Re: Tektronix 2215 Scope pre regulator repair

[Per Hansson]

I've spent the weekend playing with my preregulator board on the bench, not hooked up to the scope (waiting for fuses for that).
Anyway as described above I switched out Q933 for a IRFBC30 that I found, and it works fine.
I was curious about troubleshooting the noise people report though, when I power my scope up it's always made this "screech" or "chirp" sound.
I always thought it was the CRT but just laying bare on the bench it does the same, so I've ruled that out

What I have noticed really confuses me though, reading the manual U920 should not be switched on until Q915 conducts.
But if you look at the electrical schematic we have the large 150k resistor R911 directly after the bridge rectifier.
This feeds into C913 which is the 100µF capacitor for the PWM chip.
Then we have R912 which connects to the base of Q917, it wont be able to conduct until VR917, a 6.2v zener diode has been overcome.
When that happens current can flow to Q915's base which should then start up U920 via Q915.

But, and this is the part I don't understand: There is a resistor called R914 connected before both Q915 and Q917.
I understand it's purpose is to work as a holding circuit for them, but it's also connected to the VCC pin of U920.
So does that not mean even if Q915 is not conducting U920 can still be on via that resistor?
Because this is precisely what I'm seeing: I have ca 15.5v on the VCC pin of U920, and the preregulator is running (43v present on the output).
But after a random amount of time (sometimes 0.1 seconds, sometimes 10 seconds) the voltage drops to 14.8v.
When it does the noise disappears.
I take this as the fact that the secondary winding of T933 has come alive, and is supplying the circuit via CR913.
But this part is also confusing for me: CR913 will try to fight the power coming in the other way, via R911?

I have tried to catch this with a scope, but the waveform is haywire when the preregulator is making the noise (It's T933 making the noise).
Actually the waveform itself seems to be correct, because if I take a screenshot or hold it's stable, but it's not able to trigger on the signal.
That is until the noise abates, then it all becomes stable.
During this time, if I try to measure the frequency with a DMM, it tells me it's around 32khz, when the noise disappears it's stable at 38khz.
I figured that must mean that either C920 or R920 is bad, but I desoldered them and they measured fine.
I replaced them anyway, but it made no difference what so ever...

The thing that makes me think I'm not crazy is that if U920 really starts up before Q915 has conducted, then it will be running without a capacitor.
Because that's also part of what Q915 does: it switches in the stable VCC provided by capacitor C913, it's the whole reason for it's existence...

[Vince]

I was curious about troubleshooting the noise people report though, when I power my scope up it's always made this "screech" or "chirp" sound.

Hi Per,

I think with all your experiments, it deserves a topic of its own... it's starting to get messy/confusing with both our 2215's  being fixed in the same topic.. people will get lost and confused.  Getting hard to follow I think...   Space is cheap on the forum, just create a topic for your scope and everything will be easier for every one to follow  and help / comment on

As for the SMPS, the theory of operation is described in the service manual, have you read it ?  Most of what you say is already explained there, and your remaining question are answered there as well. Please take the time to read it.

But in short :

But, and this is the part I don't understand: There is a resistor called R914 connected before both Q915 and Q917.

As the manual says, this resistor as you can see connects the output of that 15V PWM regulated supply circuitry, to the input. It's just a little bit of positive feedback to help lock it sharply/quickly.

So does that not mean even if Q915 is not conducting U920 can still be on via that resistor?

Well no, look : the left leg of the resistor is connected to the base of Q917, which is regulated at 6,0V via the zener. So 6,9V is not quite enough to power the PWM chip, it's only half the voltage. Plus, the resistor is high value, 100K, so only 60+ uA can flow thourgh it at best... not quite enough to power the chip either.

Because this is precisely what I'm seeing: I have ca 15.5v on the VCC pin of U920, and the preregulator is running (43v present on the output).

You said it, once the regulator is running, power comes from the 43V outpuit, via CR913 which you were wondering about ;-)

But this part is also confusing for me: CR913 will try to fight the power coming in the other way, via R911?

It's not fighting anything ;-)
It's just there to power the PWM chip once the regulator/PWM chip has started. If the diode weren't there, the juice the start-up circuitry is trying to produce to feed to the PWM to get it started, would go straight to the regulator output/scope which would draw all the very modest current that's available during startup (remember it all goes through the beefy 150K resistor, so not much current available), and the PWM chip would probably never manage to start...
Well that's my understanding anyway.

During this time, if I try to measure the frequency with a DMM, it tells me it's around 32khz, when the noise disappears it's stable at 38khz.

Yes as the manual states, the regulator frequency is around 40kHz, so 38kHz sounds about right...
In my 2232 for example (exact same pre-regulator design), the PWM runs a bit faster, at 60kHz.  The inverter following it though, runs at 20kHz, same as 2215.

I figured that must mean that either C920 or R920 is bad, but I desoldered them and they measured fine.
I replaced them anyway, but it made no difference what so ever...

Probably just warm up related... I wouldn't worry about it !  You say the noise stops in just a few seconds from what I understand ? That's perfectly acceptable or any SMPS I think... when we refer to the "whining"  22XX scopes, we mean a very, very high pitch, really loud, and which last a lonnnnng time, like 5 or 15 minutes, or sometimes never goes away at all ! 

But Hakan on Tekscope (same chap you got the transformer rework procedure from) gave us an old Tektronix internal "paper" of the era, where they discuss precisely the whining 22XX issue. So even back then they the issue !  Tek wrote that it's a combination of things, but mostly a frequency tuning problem, so they advice to slightly tune the regulator frequency to put get rid of any resonance issue.

The thing that makes me think I'm not crazy is that if U920 really starts up before Q915 has conducted,

It does not, read SM again 

Because that's also part of what Q915 does: it switches in the stable VCC provided by capacitor C913, it's the whole reason for it's existence...

No C913 does not "stabilize" anything. It's Q917/Q915/ Zener that regulate the 15V supply for the PWM chip.  First this circuitry is powered at startup by the voltage building up across C913 via that big 150K resistor connected to the rectified mains, but then as soon as the PWM starts to regulate, power comes from the 43V regulator output, via that CR913 diode you talked about.

Anyway, please start your own topic if you wish to discuss this further, it's getting messy around here   
You can just copy/paste your prior messages there so people know what you already have done so far. Give us the link to your topic once created so we can follow you there, if you have more questions or what not 



As for me, back to my repair... making slow progress but making progress. Studied the schematics for hours, as well the theory of operation, several times.. with a few weeks in between each "round", so as to let the knowledge slowly seep into my tiny brain. I think I got the hang of it now, I am ready to do some probing around and produce some trouble shooting data, at last....

Hoping to start poking around and report this week.. holiday is over so time is become scarce again...

[floobydust]

I've plundered around the Tektronix pre-reg and here is what I know.

Q915 and Q917 are a discrete undervoltage-lockout (switch), as the TL594's built-in undervoltage-lockout is no good for mosfet drive.

Across C913, at 20V and higher Q915 turns on. On a cold power up, this takes around 2 seconds for start-up resistor R911 to do, charging up C913, which is that long power-on delay you see for the green power LED to light up. R914 provides lots of hysteresis.
R912/R913 are a sense voltage divider. 2235 increased it to 23V (100k->120k).

Once Q915 is on, if C913 falls below about 10V, Q915 turns off, shutting off the SMPS. U920 cannot run without Q915 supplying power.

During power-up, there is maybe 100msec for the pre-reg to start and run and then power itself through CR913. Any troubles and it hiccups.

I think the power-up chirp is normal - current-mode happening until the pre-regulator output voltage comes up, there is no soft-start.

Early Tektronix hand-selected U920's oscillator R920 to match T933/C932 resonance. They wanted PWM freq. to be precise I guess.
(I don't know what C933 does, it's too big for a snubber).

You might be seeing the PWM oscillator freq. drift during time the it takes for U920 to warm up?

Also note the 2235 changes to the power supply to tame parasitics on the mosfet - adding source ferrite bead and 39R gate-drive resistor, and some work on the feedback loop stability - if you think the problem is in the twilight zone.