Yet another nixie power supply

[newxito]

I spent some time trying to build a psu for a nixie clock with 12 x IN-18 tubes. I will need about 65mA at 170V.

I ended up using the LT3757 datasheet flyback schematic. I don’t want to share the board (it’s not a good design) but I would like to share the components I have used. Maybe this is useful for someone else that needs a little bit more current as the normal 15 to 35mA.

MOSFET:           CSD19534Q5A

Diode:                 RFN2LAM4STR

Transformer:       DA2034-AL (9A)

Sense Resistor:  0.01 ohm 

I don’t really know what’s the limit of this configuration. After blowing the 100mA fuse in the first attempt, I stopped testing at 200mA. The psu still was delivering 170V but the transformer got really hot. I had it running overnight at 100mA, no fried components so far.

The best efficiency I have measured with this thing was 91.0% at 60mA with an input voltage of 12V.

90.2% at 100mA … 70.1% at 6mA

Just for the fun of it, I designed a small (20x25mm) version with a DA2032 transformer on the top side of the pcb and the other components on the bottom side. Surprisingly, it works fine, 88.5% at 25mA, no audible noise, no heat…

[newxito]

Warning: building your own PSU can be addictive :-) 

Short update:

I changed these components in the final version:

MOSFET:            NVMFS5C673NLWFAFT1G

Diode:              RFN2LAM6STR

Transformer:       DA2033-AL (5A) (good enough for my 65mA project)

The graphic shows the efficiency measured with a 0-360V/30A electronic DC load and a regulated power supply (0-30V 5A) both with 1 mA resolution but I don’t know how accurate this equipment is.

Just to make it clear, this is not my design, I’m using the lt3757 datasheet schematic. I’m just experimenting with different component.

I don’t know yet how stressed these components are, maybe in 2 or 3 years…

[jörg]

Hi, I've build pretty the same powersupply with the DA2032 and LT3757. I love the small formfactor possible.
Unfortunately I'm experiencing singing Tubes after a while, if the PS is going warmer. About 35 degrees C.
Using coolant spray, the singing disappears.
Question: Did you have effects?

Anyway, I will investigate further.
And, thanks a lot for sharing.

[newxito]

Good point, I only did a short test with real tubes…

I’m currently modifying my through-hole clock boards, replacing the mc34063 stuff with the small DA2032 psu. When the new board is finished, I will check for singing tubes.

Maybe you can solve the problem modifying the 140K resistor with another value to change the switching frequency.

I had two identical boards, both with 140k 1% resistor, one switching with 85kHz, the other one with 110kHz, don’t know why. By the way, 110kHz seems to be a good value for good efficiency.

[David Forbes]

You might want to check the output with an oscilloscope, to see if the pwm control is oscillating. Then tune the loop response to prevent 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 view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/90008323-62eb-436f-b45a-09812f64c508%40googlegroups.com.

[Bill Notfaded]

I'm interested to also hear how you're using the 12 tubes.

Bill

[newxito]

The board has 192 HV outputs.

My idea was to build a clock with 18 digits, 6 for time, 6 for date, 3 for inside temp and 3 for outside temp. Maybe I build one with small tubes.

But first I would like to build a 12-digit clock (date and time) with 60 neon lamps as additional way to display seconds.

[Bill Notfaded]

That's really different idea... I like it. I've thought about clocks with more tubes too. Something unique is cool.

Bill

[newxito]

I would like to implement two modes for the 60 neons: a single mode and a fill mode. But I have to check the current needed by a neon lamp. Assuming 0.5 mA, this is up to 30mA additional current…