Hello, and help! IN-16 QTC troubleshooting

[barnbwt]

Glad to finally get this question in front of people who might be able to help.  Sorry in advance for the long post; just trying to be thorough for good measure :)

I very recently dove into the circuit-building hobby, specifically Nixie clocks.  The PV Electronics kit seemed like a pretty good starting point for someone with undeveloped solder skills, and even less electronics knowledge, to work at expanding both.  I would say this has definitely been the case, but it hasn't been without the expected learning curve.

I completed the low voltage and high circuits per their directions (QTC clock instructions & diagram pdf) and ran the simple voltage tests across the 5V and 175V portions, all went well.  I subsequently added the remaining components, and began the final test to light up a tube.  The seconds-place position worked great, just as I'd hoped, but moving to the minutes position didn't work out so well.  Plugging the power supply into wall yielded a number of effects, none of them good.

-Something going sizzle...

-The minutes tube either drawing voltage backward, or cathode-to--cathode, with purple flashes at its lower interior portion (this the same tube that ran fine in the seconds position)

-All RGB LEDs which had previously illuminated blue (the default start condition, I assume) were now fully lit to make a violet color (all elements full blast)

-Before promptly yanking the power supply after a second or so, I saw all these signs start to fade or draw down (brightness and sizzle)

So, the project fried, I discharged all capacitors and set to (more) carefully inspecting my handiwork.  I found what I thought was a single tiny solder bridge that may have shorted one of the resistors in series with an RGB LED and fixed it, but I wouldn't think that would cause such theatrics.  I assume that there was a short or polarity-swap in the second Nixie socket, which I guess could allow more driving current than desired and cook some things (the transistor at each Nixie that I assume is acting as a control relay, the logic IC controlling those transistors, the MOSFET supplying the high voltage, or the voltage regulator supplying the low)

-As best I can tell, the switching transistors for each tube look intact (no obvious burns, cracks, etc.) but I don't think I can test them easily while soldered in place

-The anode IC and cathode IC look intact, though again, no way to tell if they are burned internally

-I did a resistance check across all capacitors, and they are charging/discharging, though I don't have a great way to measure Faradays (just a multimeter)

-All diodes appear to still be resisting current in the right direction

-The 5V power supply circuit is still testing good

-The high voltage power supply was not amplifying (12V in, 13V out instead of ~175V)

The last point suggested that voltage amplifying transistor was shorted and non-functional, so I removed it (it easily broke apart as I did so) and replaced it.  The high voltage circuit still reads the same afterwards, as are the fully lit LEDs (no sizzle, though, and only the 'red' is illuminated after removing the solder bridge), and that is as far as my problem solving has carried me.  One point worth mentioning, is that the board is attempting to draw too much current, which causes the self-resetting fuse to disengage (the "fading" I saw initially also occurred during this subsequent testing of the HV circuit, but momentarily shorting the fuse lit the LEDs right back up).

I'm a mechanical design engineer.  This stuff ain't my bag, not yet anyway.  But as best I can tell, the logic side of the circuit diagram suggests the 'anode' IC can tell it's not getting 175V and may be lighting the LEDs fully by default (as opposed to those connections being cooked, which I'd assume would cook the LEDs themselves).  The 'cathode' IC directly meters the needed juice to the Nixies, so it is possible/likely it was damaged during the high current excursion, but there is not the visible/olfactory signs I'd expect of the distinct sizzling I caused.  I don't think the 'anode' IC could be exposed to damaging current from the Nixie portion of the board, unless both layers of switching transistors shorted source-drain (I'd think this would also cause some dramatic results in the 6 transistors involved).  So, my theory is that the problem still lies in the high voltage generation circuit...somewhere.  I suppose I could blindly start replacing stuff, but I'd like to at least chase the most likely issues first.

To anyone who made it this far, and has a suggestion, I salute you.  Many thanks, regardless.

TCB

[Nicholas Stock]

Have you checked the soldering on the connectors for the tube sockets? Sounds like you may have a short either on the minute socket or daughter board you used for the minutes. The 5V rail is functioning OK as the LED's light as you say, if the HV is only reading 12 to 13V then your assessment of the HV generator going kaput is probably correct. If you've already changed the IRFD220, then check the other parts in the HV circuit....are the Caps C3 and C4 the correct way round (an easy mistake, I've done that one before..;-)..

Have you discussed with Pete? He's the best resource for this obviously and very helpful....hope you get it sorted, they're great little clocks.

Cheers,

Nick

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/7cf1fb99-10ae-4260-9a93-402c3ee04d0e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Nicholas Stock]

Scrub the C3/C4 check, you mentioned that it was functioning OK on the seconds, so that's no an issue....apologies for that! ;-)

[barnbwt]

Well, if the IRFD fried, it stands to reason the caps could have, as well.  In fact, that is my next area of experimentation (but getting the correct caps means ordering them, which means like 20$ for about fifty cents worth of dinguses; I'm looking into rigging up similar components for testing with what little the local RadioShack still bothers to stock) since caps also go "sizzle." I did do a voltage test across all capacitor leads, and I can see a "charge" behavior from the readout, even if the numbers themselves are meaningless & changed by the attached circuitry.  Do damaged caps usually still exhibit some limited capacitance?  If they short/open upon failure, I'm definitely not seeing that on any of them.

I'll do some meter-poking of the Nixie sockets, since I can't really see what's actually happening inside the plastic sockets.  The Nixie modules, too.  If there were a short in this area, I assume the circuit would lose whatever voltage drop a lit Nixie carries, which would then result in a corresponding over-current condition?  Knowing that might at least give me an idea of how much damage I should expect.

Thanks for the advice, and I'll try hitting up Pete, too (I assume he's a busy guy, shipping out kits within five hours of ordering them, and all :) )

TCB

On Thursday, January 15, 2015 at 10:02:17 PM UTC-6, Pramanicin wrote:

Scrub the comment on C3/C4, they're obviously OK as the initial seconds and HV generation was working....apologies for that.

On Thu, Jan 15, 2015 at 7:50 PM, Nicholas Stock <nick...@gmail.com> wrote:

[barnbwt]

I'm gonna swing over the Shack tomorrow and cobble together some equivalent capacitors to replace those in the HV circuit.  There's a 220uF 16V-rated cap, and a 1uF 250V-rated cap, both electrolytic.  I suspect the high voltage cap would have been the one most exposed to the overload from the Nixie tube short.  Capacitance adds in parallel, right?  I should be able to kludge together something with what they have in stock that'll be good enough for another test, at least.  If that doesn't work, I think the next step is isolating the power supply circuits from the rest of the board entirely, so I can rinse and repeat what I've tried so far...

TCB

[Ryan]

Have you checked D5 (the UF4004)? If the caps were replaced and HV+ and GND do not measure shorted (just probe across the test point and a suitable ground), might be a good time to borrow a scope to see if there's base drive to the FET. 

[barnbwt]

Alrighty,

I got the C3 and C4 HV circuit capacitors switched out (the closest quivalent for the C4 were a considerably higher voltage rating, 35V vs 16V.  The C3 I found was identical, but a much larger metal-film version rather than electrolytic).  I also re-flowed all the solder connections in the HV circuit components left in place.  I don't think I'm changing anything, but the following appears different:

-The die-off of the circuit seems to happen a bit faster with the new capacitors (it reverted when I replaced the new caps with the better-fitting originals)

-The die-off of the LEDs does not occur when the IC2 board is not installed (circuit just stays lit, but still not generating high voltage)

I checked all the diodes, and they are still blocking current the way they are supposed to (can't measure the figures, though).  I also checked each soldered connection to the next component in the chain, and nothing in the HV circuit is open circuit (bad solder joint/etc.).  Looking at the diagram, I don't think I can isolate the HV circuit from the rest without removing a bunch of components for the LEDs and nixies, so I'm officially at a loss for the next step in troubleshooting.  Even if I had access to a scope I'm certain I don't know how to use it.  I'd obviously buy another 70$ kit before a 300$ scope, at least until that one cooks, too.

I think the fault lies outside the HV circuit (but connected to it, thus draining its voltage output to 12V), or I've managed to cook the IRFD220 again (which would be surprising, since I took even greater care to solder it without overheating this time, and the first go-around actually work until the short), or the IC2 chip is nuked and cannot activate (but I'm more inclined to think the lack of a high voltage at its sensing input is the cause for that)

TCB

[Gangyi Li]

HV sense resistors are all good? Have you tried it out with the driver IC removed? If that's all good and the FET is not shorted, might be an issue with the microcontroller.

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/b97e0a4e-c6ee-42d8-90b3-f14a7f0cc6c4%40googlegroups.com.

[barnbwt]

The only thing different with the IC2 (driver, vs. nixie-lighting) present, is that the 5V across the LEDs slowly dies off (I assume because of an additional trickle ground connection within board that is in front of the actual LEDs).  The 5V and non-amplifying 12V test points do not change whether IC2 is there or not.  I assume IC3 (the high voltage switching chip that connects the nixie cathodes) has no bearing here, since the high voltage is not even getting to it (the circuit does the same things, whether or not it is present).

The fact that 12V still shows on the far side of the amplifying circuit tells me that a short is occurring between the input power side and the output (after replacing it, I do not think the new IRFD220 chip is nonfunctional, nor the capacitors/resistors accompanying it).  This is good, since it appears there are very few 12V elements in the circuit, the rest being either 5V or high voltage.  Narrows down the possibilities.  According to the diagram, the UF4004 diode (D5) is the primary element separating the 220V output of the MOSFET from the 12V input (I am admittedly a bit confused by the circuit diagram representation, but I am guessing high voltage pulses are deflected by the inductor L1 and then flow across the diode D5 where they the remain).  I suppose I'll try replacing D5 next, with another +400V 1amp unit (so long as its activation voltage is well below 12V it should still open, right?).  The diode does show one-way metering, and a test value of 350, but I just don't see what else could be connecting the 12V and MOSFET outputs.  What would cause the MOSFET to generate a gain of 1.0?

Given the amount of damage the board is sustaining from all my picking at it, this diode swap is probably the last thing worth trying before starting over (frustrating, since I successfully hooked up a nearly-identical kit a neighbor needed help with, with no trouble whatsoever)

TCB

[John Rehwinkel]

> The fact that 12V still shows on the far side of the amplifying circuit tells me that a short is occurring between the input power side and the output (after replacing it, I do not think the new IRFD220 chip is nonfunctional, nor the capacitors/resistors accompanying it). This is good, since it appears there are very few 12V elements in the circuit, the rest being either 5V or high voltage. Narrows down the possibilities. According to the diagram, the UF4004 diode (D5) is the primary element separating the 220V output of the MOSFET from the 12V input

Nope, the 12V will flow right through the diode if the rest of the circuit is non-functioning, so you’ll see the 12V input at the output as well.

When the circuit is functioning, the FET pulls current through the inductor, then suddenly switches it off, and inductive kickback produces a high voltage spike which is then
rectified and filtered to become the HV output.

If the FET never turns on, you’ll see 12V at the output. If the FET never turns off, the inductor and FET will get hot and you’ll see almost zero volts at the output. If the FET turns on and off as it should, you should get HV.

- John

[Terry S]

A little knowledge is a dangerous thing.....

[barnbwt]

You say it better than I, but I simply mean that the fact I read 12 at the HV point means it is the FET circuit or earlier with an issue, so that eliminates all the downstream stuff from troubleshooting for now. Thanks for the explanation of the FET amplification, I understood it involved oscillating a tuned circuit and passing the peaks onward, but only conceptually. Sounds like the FET is not turning on, then. I have four FETs left, I guess I can try, try again, and see if third time is the charm. Switching the diode didn't help, as I had kind of expected from the multi-meter tests. I'm just not sure what is left that can be replaced. Is it possible a board trace was damaged?

TCB

[barnbwt]

Obviously. That said, what little I've learned has been in the process of troubleshooting here, so my ignorance wasn't very useful, either. Which is why I'm working to fix that.

[Gangyi Li]

What's important is ic2 receives power from the 5v regulator . The oscillator is connected and drive + hv sense is connected to the rest of the circuitry. If those check out (continuity) and the FET is not shorted, but something is still causing the 5v regulator to overheat and go into shutdown, likely Ic2 was damaged.

The only real way forward if that was the case, is to contact Pete and see if you can get a replacement. That Ic has to be programmed with code (it comes blank when bought).

On 19 Jan 2015 07:51, "barnbwt" <bar...@hotmail.com> wrote:

Obviously.  That said, what little I've learned has been in the process of troubleshooting here, so my ignorance wasn't very useful, either.  Which is why I'm working to fix that.

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.

To post to this group, send an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/f275fea9-bf5a-44ed-a144-cd367da63148%40googlegroups.com.

[barnbwt]

Since the 5V contact coming from the regulator remains constant, that would suggest it is not overheating/shutting down, right?  I'm in contact with Pete (very nice how he is willing to answer questions from people who abuse his products, lol) but I've also gone and ordered another board.  I don't think I'm capable of diagnosing or fixing whatever's wrong at this point; all the easy stuff has been tried, apart from dropping in new chips on faith and hoping they aren't cooked this time.  The board is starting to really suffer from all the solder/desolder fiddling, too (I'll have to use jumpers rather than the traces to reconnect C3 at this point).  Fresh start is probably a better use of my time & money (at 15$ per ship, replacing parts isn't exactly cost effective, either)

Here's a question; since the IRFD220 operates in a tuned circuit, would replacing the 25V rated small electrolytic capacitor with the much larger 35V rated metal film version screw with the FET's ability to amplify?  There is of course a difference in resistance between the two, after all.

TCB

[Nick]

On Monday, 19 January 2015 00:33:48 UTC, barnbwt wrote:

Here's a question; since the IRFD220 operates in a tuned circuit, would replacing the 25V rated small electrolytic capacitor with the much larger 35V rated metal film version screw with the FET's ability to amplify?  There is of course a difference in resistance between the two, after all.

Its NOT a tuned circuit  - its a boost-mode switching supply, and its most certainly not "amplifying" - its job is simply to turn completely on and off as and when instructed as fast as it can and with as low an on resistance (RDSon) as is possible.

Nick