cost they estimate was in the memory itself (which would have been hundreds of dollars per MB in the early to mid 90's), but I don't see why a color CRT would require any more memory than a NuColor display? The power requirements should be similar as well. It would need additional circuitry, but so does the NuColor shutter!

After thinking about it, I guess a small color CRT could have been expensive to produce. Most color TVs started out at 13" and color monitors at 12" so going below that would have increased cost, especially since they didn't have the economy of scale to bring the cost down. By that time Tek would have been making the NuColor shutter for over 10 years, so $500 seems like a reasonable figure.

I just think it was an interesting technology.

Re: TDS color depth... I don't think the TDS displays more than 20 different colors on the screen at any one time; however each color can be selected from any in the VGA palette. You can go into the display menu and scroll through the color list to see for yourself. This isn't a limit of the NuColor hardware, there's just no need to display more than that on this particular scope.

The proof that the hardware is capable of a full color gamut is evident when you look at other products using the technology: Tek made a NuColor computer monitor; JVC made portable color video monitors using the NuColor system, which is capable of displaying any color a normal color CRT TV can.

I do agree the technology is a dead end now and LCDs are clearly better, but like I said earlier, it's fun to speculate what if. I think, as another poster mentioned, the big disadvantage to them is how much light the shutters eat up, requiring you to overdrive the CRT. It's a shame as with 10 more years of development, I think they could have improved on that a lot. Ah well.

Oh, one more thing. Even a modern LCD still struggles to get the black levels of a 10 year old plasma. I also imagine they could have corrected a lot of the deficits, had it continued to be developed like LCDs have. No display technology that needs a backlight will ever be as good as one that produces its own light. Period. That's why I'm keeping my fingers cross and waiting for OLED sets to be perfected before I replace my plasma.

Edit: I almost wonder if the reason I like the NuColor display so much is because of the high refresh rate, which eliminates the 60Hz flicker found on most other raster scan style CRTs of the era. (I had to run my gaming monitor at 85Hz, minimum.)

[siggi]

The 5116 was the first use of Tek's NuColor technology.

Wow, that was a pretty fun read. Toward the end it says "As time goes on then, it is reasonable to assume that the color shutter will be used in a wide variety of products and applications.". Even as the cover photo in the article demonstrates how the trace intensity is skewered by the LCD shutter .

[Jwalling]

The 5116 was the first use of Tek's NuColor technology.

Wow, that was a pretty fun read. Toward the end it says "As time goes on then, it is reasonable to assume that the color shutter will be used in a wide variety of products and applications.". Even as the cover photo in the article demonstrates how the trace intensity is skewered by the LCD shutter .

+1
I had no idea that this technology was around in 1984. I have fond memories (no pun intended) of both Radio Electronics and Popular Electronics. (Still have a bunch of the old magazines) I especially remember the Cosmac Elf 1802 project - my first (big) project I ever did; don't remember what happened to it, but wish I kept it...
I always loved the back pages with all the surplus sellers such as PolyPaks, etc.

EDIT: Link  https://en.wikipedia.org/wiki/COSMAC_ELF
Mine didn't look that nice, though 
Jay

[nctnico]

I don't have the original release but I managed to find this:

http://www.electronicproducts.com/Test_and_Measurement/Oscilloscope_makers_ride_the_color_bandwagon.aspx

...The substantially lower cost of the NuColor display when compared with traditional shadow-mask displays played a big part in Tektronix being able to price the scopes so aggressively. Shadow-mask displays used in color DSOs (including their additional memory, power, and circuit needs) from Tektronix, Hewlett-Packard, and LeCroy in the past have  added about $2,000 to $3,000 to the cost of a color scope. The estimated added costs with the NuColor display is only about $500 or so, according to Martinez.

It shows that at least Tek believed it's cheaper to do NuColor than use a color CRT.

Keep in mind that NuColor wasn't originally designed for use on a raster scan display. It was designed to produce a color electrostatic CRT for analog oscilloscopes. (It also commanded a $1500 premium compared to the B&W only model when it was originally released in 1984!)

I can't remember about NuColor for analog scopes. What scope models was it used in, and what for?

The 5116 was the first use of Tek's NuColor technology.

The purpose was to allow different channels to be displayed in different colors. (And for the readout to be color coded as well.)

As for that article you linked, I imagine a lot of that cost they estimate was in the memory itself (which would have been hundreds of dollars per MB in the early to mid 90's), but I don't see why a color CRT would require any more memory than a NuColor display? The power requirements should be similar as well. It would need additional circuitry, but so does the NuColor shutter!

The NuColor display in the Tektronix scopes actually needs more memory. What the display system does is triple the display data coming from the acquisition/CPU overlay (640x480 @60Hz) and split it into 3 color channels and pull that through a RAMDAC which runs at 3 times the clock frequency of regular VGA. Hence the CRT also runs at 180Hz instead of 60. AFAIK the whole NuColor display wasn't to save cost but just to be able to use a rugged monochrome CRT with one big electron cannon instead of 3 smaller ones. After all the TDS500/600/700 series are built like tanks.

With a normal color CRT they could have used the RAMDAC which is already there for the external VGA output.

[David Hess]

NuColor displays also allow higher resolution and absence of convergence compared to a color CRT.

Modern LCDs have only just reached the DPI of old vector CRTs and LCDs with that kind of performance are not used in DSOs yet anyway.

[eKretz]

Just finished converting my 754A, FW v4.1e with no issues. Wish I could do the 2M option too, oh well, at least I've got 1M. Got to work on the calibration next.

[Jwalling]

Just finished converting my 754A, FW v4.1e with no issues. Wish I could do the 2M option too, oh well, at least I've got 1M. Got to work on the calibration next.

On the A scopes, 2M was never an option.

Jay

[eKretz]

Yup, I know. Anyone know if the processor and acquisition boards are the only difference between A and D?

[Jwalling]

Yup, I know. Anyone know if the processor and acquisition boards are the only difference between A and D?

The LV power supplies and HV assemblies are interchangeable. The processor on the D scopes is noticeably faster, and is very noticeable when switching back and forth from DPO/Instavue

Jay

[eKretz]

So switching the processor and acquisition boards might net me a 784D then. Hmm. May need to start watching eBay.

[tford]

I may have to replace my upgraded TDS744A CRT with an LCD soon because the shutter module appears to be de-laminating. For anyone that's been down this road before, does the LCD replacement have comparable resolution and ease of viewing? Thanks, Tom.

[Jay_Diddy_B]

I may have to replace my upgraded TDS744A CRT with an LCD soon because the shutter module appears to be de-laminating. For anyone that's been down this road before, does the LCD replacement have comparable resolution and ease of viewing? Thanks, Tom.

I suppose it depends on how good an LCD display you get. I converted one of my TDS5xx or TDS7xx to LCD because the CRT was no good. I was never really happy with it. I used to use it as a portable scope, because it made the scope lighter. I prefer the liquid crystal shutter / monochrome display 'NuColor'.


The challenge is you need to find a VGA 640 x 480 LCD the right size. Most of the VGA resolution screens are not that great.

Regards,

Jay_Diddy_B

[snoopy]

For those interested the original brochure for the 700D scopes

https://www.testequity.com/documents/pdf/tds700d.pdf

Quite impressive for the time and still is

[Wuerstchenhund]

Quite impressive for the time and still is

Not really, no. Not in 2001 (the date of that brochure).

[KE5FX]

Quite impressive for the time and still is

Not really, no. Not in 2001 (the date of that brochure).

Who else was doing 200K waveforms/sec?  That was really the DPO's claim to fame.

[Wuerstchenhund]

Quite impressive for the time and still is

Not really, no. Not in 2001 (the date of that brochure).

Who else was doing 200K waveforms/sec?  That was really the DPO's claim to fame.

HP already had MegaZoom, which managed higher waveform update rates without the issues of DPO (which in those scopes only worked at 1GSa/s or less which also limits the usable BW to less than 450MHz or so, could only use up to 0.5Mpts of memory so sample rates often had to drop considerably which meant further reduction in usable BW, couldn't do measurements, and suffered from loss of signal information through ineffective min/max compression), plus MegaZoom  wasn't a separate mode, it was always on.

[snoopy]

Quite impressive for the time and still is

Not really, no. Not in 2001 (the date of that brochure).

Who else was doing 200K waveforms/sec?  That was really the DPO's claim to fame.

HP already had MegaZoom, which managed higher waveform update rates without the issues of DPO (which in those scopes only worked at 1GSa/s or less which also limits the usable BW to less than 450MHz or so, could only use up to 0.5Mpts of memory so sample rates often had to drop considerably which meant further reduction in usable BW, couldn't do measurements, and suffered from loss of signal information through ineffective min/max compression), plus MegaZoom  wasn't a separate mode, it was always on.

You could turn off DPO and the earlier A and C models never had it anyway. Also the 784A was 1995 vintage and still could do 400K wfrms/s.

http://www.axiomtest.com/documents/models/Tektronix%20TDS784A%20Data%20Sheet.pdf

Not sure what HP had at the time.

What was the model and update rate of that HP scope ?

Anyways the Tek TD694 had 3GHz bandwidth and 10Gs/s on ALL channels although it didn't have InstaVu and Fast Frame (Segmented memory) that the later 700 series had.

http://lsm.epfl.ch/files/content/sites/lsm/files/shared/Equipment/Tektronix_TDS694C.pdf

http://www.testequity.com/documents/pdf/tds684-94c.pdf

[Wuerstchenhund]

You could turn off DPO and the earlier A and C models never had it anyway.

You had to turn it off if you wanted to use the full BW and sample rate, and the measurements.

Also the 784A was 1995 vintage and still could do 400K wfrms/s.

Yes, they had InstaVu, a non-graded persistence display. And similar to DPO, it came with the same disadvantages.

Not sure what HP had at the time

At that time HP came out with it's first generation of MegaZoom in the 54600 Series of scopes.

What was the model and update rate of that HP scope ?

The first one were the 54645A and the 54645D (MSO version of the 54645A).

Can't remember what the waveform rate was for the original but I believe it was somewhere in the 250k wfms/s region. The important bit however was that it was not just one large number at a specific setting (as with InstaVu/DPO), it reached high update rates in almost any setting.

Anyways the Tek TD694 had 3GHz bandwidth and 10Gs/s on ALL channels although it didn't have InstaVu and Fast Frame (Segmented memory) that the later 700 series had.

And it came with a miniscule amount of sample memory (30kpts, 120kpts as "long memory" option) which meant it could only sustain the sample rate at the shortest timbease settings and with longer time bases (i.e. anything over 300ns/s with standard memory or 100ns/div with "long" memory) the sample rate dropped like a stone, limiting the usable BW to a fraction of that 3GHz bandwidth, plus it lacks any advanced measurements or math functions that other scopes had at that time.

The TDS694C is a good example what's wrong with Tek: they build a DSO that was good or even great in one aspect but lacked in pretty much everything else. Which admittedly was better than it is today, where Tek scopes lack pretty much in every aspect and excel in none. Something that is reflected in their continued dwindling market share and plummeting sales.

The key to a good scope isn't in producing some single outstanding feature (that is only good for marketing), it's in a well-balanced approach to the various different properties that define a scope. Tek understood that for analog scopes, but for DSOs they have always seemed to struggle, and still are.

Anyways, I didn't want to derail this thread into a comparison of manufacturers. My point was that the TDS784C wasn't as  impressive for a scope of that class back then in 2001 as you believe it was, and much less so in 2016.

[nctnico]

IMHO the TDS600 series was specifically made for measurements at the highest samplerate to take single shots of high frequency signals and nothing else. The early ones are totally useless as general purpose scopes due to the lack of peak-detect but even with peak detect their memories are short. The TDS500 and TDS700 series are the models suitable for general purpose use.

[David Hess]

IMHO the TDS600 series was specifically made for measurements at the highest samplerate to take single shots of high frequency signals and nothing else. The early ones are totally useless as general purpose scopes due to the lack of peak-detect but even with peak detect their memories are short. The TDS500 and TDS700 series are the models suitable for general purpose use.

This is exactly the case.

The TDS600 series was a continuation of Tektronix's CCD based DSOs optimized for higher samples rates than were available using more conventional designs.  They are more like transient digitizers with oscilloscope ease of use than general purpose oscilloscopes.  If you wanted maximum sample rate on every channel, then this is what you used.  At the time they were produced, they had 10 times (total) or 40 times (per channel) the real time sample rate of the fastest TDS700 series.

Since they used CCDs, features like segmented memory (FastFrame) and DPO operation (InstaView) could not be included.  The early TDS600 models even lacked peak detection which is only surprising because their 2440 ancestors had it; I assume this was a difficult implementation detail that Tektronix had to figure out for inclusion in the TDS600B and later models.

I wonder if they sold a lot of these to nuclear weapon testers like the ancient 519 1 GHz analog oscilloscope.

[timb]

Speaking of FastFrame... What sort of use case is it for? There's no example in the manual and I never could really think of a situation where it would be useful. (Obviously it is useful, or it wouldn't be a feature!)

[rx8pilot]

(Obviously it is useful, or it wouldn't be a feature!)

I don't know.... a lot of features are designed by the marketing dept and totally useless.

[timb]

(Obviously it is useful, or it wouldn't be a feature!)

I don't know.... a lot of features are designed by the marketing dept and totally useless.

I suppose that's true for a lot of things, though I cant think of many superfluous features on test gear.

Doesn't your new Keysight scope have a voice control feature? I would think something like that was conceived by the Marketing Department. ("Hmmm, Siri is really helping the iPhone sell... We should do that, but on an Oscilloscope!")

That said, if implemented well I can see how it would actually be very useful.

I'm still on the fence about touchscreen based instruments though. It's been tried in many forms over the years, and no one has quite gotten right. (Maybe some of the newer, more expensive scopes that have nice, big capacitive touch screens might work well, I haven't tried them; IR and resistive touch screens of the past just don't cut it.)