Tektronix 5000 series (1970-1990 era) shenanigans

[WastelandTek]

greetings EEVbloggers,

I'm new here, but, this being the internet, there is little stopping me from diving right in and making a fool of myself.

I have recently developed an affliction (not a typo) for Tek 5*** equipment.  These units can be thought of as the little siblings of the popular 7000 and TM500/5000 frame systems.  They also share a great deal of technology with the awesome 4** portable scopes.  Produced for two decades, ending in 1991, the 5000 series occupied Tek's lowest price point in the market.  Indeed, the 5A24N plugin, at $25 1971 USD, has the distinction of being the least expensive Tek product ever.

The low bandwidth (2MHz) 5100 parts were targeted to dedicated industrial applications, and their D1* displays were sold to various third party OEMs for integration into their systems.  The 5400 parts (50/60MHz) were more general purpose low end instruments.

There were nearly 30 different plugins produced for the series including a curve tracer, an audio spectrum analyzer, and some quite sensitive differential vertical amps, as well as some fairly nice dual/delayed timebases.

I first became interested in these systems because I already had a network analyzer that used the D10 display, and I was looking for a new display unit for my home-built W7ZOI spectrum analyzer.  I had been viewing the spectrum analyzer on an Hitachi V-209 (legendary sharp trace, but dinky 3" display).  The non-storage variants of the D1* and D4* displays are nice 6.5" units, so I figured a 5000 series scope in XY would make a great display for the spectrum analyzer with the added benefit of sharing parts and "know how" with something I already had.

One charming aspect of these systems is that they can be configured either as "benchtop" or rackmount if you have, or can fabricate, the appropriate mechanical parts.  This appealed to me as my lab is largely built around the RS310C rackmount form factor.  Another advantage is that the series has been mostly, so far, free of the ravenous vintage Tek mania that has driven up the prices of the similar systems.

There are not a lot of resources specific to the 5000 series on the web.  One very nice overview has been put together by Christian Weagle at http://www.i9t.net/5000_scopes.html   .  I have attempted to contact Mr Weagle to no avail, and his page was last updated in '07 so I fear he may have gone SK, or at least SKbd.  I have saved that page locally against the day it goes 404.

There is also a useful writeup by Peter at tekwiki http://w140.com/tekwiki/wiki/Introduction_to_the_5000-Series_Oscilloscopes   .

Repair and restoration information specific to the 5000 series is rather scarce, so it is well that they share so many parts and architecture with the other Tek products of the era.  Much of use can be gleaned from 465, 485, 7603 etc. threads.

I am creating this thread as a place to seek assistance from the community in my personal repairs, with an eye toward leaving something of a resource to those who come after.  Several of your threads and handles (I'm not going to name names for fear of leaving someone out) have become very familiar territory in recent months as I have been lurking about, and I hope to collaborate in leaving something of use.

[WastelandTek]

So, to bring us up to date;

The first one I got in the shack was a 5440, so, in the early parlance, a 5403 acquisition frame with a D40 display, in the rackmount configuration.  It was an Ebay offering and came loaded with a 5A18N dual trace vertical amp, a 5A21N single trace differential vertical amp and a 5B12N dual/delayed timebase. 

A couple quick notes on bandwidth and nomenclature; 5 is obviously the series, the letter indicates the plugin type.  The suffixes are somewhat random but there is some convention.  Suffixes starting in 1 or 2 are low bandwidth plugins, 2MHz or less, while suffixes beginning with a 3 or 4 are intended for 5400 frames and are the 50/60 MHz units.  System bandwidth is, of course, limited by the weakest link, so in this case I had a high bandwidth frame loaded with 2MHz plugins.  Though I find the 5B12N quite nice, more on that later.

The Ebay listing classified it as the all too familiar "for parts not working" (translate as "whatever it is you're stuck with it") but the pictures showed it displaying its 400mV calibrator square wave with a decent looking trace.  The details elaborated "may have noisy or unresponsive controls".  Well, that sounded like precisely what I was looking for from a project perspective and the price was, I thought, reasonable so on it came.

It arrived safe and I can not fault the seller.  The controls were noisy indeed, and the entire center plugin compartment was unresponsive...fair enough considering.  All LV supply rails were dammed close to bang on, with acceptable ripple, at least for now.  The better part of an enjoyable day with wee paper strips and IPA, both the isopropyl and India pale varieties, and all three plugins were working nicely, albeit only in the outer two slots...on to the frame.

The schematic for the interface board is on pages 124 and 125 of http://w140.com/tek_5440.pdf     .  With a plugin inserted and enabled the beam finder revealed the missing trace hard up, normally off the screen, the position control had no effect.  Reseating u620 did not improve the situation.  While a fault in u620 remained a possibility attention turned to the driver pair Q610 Q614.  A quick probe revealed that Q614 contained no diodes, open every which way, patented not-a-transistor.  the parts list on page 101 list these as "sel from" mps6521, so either binned components or a matched pair, I'm guessing the latter.  an NTE cross search fingers the 6521 as...wait for it...NTE123AP, a bog standard 2n2222, the most ubiquitous npn transistor ever conceived.  I have a batch of motorola 6521s in a saved shopping cart and some day I will get them here and do some matching, but for now a quick trip to the silicon bin and the center bay came to life, oh the joy!

Here is that configuration displaying a 2MHz RF signal and its line frequency calibrator

And here it is in X/Y mode doing what I bought it for, this is the local FM broadcast band.

[WastelandTek]

Well, even though the setup above performed its intended function admirably, as a TEA sufferer, the idea of a 60MHz frame stuffed with 2MHz plugins was simply unacceptable.  Soon a new Ebay listing appeared which was quite intriguing.  The frame was listed as a 5443.  Now, there is a 5440 (single-beam non-storage), a 5441 (single-beam-storage) and the dual beam 5444...but nowhere that I know of in the old Tek catalogs was there ever a 5443.  Indeed, at the time there were a grand total of ZERO google returns for "Tektronix 5443" save the ebay listing in question.  The photos clearly showed "5443 oscilloscope" printed on the rear panel and the display was a D44, the same display as the rather uncommon 5444.  further, the unit was stuffed with 2 5A48 dual trace amps and a 5B40 timebase.  I simply had to have it.

In sober reflection, I was probably a bit hasty.  The auction was again "for parts not working", but this time the AC cable was clearly cut...never a good sign.  Also the fuse cover was missing.  Further, it was in over/under "benchtop" configuration less than ideal for my shack.  The unit seemed to be in good cosmetic condition, however, and was loaded with just the plugins I needed so the die was cast.

It arrived safe and I loaded the plugins in my working frame straight away.  Annnnnnnnnd, my frame failed to power up.  I quickly determined that one of the 5A48s would prevent the frame powering if inserted in any slot so it got set aside.

One of the things I was really eager to see was the on screen display offered by a combination of high bandwidth frames and non "N" plugins.  With the frame powered up with the two working plugins the behavior was quite odd.  Only a few sweep rates seemed to work.  Some were just blank.  Others would have the same sweep rate for several adjacent settings.  The timebase produced no OSD whatever.  One channel of the "working" vertical amp produced no OSD, and the other, bizarrely, was calibrated in...seconds.  Touching any of the red variable knobs resulted in horrible noise or complete loss of trace, as did many buttons.  A trip to the store was in order as I was not in possession of sufficient stocks of one of the IPAs for the task ahead.

Fully disassembling the attenuators and variable controls in these is fiddly work.  Extreme care must be taken to not damage the delicate gold contacts or break brittle 30 year old plastic parts.  A few bullet points on this;

>the markings on the buttons will come off with ANY fluid, and I mean ANY!  IPA, water, sweat on your fingers, touch these with any moisture whatever and you will have a smudged or blank button.  You have been warned.  I don't know if they were like this from the factory or if it is somehow a factor of age, but it is certainly the case now.  One member, robrenz, has resorted to engraving the tiny buttons on a CNC mill and filling the engraving with black epoxy which I have to say is just top notch.

>the gold contact pads and fingers clean up just like the 4** and 7*** stuff; use ONLY isopropyl and paper, no abrasives whatsoever.  No, not your special stash of infinity grit sandpaper, just don't.

>I found the best way to remove the plastic switch extensions was with an Xacto knife and a very fine slotted screwdriver.  Use the knife to gently pry up the tab while using the screwdriver to lever the extension away from the switch shaft.  I have not broken one...yet...but it is pretty scary.

>the buttons themselves are best left on the shafts.  They are sometimes quite tight and I marked the sides of a couple with the pliers before I figured it out.

>my method for cleaning the switches is to drip some IPA in with a needle applicator while operating the switch continuously followed by air in a can to force the IPA deeper and blow crud out, still operating the switch.  I do this operation twice and then finish off with a very small squirt of MG Chemicals "super contact cleaner with poly phenyl ether" the stuff with the gold band on the can.  Mr Carlson turned me on to it.  The Caig might be just as good, but the MG product is impressive.  This procedure has treated me really well.

> for the "uncal" variables operated by the red knobs I first carefully removed the aluminum operator shaft and coupler, made sure the pot was hard CW with the cal switch engaged, then loosened the four screws until the pot came away from the switch body.  I could then perform the same cleaning operation as above on the pot before cranking it hard CW again and re engaging it with the switch.  All of these were VERY noisy.  They are a pain in the ass.


Cleaning up the controls straightened those two plugins right out.  I had a great time putting them through their paces and marveling at the operational OSD, but the empty bay was bothering me so attention soon turned to the misbehaving vertical unit.

Repeat after me; it is always a capacitor, well, not always, but pretty dammed often.  In this case the -15V rail in the plugin was dead short to ground, and here is the bit of wisdom that many of you will already know; the tantalum drop electrolytics used in this era of Tek gear have a bad habit of failing short.  It took a good bit of tracing, schematic studying and lead lifting to finally track the fault to C153 a bypass cap between the emitters of Q150 and Q155 in channel 2.  A quick rustle in the cap bin and all was made right.

Here is that configuration triggering on a 100MHz signal


and here it is with the 5B12N back in displaying four signals in two timebases, note that the 5B12N is triggering on a signal over three times its rated bandwidth, verrrrrrr nice /borat

[David Hess]

I make use of the 7000 series but the 5000 series has some interesting capabilities including a variety of differential plug-ins, the 4 trace plug-in, and a better low frequency spectrum analyser.

[WastelandTek]

Thanks for stopping by David, you are one of the folks I was hoping would notice, for sure.

At this point I was feeling pretty pleased with myself.  Like Midas, everything I wave my soldering iron at turns to gold.  Time to pull the covers off the mystery frame.  It had a sticker on the back that indicated that it had been in a medical school environment, hmmmmm, Weagle does mention that these were sometimes used in EEG setups.  Perhaps the 5443 was only sold to institutions or 3rd party OEMs and this explains its absence from any retail catalog.  Also on the back was the "option 1" marking, uh-oh.  Option 1 was a money saving option that deleted the readout board, so no OSD in this one.

Off with the covers and, whats this?  Someone had glued thin acetate sheet over nearly every perforated section of the panels, leaving only a tiny bit of the bottom front as an intake.  Pretty sure Tek didn't do that.  Something to do with the medical environment?  Jeez, I hope I don't catch ebolaids from this thing.  The fuseholder was broken.  Replacing that and getting a temporary power cord hooked up was rather tight work down next to the HV board.  While the HV safety cover was off to enhance that access I probed the .6A slow blow fuse there and it was open.  I did not have that in stock, but snapped in a regular .5A.  I also noticed some rusty corrosion around one of the mounting screws of the right side vertical board that looked like it might short to a trace, so that got pulled off and cleaned up.

Moment of truth, installed my least favorite 2 plugins and pulled the switch.  A quick safety aside before I continue.  My father happened to be visiting for this part, a fine troubleshooter himself but with only a passing interest in electronics, and my partner was present as well.  The point being, I personally do not like, and do not recommend, working on high voltage without other folks around to yank the cord should the bad stuff happen, not that it ever will because we all have great respect for, and take great care when working with dangerous voltages...RIGHT?  Sorry to belabor the point, but yeah, the CRT anode in these things is -3KV and the accelerator is no less than 12KV! so as Carlson says "if you're following along at home, just take care."

Anyway pulled the switch, and heard a raspy buzzing sound.  Shut it down and had a think.  "Could be the fan" dad said, well, yes it could be.  Pulled the fan cover and jammed a pencil in to stall it and tried again.  Still buzzed, shut down.  Hmmmm, I don't much like where this is going.  As I started to remove the HV safety cover my GF, who had been hovering about for this part, was like "What are you DOING?  STOP.  Recap that shit, you're going to break something".  Yeah, she's a keeper.  I am a stubborn individual, however, and removed the HV cover and procured a length of dielectric hose.  The plan was; we would power up one more time and I would diagnose the display from the front while dad listened around the HV supply with the hose.  He did not need to be told, but was anyway, to keep his hands WELL back from the supply.

OK, deep breath, here goes.  I really did not like the buzzing, but in a few seconds there was life in the display.  The beamfinder allowed me to get both traces in pretty quickly.  There was some sweep, though not full width and one of the traces had a large blotch always stationary on the screen regardless of trace position or sweep.  The blotch could be made flat or round with the focus control.  The intensity controls had no effect.  About this time the buzzing, and one of the electron guns, stopped simultaneously.  I shut down and hung my head in shame.  The GF rolled her eyes and withdrew, and dad informed me the buzzing had been coming from T410.

Sorry, no pics of the display, this all happened in about 20 seconds.

Several days intervened with the project sitting there on the bench as a constant, depressing, reminder of my foolishness before I mustered the motivation to do what I should have done in the first place.

Jones' first law  thou shall check voltages.  I got everything except the interface board, which is soldered, disconnected from the LV supply and got a fused mains jumper on to it.  The rails were all over the place.  the 200V unreg was more like 245.  -30 I could only get down to about -34.  +30 was hard against its unregulated source rail, which itself was high at +42, and the adjustment pot did nothing.  By moving the mains jumper to the high input position I was able to get -30 in spec, just, but +30 was stuck at 38 and wouldn't budge.  All the components in the +30 regulator circuit checked out good.

These are my weapons of choice for this sort of decapping work.  Sure, a real desoldering station would be better but I don't have that.  The short 14ga copper wire as a tip in the Weller is another Carlson trick.  I tell you, that gentleman has it going on!  The iron gets super hot, like RIGHT NOW!


So I got the caps off it without any real drama, cleaned up with some RA flux, solder braid and a more civilized iron, not as beautiful as the work of some here but usable.



And let's see here 3117, that's a little low, wait, what? nano?? and ummm .14 uhhh Kilo??   oh jeez, I think we have a winner folks

[tautech]

Good write up and enjoyable read. WT. 
Following along at home............ 

As I started to remove the HV safety cover my GF, who had been hovering about for this part, was like "What are you DOING?  STOP.  Recap that shit, you're going to break something".  Yeah, she's a keeper.

You lucky guy, don't let her slip away. 

[texaspyro]

Darn new age young 'uns...  afraid of few measly kV.    Your index finger is the best HV probe ever invented.    Boldly probe where no man has probed before.  And I bet they don't even know the subtle joys of soldering in the nude.     Or have a collection of suicide cords (wall plug on one end and alligator clips on the other).  I ain't afraid of no stinkin' smoke... as long as it's not me smokin'.

[WastelandTek]

Well golly tautech, thank you and greetings.

This pretty well brings us up to date.  That cap was C932 which is between the unregulated -38V source rail and ground.  The -30V rail has a lot to do with the regulation of all the other LV supplies.  The rest of the caps measured, well, not great but OK so some mysteries remain, like why the 200V was so high, and why the -30V rail with the bad cap would adjust but the +30V would not.

For the moment I am stalled waiting for my dave_k capacitor boards from OSH Park.  I think those little adapter boards are just a brilliant idea.

[xrunner]

Great thread - good write-up. I'll be following this. 

[Amazing]

So, I have one of these... A Tek 5440 with 4 channels.

It was my personal main scope until about 2010.  I used it professionally to find faults on my digital designs up to 100 MHz.  Yes, it's only a 60MHz scope but the waveforms don't magically disappear -- they just get smaller and rounder.  Sometimes I'd slow down the clock rate to check bus timing.

I used to take pictures of waveforms with a DSLR and email them to colleagues, who would laugh at my
outdated tech.  But it was surprisingly useful for outdated technology.

The coolest thing about mine is the NASA calibartion sticker.

Sadly it's in really bad shape now.  I was just about to throw it out.  It's sitting in the living room in the "take to the dump" pile.

The problem with this one is that the old plastic is extremely brittle.  Early on I snapped off one of the long shafts that run from the front panel knob down into the guts, and repaired it clumsily.  Some of the knob skirts have cracked off and are useless.   One of the FPC-style cables complete disintegrated and someone handwired a replacement, badly.

And a bunch of internal connections are noisy -- the trace and readouts jump all over the place.  But it still technically works!

[WastelandTek]

I was just about to throw it out.  It's sitting in the living room in the "take to the dump" pile.

uhhhhhhh...what's your zip?

[djnz]

5A22N   

[David Hess]

Thanks for stopping by David, you are one of the folks I was hoping would notice, for sure.

You are too kind.

At this point I was feeling pretty pleased with myself.  Like Midas, everything I wave my soldering iron at turns to gold.

If you mess around with and study electronics for long enough, then it seems like you can reach the point where every project works or at least recognizably works on the first try.

Perhaps the 5443 was only sold to institutions or 3rd party OEMs and this explains its absence from any retail catalog.  Also on the back was the "option 1" marking, uh-oh.  Option 1 was a money saving option that deleted the readout board, so no OSD in this one.

Off with the covers and, whats this?  Someone had glued thin acetate sheet over nearly every perforated section of the panels, leaving only a tiny bit of the bottom front as an intake.  Pretty sure Tek didn't do that.  Something to do with the medical environment?

Someone over on the TekScopes@yahoogroups.com or TekScopes2@groups.io email lists may know about the model 5443 and the plastic sheet.  I have no idea about the model 5443 and you already covered the two links where I would look; did you search every Tektronix catalog where it would likely be listed?  I *have* seen a plastic sheet applied to the inside of a perforated panel to redirect airflow for better cooling in a stock instrument before so it would not surprise me if Tektronix did this.

These are my weapons of choice for this sort of decapping work.  Sure, a real desoldering station would be better but I don't have that.  The short 14ga copper wire as a tip in the Weller is another Carlson trick.  I tell you, that gentleman has it going on!  The iron gets super hot, like RIGHT NOW!

I really like the old style Weller soldering guns compared to the new cheaper ones which use screws instead of hollow bolts to hold the element in place.  I made a big rectangular loop out of 10 gauge solid wire for mine for cutting polystyrene foam cleanly and it works great.  Thinner copper wire or stronger brass or bronze had too much resistance to get enough heating.

And let's see here 3117, that's a little low, wait, what? nano?? and ummm .14 uhhh Kilo??   oh jeez, I think we have a winner folks

This is a common failure for old can capacitors.  They have a lot of extra electrolyte and usually operate significantly below their rated ripple current so instead of drying out, when their rubber seal starts leaking, oxygen or water gets in and corrodes the aluminum lead strips between the pins and the rolled up element until one breaks.

[David Hess]

5A22N   

The only problem with the 5A22N compared to the 7A22 is that the 5000 series mainframes have no vertical output so their plug-ins cannot be used to feed a second oscilloscope like a modern DSO.

[WastelandTek]

5A22N   

The only problem with the 5A22N compared to the 7A22 is that the 5000 series mainframes have no vertical output so their plug-ins cannot be used to feed a second oscilloscope like a modern DSO.

that's interesting, what are the extra 4 BNCs that I see on the back of some of the acquisition frames?

this one has them http://www.ebay.com/itm/TEKTRONIX-5440-OSCILLOSCOPE-W-3-PLUG-IN-5B40-TIME-BASE-5A18N-DUAL-TRACE-AMF-/272072117916?epid=1901516118&hash=item3f58c30e9c:g:KMYAAOSwNphWaG11

[Amazing]

that's interesting, what are the extra 4 BNCs that I see on the back of some of the acquisition frames?

Mine has those... And they aren't shown in the picture in the manual.  Maybe the labeling will help you?

Oddly the labels seem to be user-applied.  I'm wondering if it's a semi-official mod.

Also, I'm sorry, but your fun and interesting thread as made me want to keep it now... 

[WastelandTek]

that's interesting, what are the extra 4 BNCs that I see on the back of some of the acquisition frames?

Mine has those... And they aren't shown in the picture in the manual.  Maybe the labeling will help you?

Oddly the labels seem to be user-applied.  I'm wondering if it's a semi-official mod.

Also, I'm sorry, but your fun and interesting thread as made me want to keep it now... 

as long as its not going in the dumpster man

if you get a chance, pop the cover and document where exactly those BNCs attach hey?

[Amazing]

OK, pictures attached.  Looks like there are too many so I will have to do multiple posts.

Do you have those pin blocks in yours?

Oh, and sorry for the horrible quality...  My phone...

[Amazing]

More pics...

[Amazing]

One more thing.

The 5A48 has that two 4 pin connector in the module.  The wires continue to the front
of the module and disappear behind some shielding up near the front panel.

The 5A15N, 5B10N, and 5B42 do not have that connector.

Hope this helps.

EDIT:  Only one of my 2 5A48's has that connector.  the other one doesn't.

[WastelandTek]

that is really interesting

no, I don't have the pin blocks, nor do either of my 5A48s have the mate, fascinating

would you mind photographing the interior of the 5A48 at some point?

[David Hess]

that's interesting, what are the extra 4 BNCs that I see on the back of some of the acquisition frames?

I think I remember reading a discussion about that modification over on TekScopes@yahoogroups.com.

The holes in the plug-in's plastic frame were originally for the fiber optic readout on the 7000 series which was replaced by the electronic readout developed by Barry Gilbert.

[Amazing]

would you mind photographing the interior of the 5A48 at some point?

I'm a bit busy today, but last night I was reading the sites that you linked too previously about this scope.

http://www.i9t.net/5A48.html mentions a special version of the 5A48 with an option for back panel inputs.  That would make a lot of sense on a rack mount scope.  The picture at the bottom of that page is what the back of my 5a48 looks like.

Also, that site has a PDF describing the mod: http://www.i9t.net/5A48_images/5A48-mod-818H.pdf

That PDF talks about it being a coax connection.  Initially I thought it was just a wire, but looking at it more carefully I can see that it is coax.

[David Hess]

That PDF talks about it being a coax connection.  Initially I thought it was just a wire, but looking at it more carefully I can see that it is coax.

I immediately identified it as coaxial cable because it resembles the RG-174 cable Tektronix used with their Peltola connectors in the 7000 series and other test instruments during that era.

http://w140.com/tekwiki/wiki/Peltola_connector

Incidentally, Peltola sockets work for some (4mm?) probe tips.  I wish this inexpensive standard had stayed around.

[WastelandTek]

mail just came, this should help

[TerraHertz]

Interesting thread. I too like rackmount scopes, and I have a few - mostly R7903s.
I do have three 5103N (D12 and two D13), and had spotted that they could be converted to rackmount. But alas I'm unlikely to find time to do it.

[WastelandTek]

Those look great TerraHertz!  Really clean.

You called it David; ground side foil corroded in half



Oh man, this thing smells GREAT inside /sarc, so this is where that "old electronics" smell comes from, ugh.

Some illumination on nomenclature; upon reviewing the physical 5444 manual it turns out that is what I have.  5443 refers to the acquisition frame and a 5443 mated to a D44 display is a 5444.  This is pretty cool as the 5444 is sort of the holy grail of the 5000 series, being 2 scopes in one tube a-la the 7844, and makes me VERY hopeful that the tube itself is not damaged.  They were only offered in the 1976 catalog and are pretty uncommon.

The dave_k capacitor boards have arrived from OSH Park and look quite good so it is time for me to be ordering electrolytics.  For years I have been pretty liberal in my capacitor replacement choices, figuring that if some is good more must be better, and I have had pretty good luck with that.  Reading around here, however has given me pause as I see some are claiming that excessive voltage derating does not enhance reliability as the caps will quickly "reform" to the lower voltage and that higher voltage rated parts will have higher ESR to no benefit.  While these arguments do make sense to me I am torn as I have become comfortable "overdoing" it in both capacitance and voltage rating.

Here is the relevant schematic with the caps in question pointed out



here are the specs per Tek

C800 170uF 275V       (between +30 and +200V rails)
C848 10,000uF 12V    (+5 - GND)
C875 7,500uF 25V      (+20 - GND)
C876 7,500uF 25V      (-20 - GND)
C930 5,000uF 50V      (+38 - GND)
C932 3,000/5,000uF 50V  (-38 -GND)

C932 is 5K in the 5444 and 3K in the 5440 (and all the other 5xxx) supplies

here is what is in my Mouser cart awaiting your critique

C800 220uF 315V ELXS3B1VSN221MP3
C848 18,000uF 16V EKMH160N183MQ40T
C875/6 10,000uF 35V EKMH350VNN103MR
C930/2 6,800uF 50V EKMH500VNN682MA3

Am I crazy?  Is this too much?

[tautech]

IMO

C800 220uF 315V ELXS3B1VSN221MP3

Should be OK.

C848 18,000uF 16V EKMH160N183MQ40T

You need consider that you could potentially double the inrush current through the bridge at power on.....not advisable.

C875/6 10,000uF 35V EKMH350VNN103MR

Similar but not as bad as above.

C930/2 6,800uF 50V EKMH500VNN682MA3

Again, pushing things to a point that may cause problems.

Am I crazy?  Is this too much?

Not crazy, just get a handle on the consequences of selection far from what the design stipulates and the reasons why.
There is every chance it won't matter damn all but are you willing to take the risk ?
Not me. 

[WastelandTek]

Yeah, I get it that the inrush through the bridge diodes is the limiting factor.  I have always sort of assumed that the values chosen back in the day were heavily influenced by size and cost, which are not really issues any more.  The bridges look pretty beefy, and diodes are not exactly unobtanium...are there any OTHER reasons not to go hog wild with main filter caps?

Also, can I add a current limiting resistor to soften the power on blow to the bridge?

[tautech]

Yeah, I get it that the inrush through the bridge diodes is the limiting factor.  I have always sort of assumed that the values chosen back in the day were heavily influenced by size and cost, which are not really issues any more.  The bridges look pretty beefy, and diodes are not exactly unobtanium...are there any OTHER reasons not to go hog wild with main filter caps?

Also, can I add a current limiting resistor to soften the power on blow to the bridge?

NTC thermistors are often used to manage inrush current but another thing to consider is the ESR of the replacement caps. Lower ESR and higher capacitance will both contribute to higher inrush currents.

[WastelandTek]

OK, please don't think I did not really want your advise, how about this cart?

C800 220uF 315V UPS2F221MRD
C848 15,000uF 16V UVZ1C153MRD
C875/6 10,000uF 35V UVZ1V103MRD
C930/2 6,800uF 50V UVZ1H682MRD

all Nichicon 105C parts

[tautech]

OK, please don't think I did not really want your advise, how about this cart?

C800 220uF 315V UPS2F221MRD
C848 15,000uF 16V UVZ1C153MRD
C875/6 10,000uF 35V UVZ1V103MRD
C930/2 6,800uF 50V UVZ1H682MRD

all Nichicon 105C parts

Closer. 

There's still some room for you to be more conservative. 

I know if it was mine I'd be sticking close to the same values back, sure it's a hassle just to have to replace them but why add the risk of more hassle ?
Your call. I won't think bad of you whatever you decide. 

[WastelandTek]

We'll let it sit a while and hopefully accrete some more opinions.  I freely admit I tend to being an all or nothing sort of guy.  If it was up to me I'd probably build the Dyson capacitor, using all conductive material in the solar system for the plates and the rest for the dilectric....death to ripple.

[David Hess]

Oh man, this thing smells GREAT inside /sarc, so this is where that "old electronics" smell comes from, ugh.

Ethylene glycol is or was a common solvent in aluminum electrolytic capacitors and is a common element of "old electronics" smell.  I think propylene glycol was used in aluminum electrolytic capacitors rated for lower temperature operation.

Some illumination on nomenclature; upon reviewing the physical 5444 manual it turns out that is what I have.  5443 refers to the acquisition frame and a 5443 mated to a D44 display is a 5444.  This is pretty cool as the 5444 is sort of the holy grail of the 5000 series, being 2 scopes in one tube a-la the 7844, and makes me VERY hopeful that the tube itself is not damaged.  They were only offered in the 1976 catalog and are pretty uncommon.

I forgot about this.  The 5000 mainframes were also used to make other instruments like curve tracers so the top and bottom frames have different part numbers.  Tektronix also OEMed them for others to make instruments out of.

For years I have been pretty liberal in my capacitor replacement choices, figuring that if some is good more must be better, and I have had pretty good luck with that.  Reading around here, however has given me pause as I see some are claiming that excessive voltage derating does not enhance reliability as the caps will quickly "reform" to the lower voltage and that higher voltage rated parts will have higher ESR to no benefit.  While these arguments do make sense to me I am torn as I have become comfortable "overdoing" it in both capacitance and voltage rating.

Higher voltage capacitors are physically larger and have more electrolyte so they have a higher ripple current rating and take longer to dry out.  However there is a change in electrolyte above about 160 volts which yields a higher ESR but up to that point, higher voltage is always better.

C800 170uF 275V       (between +30 and +200V rails)
C848 10,000uF 12V    (+5 - GND)
C875 7,500uF 25V      (+20 - GND)
C876 7,500uF 25V      (-20 - GND)
C930 5,000uF 50V      (+38 - GND)
C932 3,000/5,000uF 50V  (-38 -GND)

C932 is 5K in the 5444 and 3K in the 5440 (and all the other 5xxx) supplies

here is what is in my Mouser cart awaiting your critique

C800 220uF 315V ELXS3B1VSN221MP3
C848 18,000uF 16V EKMH160N183MQ40T
C875/6 10,000uF 35V EKMH350VNN103MR
C930/2 6,800uF 50V EKMH500VNN682MA3

Doubling the value of C848 may be a little much.  Just use the closest or next higher standard size for replacements so 220, 10000, 8200 or 10000, 4700 or 5600, 3300, etc.  The risk if an input capacitor is too large is from excessive surge current although saturation in the transformer will limit this somewhat.

If an input capacitor fails with a short, then its associated input rectifier should be replaced.  Some Tektronix instruments had rectifier upgrades do to poor reliability.

[WastelandTek]

Well, I owe you folks an update.

Excess disposable income has been a little thin of late, and I have not been able to justify actually purchasing my parts shopping cart, which has grown rather large with parts for multiple outstanding projects.  I have, therefore, sidestepped into a related project, related enough, I think, that it fits in this thread.

Meet the Pacific Measurements (later gobbled up by Wavetek, later gobbled up by Wandel & Goltermann, later gobbled up by Fluke) 1038 system, in this case configured as a dual port network analyzer. 



It used the Tektronix D1x displays and a very similar 3 bay mainframe system to the 5K Tek line. I became addicted to the dual port network analyzer some years ago when I had a buddy that worked at a VERY swank lab and we could go in the night and use the equipment.  Once you have used one, doing any real RF work without seems like stumbling around in a dark cave with a book of paper matches as opposed to exploring with a nice big torch.  In the pic above, the system is displaying the S₁₁ and S₂₁ of a typical RF low pass filter simultaneously.  You can tweak on the DUT and observe the effects in real time.  This is VERY handy for developing an intuitive sense for circuit behavior.  Once you have used one it is really frustrating to go back to doing it any other way.

Unfortunately, another feature of such systems is that they tend to be pretty expensive, at least modern ones.  I was lucky and was able to assemble all the requisite bits over a couple years' worth of swapmeets at less-than-eye-bleeding prices.  Most of the expense of putting something like this together is actually in the autotester bridge and detectors.  These PM/Wavetek display frames have become quite affordable.

This one is in the process of being converted from the side-by-side rackmount configuration to over-under so I can use the rackmount mechanical parts to convert a nicer, late production, 1038-D14A.



I'm not sure I really recommend disassembling them this far, but this one was disgusting inside.  It had a really thick coating of that black, high voltage, soot all over in the display unit.



The deflection amps in these D10s seem to run pretty hot.  I'm going to see about getting some more airflow up here.



One of the deflection leads was threatening to short to chassis because the heat had baked the insulation off.

[Electro Detective]

Great post mate 
please reserve a ringside seat for me too    

I've got 2 of these I restored last century, 5111s I think,
one is fully loaded with two 4 channel plugins and delaying time base (killa scope),
and also a dead/donor/project rackmount version,

all still in storage (no pun intended) due to a premises move

Will have to get to them and fire up carefully asap 

[WastelandTek]

Welcome to the thread ED, thanks for stopping by.

4 channel plugins

ahh yes, the 5A14N.  I don't have one yet, they are one of the coolest plugins for sure.  Surprisingly, they come by on ebay quite frequently and are one of the cheaper plugins to get.  There are two cleanish looking examples right now for less than $40 each (not shipped-US) which I have to say is pretty dammed painful to resist.

[Electro Detective]

If they were in this part of the woods, I'd hit on them and fill up another box,
more than good enough for LF and audio work,

and that monster crt screen rocks   

You can't have too many channels.. or scopes..

or w....   

[WastelandTek]

Well guys, I have run into a snag.

Got the donor 1038-D10 all reassembled in the vertical configuration and when I applied power the HV protection neon bulbs, DS271-3, lit up and R273 started smoking.

I shut it down pretty quick and R273 still reads in spec.  The thing was working before disassembly, so it is definitely something I did.

I'm going to have a careful, patient, poke around it and have a good think.  Here are the relevant schematic and circuit description.  If anyone wants to chime in it would be much appreciated.

[WastelandTek]

oh man...

http://www.ebay.com/itm/Tektronix-5111-Storage-Oscilloscope-/122313864004?epid=1201603091&hash=item1c7a797344:g:VIMAAOSw2xRYfWfr

[Electro Detective]

oh man...

http://www.ebay.com/itm/Tektronix-5111-Storage-Oscilloscope-/122313864004?epid=1201603091&hash=item1c7a797344:g:VIMAAOSw2xRYfWfr

Delaying the "Buy It Now" thing for 20 bucks with 3 others 'Watching' is some serious risk taking   

Find out what the shipping charges are asap so the parts/donor/maybe working deal doesn't hit $200   

[WastelandTek]

Well, I managed to resist that, they wanted $80 shipping

but this happened

[WastelandTek]

Whelp, I fear that, barring a sourcing miracle, or a sudden leap in DIY engineering, that the 067-0680-00 calibrator is a bust.

The staircase works, which is nice as far as it goes, and the normalizer amp seems to work, so I got that going for me, but the pulse generator is missing the absolute unobtanium 10ma 4.7pF tunnel diode 152-0177-00.  This is one of the notorious custom Esaki devices that can not be had for love nor money, and have been the end of so many Tek scopes.  none of the commonly available soviet NOS tunnels seem to share this parts relatively stout current and very low capacitance. 

Someone had attempted to replace the tunnel diode with a bog standard 1N4152 which...isn't going to work.  In their defense, there is a typo in the 067-0680-00 manual which does, in fact, call out a 1N4152 as CR280.  This is clearly incorrect, but "fortunately" this fixture is nearly identical to the earliest versions of the 7000 series calibrator 067-0587-01, note that the -02 and -03 revisions of the 7000 calibrator are completely different.  The 067-0680-00 schematic indicates a 10mA tunnel, so I am fairly certain that 152-0177-00 is indeed the part.  In this online picture you can just make out the characteristic gold DO-17 case of the device in question.  I can only hope that my missing one was scavenged to fix a much nicer piece of gear.



I had high hopes for the calibrator, they are pretty uncommon.  There was a typo in the ebay listing and the buy it now price was the lowest I have seen, plus I was able to bundle shipping with the other lot of 3 plugins pictured above.

The beauty of the calibration fixture, as I understand it, is that it supplies very fast risetime pulses directly to the mainframe input pins allowing tuning of the square edge compensation adjustments in the vertical amps in the frame.  Without it you must apply pulses through a vertical plugin, in my case an uncalibrated plugin, which multiplies the unknowns and complexity of adjustment.  I am thinking that my fall back position might be saving up for a Leo Bodnar pulser and try and figure out how to couple it direct to the frame through a JAMMA cable, though I'm not sure the impedance is going to match...and the length of the cable...oh never mind, that's not going to work either...

 

maybe a little PCB with the 2 card edge pins, a matching network, and a BNC for Leo's pulser...and get a leprechaun to reach in the frame and hook it up?  I'm pretty sure leprechauns are easier to find than tunnel diodes.

[alm]

Is the pulse transition time that critical for the 5000 series? I believe the fastest mainframe was 50 MHz, and that was a special (5400?) series that had its own dedicated plugins. I can not imagine the skin effect loss from a piece of coax soldered to a header connector would be visible at all. Obviously you would need the correct impedance and might need a differential signal.

[WastelandTek]

Greetings alm, welcome to the thread.

I am working on a 5440 and a 5444 here, so the 50/60 MHz stuff.  It is entirely possible that Tek went a little overboard with the risetime.  Near as I can tell the tunnel should rise on the order of 400pS which, at 875MHz bandwidth, does seem a bit quick for the application.  I kind of suspect that they were binning the diodes from GE and the nicer ones became 152-0177-01 or 152-0177-02 and they had a pile left over that they used here, but who knows.

There are 10mA tunnels available with higher capacitance

1n3719 10ma 27-50pF

1I305A 9.2-10.4 ma  <30pF

1I305B 9.6-10.8 ma  <30pF

I might get a batch of 4 of the Russian ones for $15 and see what happens.

The first stage on the mainframe interface board is indeed a differential pair, 470k to ground.

[David Hess]

The 5000 and 7000 series calibrators all use some variation of a transconductance output circuit for the pulse output with diodes to disconnect the edge from a shunt termination.

The tunnel diode used in the 500 MHz 7000 standardizer and the 50 MHz 5400 standardizer (they both use the same design) is *not* part of the output circuit.  Instead, it sharpens the pulse from the clock generator before it enters the differential amplifier chain.  It could be replaced with any 10 milliamp Ge tunnel diode or with a 10 milliamp GaAs tunnel diode if resistor divider R288/R289 is adjusted to take into account the change in voltage step.

Alternatively, I doubt the 50 MHz standardizer needs that sharp of a pulse at the input of the differential amplifier chain anyway.  The tunnel diode could be replaced with a resistor and a resistor placed across C125 to create the proper amplitude at the base of Q282.

Or even better, I might actually replace the tunnel diode with a small signal schottky diode or 2N3904 base-emitter junction to control the amplitude at the base of Q282, add the resistor across C125, and adjust other resistor values at the input to get the right levels at the collectors of Q310 and Q312.

[WastelandTek]

Well....

Clearly I have come to the right place for help.

1N3719 are $30 for 2 on Ebay right now, if you don't think that an order of magnitude more capacitance is an issue that is probably the easiest solution

[WastelandTek]

So the DO-17 cased 1N3719 tunnels for the calibrator arrived.

I have read dire warnings about the thermal fragility of these Germanium devices.  I don't have any specifically low temp solder here, mainly just 60/40.  I can clamp some heat sink to the body.  What do folks think as far as iron temp; hot and fast or cool and slow?

[dave_k]

Clamp each leg with a hemostat to sink the heat, and go hot 'n' fast.

[WastelandTek]

Well a warm welcome to the thread dave_k!!  Your thread figures large in the inspiration for this project.

If you read above you will see I have your capacitor adapter boards here, thank you so much for the gerbers!

[dave_k]

Nice .. I hope the boards work out for you

Reading your thread makes me want to resume working on my collection of 7603 and 7613 units.

[David Hess]

I have read dire warnings about the thermal fragility of these Germanium devices.  I don't have any specifically low temp solder here, mainly just 60/40.  I can clamp some heat sink to the body.  What do folks think as far as iron temp; hot and fast or cool and slow?

This applies to any germanium device like the 1N270 diodes that I like to use.

I wrap rubber bands around the handle of my needle nose pliers and use them as a heat sink on the lead between the body of the diode and the soldering point.

The heat capacity of the soldering iron tip makes more difference than the temperature.  With a larger heat capacity, the temperature can be lower.  With that in mind, I use my fattest soldering tip which will fit.  Then soldering is so fast that I just touch the already fluxed and soldered pad momentarily.

[WastelandTek]

Whelp, got the tunnel in it and still no pulse generator output.

All the transistors and diodes in the pulse generator appear to have the requisite semiconductor junctions in them just probing around.

I guess it's time to cut the notch in my new JAMMA cable card edge and have a look for DC getting to those transistors and the clock signal actually making it to the tunnel diode.

[PrecisionAnalytic]

Meet the Pacific Measurements (later gobbled up by Wavetek, later gobbled up by Wandel & Goltermann, later gobbled up by Fluke) 1038 system, in this case configured as a dual port network analyzer. 

The Tekwiki for the Pacific Measurements 1038 system is moving forward at full blast.   

I figured I'd promote for membership (enjoy yourself... I'm not nit picky) and have a central source for resources, experiences, history or whatever insight they'd like to share regarding Pacific Measurements and Wavetek Systems.  Thanks in advance for your time and support!   

https://www.facebook.com/groups/698747963909529/?hc_location=group

[WastelandTek]

sorry, no facebook here

I will have a look at your tekwiki pages.

[PrecisionAnalytic]

sorry, no facebook here

I will have a look at your tekwiki pages.

That's OK... I think I have everything linked on the Tekwiki other than the still elusive N10, NS20 and NS201 documents. I do have a Handbook for the NS20 and NS201 that for some reason, until I just typed this, I forgot about scanning.

Have a C10 with a manual in the mail and will scan and update the Tekwiki once I receive accordingly... which gets me thinking I can next update with the photos I have for.

Made a new Tekwiki page for the what looks like 5000 series EDX-1 that I haven't found any documents yet regarding:
http://w140.com/tekwiki/wiki/EDX-1_Electromyograph_Storage

I also made an update to the Wikipedia JAMMA page:
https://en.wikipedia.org/wiki/Japan_Amusement_Machine_and_Marketing_Association