Dead Tektronix 2235

[Rickenbackerman]

Hey guys, first post here...  I left my Tek 2235 powered up overnight accidentally a few days ago only to find it dead the next day.  No power LED, no smoke, no funny smells, no sounds, no nothing.  Luckily it didn't decide to catch fire or something and burn the house down!  I know the PSU is the likely culprit here as most of these have been gone through already but mine looks to be 100% factory stock.  I just made the digi-key parts order under NFM's (excellent) video on the popular video channel.  Hopefully a full recap as well as the 050-2242-03 service bulletin upgrades will bring it back to life.  Stay tuned...

[bd139]

Some notes here which may be of assistance. Good luck

https://www.eevblog.com/forum/testgear/tektronix-2235-repair-thread/

[Rickenbackerman]

Thanks bd139, yeah, I saw your thread earlier while doing some poking around in here...  it looks helpful for sure.

I just pulled all of the 840uF mains caps and C906- the big 75uF can - as well as C940 which I think was replaced at some point.  It was a 50V 1000uF nippon chemi-con.  Don't want to pull too much out as I generally like to replace as I go and parts aren't here yet. 

I do think I got the fan area sorted out though - the pad areas on mine are way over right next to the side of the chassis, and comparing to the skit that NFM has up it looks the same.  Not too worried about the fan yet though since it won't even power up yet!  Ha!

[Rickenbackerman]

So far I replaced all of the PSU caps (I think).  My attempt at taking NFM's digi-key parts list and cross referencing it to corresponding schematic parts as well as what I've done so far below:

1) 120uF 450V radial - C906 - replaced
1) 33uF 200V radial - C954 - replaced
2) 2.2uF 50V radial tantalum - not replaced yet but I'm pretty sure these go on the attenuator board
2) 1.0uF 50 axial - ?
3) 2200pF 1kV radial - C902, C903, C904 - replaced
3) 2.2 uF 100V axial - C90, C91, ?
1) 0.15uF 1kV radial film, C900 - replaced
2) 10uF 160V radial - C942, C943 - replaced
4) 3.3uF 50V axial - ?
1) 1uF 100V radial - C944?  I see it on the skit and the screenprint on the back of the board but it doesn't seem to exist on the front of the board?  It should be on pg 162 of the service manual and I don't see it there either?
7) 1000uF 50V radial - C960-963, C968, C970, C940 - replaced
1) 100uF 50V radial - C925, replaced (this one sucked ass! - damn chassis in the way!)
1) 270uF 63V radial - C956, replaced
10) 22uF 50V radial - ? (I see lots of little ones that I haven't gotten to yet)

I figured I'd plug it in and see what's up.  Nothing but "tick tick.......    tick tick.....   tick tick....."  I don't think I even got that before but maybe I did, I dunno.  I didn't have the case off like I do now.

I haven't done the service bulletin mods yet.  That's next.  Gotta get the power supply going again before I screw with anything else!


edit 5 minutes later - I see C944 now!  Tiny little bastard!  I also see three other tiny little axial bastards on the main board...  back at it soon...

[bd139]

Good luck. Power supply start up is the usual pain in these. The supply is in two sections, a preregulator and a secondary SMPS. The secondary SMPS will come up on 43V DC at two test points on the board. Always worth checking with that first. The preregulator is the trouble on these as it likes to explode in your face I find. If I ever get my hands on another one I'm bypassing the entire thing and sticking a 48V MeanWell unit in there instead!

[Rickenbackerman]

Thanks bd, I replaced C944 tonight and C200 and C201 while I was at it (two other tiny axial bastards - these are a bitch!  I kinda butchered two pads - dammit! - but everything checks out as far as I can tell).  Didn't have a whole lot of time.  I was really hoping C944 was the culprit as it looks to be critical judging by the schematic.  Plugged it in and got the same "click click....   click click...."  and no power LED...  so NOPE!

Next time I work on it I'll dig around in the service manual and look for those test points like you said.  If those look ok I'll give the service bulletin PSU upgrades a go...

I know what you mean about bypassing - I have a basement recording studio with a Soundcraft Delta 16 channel board from the 80's I think?  Great board but the factory PSU is a total joke.  I built a little birch 13-ply box housing two Power One units - one for the board and one for 48V mic power.  The factory PSU now sits on a shelf where it belongs!  But that is an outboard standalone supply, unlike this damn thing...

edit - re-reading your thread now for knowledge/inspiration...  the back and forth between you and David Hess is so far over my head it's not even funny.  I have an engineering degree and own and maintain 1800+ watts worth of vintage tube amps (which is the reason I bought this damn scope!), but fixing this thing might be WAAAAAAAY beyond my dumb ass...  I won't give up yet tho!  Not yet!

[WattsThat]

Sorry to be negative here but why? It’s a simple question. Why?

The scope is dead. And you just start replacing caps? Why?

Have you measured the power rails? Check the ripple voltage? Anything? It’s all in the service manual. The voltages are easy to measure.

I simply don’t understand why people just start ripping what are perfectly fine caps off a board in mass. Get it working first. Check the ripple voltages. That’s when you replace the caps that need replacing, if any. Sure, if there are signs of leakage, they go. If the labels are cracked from heat, they go. But as the first step in a repair process to start ripping parts off a board? That most times creates more problems than it solves.

I wish you success but what if it turns out to be a bad hv transformer? Then what do you do with all those new caps?

[Rickenbackerman]

Why?  1985.  That's why.

[Rickenbackerman]

Sorry Watts, not trying to be negative either, but at last count I've got a...:

Marantz 2245
Marantz 2270
Sansui 8080DB
Dynaco ST-70/PAS-3
Scott LK-48B
Fisher 400
B&K 747B tube tester
Heathkit IG-18
Heathkit cap tester - can't remember the model - the one with the "magic eye"

Not to mention all my bass amps (anyone else own three 70's Ampeg SVT's?  Hiwatt 200?) and guitar amps that I've fixed which have all done just fine with my method - the shotgun one.  If I find an electrolytic older than 20 years, it goes.  No questions.  I've done well so far.

edit - I know I'm new here.  Don't want to step on any toes but I think I "kinda" know what I'm doing...

[bd139]

Yep. I’m not normally a cap swapper myself but the 2235 has some really hooky ones in it. You build up a list of “potential targets” in gear after a while and it pays off.

As for duff ones I received an HP power supply yesterday with a really top end Sprague 32D compulytic which was open circuit. Dated 1989. This does happen 

Testing isn’t even conclusive. I’ve had ones that report good on my ESR meter and my LCR meter and piss out their electrolyte almost immediately.

[floobydust]

I use this 2235/2235A PSU electrolytic capacitor list, and replace them without troubleshooting down to the component level. That is a waste of time when you know the old Spragues have dried up and Sangamo's cracked and pissed. So far, I always find low value or leaky or open caps in 2235's.

Some were upsized for the 2235A so I do that in 2235's - the 840uF to 1,000uF.

[Rickenbackerman]

Oooooh - flooby - I see your list - thanks!  I will save it.

bd I've had ones that test good as well on my Heathkit dude and do not pass the mustard in-circuit.  So maybe I'm not nuts to shotgun...  Let's see if we can get this damn thing working!  Back in a day or so...

[BravoV]

I use this 2235/2235A PSU electrolytic capacitor list, and replace them without troubleshooting down to the component level. That is a waste of time when you know the old Spragues have dried up and Sangamo's cracked and pissed. So far, I always find low value or leaky or open caps in 2235's.

Some were upsized for the 2235A so I do that in 2235's - the 840uF to 1,000uF.

Wow thank you  , sorry abit off topic, especially to bd139, as with my recent acquisition of "dead" 2215  , this list should be handy right ?

[beanhauler]

For anyone trying to fault find this scope i think the procedure listed for the military version is probably the best plan.
Lift up the resistor on the emitter of the two push pull transistors R949 listed a 1 ohm.
put the scope on a variac if you have one and slowly wind the ac up to about 80v monitoring the voltage between 950 and 940.
If it comes up to 43V then the pre reg is fine, If no 43V then the pre reg is the issue.
if you have 43V now start looking around the power transformer.
In its click click overcurrent state there is no point checking around the pre reg as the voltages are constantly being set and reset by the overcurrent sense around the 0.2 ohm resistor which resets the pre reg chip.


Check all the diodes on the o/ps and measure the resistance as follows:

-8.6 TP500 114 OHMS
+8.6 95 OHMS AT W975
+5 330 OHMS AT W 985
+30 905 OHMS AT W965
+100 12.5K OHMS AT W966

IF YOU FIND ANY REALLY LOW THEN CHECK THE CAPS BY LIFTING ONE LEG.STILL LOW THEN INVESTIGATE FURTHER BY LIFTING THE LINK OR LINKS IN THE CIRCUIT CONCERNED.

If that fails you may have a shorted turn in the transformer or a cracked ferrite former.
Only way out may then be to look for another transformer.
The former on these has 22 pins and a very imperial inch layout so any attempt at a wind with a different former would be difficult at best.

The HV wind is 1025 T #39 for a voltage of -2050V and i know someone has attempted it and succeeded but its probably difficult.
+30 is 16T #23
+100 37T #31
+5 3T #24
+8.6 and -8.6v 5T #24
filament 6.3v 3.5 T #24

The main core is also encapsulated with resin so if you just have shorted turns then it would be very difficult to rewind without rewinding it completely i guess and re-encapsulating in a pressure chamber. Otherwise you would probably have corona issues and flashover at 2K plus.

I am waiting for another transformer to try as i have a broken ferrite core and have exhausted all the above.
I will post the result after i fit it....who knows if it will. I am not sure when i will give up.

Actually if it does not fix it then the problem could be the inverter itself or the feedback self oscillating transformer although its very low current and unlikely to break. watch this space.

Like the idea of seeing if it will run on a 43v independent supply but you would need to monitor the supply current.
It may be that it will run on the 43 with more current if the no of turns has reduced in the primary 20T+20T #25 but the core may then start to saturate due to the reduced number of turns and overheat as the resistance is now moving towards the resistance of the wire itself .
 20T of #25 looks like 0.2 ohms at a path of 94mm ish which is an instant short .


Continuing my quest on the 2235 scope and while waiting for another transformer i decided to tear down the old one.
Waded my way through the first two winds on the first layer which has plus and minus 8.6 with 5 turns and wound next to it plus 5v with 3 turns all Bifilar and hit the first layer of heavy duty tape...what is that stuff? about 3 mm thick and ran in to the HV wind. This HV wind actually straddles the pins on the former from pin 8 which is grounded to an exit point at the top of the transformer where the #39 wire is soldered to a HV wire before exiting.
On looking at the connection between the very fine wire and the lead out i noticed that around the solder connection it was very black and sooty...is this likely my problem i wonder. If it is i wonder if the multiplier has been affected.....time will tell
 

[akimpowerscr]

Why?  1985.  That's why.

I totally agree with WattsThat.

Some people have developed a real paranoia about electrolytic capacitors, and in the majority of cases this is not justified.

Systematically changing all the electrolytic capacitors is a strategic error in the diagnosis and in the troubleshooting, for several reasons:

1) there is a risk of introducing new failures: damaged printed circuit tracks, assembly or value errors, new capacitor defective or needing to be reformed, ....

2) small variations in ESR or capacitance generally have no influence on most analog circuits

3) The life of an electrolytic capacitor is not measured simply in years but in time and temperature of operation.
It also depends on how well it is built.

To conclude that an electrolytic capacitor needs to be replaced just because it is "old" is a fundamental mistake.

I have for example a Marantz 2285B receiver from 1974, all its electrolytic capacitors are the original ones except one and it still has all the original performance.

Only those electrolytic capacitors which are visually defective (swollen case, liquid leakage), which are measured as being short circuited or which are known to be generally defective such as certain brands and years, or those which are nearby from a heat source should be systematically replaced.

Apart from these exceptions, it is necessary, before any component replacement, to make a precise diagnosis of the failure.

[akimpowerscr]

For anyone trying to fault find this scope i think the procedure listed for the military version is probably the best plan.
Lift up the resistor on the emitter of the two push pull transistors R949 listed a 1 ohm.
put the scope on a variac if you have one and slowly wind the ac up to about 80v monitoring the voltage between 950 and 940.
If it comes up to 43V then the pre reg is fine, If no 43V then the pre reg is the issue.
if you have 43V now start looking around the power transformer.
In its click click overcurrent state there is no point checking around the pre reg as the voltages are constantly being set and reset by the overcurrent sense around the 0.2 ohm resistor which resets the pre reg chip.


Check all the diodes on the o/ps and measure the resistance as follows:

-8.6 TP500 114 OHMS
+8.6 95 OHMS AT W975
+5 330 OHMS AT W 985
+30 905 OHMS AT W965
+100 12.5K OHMS AT W966

IF YOU FIND ANY REALLY LOW THEN CHECK THE CAPS BY LIFTING ONE LEG.STILL LOW THEN INVESTIGATE FURTHER BY LIFTING THE LINK OR LINKS IN THE CIRCUIT CONCERNED.

If that fails you may have a shorted turn in the transformer or a cracked ferrite former.
Only way out may then be to look for another transformer.
The former on these has 22 pins and a very imperial inch layout so any attempt at a wind with a different former would be difficult at best.

The HV wind is 1025 T #39 for a voltage of -2050V and i know someone has attempted it and succeeded but its probably difficult.
+30 is 16T #23
+100 37T #31
+5 3T #24
+8.6 and -8.6v 5T #24
filament 6.3v 3.5 T #24

The main core is also encapsulated with resin so if you just have shorted turns then it would be very difficult to rewind without rewinding it completely i guess and re-encapsulating in a pressure chamber. Otherwise you would probably have corona issues and flashover at 2K plus.

I am waiting for another transformer to try as i have a broken ferrite core and have exhausted all the above.
I will post the result after i fit it....who knows if it will. I am not sure when i will give up.

Actually if it does not fix it then the problem could be the inverter itself or the feedback self oscillating transformer although its very low current and unlikely to break. watch this space.

Like the idea of seeing if it will run on a 43v independent supply but you would need to monitor the supply current.
It may be that it will run on the 43 with more current if the no of turns has reduced in the primary 20T+20T #25 but the core may then start to saturate due to the reduced number of turns and overheat as the resistance is now moving towards the resistance of the wire itself .
 20T of #25 looks like 0.2 ohms at a path of 94mm ish which is an instant short .

Very good diagnostic strategy but I prefer to leave the pre-regulator test for later, not to supply the oscilloscope by the mains, but rather to supply C940 directly at 43Vdc with current limitation at 1.5A by an external dc power supply.

It is much safer, from all points of view, because we do not work under dangerous mains related voltage.

[bd139]

Also good for the nerves because the preregulator is a bit explodey 

[akimpowerscr]

Regarding the pre-regulator, I would like to add a point that few know about.

I read in another topic that some consider that the 2235 is not running hot because it consumes little.
This is indeed the impression we have because the case does not heat up much.

However, the problem is that the airflow in the oscilloscope is very poor.

On the contrary, some components, especially of the pre-regulator, operate at a high temperature, too high in my opinion.
This is particularly the case with the T906 inductor and the Q9070 mosfet.

This undoubtedly has a very negative influence on the reliability of the pre-regulator.

That is why I strongly recommend installing a fan in your 2235.

Everything is planned for this, see the 2236 service manual for the information needed to install a fan.

[bd139]

I think the main issue with the preregulator is the MOSFET they specified was fairly tight on the SOA. If you swap it for a newer one
with a larger SOA window in this use case (I think I used an IRF820) it's less likely to explode.

Oh also in the early ones the MOSFET is attached with a crappy molex connector. That's not going to help.

[akimpowerscr]

Yes, but this is not a major concern because it is easy to repair and we have now MOSFET far better than in the past.

Those Molex conectors are mostly almost cooked and it is obvious that the wires must be soldered.

But the most severe failure is when inductor T906 is cooked for working at a too high temperature for a long time : you will never find a spare part to replace it.

[bd139]

Yes indeed. On your last point, the only answer is to buy more than one scope 

[akimpowerscr]

Yes indeed. On your last point, the only answer is to buy more than one scope 

Or to install a fan before this happen !!!!

NB: I have installed a fan on my own 2235.

[floobydust]

No need for a fan at all, even Tektronix concluded that with the 2235's.
I like that they are quiet and not sucking in dust through the switches and the PSU has less load - with no fan. My preference.

[jdragoset]

This guy has numerous 2235 "parts mules" that many of power up with EHT.
alltest_surplus

[beanhauler]

This alltest place is amazing. Shame its in the US as the shipping is so expensive to the UK