• Please review our updated Terms and Rules here

Tektronix 4010 restoration

thunter0512

Veteran Member
Joined
Sep 27, 2020
Messages
839
Location
Perth in Western Australia
Last Sunday I picked up a very dirty and corroded Tektronix 4010. I spent two days on disassembly and cleaning and now I am looking for advice on getting the terminal going again.

Roland Huisman had some of the same issues as documented in this 3 year old thread: https://forum.vcfed.org/index.php?threads/tektronix-4010.77121/

I have removed the power supply from the pedestal and identified it as a "Heavy Duty (HD) Power Supply" used if you have some of the optional extras which require more power.

Roland's Tek 4010 had a so called "Low Voltage (LV) Power Supply).

I would like to reform the 4 large electrolytic filter caps (C33 dual 150 uF, C5 and C205 9600 uF, C255 86 mF - (yes milli-Farad)) in my supply before attempting to power on the supply for stand alone load testing.

Unlike other power supplies I worked on it seems very tricky to isolate the soldered in capacitors (C33, C5 and C205) to allow for individual slow re-forming via a current-limited bench supply.

It appears that I have to completely disassemble the power supply to be able to desolder the 3 caps (C255 has screw connectors but the PCB still has to be removed to get C255 out). Roland too complained about the "worthless construction".
It seems I have to remove the rear hit sink to access the bolts holding 3 of the filter caps, unsolder at least some of the transformer's heavy duty secondary leads, disconnect a few connectors going to the heatsink, remove the supply PCB, and possibly more I haven't discovered yet before I can even think about removing the filter caps. This is surprisingly complex to just isolate or remove the filter caps.

I don't believe in variac based "reforming" but wonder if this is my only viable option.

In my experience reforming of individual caps via a current limited bench supply works well, limiting current to a say 10 mA and slowly increasing the voltage from 1V to the rated voltage pausing after each increase until the current drops to zero or close to zero.

Of course this only works if you can isolate at least one pin of the capacitor, otherwise something else in the circuit will draw current.

I think it would be irresponsible to just apply mains power to the power supply and risk blowing up the very large and expensive (unobtainium?) electrolytic caps.

Another problem is that the Tek 4010 maintenance manual on Bitsavers has no parts list for my "Heavy Duty (HD) Power Supply" (see https://bitsavers.org/pdf/tektronix/401x/070-1183-01_Rev_B_4010_Maintenance_Manual_Apr_1976.pdf).

Any thoughts or advice or suggestions?
 
Last edited:
Can you get at the bridge rectifiers?

If so, I suspect you can reform the majority of these capacitors without removing them if that is tricky...

Dave
 
Thanks Dave. I think removing the bridge rectifiers would help, but does not completely isolate all caps. Also removing some of the rectifiers requires removing the PCB which is the difficult part.

Instead I made an attempt at removing C33 I thought would be the hardest to get out as it has six pins. I managed to disassemble enough so that I could lift and tilt the PCB a little so that there was a chance to get the capacitors out.

I used a desoldering iron with vacuum pump to remove most of the leaded solder on the 6 pins of C33. I then used "chip quik" bismut based low temperature solder on all 6 pins and then heated them gently with a temperature controlled SMD heat gun to soften the bismut & leaded solder alloy. It appears that the 6 pins are press-fit into the PCB.

At this point I gave up trying to remove the caps for reforming.

I removed all the bismut alloy with my desoldering iron and then resoldered the 6 pins with leaded solder. I then re-assembled the PCB, cap holders, heat sink etc.

I could not re-fit two of the bolts holding the plastic cap holders because the "captive" nuts were not (or no longer) captive and fell out. Due to the mechanical arrangement it is impossible to fit the nuts once all caps are soldered in.

I am now using a variac to very slowly increase the mains input from 1V to the nominal Australian 240V. I have replaced the mains fuse with a 500mA version for protection. As I am writing this I am at 115V and nothing gets hot and the fuse hasn't blown.

I much prefer doing this on an isolated electrolytic cap with a current limited bench supply as with the current limit self-healing can occur without overheating and destruction of the cap.
 
Last edited:
You don't need to remove the bridge rectifiers for SOME of the capacitors.

For example, C255 if you remove fuse F145 that (effectively) isolates the capacitor because you are going to supply a positive voltage to the + side of the capacitor and the diodes in the bridge rectifier will be reversed biassed - so you are good.

I can see a few more capacitors where you can do the same trick - but I can't see the full schematic (so I can't say this is true of every capacitor).

Dave
 
I had the Tek 4010 supply running for a few hours now at the full 240V mains power so the electrolytic caps should be fully reformed.
During the reforming process it run idle with no load and provided stable -15V, +15V and +5V.
Subsequently I loaded the +/-15V rails with 30 Ohm and the +5V rail with 5 Ohm resistors and all rails were nice and stable and everything stayed nice and cool.
This means the 4010 supply should be ready for action.
 
I had to replace a number of shorting and popping tantalum caps on the boards in the pedestal.
After these stabilised I reconnected the display cables and the 20V unregulated supply fuse blew.
The culprit was another shorting tantalum C41 22 uF on the HV and Z Axis PCB. I replaced it with a standard electrolytic cap and the fuse no longer blew.
All further trouble was on the HV and Z Axis PCB:
After about one minute the just replaced C41 exploded. It was rated 35V and the nominally 20V unregulated supply line was at about 21V. Polarity was correct. Weird!
I cut out the remnants of C41 as on the other end of the 20V supply there is a large cap so it should limp ok without C41 - or so I thought.
When I reapplied power, after a few seconds something else popped and let out a small amount of smoke, but I couldn't find the culprit.
While probing around I touched one of the glow lamps and the leads just broke off.
When I repowered the unit a large 4700 pF 6kV ceramic cap started arcing in the top center. It is used in the voltage doubling circuit.
I now suspect that it was this ceramic cap which popped earlier but the damage was not obvious.
Mouser has suitable glow lamps and 4700 pF 6kV ceramic caps, so it looks I will have to spend some money and wait.

The Tektronix 4010 is remarkably complex and not easy to work on. I don't fully understand the chopper circuit around Q101, Q103 and Q105. There is also an undocumented cut PCB trace coming from one of the leads of Q103.
I wonder if some else in the past has made a repair attempt on the HV and Z Axis PCB.
 
There is quite a high failure rate of tantalum bead capacitors even when new!

I placed a contract for over a thousand remanufactured boards at work - each with tantalum bead capacitors around the +15V and -15V regulators. We had quite a few fail short circuit when they were burnt-in. Most caused the power supply to crow-bar when initially turned on. Some went short circuit during the burn-in testing.

Due to the high failure rate, the component manufacturer was contacted - but we were still within the quoted infant mortality curve...

In retrospect, we would have designed the burn-in rig differently if we knew we were going to have these problems with the 'smurf grenades'...

Still, all sorted now!

@Hugo Holden came up with an interesting observation regarding whether the tantalum bead capacitors explode or not! Well worth a read.

Dave
 
[offtopic alert, though I am following this thread with interest]

@daver2 , you seem like the kind of person who might have a list of tantalum bead capacitor values for 4010-series terminals and 405x-series computers... I have a 4006 whose HV and Z-axis board has popped a tantalum or two, and a 4051 that I haven't dared power up yet. Is there a handful of values used in these machines that I could buy in bulk --- that way I can replace failures one by one, or perhaps replace every capacitor in one fell swoop? (I'm not sure what would be wisest.)


Slightly more related, a friend sent me this Reddit thread where a mysterious catastrophe has attacked the display: the entire coating of the screen is gone. It's a window now. I sent it to some folks at the VintageTEK museum and their guess is that the seal between the CRT bell and the faceplate failed: this caused a gust of wind to sweep the coating completely away. (By contrast, a vacuum failure at the neck usually focuses a blast at the centre of the screen, which creates a kind of "splat" scour pattern in images I've seen.
 
Well, the parts list exists for the various machines (it certainly does for the 405x series of machines) so it is certainly possible to go through the parts lists and identify the capacitance and DC working voltage for the various parts. If there are common parts that appear (queue spreadsheet) then you can purchase some stocks.

These parts are so common, though, that I suspect it is not worth it... Your choice of course.

I think someone has removed the CRT from that machine. Breaking the vacuum does cause the phosphor to become dislodged from the screen - but not like that! I wouldn't even have thought that a break elsewhere would have given such a clean effect...

The wisest way is with the computer outside (on a nice hot day like we are having) with an extension lead going inside where you are!

Seriously, break the power supply down into manageable pieces if at all possible.

Dave
 
That was my thought too, but the VintageTEK folks thought it really could happen this way --- they said that the coatings on 4010-era DVBSTs were not bonded so well --- they certainly were fairly complicated, structurally speaking. I still am not sure what to think, but I hope I never find myself testing the theory!
 
Not at the moment. That's why I left it as an exercise for the reader...

Basically, the power dissipated within the tantalum bead is at a maximum when the internal impedance of the tantalum capacitor matches the source impedance of the power supply.

If the impedance is less than that of the PSU, the current increases - but the power dissipation reduces. This tends to not destroy the capacitor - but can cause resultant damage to PCB tracks etc. unless the power supply protects itself.

If the impedance is higher than that that of the PSU, the current reduces - and also the power dissipation reduces. This also doesn't destroy the capacitor - but there will still be an increased DC current flow in the capacitor (and from the PSU) as a result.

The 'window of eruption' occurs when the power dissipation is at a maximum - or either side of the maximum. This (obviously) depends upon the construction of the capacitor.

You can plot a graph of the power dissipation within the capacitor by assuming that the PSU is a constant voltage source V with an internal resistance of R1. You then vary the resistance of the capacitor (R2) from 0.1 Ohms to (say) 3*R1 and plot the graph of power dissipation within the capacitor from Ohm's Law [where: I=V*(R1+R2) and P= I * I * R2].

You should observe a peak where R1 = R2.

Dave
 
I have removed the HV & Z-Axis board and will try to slowly power it up with a current limited bench supply.
I have already found the resistor which yesterday made the smaller pop sound and emitted a bit of smoke.
I will probe the chopper circuit with a scope to see what is going on.
 
On the HV & Z-Axis board R170 was open circuit so the +15V (DEC) rail was floating. This caused the HV regulation to shut down the chopper circuit and resulting in no HV generation. With a new R170 fitted I get HV, but at crazy high levels of 11 kV rather than the approx 4 kV I expect. Something is still not right with the regulation. I noticed that voltage adjustment trim-pot R82 which is meant to be 50 kOhm is only 14 kOhm. Maybe there is also some trouble in the circuit around Q107 (n-channel JFET). Unfortunately Tektronix have used a lot of Tektronix numbered parts which makes replacement tricky.
 
On the HV & Z-Axis board R170 was open circuit so the +15V (DEC) rail was floating. This caused the HV regulation to shut down the chopper circuit and resulting in no HV generation. With a new R170 fitted I get HV, but at crazy high levels of 11 kV rather than the approx 4 kV I expect. Something is still not right with the regulation. I noticed that voltage adjustment trim-pot R82 which is meant to be 50 kOhm is only 14 kOhm. Maybe there is also some trouble in the circuit around Q107 (n-channel JFET). Unfortunately Tektronix have used a lot of Tektronix numbered parts which makes replacement tricky.
Check out this link on TekWiki for the various Tektronix Common Design Parts cross reference catalogs:

https://w140.com/tekwiki/wiki/Common_Design_Parts_Catalogshttps://w140.com/tekwiki/wiki/Common_Design_Parts_Catalogs
 
I have got the bulk of the HV circuit working on the bench powered from external power rails of +20V, -15V and +15V. The nominally 3.85 kV is adjusted to about 3.88 kV and nice and stable.
Here are a few photos:

IMG_20230607_214416032.jpg

IMG_20230607_214427669.jpg

IMG_20230607_214441601.jpg

You can click on the thumb-nails above to see the full size photos.

I still have to test and if necessary fix the following subsystems on the board: Focus Supply, Intensity Control and Z Signal Amplifier.

After testing in the actual Tek 4010 I will have to order a few parts I substituted with working but not quite right parts. Most notably C93 the large orange 4.7 nF 6 kV cap which arked I replaced with a much smaller (blue) 3 kV part which works fine as it is, but I will replace it with a 6 kV part from Mouser. I will also replace all the glow lamps where the leads have corroded and have become very fragile. It makes sense to wait until and use "close enough" parts until I have discovered what else I need to save on shipping. Perth in Western Australia is a nice place to live, but the two local electronics store chains (Jaycar and Altronics) are a joke. Sadly even on the much more populated east coast of Australia there is not much more (Element14 and RS Components are pretty useless these days). Our US friends have Digikey, Mouser, Jameco etc. I envy you.

Edit: I got a bit nervous about the 4.7 nF 3 kV ceramic cap running 30% above its rated voltage. I dug through my parts in the shed and found a few more 4.7 nF 3 kV caps of the same type and combined 4 of them into one 4.7 nF 6 kV cap. It looks a bit dodgy, but will be fine until I place my order with Mouser for whatever else I need.
 
Last edited:
I have re-fitted the HV & Z-Axis board and after power-up I get a bright vertical line along the left side of the screen (and also a dimmer curved line to the right of it).
It responds to the Page keyboard press with a bright green flash which erases the screen, but immediately after the vertical line is back.
I cannot yet explain the reason for the vertical line which appears immediately after power on.
Has any Tektronix 4010 person here got any ideas about what could be causing this?
Here is a photo (click to see full size):

IMG_20230608_225012390.jpg

Can anyone offer any advice?
 
The first thing to do is to divide and conquer...

Is your problem generated within the X and/or Y deflection amplifiers or in the logic cards?

Look on the schematics for the XANA and YANA signals and have a look at those with an oscilloscope.

From the look of your display, the XANA signal appears that it should be a fixed value whilst the YANA signal is varying. If this is the case when you look at those signals, the problem is in the logic cards. Otherwise, the problem is in the deflection amplifier(s).

If it is the XANA and/or YANA signals - we can work our way backwards to the X and Y latches and then further backwards as appropriate.

EDIT: Just looking at the schematics, the X and Y DAC latches are constructed from 74193 ICs. These have LOAD, UP and DOWN inputs. It is just possible that the Y DAC is being clocked by the logic - but we will get to that shortly by our detective work...

EDIT2: There are two signals that affect the Y deflection amplifier - YANA and YMAT and two that affect the X deflection amplifier - not unsurprisingly XANA and XMAT.

The ANA signals come from the main DACs. The MAT signals come from the alpha generator for the characters. From what I remember, the MAT signals should only vary the ANA ones by a small amount (as each character is composed of a maximum of 8x8 pixels). This is why I would suggest starting by looking at the main X and Y ANA drive signals.

Dave
 
Last edited:
Back
Top