The worst products from HP and Tektronix.

[JM1366]

Horrible waste of time, but I need some entertainment.

I think we can agree that most of what HP and Tektronix sold was pretty good kit, but this is a place to talk and laugh about the things that should never have made it to market.

[H713]

I think the 491 makes the list of Tek dogs, though I've never actually seen one in operational condition (I have seen a few dead ones though).

From the Tekwiki:
"Due to severe performance limitations, the 491 was replaced by the completely redesigned 492 in 1980."

Ouch. Double ouch if you spent the equivalent of $40,000 one one back in the late 70s.

[hvontres]

I nominate my TDS460.... a 400Mhz analog Front end with a 100Ms/s Digital backend... but at least they have the "ET" mode.

[srb1954]

HP 3458A, which is both the best and the worst high-end DMM that they made. When you get a good one it has outstanding performance but if you get a bad one it has really terrible drift in the ADC.

What is worse HP/Agilent apparently hadn't got round to fixing the problem in the design even after 25 years from the product's introduction. I am sure that state of affairs would not have been allowed to persist if Bill and Dave were still around.

[wn1fju]

I'll restrict my nominees to pre-1995 stuff because that is mostly what I have:

For HP, they had a period where they decided to compete in the low-end market.  For example, the 3435A, 3465A and 3466A are 3-1/2 or 4-1/2 digit DMMs that are cheaply built and not worthy of HP's state-of-the-art reputation.  Similarly, the 5300A and 5300B are "mainframes" that accept a variety of plug-ins, mostly counters.  Again, they are low-end stuff and if you want a counter, then buy a real one.  The HP (Germany) 811x pulse and function generators are very nice - when they work.  Almost everyone I see offered on eBay is throwing errors indicative of faulty HP-custom ICs that are, of course, unobtainable.  Finally, you would think when HP developed their 6-1/2 digit DMM HP 3457A, it would be better than the preceding 3455A and 3456A.  But I find the 3457A slow, noisy, hard to see the LCD display, and generally a "dog."

For Tektronix, their stuff seems electrically excellent.  But mechanically, that is another story.  Repair and maintenance of some of their earlier scopes (the 7000 series or the 22xx/24xx series) is a chore.  Cam switches, endless brackets and screws, stuff packed into the chassis like sardines, etc.  And I am not a big fan of the 500-series mainframe and plug-ins.  Although I suppose it is nice to have a 6-bay mainframe packed with a variety of plug-ins, I find the individual plug-ins to be somewhat lacking in performance and I would rather use a dedicated piece.

Just my opinion...

[precaud]

For HP, they had a period where they decided to compete in the low-end market.  For example, the 3435A, 3465A and 3466A are 3-1/2 or 4-1/2 digit DMMs that are cheaply built and not worthy of HP's state-of-the-art reputation.

I agree on your other calls, but am very surprised to see the 3466A on the list. I bought mine new in 1981 ($995 !!), it went back for a warranty repair 9 months later, and has worked flawlessly ever since. It has been my go-to daily driver DVM for decades - still is. Have not even had to tweak the cal adjustments or clean the switches. Excellent 100kHz+ TRMS converter. Very useful 20mV/20Ω FS ranges. Real 1200V DC max input rating. Fast autoranging. Nice big LED's. What's not to like...

[WattsThat]

I’ll second the Tek 2245/6/7 series. While they are decent scopes when functional, service is an absolute bitch, especially the main board on the bottom. Good luck extracting it to do any required work.

[chronos42]

I’ll second the Tek 2245/6/7 series. While they are decent scopes when functional, service is an absolute bitch, especially the main board on the bottom. Good luck extracting it to do any required work.

I agree, I have repaired several 2246 in the past, repairing the mainboard is a PITA when the board has to be extracted. Even the CRT must removed to get out the board. I really like older Tektronix stuff, but most of these instruments were never easy repair because of their mechanical construction.

[Neepa]

The usual low hanging fruit with HP would be the 16500 Logic Analyzer. Featuring that crap "touch" screen using a IR sensor grid. And then only featuring a single rotary knob for its user interface aside from that.

HP must have sold a keyboard two weeks after each sale of a 16500.

[Vincent]

I don't care if the Tek 491 is a design fail or something, it's a solid 10 out of 10 for the looks and I would adopt that boat anchor anytime if I found one for (relatively) cheap. And bonus points for the repair challenge if it's actually not working at all. 

[vk6zgo]

The Tektronix 650 series Picture Monitors were seriously dire.

The Tek 520 Vectorscope was OK when it worked, but when it got a bit old, it was a nightmare to work on.

The RM529 Waveform Monitor was good when it was new, but was compact & full of tubes, which was a poor combination when it came to reliability.
Add this to fairly short lived CRTs, there is little wonder why we sighed with relief when the 1481R came out.

HP?
I always found their 'scopes a bit more "fiddly" to use, though I quite liked the 180 series.

[jonpaul]

Tektronix 2710, 2712 portable Spectrum Analyzers, horrible membranes panel, bad interfaces design, easy to blow out inputs mixer

Jon

[joeqsmith]

The usual low hanging fruit with HP would be the 16500 Logic Analyzer. Featuring that crap "touch" screen using a IR sensor grid. And then only featuring a single rotary knob for its user interface aside from that.

HP must have sold a keyboard two weeks after each sale of a 16500.

 
I liked the 16500C.  The system still drove like their earlier products.  It was that last system I looked at from Keysight that looked like they had fired all their software engineers, hired all new managers with bright new ideas how things should work and never even knew what the product did.  It looked like softheads were the targeted users. 

 
For Tek, that 78nn DSO.  Nice when it would stay running but more down time than up time.   

For HP, I can't think of anything.  May explain why most of my home is fitted with HP.   
 

[EEVblog]

Can we include the HP name changes? 

[Tomorokoshi]

...
The HP (Germany) 811x pulse and function generators are very nice - when they work. 
...

I bought a couple 8011A pulse generators with the counter options for a specific project. I also have an 8010A dual-channel pulse generator. Both are from HP Boblingen in Germany. While they are constructed quite differently from each other, they are each assembled very strangely. The 8011A style is troublesome to work on and the user interface is one of the worst. Completely unusable without actually monitoring the output with an oscilloscope. The 8010A is actually somewhat elegant, except that all the connections between the many circuit boards are through harnesses that use silver-plated contacts to gold-plated pins. The silver tarnishes to black and basically disconnects.

[T-R]

Keysight EXR/MRX: loud like jumbo jet.

They replace power supply with fanless in service center as modification of product. But it is still to loud for me. If I compare with S series.

[T-R]

Handheld products, like DMM's. OLED problems and other fails. It is not good for brand like Keysight.

[Gavin Melville]

Tek: TDS540.   Somebody went and bought a truckload of crappy capacitors to make these.  It cost a fortune in 1991 new, from memory about 35k.  Noisy front ends.   No schematic then, but I think you can get one now. Can be restored to life by changing all of the capacitors, 100s of them.

2246:  Unreliable power supply, but otherwise good, full of unobtainable hybrids.

Keithley (Tek now). 2450 SMU.  Very expensive piece of kit, crashes all the time, never writes the error to the log. Tek tech support always asks what’s in the Log.  Firmware updated, no difference. Too new to be this bad.

Tek: DPO5034.  UI from hell, changing from Auto trigger to normal needs a mouse.  Touch screen alignment only so-so. Otherwise a very good scope.

HP35670A Dynamic Signal Analyser, the soft keys up the right side of the screen fail. My employer has half a dozen of these, all have the same fault. I’m not sure how they built it, but suspect the sealed soft key strip was jigged in space and the rest of the instrument built around it.

[edit] HP1650.   Forgot this, expensive (16k), 1000 samples deep, for the time a sort of functional logic analyser.  Used a proprietary floppy format, and none of the original disks will now boot.  If they would I could make another boot disk. Not PC compatible in any way. Not worth the effort, and the UI from hell.

[AVGresponding]

Handheld products, like DMM's. OLED problems and other fails. It is not good for brand like Keysight.

Couldn't agree more, with this. I have several HPAK handhelds, and the only one I like wasn't even made by them, but by Escort (U1401B).

[MadTux]

HP 8112A/8116A: No PLL locked instruments build around unreliable ASICs by German design team, which sucked.
HP 8082A: Same garbage as 8112A/8116A, but with added suckiness with sliding switches, that destroy the PCB contact traces over time.
HP 8656A/8657A: Cheap design, construction and poor performance, slightly better on HP 8657B....
Forgot: HP3457A: Display sucks, cheap plastics, bad keyboard, stupid preamp/attenuator design and +/- 3V full range ADC that gives a noisy instrument.
HP3458A: Soooo much fun, if you have failed U180 ;-D

Tek: 7L13, 7L14: Fricking spectrum analyzer that is squeezed into 3 plugin compartments, with no regards to serviceability whatsoever...
Anything with storage tube nowadays, was the only technology back then, but totally outdated nowadays and a pain to use
Their CCD based ADC, 2kPts DSOs, not great released, equally outdated nowadays as deep memory is about the only reason why I would use a DSO over a classic CRO (Tek 7000)

Keithley: Poor plastic quality on their old, brown instruments.....

Classic Fluke: Most structural parts are made from cheap plastics that fall apart, over time (8505A/8506A, Fluke 5440B...)
And their stupid delete/substitute options!!!

[mawyatt]

HP 8112A/8116A: No PLL locked instruments build around unreliable ASICs by German design team, which sucked.
HP 8082A: Same garbage as 8112A/8116A, but with added suckiness with sliding switches, that destroy the PCB contact traces over time.

One of those was to replace the hp3325. We used lots of 3325 back in the day and needed another, hp rep told us about the "new" replacement, so we ordered it. Turns out we didn't actually use it until a year later and discovered the output waveform was horrible (lots of HD, spurs, and Phase Noise). Sent in for repair only to find it was ok and were told "ours was better than most" So contacted rep for return, he indicated that we couldn't return it because of the length of time since receiving, even tho we explained we never got to use it.....didn't matter hp policy was policy so we were stuck.

We ordered a Stanford Research function generator and stuck the HP on the self never to be used.  Later we got a call from the HP rep asking about an order for ~$1/2M (MW Gen, SA, VNA and such) that was in the pipeline, we told him that we had placed a "hold" on the order and were evaluating some R&S equipment. A couple days later he showed up with a new hp3325 for no charge 

Mind that we were at a large international technology company that purchased ~$50M/yr from HP and Tek alone, so somehow the rep found a way to "bend" the hp company policy

Anyway, that "new" hp generator was a total PoS, absolutely worthless for our needs, and how this was allowed to become an hp product is still unknown to us

Best,

[TERRA Operative]

I have a couple of the HP 811x function generators, one has the traditional bad custom chip, the other works and is all calibrated. The output is most certainly not impressive at all, and without any form of accurate frequency reference internal or external, I have no intention of using it... I'll keep to my 3325A thanks.... The 811x squarewave output is more of a bad sinewave for a concerning portion of it's output range.....

I do like the Tek TM500/5000 series modules though. I don't bother with the toy multimeters or frequency counters, they don't have nearly enough resolution for my taste, but the calibration oriented modules are great IMO (SG503, SG504, PG506, SG5030, etc)
I just wish the plastic corners on the front panels weren't so brittle and prone to cracking.
On the earlier units, the front plastic was held in place by countersunk screws at each corner under the aluminium faceplate. Not a good idea to put outward pressure near the edge where it gets knocked a lot, cracks are rampant...
Later models changed the design which improved things, and larger later units used aluminium in place of the plastic which is the best solution IMO.
I'll be getting a replacement plate for my SG504 machined in aluminium as part of its restoration. No more cracks!

[ci11]

I’ll second the Tek 2245/6/7 series. While they are decent scopes when functional, service is an absolute bitch, especially the main board on the bottom. Good luck extracting it to do any required work.

I would offer a counterpoint on this - the Tek 2245/6/7 are far easier to live with than the venerable 2465/7 due to not using the real PITA in Tek scopes - the infamous Dallas chip and leaking tantalum caps. My 2247A's are a joy to use They are keepers for general AF work and especially for monitoring RF modulation. That well-honed, excellent UI - including the adjustable orange graticule - is very tough to beat. Mine have been working for years without complaint.

[TERRA Operative]

I'll go even one more step back to my 2235. I just calibrated it and it works great. Super simple to use and after the usual service upgrades, it's rock solid.
I'll be eventually selling off all my 24xx series scopes and keeping this 2235. No weird chips or RAM backup batteries or complicated functions, just a few dials on the front and a nice sharp trace.



Oh, talking about crappy scopes, the Tek TDS694C..... Due to manufacturing 'reasons', the trigger chips often aren't thermally bonded to the PCB so well, and Tek neglected to install heat sinks, so these unobtainium chips that run reaalll hot tend to die.

Luckily, the seldom used 'Other' (or External) trigger chip can be taken out and swapped with the faulty chip if one doesn't mind the small loss in functionality.

A shitty failure point for a screamer of an early 2000's 3GHz, 10GS/s scope, all for the want of a little heat sink...

[Wallace Gasiewicz]

How about the Hp 8165 "Signal Source"
They did not even call it a function generator.
I have two, if anyone is interested.
Just weird combination of different technologies including Emitter Logic....Capacitor discharge, crystal generator and on and on
A real dream (a bad one) to work on.