Maker Pro
Maker Pro

Newby tired of burning component

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
Burning is learning

Are you saying that you tried to measure current across a voltage source? It's kinda amazing your meter survived.

Anyway, the transistor tried to do the right thing by sacrificing itself to try to protect the fuse. (That's simply an amusing way of saying that semiconductors can die a lot faster than fuses can blow.)

If you want to limit the current then you're going to need something to do that, it won't just happen.

A simple way involves a low value resistor in series with your load that turns on a transistor when the voltage across it hits about 0,6 volts. As the transistor turns on, it shunts current away from the base of the pass transistor leading to a lower voltage output (and thus hopefully a lower current requirement)

When you post the circuit I'll try to show you how to add current limiting to it.

Note that current limiting is no panacea. You may have to cope with the entire power being dissipated in your pass transistor (for a dead short). And that's likely to make it get very hot, very quickly.

Hi,

Yep that's exactly what I've tried ;)
The meter survived but a fuse, a potentiometer and a 2N3055 burned (F1, R1 and T1 in the first circuit included).
It might sound stupid but I'm learning a lot this way and I've tried not to blow the same component over and over (which is insanity) but a different one each time :D

Ok seriuously, I've rebuild the thing and re tested both circuit. The first one is behaving as expected, supplying 0-12V and 0-50V (with a running fan cooling heatsinks for B1, T1, IC1 and IC2).
When supplied by the first circuit, the second circuit produce a nice square wave pulse ( in the scope I just bought) when connected to the Emiter of T1 (BU508D), and I'm able to control pulse frequency with R1.

The problem arise when I connect T1 Emiter to Pin 1 of the flyback (DCF 2077A) and Pin 2 to the 53V supply from circuit 1.

With nothing connected between Pin 3 (HV 30KV) and Pin 4 (0V) of the flyback (no meter or anything :p ) it shorted the circuit ( I've only partially burned R1 of the supply circuit this time)!
I understand that I've put 53V to ground in a sense but isn't it the way to feed the flyback?

Anyone sees a major flaw in the circuits design? If not I'll simply re check everything again to make sure I didn't simply connected something where it doesn't belong!

I might have misidentified the flyback pin so I've included the pin diagram. Pin 1 - 2 are the same in the circuit and the diagram.
Pin 3 is a big rubber plug so it is easy to identified.
Pin 4 in the circuit should be pin 6 on the diagram ???

Thank in advance for any advice!!!

PS I need current to obtain 30KV in the flyback so I dont want to limit it!
 

Attachments

  • Power supplys with fan.png
    Power supplys with fan.png
    20.1 KB · Views: 472
  • H Voltage output 30KV.png
    H Voltage output 30KV.png
    15.5 KB · Views: 313
  • flyback.png
    flyback.png
    67.3 KB · Views: 3,087
Last edited:

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
Ok, I've re check everything and i've put a 1k resistance and a LED in place of Pin 1 and 2 of the flyback to make sure I had no short elesewhere... it seems to work ok although I can't say if it is pulsing or not!

I'll retry the flyback has it is if nobody has anymore comment soon!

Thanks...
 

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
It seems you have no decoupling capacitors on the 0-60V?
Next, what is the pulse-width you're feeding to the BU508?
There will be a pulse current each time the transistor conducts, growing larger the longer it conducts, so the pulse length needs to be precisely timed and very short.
The power supply can't deliver anywhere near the current level of this pulse so a big decoupling cap has to be present to deliver this current.
The length of the pulse determines the peak HV and the repetition time determines the power [W] that can be had.
It's sufficient with 1-2Hz repetition rate to build up and maintain 30kV (no-load) and this only needs to draw 1W from the power supply.
I don't think a single 555 is the way to go.
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
It seems you have no decoupling capacitors on the 0-60V?
Next, what is the pulse-width you're feeding to the BU508?
There will be a pulse current each time the transistor conducts, growing larger the longer it conducts, so the pulse length needs to be precisely timed and very short.
The power supply can't deliver anywhere near the current level of this pulse so a big decoupling cap has to be present to deliver this current.
The length of the pulse determines the peak HV and the repetition time determines the power [W] that can be had.
It's sufficient with 1-2Hz repetition rate to build up and maintain 30kV (no-load) and this only needs to draw 1W from the power supply.
I don't think a single 555 is the way to go.

Hi and thanks a lot for your help Resqueline.

Do you mean that C1 ( 9400 uF 50V) in circuit 1 (Power supply) is not enough or that I need an other one somewhere? I'm able to live with only 50V instead of 60V if that's what you mean.

Next I don't know the pulse width but I will try to mesure it if I must. My undestanding is that it is controled by potentiometer R1 in circuit 2 (HV output 30KV). At least I can see the pulse at the BU508 emiter pin and it varies when R1 change.

What would you suggest instead of a 555 timer to trigger a pulse in the BU508?, Again it seems to pulse at the emiter end of it!

Thanks in advance!
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
I've measured the pulse width at the emiter pin of the BU508. It range from 5 to 20 useconds, i.e. between 200kHz and 50 kHz if i'm not wrong!
 
Last edited:

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
C1 needs to be there, but there also needs to be one at the emitter of the 2N3055. It has a limited base drive & gain and can't deal with pulse currents.
It could even have been destroyed due to inductive kickback voltages reaching it from the flyback transformer. A cap also stops that.
The capacitors used must be rated for the voltage applied to them.
Yes, pin 4 in the diagram means/connects to pin 6 on your transformer.
And yes, the frequency is not important. Electrosurgery for example usually uses around 500kHz but can also use 27MHz.
You refer to the emitter of the BU508, but it's the collector that pulses and is connected to the transformer and the emitter is grounded, right?
If you see pulses at the (grounded) emitter then you grounding layout is not good.
The pulse width is perhaps ok, and I guess the HV can be adjusted by way of R1 since I guess it affects the pulse/pause relationship. What are the figures at min/max?
I would have liked to have one astable to set the operating frequency and one monostable to set the pulse time. That would give a better control of the power drain.
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
Errors drives the mind to find solutions!

C1 needs to be there, but there also needs to be one at the emitter of the 2N3055. It has a limited base drive & gain and can't deal with pulse currents.
It could even have been destroyed due to inductive kickback voltages reaching it from the flyback transformer. A cap also stops that.
The capacitors used must be rated for the voltage applied to them.
Yes, pin 4 in the diagram means/connects to pin 6 on your transformer.
And yes, the frequency is not important. Electrosurgery for example usually uses around 500kHz but can also use 27MHz.
You refer to the emitter of the BU508, but it's the collector that pulses and is connected to the transformer and the emitter is grounded, right?
If you see pulses at the (grounded) emitter then you grounding layout is not good.
The pulse width is perhaps ok, and I guess the HV can be adjusted by way of R1 since I guess it affects the pulse/pause relationship. What are the figures at min/max?
I would have liked to have one astable to set the operating frequency and one monostable to set the pulse time. That would give a better control of the power drain.

Hey there,

Nop I said the emiter cause I've reverse the pin from the diagram (wich is ok) to my board where the collector is to the ground. I suppose that could explain a lot!! I've retest on the bread board with the correct setting for yhe BU508 and I will switch the emiter- collector on the board... fun fun!

I've add a 1000uF 60V capacitor at the 2N3055 emitter (see new diagram), I hope it's enough?

The min max figure are 5 usec to 20 usec (rechecked with the correct BU508 setting).
You mean that it would be better with two 555 timer?

Again, very helpfull and enlighting... thanks!
 

Attachments

  • H Voltage output 30KV b.png
    H Voltage output 30KV b.png
    10.8 KB · Views: 316

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
What do you mean, reversed the pin? Were you driving the transformer from the emitter? If so that's nuts (unless you have a pulse transformer driving the base).

Yes, I guess 1000uF should be enough.

Yes, I got the 5-20us pulse figures (or is that pulse, and not period?), but what are the pause figures?
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
What do you mean, reversed the pin? Were you driving the transformer from the emitter? If so that's nuts (unless you have a pulse transformer driving the base).

Yes, I guess 1000uF should be enough.

Yes, I got the 5-20us pulse figures (or is that pulse, and not period?), but what are the pause figures?

Hi,

What I mean is the BU508's collector was to the ground, the base to the 555 pin 3 and the emitter to pin 2 of the flyback.... no wonder it blew I suppose... nuts :p

My figures are:

Min 6 us pulse -- 10 us pause -- so I guess 6 us pulse at 100 kHz frequency and
Max 30 us pulse -- 80 us pause - 30 us pulse at 12.5 kHz Frequency
if I'm not mistaken.

Thanks,
 

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
Hehe, it surely didn't help.. ;)
But actually, a transistor will still work with reversed collector - emitter, although with a much reduced gain (and I don't know what else suffers) so it won't work well.

Ok, the pulse lengths are no too bad, and you have 27% duty cycle at minimum frequency and 37.5% at maximum frequency.
It'll surely work but it's a somewhat limited setup. I'd want two 555's, the first setting the frequency and the second (in monostable mode) setting the pulse width.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
(and I don't know what else suffers)

The big one is the difference in reverse voltages that the cb and be junctions can withstand.

the base emitter junction is likely to have been somewhat compromised if the voltage across the transistor ever exceeded about 10 to 15 Volts
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
The big one is the difference in reverse voltages that the cb and be junctions can withstand.

the base emitter junction is likely to have been somewhat compromised if the voltage across the transistor ever exceeded about 10 to 15 Volts

Hi steve,

If you're talking about the 2N3055... I' had to replace it! It was fry and had no more resistance between either pin B-E, B-C or E-C!
The BU508 seem to be OK though. I have the cct on a breadboard and I compared the scope reading with the actual circuit board (after having reverse the emitter-collector pin correctly). They pretty much compare so I kept it!

I've also replace the potentiometer and the fuse, obviously!

I'll be retesting every component apart, then reassemble to box and try to make it work today! I still don't know how to be sure I have high voltage from the flyback (I will not put my meter in between :eek: ) but I suppose if nothing blows it will be a good start.

I'll then proceed to feed an asymetric capacitor with it and see what append!

Thanks a lot for your comments
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
I've finally put everything together and powered it. First with Switch 1 Off (see the whole cct attached as one file ), no problem, then with Switch 1 On.
Gradually turning R1 On made it blew (again) :D
I've rechecked everything (again), put it back together (again) and tested it without C7, which had been added to the cct to "prevent inductive kickback voltages reaching T1 from the flyback transformer".
The cct seems pretty happy without C7 so, I've added it back and blew R1 (again) :(

This time I'm really tired of blowing component put I'm still open to suggestion. Either I run it without C7 and take my chance or...maybe a Zener somewhere could be the solution, although I'm not really sure where or even why at this point (remember I'm a newby) :rolleyes:

Any idea??

Thanks in advance,
 

Attachments

  • Power supplys 0-12V + 0-50V + 30KV with fan.png
    Power supplys 0-12V + 0-50V + 30KV with fan.png
    32.9 KB · Views: 223
Last edited:

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
There's one "strange" thing I noticed earlier but forgot to mention, and that's the 12V regulator being supplied from the 0-50V supply. That won't work at low settings.
Increasing the voltage the 555 will wake slowly (or suddenly) from "the dead" and who knows what it does at that instant.
Why not supply the 555 from the "fan" supply so it's always active and becomes independent of the HV setting?
Does anything make a sound or get hot before it blows?
Running w/o C7 makes the 2N3055 act as a current limiter for the pulses. If the circuit works & survives without it then just leave it out.
Otherwise, a picture of the actual setup might aid diagnosis since an unfortunate layout may corrupt the function due to the fast high current pulses.
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
There's one "strange" thing I noticed earlier but forgot to mention, and that's the 12V regulator being supplied from the 0-50V supply. That won't work at low settings.
Increasing the voltage the 555 will wake slowly (or suddenly) from "the dead" and who knows what it does at that instant.
Why not supply the 555 from the "fan" supply so it's always active and becomes independent of the HV setting?
Does anything make a sound or get hot before it blows?
Running w/o C7 makes the 2N3055 act as a current limiter for the pulses. If the circuit works & survives without it then just leave it out.
Otherwise, a picture of the actual setup might aid diagnosis since an unfortunate layout may corrupt the function due to the fast high current pulses.

Yep I got you... I'm adapting the cct from someone else (added fan etc) and I see what you mean with the 555 powering up with R1. I've changed the cct so the timer (555) is powered up with a Switch (S1). The NPN Transistor T1 (2N3055) is on a second one (S2) and is still control by R1.(see attachment).

I've chose to use the main Transformer power line instead of the center tap of the fan as suggested because itsn't fully bridged and I was worried about fluctuation in the timer's current supply this way (the fan is somehow unstable in this regard even with a cap although I don't really care about the fan stability).

C7 is still in the attached cct but not in the actual board. Any idea why it seems to have reverse R1? Ri made a frying sound and it's the way I usually detect problem (smell is also a good indicator that something is geting hot :eek:). I say it blew but in reallity it partially fried, resulting in skew resistance values between Pin 1 2 and 3 of R1.

I would prefer having C7 if it is necessary but ...

As for a picture of the actual cct... I supposed you earn the right to see it although you'll be laughing and it will not help you help me... go ahead and enjoy!!

Thanks...
 

Attachments

  • Power supplys 0-12V + 0-50V + 30KV with fan.png
    Power supplys 0-12V + 0-50V + 30KV with fan.png
    35.9 KB · Views: 225
  • HV 30KV (allegedly).JPG
    HV 30KV (allegedly).JPG
    149.1 KB · Views: 238

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
Hehe, I had an image in my head of a crows-nest and I was not disappointed.. ;)
The thing with such layouts and fast high-current pulses is that the inductance in the wiring can create all sorts of undesirable side effects.
I don't like the R3 way of supplying the 7812. It gives very little leeway in the current the 7812 can deliver before being starved of input voltage.
It would be better to add a 220uF cap from the existing one and up to positive. The 7812 supply should then be entirely symmetrical. It is fully bridged.
If the pot is already cooked I guess it'll have to be replaced. I see it's a slider pot, but are you sure it's a linear type and not from a volume control?
It would also be highly beneficial to replace the 2N3055 with a darlington, for two reasons. To relieve the pot, and to enable a higher current capacity.
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
Hehe, I had an image in my head of a crows-nest and I was not disappointed.. ;)
The thing with such layouts and fast high-current pulses is that the inductance in the wiring can create all sorts of undesirable side effects.
I don't like the R3 way of supplying the 7812. It gives very little leeway in the current the 7812 can deliver before being starved of input voltage.
It would be better to add a 220uF cap from the existing one and up to positive. The 7812 supply should then be entirely symmetrical. It is fully bridged.
If the pot is already cooked I guess it'll have to be replaced. I see it's a slider pot, but are you sure it's a linear type and not from a volume control?
It would also be highly beneficial to replace the 2N3055 with a darlington, for two reasons. To relieve the pot, and to enable a higher current capacity.

Hi,

Since I had it all ready set the way it was I figured why not try it anyway! The slider pot was temporary and R1 was replace by a A10K pot (which is log and not linear???) since I had no more 4,7K pot.

Surprisingly it worked for a time! I know because I generated sparks between Pin 0 - 3 , Pin 0 - 4 and Pin 2 - 3 of the Flyback (tracks spacing were not width enough) and I could control sparks frequency using R6. It took my a while to find what was making that weird sound (and smell from the coating I'd applied) and maybe this is what I previously though to be a short in R1.

Anyway I fixed the track spacing and plugged it back... it still worked... until R1 resistance value became erratic and... well I need a new lot of pots for my experiments (and I'm going there this morning).

So... I'll replace the 2N30555 with a MJ2501 10 Ampere Darlington Power Transistor (hopping it will do)!
Will I still need C7 then???
Do I necessarly need linear pot and should I take high wattage??

As for the 7812 setup.. I understand and I need to re design the whole thing here. I have two 7812 when I only need 1 (without R3 but with 2, say 330uF 25V, caps on each side to be symetrical).
I'll also get ride of some cable... its like designing a computer program (which is what I do), when its done you need to rewrite it from scratch!!!

I'll let you know what happend!

Thanks...
 

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
Log pot's will give you a non-linear voltage output; very little change to begin with and a very large change near the end of the travel. I'd prefer a rotary linear pot.
The log's don't stand as much power as linear pot's however, and a 4.7k pot already dissipates 0.5W at 53V, and that's without the extra load from the transistor.
Go for something beefier than minimum. A wirewound multiturn (10 turn) pot would be the best, but they're pricey.
I figure the MJ2501 will do nice, and reduce the need for C7. Only an oscilloscope can reveal if C7 is still neccessary, but I would surely prefer to have it in place.
Don't overVolt your cap's, better to stay on the safe side of the redline, use at least 35V units in that series connection.
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
It's alive!

Log pot's will give you a non-linear voltage output; very little change to begin with and a very large change near the end of the travel. I'd prefer a rotary linear pot.
The log's don't stand as much power as linear pot's however, and a 4.7k pot already dissipates 0.5W at 53V, and that's without the extra load from the transistor.
Go for something beefier than minimum. A wirewound multiturn (10 turn) pot would be the best, but they're pricey.
I figure the MJ2501 will do nice, and reduce the need for C7. Only an oscilloscope can reveal if C7 is still neccessary, but I would surely prefer to have it in place.
Don't overVolt your cap's, better to stay on the safe side of the redline, use at least 35V units in that series connection.

Hi,

I've change T1 for a MJ2501 and put C7 back along with a new pot for R1. I hadn't had time to redo the 7812 setup for the fan/pulse supply (i'll simplify everything on a new cct) but it seems to be working smoothly never the less.

Thank you so very much for you help... I've learn a lot with this!! :)

Bye for now...
 

bibiz999

Feb 2, 2011
43
Joined
Feb 2, 2011
Messages
43
Me again...

Would it be ok to put a spark plug across the HV and 0V in order to test the cct or am I crazy? I would put a kind of telegraph On/Off switch to see if it sparks!

Thanks in advance...
 
Top