IBM 6734-ABO Not working.

[PlainBill]

Re: IBM 6734-ABO Not working.

Ok. I verified my little test setup using a known good monitor. I removed one of the lamp leads from a known good monitor and plugged in my test setup. The lamp came on. So i know my testsetup works.

Now when i go to the "broken" monitor i get "no light flashing" at all from any of the 4.

OK, let's try this. One of two things are happening. Either the Over-Voltage Protection (OVP) or the Open Lamp Protection (OLP) circuits are kicking in. The OLP circuit on the BIT3105 shuts down the inverter if an open lamp is detected. The OVP appears to work by reducing the output voltage. I'm assuming the actual problem is a bad component in the OVP sense circuit. so lets see if we can isolate the problem.

I've marked V1 through V4 on the picture. These represent the 4 voltage sense lines. If you trace V4 back, you see it goes to C242 and D242. Short across C242 and test. Similarly V3 goes to D241 and C241. Leave the short across C242, short across C241 and retest. At some point hopefully the CCFLs will stay on.

PlainBill

[Welchs101]

Re: IBM 6734-ABO Not working.

I think i understand what you want me to do.

But i wanted to ask a few questions to help me understand why we are doing this....learn a little.

Note: I measured the resistance of the capacitors in question and they all "seem" to be ok....all measuring around 1.4MegOhm.

If i short the caps how will this some how turn the monitor on?

Also, why do you want me to short C242 and C241 if shorting C242 does not leave CCFLs on?

[PlainBill]

Re: IBM 6734-ABO Not working.

I think i understand what you want me to do.

But i wanted to ask a few questions to help me understand why we are doing this....learn a little.

Note: I measured the resistance of the capacitors in question and they all "seem" to be ok....all measuring around 1.4MegOhm.

If i short the caps how will this some how turn the monitor on?

Also, why do you want me to short C242 and C241 if shorting C242 does not leave CCFLs on?

The basic design of an inverter is a constant current AC source. A CCFLs start voltage is much higher (roughly 1400 volts) than it's operating voltage (maybe around 700 volts). So when the inverter starts the output voltage will rise until the CCFLs 'ignite'. If a tube doesn't ignite this would result in an transformer or wiring arcing. The BIT3105 inverter controller has a feature called Limited Ignition Voltage (which I call Over Voltage Protection). In essence, the controller will limit the voltage to the CCFLs to a 'safe' voltage, let's say 1600 volts for CCFLS that should ignite at 1400 volts.

The way the do this is they use a voltage divider to provide a small fraction of the output voltage to a diode, which converts the AC to DC. That DC is fed to the 'clamp' input of the BIT3105. If you look at the 'Application Information' schematic in the datasheet for the BIT3105 you will see the components in the upper right hand corner. Yours is slightly more complicated because you have 4 outputs, but the theory is the same.

Usually we think of a voltage divider as two resistors in series. For example, if we have a 990 Kohm resistor in series with a 10 Kohm resistor, and we applied 100 volts across the pair of resistors, the voltage across the 10 K resistor would be 1 volt. This design uses a pair of capacitors, because a capacitor that can tolerate 2 KV is a lot cheaper than a resistor that can drop 2 KV.

C308 and C241 form one voltage divider. C309 and C242 form another. There are two more on the other end of the board. NOW, there are several possible reasons the OVP is kicking in. We can ignore the possibility the output voltage is getting too high. One is that one of the capacitors in the voltage divider has drifted in value or is leaky. That means the voltage across the SMT cap would be too high. The vast majority of the voltage is dropped across C308, C309, etc. Very little voltage is dropped across C241, C242, etc. This means that if you short across C242 the voltage dropped across C309 goes from 1398 volts to 1400 volts - not enough to damage a 2KV capacitor. But by shorting across C242, that output is taken out of the OVP circuit. In other words, a bad C309, C242, or D242 will not have any effect on the inverter controller.

Now, I want you to progressively short out the capacitors for a reason. There are three possible scenarios:
1. A single 2KV cap is leaky. Shorting out only one cap at a time will catch that problem.

2. More than 1 cap is leaky. Shorting out one at a time will not catch that.

3. All caps are good, something closer to the BIT3105 is the actual fault. If that is the case, when you have shorted all caps, it STILL won't work. But we will be able to work toward the BIT3105 until we identify the culprit.

Additional points.

The voltage divider pairs are C306 and C231; C307 and C232; C308 and C241; C309 and C242. In each case, short across the tiny SMT cap.

All CCFLS must be hooked up while doing this. You are disabling an important part of the protection scheme.

Use very short leads - a short length of resistor lead bent into a U shape would be best.

DON'T run the monitor with shorts in place for a long time.

If the monitor starts working after you have shorted the third cap, remove the first and second short (one at a time) to see if it still works.

A bad solder joint on the SMD caps could also cause this problem. So I won't feel bad if you decide to resolder them first.

PlainBill

[Welchs101]

Re: IBM 6734-ABO Not working.

Thanks Bill. I will probably have to read that a few times to digest all the information but i think i understand the main points.

Ok. So just to be clear on the order of shorting the caps.

Should i short

C306 and C231

test

then short out C307 and C232;

test

then short out C308 and C241

test

then short out C309 and C242.

Note: I dont remove any shorts......as i proceed.....right?

[PlainBill]

Re: IBM 6734-ABO Not working.

Thanks Bill. I will probably have to read that a few times to digest all the information but i think i understand the main points.

Ok. So just to be clear on the order of shorting the caps.

Should i short

C306 and C231

test

then short out C307 and C232;

test

then short out C308 and C241

test

then short out C309 and C242.

Note: I dont remove any shorts......as i proceed.....right?

AGHHH!!!!! NO!!

Short C231, then test.

If the lamps flash, then go off, short C232, then test.

If the lamps still flash, then go off, short C241, then test.

If the lams still flash, then go off, short C242, then test.

If the lamps still flash, then go off, report and I will come up with a new test. At that point we know the problem is not any of the lamps or the capacitors.

If the lamps stay on after you short a capacitor - for example C241, leave that short in place and remove the short on C231, and retest. If the lamps still stay on, remove the short on C232, and retest. If they still stay on, we know that the problem is on the C308, C241, etc path, and we can concentrate on that path.

PlainBill

[Welchs101]

Re: IBM 6734-ABO Not working.

will do.

[Welchs101]

Re: IBM 6734-ABO Not working.

Update:

I had some interesting results this morning. I shorted C242 first since i already had this one shorted. Powered up monitor and it did not come on. So i then shorted C231 (with C242 still shorted) and the monitor still did not come on. I had trouble soldering one of the leads of C231 so i resoldered the short across C231 and monitor CAME ON! I thought wow......so i removed short across C242 and monitor did not come on. I put the short across C242 back and monitor did not come on. THis was strange.

So now i decided to "start" over. I removed all the shorts and i re-flowed several solder connections that looked like they could use some additional solder. I even added some solder to the "ground" locations where the screws are located because they looked like they really needed some. Plugged everything back in with "NO caps shorted" and the monitor "CAME ON". I thought.....huh....perhaps it was just something i reflowed. I then removed the board and added some additional solder to the ground locations like i did before.......powered monitor back on and .......NOTHING. Monitor would not come on.

Now, i am thinking i perhaps should still try shorting caps like you said but something strange seems to be happening. I am not sure what....any ideas?

[Welchs101]

Re: IBM 6734-ABO Not working.

Update: Some good news.

So after the last update i was really kinda lost as to what to do. At the time i thought for sure i had a "loose" connection somewhere. So i re-soldered a lot of nodes that i thought needed it. But it did not help.

At this point i was totally confused as to what to do. So decided to go back to what Bill said and start over.

I shorted C231. Powered up monitor and nothing happened.

I shorted C232 and C231. Powered up monitor and ........IT WORKED!

I removed short C231. Shorted C232. Powered up monitor.......IT WORKED!

I removed short C232 (so no shorts present). Powered up monitor........nothing happened.

I shorted C232 only. Powered up monitor..........IT WORKED!

So i guess there are two things to be said here.
1) When all else fails and you dont know what to do.........go back to where you started.
2) Ok....so now with C232 shorted LCD comes on........what should i do now.

[PlainBill]

Re: IBM 6734-ABO Not working.

Update: Some good news.

So after the last update i was really kinda lost as to what to do. At the time i thought for sure i had a "loose" connection somewhere. So i re-soldered a lot of nodes that i thought needed it. But it did not help.

At this point i was totally confused as to what to do. So decided to go back to what Bill said and start over.

I shorted C231. Powered up monitor and nothing happened.

I shorted C232 and C231. Powered up monitor and ........IT WORKED!

I removed short C231. Shorted C232. Powered up monitor.......IT WORKED!

I removed short C232 (so no shorts present). Powered up monitor........nothing happened.

I shorted C232 only. Powered up monitor..........IT WORKED!

So i guess there are two things to be said here.
1) When all else fails and you dont know what to do.........go back to where you started.
2) Ok....so now with C232 shorted LCD comes on........what should i do now.

IIRC, there are three possible components (I'm not really in shape to start looking at the pictures), the diode next to C232, C232, and C307. C307 is the most likely cause. An easy test Would be to swap C307 with C306, then test. If the lamps STILL come on, C307 is good.

If the lamps don't come on, short C231. If they now come on, remove the short from C232 (leaving C231 shorted). If they still come on, you have proved the problem is C307. Depending on cost, it might be a good idea to replace all of them.

PlainBill

[Welchs101]

Re: IBM 6734-ABO Not working.

thanks bill. I will take a look at them tomorrow and let you know.

Hey what does IIRC stand for?

[retiredcaps]

Re: IBM 6734-ABO Not working.

If I recall correctly.

Hey what does IIRC stand for?

[Welchs101]

Re: IBM 6734-ABO Not working.

Update:

Recall that i shorted C232 and monitor powers as it should.....ie....bulbs power up and i can see image on screen.

I swapped C306 for C307 to see if C307 was the problem. I kept C232 shorted. If C307 was the problem then monitor should not turn on. Turns out monitor did turn on after i swapped C306 and C307.

So i guess this leaves either the diode (D231) or surface mount cap (C232).I have enclosed a pic showing D231 and C232. Note: C232 is shorted in this picture.

Not really good at soldering surface mount caps such as this or such diodes. Any suggestions?

My next step is to swap caps.....and if this does not work swap diodes......is this right?

[Welchs101]

Re: IBM 6734-ABO Not working.

Update:

I removed the short across C232. I then measured the diode and capacitor using a DVM. I compared the measurement results with those of the diode and capacitor for C231, C241, and C242. I determined that the capacitor itself C232 "was different". I removed C232 and C231. When i removed C232 it "fell apart". The metal on one end of the cap just came off. I dont think i heated it up too much when removing..........i think it was just bad that way. I also removed C231 and i did this so i could measure C231 and know what cap value i would need to replace C232.

C231 measured ~ 35nF.

I could not find an exact replacement among my various spare parts but i did find a 45nF cap.........do you think this is ok to try as a replacement.

[Welchs101]

Re: IBM 6734-ABO Not working.

Update:

Replaced C232 with a ~41nF. IT WORKS!!!!!!!!!!!!!!!!!!

Bill, you fixed another one.

I did have a few questions (if you have the time):
1) How did you know the V1, V2, V3 and V4 you showed in the pic were the "sense" lines?

2) How did you know the problem was in this area? I am continued to be amazed that i provide limited info and "bam" your able to hone in on the problem area.

[PlainBill]

Re: IBM 6734-ABO Not working.

Update:

Replaced C232 with a ~41nF. IT WORKS!!!!!!!!!!!!!!!!!!

Bill, you fixed another one.

I did have a few questions (if you have the time):
1) How did you know the V1, V2, V3 and V4 you showed in the pic were the "sense" lines?

2) How did you know the problem was in this area? I am continued to be amazed that i provide limited info and "bam" your able to hone in on the problem area.

First, I am glad you have it working. I would STRONGLY recommend replacing the cap with one of the correct value before turning the monitor out into the wild.
Nex, as far as your questions, here is the process involved. By looking at the parts and the datasheet for the BIT3105 I realized this inverter employed IVL (Ignition voltage limiting) and OLP (open lamp protection). If you remember the (cough, cough) earlier test, we established that the IVL was shutting the inverter down. (I missed a clue, more on that later.) By looking at the high voltage caps (C306 - C309) It was obvious that they were half of capacitive voltage divider circuits. Also, from inverter schematics it was obvious that this was a 'Shut down if any one is bad' design.

Now the next step was to disable the IVL for each output. The first idea was to simply unsolder each high voltage cap, but they might also be involved in 'tuning' the outputs. Then I realized that by shorting the SMD cap I could accomplish the same thing. Also, if you remember the description of a capacitive voltage divider, the failure would require the blue cap to increase in capacitance (or become leaky) or the SMD cap to decrease in capacitance (or go open).

I missed an important clue. The CCFLs were flashing dimly. OLP is designed to catch a lamp that is drawing too much current. It would probably require a lamp to conduct fully. One test I didn't think about - if you had measured the output voltage of the transformers you would have found they were still active, just at a voltage too low to keep the lamps operating.

Remember, troubleshooting 'Two seconds to black' problems is a work in progress. In general, I have you do a bunch of tests until someone finds the problem; then I act like I knew what I was doing all along.

PlainBill

[Welchs101]

Re: IBM 6734-ABO Not working.

thanks Bill,


I re-visited the BIT3105 schematic to see if i could see the same thing that you mentioned........i did not. I guess i still have much to learn...but thats what makes this a fun hobby.

Quick question: The cap i used as a replacement actually measured 41nF. The cap that is "good" from the original circuit measured ~36nF. Thats only a 5nF difference............do you still think i need to get a new cap?

[PlainBill]

Re: IBM 6734-ABO Not working.

thanks Bill,


I re-visited the BIT3105 schematic to see if i could see the same thing that you mentioned........i did not. I guess i still have much to learn...but thats what makes this a fun hobby.

Quick question: The cap i used as a replacement actually measured 41nF. The cap that is "good" from the original circuit measured ~36nF. Thats only a 5nF difference............do you still think i need to get a new cap?

Need? No. But I strongly recommend it. At a minimum, compare the in-circuit capacitance of the three original caps C231, C241, and C242 to the in-circuit capacitance of the replacement. If they are within 10% I wouldn't bother changing the replacement.

You have to take into consideration my attitude on safety features. Some people take the attitude that fuses are there to annoy them. If a 15 amp fuse blows, they will replace it with a 20. Or a 30 if that's handy. I'll replace it with a 15 and see if I can move something off the circuit. If I can't, I'll run a new circuit.

The other story I like to mention is Eastern Airlines. They had a bad relationship with their mechanics union. As far as I know, this was something both sides were responsible for. One of management's techniques was to cut costs as much as possible, to the point of forcing the mechanics to provide as much of their own equipment as possible. A lot of routine maintenance of the airplanes was done at night.

"Lights? Let them use flashlights - Make them buy the flashlights!" "Buy a flashlight? Heck, no!! I'll use the landing lights. Or work in the dark."

On May 5th, 1983 Easter flight 855 lost oil pressure on all three engines, and with a great deal of luck, managed to limp back to Miami International. It flew the last 13 minutes with NO engines operating. The cause of the failures? Three O-rings, not installed because the mechanic was working in the dark. The consequence - bankruptcy for Eastern, loss of jobs for the mechanics.

Now, I KNOW the chances of damage if the output voltage goes too high are small. But the caps cost about a nickle each from Digikey, so I SUGGEST ordering 10 (minimum quantity) on your next order and holding on to the monitor until they arrive.

PlainBill

[Welchs101]

Re: IBM 6734-ABO Not working.

Bill,

I understand your argument and i agree.

Can you help me with something. The capacitance i measured was around 35 to 36nF. But i have no ideas as to the voltage rating. Any suggestions. Also, i need to measure the dimensions of the cap but again i am not sure of the voltage requirements.

Thanks again for all your help. Hope I can learn even more.

Also, how did you know that "It was obvious that they were half of capacitive voltage divider circuits. Also, from inverter schematics it was obvious that this was a 'Shut down if any one is bad' design."

I know your busy so if you dont have time ot answer i fully understand.

Thanks again!!!!!!!

[Welchs101]

Re: IBM 6734-ABO Not working.

One more thing: What type of capacitor is this surface mount........are they film, ceramic, aluminum? I am just not sure.

[PlainBill]

Re: IBM 6734-ABO Not working.

As far as the caps are concerned, we can infer the voltage from the design. The cap is tied to the anode of the diode, the cathode goes to (probably) a second diode, then to the clamp input of the BIT3105. The BIT3105 will start limiting the output voltage when the clamp input rises above .2 volts. So .2 + .7 + .7 means that the cap must be able to withstand over 2 volts. In other words, voltage isn't a factor. I'd say anything with a voltage rating of over 10 volts is more than adequate. I suspect you might find the lowest voltage rating available is 50 volts.

Measure the package using a small metric ruler, or calipers if you have them. Lay the ruler next to the cap and take a picture if that will help. That should let you identify the package.

You measured 35 nF. I suspect the actual value is 33 nF, a standard value. Given the package type, the height, the value, and an estimate of the voltage, you should be able to find several that meet those specs. Also, if you DON'T find any values available other than 22nF, 33nF, and 47nF (standard values I believe) you get even more confident of the value.

I've already mentioned the application note for the 3105. Here are a couple more schematics we will examine in detail.

First, we'll look at the FAN7413. On the right side look at C10, C14, D4, C11, and C30. Below that, the identical circuit is C12, C29, D3, C13, C15. Note that the cathodes of D3 and D4 are tied together, and to R14 and to the OLR input of IC1. (OLR is Open Lamp Regulation, Fairchild's equivalent of BIT's IVL - Ignition Voltage Limit).

What does this do? From the datasheet, OLR kicks in whenever the OLR input goes over 2.0 volts. Notice that there are no capacitors between the cathodes of D3, D4 and IC1. That means the circuit reacts to each output on an individual basis. In very simple terms, it is an analog OR circuit.

Note, there is one difference, R11, R12, R13, R26 do not have an equivalent on your board.

Let's take a look at the second inverter schematic. This uses an OZ964 controller. On this one, C8, C13 form a capacitive voltage divider. CR2 limits the negative peaks, while passing the positive peaks to R14, C16, and the OVP input of the OZ964. Like the FAN7413, the OZ964 begins limiting output voltage when the OVP input exceeds 2 volts. Unlike the BIT3105 and the FAN7413, this design employs C16, which means the OVP will react more slowly.

PlainBill