Beckman Industrial 9102 scope problem

[kf7pcl]

So I ended up picking up a Beckman Industrial 9102 20MHz scope for a very low price. It was sold with a known issue of "losing trace after warming up".

I noticed the scope is almost completely unmentioned online. I couldn't find a manual or schematic.

I turned it on and it seems to operate okay except the trace is somewhat dim. I have to put the brightness almost all the way up in order to get a moderate brightness trace. Also, if I increase the brightness pot past about 80%, it actually gets slightly dimmer.

After just a 2-3 of running, a heatsink with three transistors and a 1.2K resistor becomes too hot to touch. It starts smelling like hot components not long after and I measured the resistor exceeds 100C.

The transistors seem to be part of the voltage regulation circuit for the power supply. But I don't have a schematic and there aren't even any labeled voltage test points.

Any ideas?

[floobydust]

Welcome to the forum 

It looks like one of the South Korean Hung Chang/Protek scopes, just like model 5504 but a 20MHz version, not 40MHz. They all use a similar circuit, a bit difficult to repair.

Start with finding the schematic/service manual. I could not find one; EEVblog member Armadillo showed pics of it (5504) here: https://www.eevblog.com/forum/repair/hung-chang-5502-oscilloscope-problem/msg1132684/#msg1132684

Hung Chang 5502 20MHz  Dual trace, Single Timebase
Hung Chang 5504 40MHz  Dual trace, Dual Timebase (Beckman 9102?)
Hung Chang 5510 100MHz Dual trace, Dual Timebase

Hung Chang 5602 20MHz  Dual trace, Dual Timebase, On-screen readout, aka Beckman 9202
Hung Chang 5604 40MHz  Dual trace, Dual Timebase, On-screen readout

https://www.eevblog.com/forum/repair/hung-chang-5510-oscilliscope-repair/

Mods maybe move this thread to Repair?

[kf7pcl]

Thanks! I wondered if it was a rebrand. It is indeed a Hung Chang.

I couldn't find the correct manual but I did find one for the 5602 which is certainly better than nothing

[floobydust]

I would guess something is overloading the 140VDC power supply, which is then overheating.
It's usually the deflection amplifiers (transistors) or some other component like a capacitor. Don't run it too much and keep a careful eye what happens, the beam may not be going  out, but rather off screen. If it's over 140V then stop now.
The Korean scopes, most parts are run at their max. rating, so any failures can cause a bunch of failures in other parts as a domino effect.

The 140VDC regulator is extra confusing because the pass-transistor is on the -ve (GND) side and an assist resistor is used to get more current out. Both may be your hot parts.
5602 schematic

[kf7pcl]

I would guess something is overloading the 140VDC power supply, which is then overheating.
It's usually the deflection amplifiers (transistors) or some other component like a capacitor. Don't run it too much and keep a careful eye what happens, the beam may not be going  out, but rather off screen. If it's over 140V then stop now.
The Korean scopes, most parts are run at their max. rating, so any failures can cause a bunch of failures in other parts as a domino effect.

The 140VDC regulator is extra confusing because the pass-transistor is on the -ve (GND) side and an assist resistor is used to get more current out. Both may be your hot parts.
5602 schematic

Yes, those are indeed the hot parts.

I tested the regulated rails:

+12V: 11.99V
-12V: -12.00V
+5V: 5.15V
+140V: 158.3V

Obviously the 140V rail is way higher than it should be. What should the next action be? Check the pass transistor?

[james_s]

Check the pass transistor and whatever drives it, does it use a regulator IC or just a zener reference? 158V is about what I'd expect if it's just running balls out and not regulating at all.

[floobydust]

This repair thread is for different scope but same (Korean scope) circuit prevails: https://www.eevblog.com/forum/repair/meguro-mo-1252a-25mhz-analog-oscilloscope-nothing-on-crt/msg1531487/#msg1531487

The pass-transistor may be shorted, I'd pull it and test. 2SB861 PNP 200V 2A 30W TO-220 BCE pinout
For a sub, I would consider tougher MJE15033 PNP 250V 8A 50W TO-220.

[kf7pcl]

Check the pass transistor and whatever drives it, does it use a regulator IC or just a zener reference? 158V is about what I'd expect if it's just running balls out and not regulating at all.

The input to the regulator is actually about 200VDC (peak voltage from the 140V tap on the transformer)

It's a voltage divider referenced to -12V driving an opamp.

This repair thread is for different scope but same (Korean scope) circuit prevails: https://www.eevblog.com/forum/repair/meguro-mo-1252a-25mhz-analog-oscilloscope-nothing-on-crt/msg1531487/#msg1531487

I'm beginning to think I'm chasing my tail. All the component values are identical to that scope which has a +160V spec. If the rail should actually be +160, then 158 is just fine. I also noticed, one of +140V markings on the schematic has (160V) in parenthesis.

Since the schematic covers two models, I am wondering if one has a 140V spec and one has 160V. The difference between 9.1K and 10K on R1019 might set that. Mine is the 9.1K

So far I have checked all the resistors and the diode and they check out. I measured Q1003 in circuit and it isn't shorted but I can't test further than that in circuit. The voltage differential at the input of the opamp is only ~1mV. Wouldn't that indicate the regulator is at it's target voltage?

[floobydust]

I see the +140V rail has different resistors for the 20MHz and 40MHz versions, math says 144V or 158.2V so your 158V seems fine. Another sch page says "+140V/+160V". You must have a 40MHz build.

That leaves the HV and CRT bias to look at for the brightness problem.

Which CRT does it have?
CRT Toshiba 150BTB31A CRT is 20MHz, operating -2kV, cutoff is -25 to -65V.
CRT Toshiba 150CTB31  CRT, operating 10kV so I'm guessing +8kv and -2kV.

The HV might be sagging and DC-DC transistor making the heat then, Q2001 2SD880, assuming it's running full tilt.
The CRT cathode and grid are at -2kV so a HV probe would be needed or a string of resistors to make one, to measure CRT voltages.

Or look at the HV regulation circuit, the feedback voltage at IC2001 op-amp output will tell you if the HV regulator is full out.

My guess would be a bad HV multiplier ceramic cap or diode.

I don't know anything about the HV multiplier module M1001 (in 40MHz scope) to (I think) make +8kV.

There's other threads here with the same problem on these Korean scopes- trace fades, dim, goes out but I did not see the fix yet.

[kf7pcl]

Which CRT does it have?
CRT Toshiba 150BTB31A CRT is 20MHz, operating -2kV, cutoff is -25 to -65V.
CRT Toshiba 150CTB31  CRT, operating 10kV so I'm guessing +8kv and -2kV.

It's the 150BTB31A. A little confusing since the other parts of the circuit seems to be the 40 MHz version.

The HV might be sagging and DC-DC transistor making the heat then, Q2001 2SD880, assuming it's running full tilt.
The CRT cathode and grid are at -2kV so a HV probe would be needed or a string of resistors to make one, to measure CRT voltages.

Unfortunately I do not have a HV probe.

Or look at the HV regulation circuit, the feedback voltage at IC2001 op-amp output will tell you if the HV regulator is full out.

The output pin of IC2001 is +3.77V w.r.t. GND

I measured the voltage across R1014 and got 42V. 42²/1200 ohms gives 1.47W which is within its two watt rating. But it still concerns me that it gets to 110C+ within just a couple of minutes running.

I measured the current usage from the +140V tap on the transformer and got ~63mA. The transformer is rated for 80mA so I guess that isn't too unreasonable?

[floobydust]

The scope seems to be a mix of 20MHz and 40MHz parts. I don't think the 20MHz CRT even has a 2nd anode, unlike the 40MHz CRT.

I can only guess at the sections to look at with the tools you have. It's a bit tough to find the section that is misbehaving.

IC2001 looked like a good voltage controlling the HV DC-DC converter.
The 140V/160V rail seems a bit high for current, it powers the H and V deflection and beam intensity control circuits. Assuming it stays up at 158V, then the trace fading problem is somewhere else.

Another place to look is the beam current circuit.
It runs off the +140V (+160V) rails and could draw excessive current from that rail. Q3001-Q3006. That section is to modulate the -2kV (using only 300V transistors) with CRT grid bias. You could take some measurements after the intensity control VR3002 on the transistors and look for drifting voltages as the trace dims. Note the Z-axis input is a duplicate circuit.

For measuring the scope's HV power supply, you can make a poor man's HV probe with a string of 10MEG through-hole resistors inside a straw or pen. I take 1/4W parts past 500V without problems, so use 4-5 in series. With power off, I put it in place and power on to take a reading on a DMM. If the (DMM) ground/alligator clip lifts doing this, there will be drama. Knowing the DMM input resistance, you can calculate the divider ratio and work backward to the HV value. This would tell us if the -2kV supply is steady.

If the focus also goes way fuzzy with a dimming trace, that I think points to the HV fading. If instead it was the beam current (control), the trace could get dim but remain largely in focus.

[kf7pcl]

The scope seems to be a mix of 20MHz and 40MHz parts. I don't think the 20MHz CRT even has a 2nd anode, unlike the 40MHz CRT.

It doesn't

The 140V/160V rail seems a bit high for current, it powers the H and V deflection and beam intensity control circuits. Assuming it stays up at 158V, then the trace fading problem is somewhere else.

I try not to run it too long but the 160V rail sags slightly over time (down about 1V after 10 minutes) but I don't think that's enough to cause a problem.

For measuring the scope's HV power supply, you can make a poor man's HV probe with a string of 10MEG through-hole resistors inside a straw or pen. I take 1/4W parts past 500V without problems, so use 4-5 in series. With power off, I put it in place and power on to take a reading on a DMM. If the (DMM) ground/alligator clip lifts doing this, there will be drama. Knowing the DMM input resistance, you can calculate the divider ratio and work backward to the HV value. This would tell us if the -2kV supply is steady.

I managed to find some 10M resistors and strung 6 in series. The actual sum came out to be 59.6M. The meter input impedance is 10M
I measured -292V which if my math is correct means that it's -2.03kV. That seems fine to me.

If the focus also goes way fuzzy with a dimming trace, that I think points to the HV fading. If instead it was the beam current (control), the trace could get dim but remain largely in focus.

I havn't reproduced the problem with the disappearing trace lately. Maybe because I'm not running it long enough though.

I found another hot part: Q812 which seems to be a horizontal CRT driver is getting up to ~90C. It's neighboring Q809, Q810 and Q810 are in the 60s/70s

I attached a picture which shows the trace brightness. This is as bright as it gets

[floobydust]

It sounds like all the power supply voltages are good. Kind of need another clue.

What you need is a faded dimmed trace, in X-Y mode measure some voltages to compare to when the trace has good brightness.

If it's temperature related, hot air or freeze-spray (not on the HV section OK, frost conducts!) might be worth trying.


For the horizontal deflection transistor running hot, with the (H) POS ADJ control centered, I would make sure VR807 is reasonable and VR806 if you like to adjust for x10MAG centering. In X-Y mode you can measure the X and Y plates voltages. Each +,- amp is symmetrical. 150BTB31A X-axis 19-21.6 VDC/cm, Y-axis 10.8-13.2 VDC/cm and a 14cm screen.

I think if a V or H deflection is not balanced but the front panel control is cranked, it might cause the (defl amp) bias current to be upset. Or if trying to cover for a lazy deflection transistor or DC offset in the sweep oscillator with the front panel control.

I thought if the deflection plates (both in the pair) had low voltage, the acceleration would be a bit less, but the astigmatism would go off too.

Maybe look at C2025 (CRT pin 12) I think that's the 1st acceleration anode. Cap should be over 100V rating

Otherwise, it might be the beam blanking is drifting and upsetting the intensity (grid bias), and not easy to troubleshoot. Or try inject a signal on the Z-axis and see how it does.

The CRT bias tells a lot, but measuring it is too difficult as -2kV common-mode exists and a DMM would be dangerous.
If instead you measure the (HV) cathode (CRT pin2) and then control grid (CRT pin3),  the difference is the CRT bias, beam cutoff is typically -25 to -65V for the 150BT131.
It might tell, with a maxxed out intensity control if the CRT is biased weak.

[kf7pcl]

For the horizontal deflection transistor running hot, with the (H) POS ADJ control centered, I would make sure VR807 is reasonable and VR806 if you like to adjust for x10MAG centering. In X-Y mode you can measure the X and Y plates voltages. Each +,- amp is symmetrical. 150BTB31A X-axis 19-21.6 VDC/cm, Y-axis 10.8-13.2 VDC/cm and a 14cm screen.

The front panel horizontal is almost centered to bring the spot to center. Adjusting VR807 has absolutely no effect on spot position.

Q912 exceeded 100C according to my probe! When I put the scope in XY mode, the transistor cools down by about 20C

I think if a V or H deflection is not balanced but the front panel control is cranked, it might cause the (defl amp) bias current to be upset. Or if trying to cover for a lazy deflection transistor or DC offset in the sweep oscillator with the front panel control.

The front panel control is almost centered.

I thought if the deflection plates (both in the pair) had low voltage, the acceleration would be a bit less, but the astigmatism would go off too.

TBH, I'm not quite sure what this means.

Maybe look at C2025 (CRT pin 12) I think that's the 1st acceleration anode. Cap should be over 100V rating

It has a 50V rating

Otherwise, it might be the beam blanking is drifting and upsetting the intensity (grid bias), and not easy to troubleshoot. Or try inject a signal on the Z-axis and see how it does.

What is the normal amplitude range for Z then?

The CRT bias tells a lot, but measuring it is too difficult as -2kV common-mode exists and a DMM would be dangerous.
If instead you measure the (HV) cathode (CRT pin2) and then control grid (CRT pin3),  the difference is the CRT bias, beam cutoff is typically -25 to -65V for the 150BT131.
It might tell, with a maxxed out intensity control if the CRT is biased weak.

I measured the voltage on pin 2 and 3 of the CRT. But every time I put the probe on the grid, it causes a Heisenberg issue: The CRT becomes unusually bright and the intensity control has no effect at all.

I got 292.9 on the cathode and 289.2 on the grid. That's a difference of 3.7V or 26V when you consider the divider.

[Allen49]

I ran across a website at userrmanual.com that has 100,000 manuals and will sell you a manual for the Beckman 9102 for about $10.00.