Tektronix 2235 power supply

[Christophe]

Dear all, I have a Tektronix 2235 that I bought in 1986. I use it time to time to fix audio equipments. Last time, the front LED suddenly went off and I could hear the power supply chirp every second or so. I open the scope and found that the pre-regulator goes into over current protection by sensing 200 mV over the 0.2-ohm sense resistance. There is absolutely zero volt on any of the secondary windings, the neons are not lit. So it seems that the pre-reg. does its job but the problem seems located in the inverter (I replaced the TL594 in case). I lifted the 0.2-ohm resistance and tried to apply 40 V across the 1-mF capacitor with a 1-A current-limited dc power supply. Around 8-10 V, the max current limit is hit and nothing happens. So I am left to believe that the inverter is guilty but given the rather busy PCB with heatsinks, I would like to go straight to the faulty guy rather than unsoldering the TIPs and the transformer. If anyone has some idea what part I should change first (the TIP-31 seem ok with the Vf-meter at least) I would appreciate! As a side comment, the power supply schematic is interesting and done in a clever way, probably to purposely give hard times to the service man...  Merci in advance for your help. Cheers, Christophe.

[dfmischler]

Did you try the "Power Supply Isolation Procedure" outlined in the manual?

[Christophe]

Thank you. Unfortunately, the manual does not give details on the inverter itself. The troubleshooting procedure describes how to isolate a faulty section in the secondary but there is nothing specific to fix the inverter. I confirm that both TIP31 are alive but I have a doubt regarding the saturable drive transformer. These drive XFMR rarely fail but this one is almost 30 years old and looks "tired". I measured 2 x 14 µH for the primary and around 9 mH for the secondary. Don't know if someone has the specs for this part? Thanks.

[rf-loop]

Thank you. Unfortunately, the manual does not give details on the inverter itself. The troubleshooting procedure describes how to isolate a faulty section in the secondary but there is nothing specific to fix the inverter. I confirm that both TIP31 are alive but I have a doubt regarding the saturable drive transformer. These drive XFMR rarely fail but this one is almost 30 years old and looks "tired". I measured 2 x 14 µH for the primary and around 9 mH for the secondary. Don't know if someone has the specs for this part? Thanks.

There is some diodes (400x or similar) and some electrolytics (typical "not start - chirp") fail is in secondary side) what is extremely typical reason for this "chirp" not start problem. In some cases also power led blink.  (repaired many of these)

[Christophe]

I've thought of the secondary but I am not sure it would prevent the inverter from starting. I don't have pulses on the primary side at all, the pre-reg. works fine but that is all. Unless you observed that in presence of a secondary short circuit the pre-reg does not even start?

[rf-loop]

I've thought of the secondary but I am not sure it would prevent the inverter from starting. I don't have pulses on the primary side at all, the pre-reg. works fine but that is all. Unless you observed that in presence of a secondary short circuit the pre-reg does not even start?

Tek 2213 - 2236.
It try start but stop. ("chirp" voice)

mostly secondary as I told. (but not of course always.)

And this short "Chirp" usually just means it try start (chirp) shut off... and it try agen (chirp)...etc.

Also you can not easy measure in circuit these due to its principle.
Short circuit cap in secondary side can absolutely sure affect it do not start. Tektronix guys are not stupid when they design these.
All what I have seen these 95% is just this secondary side caåacitors and rectifiers problem.
Diodes, may be short (or open) electrolytics, may be short or loosed capacitance and ESR super high.

Also, not all cases it do not prevent startup. It can also start up if there is error in PSU outputs.
There is jumpers and you can isolate PSU and measure.

After you get is start, look -8?? V (not remember exact value). DO NOT adjust it if it is not clearly out of service manual value. If you change it from value what it have been when it was calibrated, it need calibrate agen. And this need do exactly as TEK service manual tell and also order is important.  So, do not change any adjustment until really know it need adjust.
Look that ripple and voltages are inside service specifications.

When first electrolytic fails there, it is sign for change least all in secondary side, they are old and in this time Tek use sometimes poor electrolytics there (perhaps designed for reduced life time).

Note, safety! Primary side and HV circuits may have enough for kill you even after power plug is out.

[Christophe]

Hello Rf-loop, ok, so according to your experience, a short on the secondary side could block the inverter and force the TL594 to enter an over current condition. What really surprises me is that I have observed any of the output windings in the secondary side while the power supply hiccups and I don't see a single mV of voltage rise. Usually, when the output of a multi-output supply is shorted, hiccup mode gives a little voltage on the other sound windings. But here nothing. It is perhaps due to the fact the inverter does not even start in case of short circuit? Rather than opening the jumpers, can I individually bias output by output with a dc power supply and see if the output is shorted or not? Thanks.

[markce]

Recently I did quite some work on a 2236 that had more and more trouble stating (slow and chirp).
I ended up exchanging most of the cap's in the PSU section, only a few were actually less than optimal. Some were not 105C, and given
their age, I retired the others. I also replace a number of carbon composition resistors that where close or out of spec.
As a start, first check C925 and R926. Mind you, switching cap's should be high quality low ESR: Nichicon (PM, PW) or Panasonic (FR, FM)

[Christophe]

Thank you for the reply. I will check these caps. and will try to individually bias the dc outputs one by one while the primary side is off. If there is a frank short circuit, I hope to see it quickly. I have to re-solder the parts in the primary side first, will do tomorrow!

[markce]

For checking capacitors in a SMPS, you would normally use an ESR meter.  If you have a signal generator doing 30kHz into 50ohm, with adjustable DC offset, you could use that (series resistor divider (fe. 47+3 ohm)  and AC mutlimeter).

[Christophe]

Thanks, sure, but what I want to do is to individually bias the outputs to see if at least they accept their nominal voltage without excessive current flowing in the dc power supply. I assume that I can detect which one is shorted using this method as the ohm-meter does not reveal any frank short.

[markce]

individually bias the outputs to see if at least they accept their nominal voltage without excessive current flowing in the dc power supply

You could do this, but to be sure, you will need to test all secondary supplies in one go as well.

[Christophe]

Hello Markce, I did this test but I could not spot any issue. I individually biased all individual outputs (-8.2 V, 8.2 V, 5 V, 30 V and 100 V) and the consumption seemed to reasonable on all of these lines. I could not bias them all at once, not enough power supplies  I could even see the front LED well lit when the 5-V line was biased. The only thing that I did not open/test is the voltage multiplier that biases the tube high-voltage electrode. Then, I open/lifted all jumpers included the series inductances 960/961 and the rest of the available jumpers so that the secondary is disconnected. Same problem in the power supply. I then looked at the waveforms driving the primary side of the transformer 948 and when nodes 3/4 go up (at the hiccup pace), then 2 and 5 also go up in phase, so there is a net 0 V across the primary and no wonder I observed nothing in the secondary outputs. I was suspecting the drive transformer T944 but I doubt it can fail, there is little power in it. Based on these observations, it does not seem that the issue is located in the secondary, unless the multiplier U975 messes up everything, but I doubt. Next step is to either drive the primary side of T948 with a MOSFET and a clamp network at low voltage to see if the secondary (T948) lives or externally drive Q947/948 with two 180° delayed square signals, also at low voltage and limited currents.
Any suggestion or comments will be appreciated! Thanks.

[markce]

As far as I know Tektronix used quality caps in the 22xx scopes, so I would not expect blown or leaky caps. But they are 30 years now.
A failure mechanism I've seen is increase in ESR, drop in capacitance and some increased leak current. So I would not expect a complete short.
BUT As a result the primary self oscillation could stop.
Some resistors in the power supply are carbon composition type (looks like small cigars), and I've seen them increase value by 20% and more over time when stressed enough (high voltage or peak current). THIS IS EXACTLY WHY THIS TYPE WAS USED, they do not fail, but wear out.
In my scope, the resistor I mentioned before is of this type and is needed to start the control IC.     

[markce]

Ok, practical approach:
Check R926, if +/- 10% leave it, else replace (I use a MOX resistor as replacement).
If you are able to check ESR of capacitors:
- C925 : 100uF
- C940: 1000uF
- C944: 1uF tantal
Look-up the typical ESR on the Nichicon www site.

If you can't really do low ESR measurement, replace all these cap's. F.e. Mouser has them.
You would want to do this anyway if you continue using this scope, and you will...
Let us know where you are after this!

If you are going to order cap's online, I can advise you that one of my next steps would be to replace all secundary PWS cap's.
The power supply is unregulated, and quality is very much dependent on capacitor quality...
Ordering these cap's also might save some costs.

[Christophe]

Markce, thank you for these suggestions, I will consider them. R926 that you mention concerns the upper pre-reg. This one starts up ok and hiccups nicely. It is the one that goes into protection because of too much current (200 mV/0.2 ohm) that flows in R907. As I explained, I have isolated the inverter by lifting up R907 and applying less than 40 V between TP940 and TP950. I have read in other posts that the inverter must crank in this mode. My problem is that both Q947/946 are on at the same time and none of them block as it should once the drive transformer saturates. This is the reason why power the inverter in my case puts the supply into its 1-A current limit and limits the voltage at 10 V. I have lifted pin 4 of T944 and I applied small voltage pulses: Q947/946 react ok when the bias is applied and conduct as they should. Voltage at pin 4 of T944 shifts the average BH operating curve and pushes it closer or farther from saturation hence playing on the inverter operation. So if the caps you mention are bad, they surely will affect operations but an ESR shift would create the start-up current limit that I observe according to your experience? Have you seen a defective drive transformer already? Thank you.

[markce]

Your test supply is too small I think. You need 40V/3A, or more.

[Christophe]

I will try with higher current capability. I checked the rear panel of the 2235 and they say max power in watts is 40. It roughly corresponds to the 1-A limit set by R907. But the secondary-side in-rush current could potentially put the dc source in limit. I will try with the 3-A source you suggest and will see what it does. Funny, the power summary given by schematic values (5.2 V/6 A, -8.6 V/7 A, 8.6 V/3 A and 100 V/ 1 A is way above the 40 W number in the rear panel. More news next week!

[rf-loop]

Funny, the power summary given by schematic values (5.2 V/6 A, -8.6 V/7 A, 8.6 V/3 A and 100 V/ 1 A is way above the 40 W number in the rear panel. More news next week!

No.
1A,2A,3A,6A,7A

they are "address" where these lines are going.

Also in first post you tell that it "chirp" when it try start.
Mostly but of course not exactly allways it means that primary try start. In start up phase it go to "error" and stop. After short time it may "chirp" agen becouse it agen try start.

I have repaired several of these where have been this "cirp" without power up, and really (in my case) all was secondary problems and electrolytics and/or some of diodes in secondary.

(yes I have also seen broken T948 winding just near solder point. but it was different case)

Of course problem may be in many places in individual unit but I still recommend you check these secondary outputs becouse it really is common known problem in these.

Look diodes CR954 - 957, CR960-963, CR967 and 979 970.
Then test every electrolytics after these diodes, there are exatly 8 of these.
They may be "open" or "short" or normal. If someonme tell they do not go to short. I can proof they can.  Also diodes may fail to "open" or "short". Measure them. Pic up other end of diode and measure.  If you find failed diode (short or open) specially then look this circuit electrolytics.

As someone tell. I recommend change all these secondary output electrolytics, if fail or looks ok. They are not good and they are old. (and depending where it is manufactured, and what lot, there may be moderate good capacitors or not so good.)




--------------  more about what are these 7A and 3A etc.
Yoy can see in pictures example this (if you look sheet 7 and there 8.6V power output.):

+8.6V
TO
3A
8  (around this 8 is small box)

look sheet 8, it is going there. (power distribution "map"
Then find location 3A there (in sheet

This address system is good very common and now you easy find where they go.
(exept that in my service manual there  is small error and destination is in location area 4A, so it need also be wakeup when use these manuals, there are also small errors what make some times things more fun.

[Christophe]

Hello rfloop, thank you for these advices and the comments about the addresses, it makes sense given the 40-W nominal power and 1-A current limit of the pre-regulator. I can try to change all these caps and diodes, but I would like to further comment: an ESR going up in an output cap. usually increases the ripple across the cap. and can potentially trigger the OVP circuit but it also increases the dissipated power in that cap. Here, the circuit runs open loop as there is a local feedback around the rectified collector voltages in the primary side only. Given the push-pull configuration (buck-derived), the circuit is blind to OVPs on the secondary side. Also, I dc-biased all outputs individually and I all current consumptions seemed within the right boundaries. I assume that a shorted capacitor would have manifested itself via an over-current condition, no? However, it is true that I did not individually Vf-test all diodes, will do.

I found the diodes references in the schematic but could not locate CR979 (found CR879 though). Could you please kindly let me know in which part of the sketch this diode is?

The inverter is a self-relaxing converter based on the saturable core transformer T944 and the middle point is adjusted by the pre-regulator to around 40 V. In a healthy power supply, could you  please confirm (or disagree) that lifting R907 lets you separately power the inverter with a 40-V dc source applied between TP940/950? Also, when you repaired the chirping power supplies, did you notice also absolutely 0 pulse on any of the secondaries, indicative of a non-starting inverter? In these cases, output caps/diodes replacement was the solution? Because this is what I have and I don't understand that absolutely no pulse goes through at power up. Unless T948 primary is a complete short?

Anyway, as you recommended, I will order 8 new low-ESR caps and replace them anyway. The scope is almost 30 years.

I will happy to read your comments! Thank you.
 

[rf-loop]

oops, CR979 was type mistake, 970 of course.

You have measured every seondary side diodes?

(And of course you have measured primary side diodes CR946 and 947?)

If I have listen this "chirp" I first suspect that primary try starting if I hear right voice. (and as you tell this chirp, I first think this is just in your scope)

Perhaps we talk different "chirp" and then if need suspect primary is not starting it need of course look there.  In normal case I look with oscilloscope what is going there. But all this was years ago and only result what I remember exatly is that secondary caps and diodes habve been typical case.

If you desing to run it with some external power, please read schematics carefully! Do you have right schematics at all?
And if you use oscilloscope with power supply, read service manual note for scope GND connection and/or isolation.




 

[Christophe]

Hello rf-loop, I will measure diodes and replace caps. tomorrow and will let you know. Yes, the chirp is the power supply going into hiccup. I have looked at signal around the pre-regulator and the voltage across the 0.2-ohm R907 jumps to a heavy Continuous Conduction Mode of operation and is clamped to 200 mV, meaning 1 A on average is delivered. What fails to start is the inverter. I can't manually crank it with an external dc source after I lifted R907 to decouple it from the pre-regulator but the source hits the 1-A limit, just as the pre-regulator does. With these self-relaxing inverters, it is very difficult to know the source of the problem, unlike a classical PWM-based type of power supply where you can trace the signal in an easier way. As both TIP31C are ok, as well as the surrounding diodes, either the saturable drive transformer passed (which I doubt), the main transformer has its primary shorted (again, very rare to see a transformer shorted, especially when driven by a current-limited source) or there is a frank short on one of the output that I can't spot by individually dc-biasing these outputs. Tomorrow another week starts and I will continue my search, letting you know the outcome. Cheers.

[Christophe]

Ok, today I checked all output diodes and they were all ok. I unsoldered the whole transformer secondary side (including the wires to the filament and the HV multiplier) and I removed the driving TIP31C transistors. So T948 transformer is left with a fully open secondary and its primary middle point connected to TP940. I connected two 200-V/10-A MOSFETs between both primary sides to gnd and gates are tied to their source via a 10-k resistance. Then I applied 15-V/4-µs wide pulses on the gate of one MOSFET with a 15-V bias between TP940/950 while the other MOSFET is left blocked. I also installed a diode from the drain to TP940 as a free-wheel diode. I expected to see the drain collapsing to almost gnd for the 4-µs duration and going up afterwards. I was also expecting to probe voltages on the secondaries since all is open. Instead, I saw 5-A current pulses in R947, a hot MOSFET and a drain barely touching ground for a less than 100 ns, immediately going up due to the rds(on) quickly increasing. I've done the same on the other winding: exactly the same observation, no sign of life on any of the secondaries and almost impossible to bring the drain to ground by the MOSFET. My conclusion, the transformer primary is dead  The reference is 120-1348-00 and I am currently looking for a spare. I found one on eBay from a company in Greece, has anyone used this company before? Thanks.

[dfmischler]

I found one on eBay from a company in Greece, has anyone used this company before? Thanks.

I bought the ganged coarse/fine horizontal position pot for my Tek 475A from qservice_rhodes.  I can't remember how long shipping took, but I left positive feedback, and the part fit and worked perfectly.

[rf-loop]

I found one on eBay from a company in Greece, has anyone used this company before? Thanks.

If you mean Qservice. Exellent service.
I have used it many many times, always satisfied.