Sorry I do go on just me unfortunately. Hope you have plenty of paracetamol ha ha. I could just swap the TO3s over but I will wait for the sil pads to turn up. By way thanks for info on that one.
Working on I V/R = 420mA
Voltage drop 45v input drop to 11.5v so 34.5v drop = 420mA. Think that's correct
Voltage drop 45v input drop to 11.5v so 34.5v drop = 420mA. Think that's correct
Yep something loose in that case on tapping again it dropped straight to 1.6mV. Then shot within say 10 seconds straight back up to 46v now started humming as well so immediate shut down. Thats because I messed around with trimmer yesterday so obviously pushed the current right up. Was around 26v previously.
Anyhow I'll wait for these pads and I'll switch the TO3s around see what happens.
Anyhow I'll wait for these pads and I'll switch the TO3s around see what happens.
There are two issues I can think of with the symptoms:
1) If something's loose inside the N-channel MOSSIE (2SK135 / BUZ900), the more likely symptom would be 'full negative swing', rather than positive.
2) The '10 seconds or so' time delay before it swings, hints at a leakage current charging a Gate capacitance. That would make it a highly unusual failure.
I suggest using your Variac to provide, say, half normal supply voltage, then measure some other spots in the circuit. Make sure to note with each measurement whether the output is where it's supposed to be, or 'latched, full-rail'.
There are at least a dozen places in that circuit, going all the way back to R3 at the input, where a cracked foil path, or cold solder, or oxide plating on a component lead (very common when a unit has been sitting for a decade or three) could cause the nearly-full-rail output -- it is simply falling out of 'closed loop'.
Since everything except TR1 and its friends are balanced, almost any other connection could be at fault. If you find a transistor turned off (less than 0,5 Vbe or so) you can stop blaming the MOSsies - I'll bet they'd be grateful.
Regards
1) If something's loose inside the N-channel MOSSIE (2SK135 / BUZ900), the more likely symptom would be 'full negative swing', rather than positive.
2) The '10 seconds or so' time delay before it swings, hints at a leakage current charging a Gate capacitance. That would make it a highly unusual failure.
I suggest using your Variac to provide, say, half normal supply voltage, then measure some other spots in the circuit. Make sure to note with each measurement whether the output is where it's supposed to be, or 'latched, full-rail'.
There are at least a dozen places in that circuit, going all the way back to R3 at the input, where a cracked foil path, or cold solder, or oxide plating on a component lead (very common when a unit has been sitting for a decade or three) could cause the nearly-full-rail output -- it is simply falling out of 'closed loop'.
Since everything except TR1 and its friends are balanced, almost any other connection could be at fault. If you find a transistor turned off (less than 0,5 Vbe or so) you can stop blaming the MOSsies - I'll bet they'd be grateful.
Regards
Last edited:
OK, I didn't explain that very well ..
Whether the output is being driven, high -- or the sinking drive is dropping out -- are two somewhat different problems to hunt down.
Cheers
Whether the output is being driven, high -- or the sinking drive is dropping out -- are two somewhat different problems to hunt down.
Cheers
Last edited:
Okay Rick ill go through it all again re set the trimmer mid point. Like say started to buzz new problem but that's no doubt due to messing with trimmer.
Interesting re track on board as I did have to make a couple of repairs not professionally just used sticky back copper track tape. Which I had to do round the mosfets. But continuity is good so can't see that causing the problem. If it is then i would new a brand new board. Beyond my capabilities don't have gear to do that.
But ill go along with what you've said. I'm assuming I need to leave those 109r resistors in?
Interesting re track on board as I did have to make a couple of repairs not professionally just used sticky back copper track tape. Which I had to do round the mosfets. But continuity is good so can't see that causing the problem. If it is then i would new a brand new board. Beyond my capabilities don't have gear to do that.
But ill go along with what you've said. I'm assuming I need to leave those 109r resistors in?
Well eye opener look at this.
TR1 0.609 VDC
TR2 0.60 VDC
TR3 - 9.13 VDC
TR4 11.78 VDC
TR5 12.05 VDC
TR6 0.71 VDC
TR7 0.68 VDC
So TR3 TR4 & TR5 well out of spec for sure. So whats causing such high readings?
TR1 0.609 VDC
TR2 0.60 VDC
TR3 - 9.13 VDC
TR4 11.78 VDC
TR5 12.05 VDC
TR6 0.71 VDC
TR7 0.68 VDC
So TR3 TR4 & TR5 well out of spec for sure. So whats causing such high readings?
So TR3 could be that 100uf cap and 22k resistor?
TR4 & TR5 100R resistors plus that R13 which is 6.8k but obviously 12k on schematic & the 4n7 then onto gate resistor 220r in on schematic again 470r plus then trimmer.
Also just to make sure leads are right way round on meter for measuring VDC so NPN Base +ve and Emitter-ve PNP reverse leads round.
TR4 & TR5 100R resistors plus that R13 which is 6.8k but obviously 12k on schematic & the 4n7 then onto gate resistor 220r in on schematic again 470r plus then trimmer.
Also just to make sure leads are right way round on meter for measuring VDC so NPN Base +ve and Emitter-ve PNP reverse leads round.
I think C1 is wrong way round as I'm getting a -ve reading -150 odd mV otherside I'm getting +ve 32mV which is within ball park figure.
I could flip it see what happens could be cap marked up wrong. Think you suggested flipping it way back now. Make sense as it's at beginning of circuit. Which is gonna affect the circuit up line if its feeding negative as opposed to positive.
I could flip it see what happens could be cap marked up wrong. Think you suggested flipping it way back now. Make sense as it's at beginning of circuit. Which is gonna affect the circuit up line if its feeding negative as opposed to positive.
Yes, leave the 100r resistors in do not put the fuses back in yet.
Those figures don't make any sense.
420mA would give 17.64 watts and a voltage drop of 42v
The 100r 2w resistors would burn out
For 100mA (which is more like what you would expect) through 100r you would get a 10v drop and 1watt dissipation.
C1 is the correct way round.
Sid.
IIRC that was C3 that I suggested flipping -- if you worried that it might be a problem. Since C1 has 100k to ground on the other side, it can't be causing the offset (unless something with a DC path to ground is connected at the input).
If everything was 'normal', there would be only a few to couple-10's of mV either side. The only reason to place the polarity one way or the other (C1 and C3 only) is for it to stay 'correct-polarity' when the circuit powers up or down. As long as it's running it doesn't matter. Many designs specify a non-polar part in those locations. Best not to worry about it.
Yah, sorry my awake hours aren't too compatible with yours ..😱
I could start checking in during that hour or so I'm seeing my super-wife off, on her way to work -- should be about your 11:something-AM. Mostly only weekdays, though.
Now, about those voltages ..
If you're sure that you're always measuring Emitter-Base, we got problems.
All the TO-126's (probably TR4, 5, 6, 7) that I've used have the collector in the middle. That may or may not have been the way of the originals. Since it seems to WANT to hold its output near ground, we can probably safely assume they are installed correctly. The one exception would be if the mV readings are simply 'floating' and have little or no 'oomph' behind them.
Let's take a bit different tack ..
This intermittent is enough of a threat to those lovely, hard to replace, output MOSFETs, let's protect them by taking them out of the circuit. No need to dismount them. (Yay!😉)
With the power off:
1) lift the MOSFET end of each Gate stopper resistor (R15, R16), and connect them together
2) connect a 10k resistor to ground from their junction
3) run a jumper from that node to R10 (which closes the feedback loop)
4) fit a 4k7 to 27k resistor from the original output node to each Gate -- 2 resistors (necessary to hold them off)
Now you can ease up the variac and take some measurements. Now if a probe slips it's not a catastrophe.
And don't worry about polarity, just make sure you're measuring Base-Emitter voltages.
And while you're at it, maybe get a measurement across each of the 100R resistors, R6, R11, R12, and R14.
Regards
edit: Kwap - I'm too slow again; gotta go back and study sidsparks's posts .. he probably already covered mine .. 😱
edit2: Now that I see how long this is, I'll understand if you choose a different approach -- won't bother me even a little.
If everything was 'normal', there would be only a few to couple-10's of mV either side. The only reason to place the polarity one way or the other (C1 and C3 only) is for it to stay 'correct-polarity' when the circuit powers up or down. As long as it's running it doesn't matter. Many designs specify a non-polar part in those locations. Best not to worry about it.
Yah, sorry my awake hours aren't too compatible with yours ..😱
I could start checking in during that hour or so I'm seeing my super-wife off, on her way to work -- should be about your 11:something-AM. Mostly only weekdays, though.
Now, about those voltages ..
If you're sure that you're always measuring Emitter-Base, we got problems.
All the TO-126's (probably TR4, 5, 6, 7) that I've used have the collector in the middle. That may or may not have been the way of the originals. Since it seems to WANT to hold its output near ground, we can probably safely assume they are installed correctly. The one exception would be if the mV readings are simply 'floating' and have little or no 'oomph' behind them.
Let's take a bit different tack ..
This intermittent is enough of a threat to those lovely, hard to replace, output MOSFETs, let's protect them by taking them out of the circuit. No need to dismount them. (Yay!😉)
With the power off:
1) lift the MOSFET end of each Gate stopper resistor (R15, R16), and connect them together
2) connect a 10k resistor to ground from their junction
3) run a jumper from that node to R10 (which closes the feedback loop)
4) fit a 4k7 to 27k resistor from the original output node to each Gate -- 2 resistors (necessary to hold them off)
Now you can ease up the variac and take some measurements. Now if a probe slips it's not a catastrophe.
And don't worry about polarity, just make sure you're measuring Base-Emitter voltages.
And while you're at it, maybe get a measurement across each of the 100R resistors, R6, R11, R12, and R14.
Regards
edit: Kwap - I'm too slow again; gotta go back and study sidsparks's posts .. he probably already covered mine .. 😱
edit2: Now that I see how long this is, I'll understand if you choose a different approach -- won't bother me even a little.
Last edited:
Yes 3 legs 2sa looking from back the curved side not flat side 3 base 2 collector 1 emitter
Bf479 1 emitter 2 collector 3 base.
Bf479 1 emitter 2 collector 3 base.
Wrong Bf 470 ,& BF 469 as per BfF 479 put up there. Also note that if measuring pnp or npn you need to swap the +,ve and -ve leads around depending which type your measuring.
Now it's also not easy testing them in circuit either high probability of shorting them considering thete placement on board so I'm having to use non insulated long thing probes so don't wanna be in there for long.
Now it's also not easy testing them in circuit either high probability of shorting them considering thete placement on board so I'm having to use non insulated long thing probes so don't wanna be in there for long.
Unless these are newer models but dated 82 so don't think the lead outs would be changed around!
Those resistors are only 2 watt but can withstand up to 700v. So there pretty tuff little buggars. But unit not left on for hours. I'll take some more readings. The C2 was giving me a -ve reading now its +ve after change which is what it's supposed to be. Another thing now as you rank up voltage C2 changes whereas C2 on okay board just dits at the same value. Why would it do that?