开云体育

Hi,
I found a Tek 2236 scope in the garbage. After running it for a few minutes, the power supply failed. Powering the scope from a 42 volt supply caused the NPN transistors to fail.
I replaced the transistors with the MJE15030 (the hFe was measured to be 40 for both transistors). The scope powers on, but when the transistors get hot the power supply stops oscillating and the scope shuts down. Any ideas as to what could be happening here?

Also, when the scope powers on, the horizontal trace doesn't go all the way to the end. Increasing the intensity causes the trace to expand on the screen, which shouldn't happen. The trace goes from being smaller than the CRT to extending past the edges of the CRT when the intensity is increased.

The scope is from 1984 and the serial number is for the earliest revision.

Increasing the intensity increases HV PS drain, which is causing the HV to sag. Reduced HV expands the trace.

So, yup, you've got PS troubles.

Look first at the usual suspects (e.g., bad bypass caps -- tants especially), then move from there.

--Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
420 Via Palou Mall
Stanford University
Stanford, CA 94305-4070

The 2200 series have problems with the rectifier diodes going bad on the secondaries. Use something like 1N4937 diodes as replacements. A very fast or higher speed is necessary as replacements. The electrolytics in the supply should be replaced with low ESR types such as Nichicon ULD, UHE or UCY. Increasing the capacitance is fine, i.e. 270/330mfd for a 220mfd. Mount the new condensers above the board to allow air flow under them. The main condenser from the mains can be a Nichicon LGR type. The tantalums in the supply can be replaced with film types. The tantalums would be ones like 4,7mfd. Replace the RIFAs! Use something like X1 or X2 and Y1 or Y2 types. Remove the label and rubber top to oil the fan while in the supply. Tom Lee is right about checking other condensers.

Mark

Does anyone know what the power supply should draw when connected to an external power supply?
At 40 volts my power supply is drawing a little over 1.2 amps.

I have also replaced the capacitors in the power supply already, and I tested all diodes too (none of them were dead).

Hope you get your 2236 working again! It's a great little scope, I've used
one for the past 7 years as my daily driver at work. :-)

/mr

It appears that there is saturation and ringing on my transistors.


Here is a plot of the base and emitter of my transistors. The yellow trace is the current through the emitter at 1A/V (It is across a 1 ohm resistor). The green trace is the base voltage. As you can see, the transformer saturates until the transistor draws 5 amps. Then, ringing occurs (and it looks like Vbe breakdown might be happening). By turning the 8.6 volt adjust potentiometer from the highest to the lowest (I had to turn it up all the way to get the system to oscillate correctly at startup), the transistor saturates less. I tested the NPN transistor Q944 in the startup circuit, and it is working correctly. It seems like the startup circuitry isn't providing enough current to start up the power supply. I wonder if this is because the hFe of my power transistors is too low (it is about 40).

I don't know how you're probing those nodes, but I suspect the ringing you see is measuring error. It is tricky to get good CMRR in those kind of measurements. Try looking at one of the low voltage secondary windings on the output transformer, like the 8.6 V one. When things are right, it should be a very nice trapezoidal squarish wave with rounded corners. First find that good wave, and if it's not, then there's trouble, and you may be able to diagnose it from there - as long as you can have it running somehow, which seems to be the case.

The way it's supposed to work is that the base drive transformer saturates in each half-cycle to determine the oscillation frequency. It's basically a Royer type converter, but not using the big output transformer as the saturable device, which would waste lots of power, and make the waves ugly, It's really pretty slick, needing only the little transformer's saturation to make it work, and it also shorts out the primary of the big transformer at the right times to help clean it up even more.

The transistors are supposed to never saturate, but be close, or drop up to a few volts, depending on which of two different control schemes are used. One is where the switching transistors' drive is adjusted to drop the overhead to provide regulation. They have to then dissipate all that "series" power, so they do both linear regulation and switching. The other scheme uses a separate linear pass element to control the overall supply to the chopper stage, and it has to dissipate the excess power instead, while the switch transistors want to be just near saturation, to minimize their dissipation. I believe the switch transistors in all models needed to be selected or matched somewhat, to keep things from getting too lopsided, which could cause failure.

The chopper was to operate more or less the same either way, but can't exactly, with the different conditions. The outcome at the secondary should be about the same either way, nice clean waves with slow edges and rounded corners to minimize HF noise.

Ed

I replaced the transistors with a more matched pair (hFe for both transistors was 73 at a Vbe of 550 mV).

The circuit still does not start up reliably. There is another mode of oscillation at a few hundred kHz which the circuit sometimes falls into. I had to crank the 8.6V potentiometer all the way to get it to start.

Here are the oscilloscope captures.




There is still ringing.

The scope shots in the last post were taken on the output

I probed the collector and emitter voltages and it looks like the transistor is in saturation.




The first two show the transistor emitter and collector voltages on the input side. You can see that the collector voltage (trace in yellow, with a 10x probe) goes all the way down to zero, so the transistor seems to be hard switching.

The third one shows what happens when the scope fails to start properly. The power supply oscillates at 1 MHz.

Did you check or replace Q944? It's possible for it to be damaged during output transistor failure. If Q944's B-C junction blows open, it's possible for Q939 to supply enough base drive through R944 and Q944's remaining B-E, to start up and run poorly, without the current gain it would normally provide. Measure Q944's collector voltage to make sure some reasonable current is flowing through R945. Also check the base resistors R946 and 947.

Ed

I have checked (and replaced for good measure) Q938, Q939, and Q944. The base resistors R946 and R947 are fine too.

One weird thing I've noticed is that the voltage on the emitter of Q944 is negative. The base is at around 0V, while the emitter is at -0.7V. Measuring the collector voltage of Q947, it appears to go to -5V. I'm not sure why that is happening.

What reference point are you probing from? For now, forget the scope and do some DC DVM measurements while it's running. Clip the + lead on TP940, and leave it there for the following measurements.

Put the - lead on TP950. You should get the voltage applied from the external supply, roughly 42-43 V.
Then on the junction of CR946 and CR947. You should read around 40 V if it's regulating. If it is, then the regulator section is probably OK.
If not, write down the value, then check the bottom of VR943, which should be about 6.2 V as marked. If it's OK, check the emitters of Q938 and Q939.
That node should read around 5.5 V. Q938's base should read very close to the VR943 voltage when it's working, but won't be now if it's not regulating.
It should be close though, even if wrong. Calculate the expected voltage at the junction of R939 and pot R938, as a simple divider, using the previously measured voltage from CR946 And CR947. Then compare it to the voltage measured there. It should be fairly close, and if not, check C942 for leakage. Also feel C943 to see if it is hot, just in case.
Now check Q944's collector. It can show a very wide range, but should be maybe 38 to 1 V, indicating some current, and linear operation.

Now clip the DMM's - lead on TP950, and use the + lead for measurements.
Check Q944's emitter. This can get kind of weird because the B-E junctions of the output Qs act as rectifiers, and the transformer combines them and makes it symmetric, and it's superimposed on the R949 voltage, which varies with load. I think Q944's emitter can go from a little bit negative to a little bit positive, depending on conditions. The base should read that value plus the usual 0.7 V or so above the emitter.
I think normally, Q944 is only driven hard during startup, to activate the output Qs. Once the Royer operation is attained, T944 provides the base power, and Q944 adjusts the regulation point for the DC voltage drop - remember in this scheme the Qs are both linear regulators and switches.

Check the R949 voltage, which shows the main current, which should agree fairly closely with the output current shown on the external supply.

Anyway, that should answer some questions. Good luck.

Ed

My reference point was TP950, and I was measuring the voltage at the junction of CR946 and CR947. I got that this junction was at around -5V compared to TP950. This seems to be because the collector voltage of the transistors was going negative.

Here are my measurements:
Input voltage: 39.57V
Q944: C = 10V, B = -0.3V, E = -1V
Q939: C = -0.3V, B = 33.41, E = 33.91
C943 negative: -1.9V
Q938: C = 5.58V, B = 33.28V E = 33.90V
C942: 33.83V across

If I crank the input voltage to 42 volts, I get 1.9V on the emitter of Q944.

I meant that I get 1.9V on the bottom of C943 with 42 volts (and with the potentiometer turned a little).

It looks like the regulation is working, but it still doesn't start up properly. To start it up, I have to start at 35 volts and slowly crank it up to 42 volts. If I crank it up more the power supply stops oscillating.

Have you already checked/replaced C940 and R949? I assumed yes but not sure.
What is the voltage across R949 (0.51 ohms)? Measure on a DMM, not scope, then calculate the current. Does it agree fairly closely with the external supply's current meter (assuming it has one)? If you have a DMM that can measure DC up to 2A or more, you can get a much more accurate reading. Or, you can apply the plus supply output to the top of R907 (0.2 ohm) instead of the +bus (TP940) so you can measure its voltage drop to get the current.

I believe the total running DC power for all the 2236 guts should be maybe 45-50 W. If the regulated voltage on the chopper is say 40 V, then the main current should be somewhere around 1.2 A plus some more to cover losses, say another 20 percent, so around 1.5 A ought to do it. If it's grossly different from that either way, that says something more about what's going on. This has to be measured when the 8.6 V is properly set, if it's possible to keep it running there.

Does the external supply have foldback current limiting? If so, where does it engage (trip)?
What type of capacitor is installed in the C944 spot? It should be a plastic film or ceramic type. Don't go by the parts list - see what's actually in there (in case someone has changed it in the past).

On looking at the power stage again I realize that Q944 has to supply all of the average base current for Qs 946 and 947, since there's no other reverse polarity return in the base circuit. So, Q944 really does have full control of the stage.

BTW, does the fan run when the supply is up? I don't recall any mention of it in the discussion. If not, try disconnecting it. More later.

Ed

C944 was ceramic, but pulling it out revealed that it was leaky and high ESR.

I swapped it out and the power supply seems to work better now. It now runs off mains when the potentiometer is turned all the way up (but the 8.6V rail is at 9.53V instead). I think I might have destroyed some transistors during testing though.

I suspect it may be because the fT of my transistors is too high. The MJE15030 has 10 times the fT of the TIP41C (30MHz vs 3MHz). This could explain why the MJE15030 is able to oscillate at 1MHz while the TIP41C can't.

I got the TIP41C transistors and I swapped them in.

The power supply no longer oscillates at 1MHz, but the -8.6V rail is at -8.2V when I feed in 42V, and that's with the potentiometer turned all the way up.

Would it be a smart idea to bodge a 10k resistor in parallel with R937? Is there a better way of getting the voltage to match -8.6V? The current consumption is 800 mA, which seems a bit low. When I plug in the scope, it still doesn't power up.