Filament transfomer current for driving six NIMOs?

[Dekatron42]

Hi,

I need help with how much current a regular mains transformer would need for the center tapped filament winding for six NIMOs.

The filament for a NIMO is specified at 1.1V AC or DC +/-0.15V at some 0.2A. The cold resistance of the filament is some 2.1 Ohm.

Do I need to specify some overhead for the heater transformer to be able to power six NIMOs or do I just need 1.2A current capability (6*0.2A)?

/Martin

[John Rehwinkel]

> I need help with how much current a regular mains transformer would need for the center tapped filament winding for six NIMOs.
>
> The filament for a NIMO is specified at 1.1V AC or DC +/-0.15V at some 0.2A. The cold resistance of the filament is some 2.1 Ohm.
>
> Do I need to specify some overhead for the heater transformer to be able to power six NIMOs or do I just need 1.2A current capability (6*0.2A)?

1.2A should be sufficient. The cold filaments will attempt to draw more, but the transformer will be unable to supply it, so the voltage will drop initially, which is what you want: this will give a gentler warmup than smacking them with full voltage would. It is, of course, possible to add more circuitry for an even softer start, but a transformer sized for the equilibrium load is pretty simple and works well.

- John

[gregebert]

Congratulations on acquiring 6 NIMOs !!

I've gone thru all the NIMO filament analysis and came up with the following, which I have implemented in my design

1. Transformer is 2.5V CT, 3A. Triad F301X.. It has dual primaries so it can operate on 115 or 230 VAC mains.

2. Series fuse 250mA (BEL C1F250). One for each NIMO

3. Series resistor; empirically I'm choosing 6.8ohms.

The filament is the most-critical reliability-impacting element of your NIMO tubes, so be extra careful to make sure you design with this in-mind.

Filaments usually fail due to power-on surge-current, just like conventional incandescent bulbs. The surge-current is the result of the lower-resistance of a cold filament being the only limiter of current. For example, measured NIMO-tube filament resistance varies from about 2.8 ohms cold, to 7.0 ohms when stabilized at 200mA. If you use a constant voltage to power the filaments, you will have significant surge current. But if you add some series resistance and drive from a higher supply voltage, the surge current will be less. With the components I listed above, the surge current is about 260mA.

The fast-blow fuse will protect the NIMO from line-surges, or other uncontrollables.

I'm almost done with my NIMO clock, and I will collect waveforms of the power-on transient for each filament once I get my other tubes (I have 1, need 5 more)

[Kevin A.]

I like the discussion here. I'm working on a 2 Nimo clock w/VFD for myself, and a 6 tube clock for a friend. I'd like to add that I found these power supplies which are really small and quite trick for a low profile nimo design, needing between 1500-2000 vdc. They aren't cheap, but it's the latest and greatest and would really make for a slick and fully controllable design: https://www.digikey.com/product-detail/en/xp-power/AG20P-5/1470-3211-ND/5873521

[Tony]

The old-school method used was a thermistor in series with the heater
chain, chosen with a suitably high resistance when cold which would
slowly drop as it warmed up to a low operating value.

Tony.

[Dekatron42]

Thank you all for your input!

@jrehwin - Thanks for explaining how "a transformer sized for the equilibrium load" works in a circuit like this!

@gregebert - Do you have to take the fuse resistance into account in your circuit to maintain a centered heater voltage (I guess so as the fuse resistance can be several ohms with high speed low current fuses)?

@Kevin A. - I've seen those small EHT supplies and they are very nice except for the price, it is simply too high for me, and I'd also like to be able to repair the design myself so I'll go with a home built diode doubler design.

@Tony Adams - I have only used that kind of protection on the primary side of a transformer but will ask manufacturers if they have anything that would work on the secondary side with these low voltages/currents.

/Martin

[gregebert]

@gregebert - Do you have to take the fuse resistance into account in your circuit to maintain a centered heater voltage (I guess so as the fuse resistance can be several ohms with high speed low current fuses)?

Eventually, yes. Right now I have safe-enough values to get the filament warm enough so that the tube should function. After I get more tubes and have the I-V data plotted for their filaments, and also confirm the brightness is correct (nominal 1850V anode supply @30uA per tube), I will fine-tune the series resistor value. The goal for me is to use somewhat less than the spec value of 200mA of filament current at the highest line-voltage I record at my house over a weeks' time, and verify that the series resistor is sufficient to view the tube during low-line-voltage periods. 

So far, I have only energized the filament of my one-and-only NIMO tube once, and that was to gather IV data.

I'm still tuning the HVDC inverter and have not actually fired-up the tube yet; going very, VERY slowly to make sure I dont risk any damage to the NIMO. Everything else is ready-to-go. Only after I have several days of clean operation of the HVDC supply during no-load and 150% overload will I attempt to fire the tube. I just found another bug in the FPGA that controls the inverter. It's a new design of my own, and probably way more complicated than it needs to be, but it's regulated and allows software control of the voltage, contrast, and even the tube current.

[Kevin A.]

This is a simple power supply I'm devising for a pair of Nimo's. 12 Volts in main power, HV power supply, NTC thermistor "damped" voltage divider circuit from 12V to produce 1V @ around 200ma for each tube filament. Just a quick sketch, think this would work well without being TOO complicated

[Kevin A.]

*That would be 12 Volts DC in.

[gregebert]

I recommend small series resistors for each anode; I'm using 100K SMT devices and with a 30uA load-current per tube that's a negligible 3V drop across the resistor.

It allows you to measure tube-current, but be careful as it's 2kV.

Also helps reduce the zap if something shorts-out unexpectedly, such as when you are probing around and your hand slips ?

I have an A/D converter under software control that measures the current thru the filament-transformer center-tap to GND. This allows you to measure all tubes (or individual ones if you shut off the others).

[Dekatron42]

@gregebert: Isn’t there a risk of flash-over if you don’t use high voltage resistors for the anode connection?

[Dekatron42]

@jrehwin: If I order a transformer with 1.1V center tapped transformer with 1.2A rating for driving six NIMOs, do I have to load it with 5 ”NIMOs” if I want to test just one NIMO?

[gregebert]

On Tuesday, October 23, 2018 at 4:59:29 AM UTC-7, Dekatron42 wrote:

@gregebert: Isn’t there a risk of flash-over if you don’t use high voltage resistors for the anode connection?

These are series resistors which normally have only a minimal voltage-drop across them, typically 3V. No flashover risk for those. Just be sure to provide adequate creepage & clearance spacing between those resistors and anything that is low-voltage.

I do have a 100Meg resistor which is physically large enough to prevent arcing. Since it is for a voltage divider, it has almost 2kV across it. If I used an SMT device, it would be risky not only for flashover, but also for tin-whisker growth.

[GastonP]

You don't need to put all of the tubes if you use limiting resistors (or some other protection device) in series with the filaments. Just the one you calculated for normal use will do.

This case (very expensive, unobtanium tubes that will be gone forever if zapped) fully justifies a slow turn on regulated DC power supply instead of an "economical" solution.

Gastón

[GastonP]

Hmmmm... a transformer short circuit current, which is the maximum current that it can supply, is not the design or specified one, which in this case would be 1.2A and which is the current it can supply indefinitely while maintaining all of the design parameters.

The ideal transformer has, of course, infinite output current, but as the real transformers coils are made of copper, impure, there is its resistance that limits the current output (and of course heats the transformer). I'm dismissing losses due to the iron core as we are talking of the transition start up, which is likable to a short circuit.

One can calculate the short circuit current as the transformer secondary design voltage divided by the secondary total DC resistance, that would be the secondary DC resistance plus the primary resistance reflected on the secondary (the primary DC resistance divide by the transformer ratio).

In this case, the transformer resistance is Rsec + Rpri * Vsec/Vpri, and the short circuit current would be Vsec/(Rsec + Rpri *Vsec/Vpri).

The tube startup current without limiters of any kind (assuming there is only one being fed by the transformer) would be then Vsec/(Rsec + Rpri*Vsec/Vpri+Rfilament)

So, the transformer actually supplies several times the specified current at startup, given the very low cold resistance of the filament. This accounts for the "flash" one sees in some receiving/transmitting tubes.

The moral is: one needs to put some limiting element when feeding delicate tubes.

Gastón

[Dekatron42]

A simple solution then would be to use one resistor, or two resistors one in each branch, of the heater chain to maintain a centered reference compared to ground even during startup, and a relay shorting that/those resistors after some time - just like slow starters for toroidal transformers are designed. This could then be used with AC and DC heaters. A more complex design would use some extra components with a voltage regulator ramping up the voltage, but would only work on DC designs.

Using the same circuit combined with an inrush limiter on the primary side of the transformer supplying the heater voltage would also limit the startup voltage spikes, adding a MOV would protect against over voltages.

Does this seem like a usable design?

/Martin

[gregebert]

If you want absolute minimum energy consumption, then you will want to ramp the power supply, etc. I looked at a lot of tradeoffs and calculated the cost of energy for each.

Circuit/component costs were not modeled.

I have the option of soft-starting my 2.5V filament transformer, because it's controlled by an opto-triac by the onboard FPGA, which in-turn, has AC-line sync signals, but I did not think it was necessary given the huge reduction in surge-current with a series-resistor. Basically, the FPGA would trigger the opto-triac just before the AC-line made a zero-crossing, then on the next cycle a bit sooner, and so-on for about 10 seconds, by which time the opto-triac was triggered for the entire AC-line cycle and the filament is fully heated.

I have yet to define how software will decide when to turn filaments on vs off based on PIR sensor activity, time-of-day, etc. Leaving filaments on for extended periods vs power-cycling them has a tradeoff point, but I dont know where that lies. Also, phosphor lifetime is finite, which I'm assuming is 25,000 hours per the MTBF in the datasheet. An older NIMO datasheet shows 10,000 hour MTBF (ouch, that's just over a year when running 24/7).

	Total Filaments	Current (Amps)	Volts	Efficiency	Power	Tube Life (hrs)	Elec Cost Kwhr	Filament Energy Cost
Perfect solution	6	0.2	1.1	100%	1.3	25000	0.1	$3.30
7v Supply	6	0.2	7	100%	8.4	25000	0.1	$21.00
10V supply cur limit	6	0.2	10	100%	12.0	25000	0.1	$30.00
5V curr limit	6	0.2	5	90%	6.7	25000	0.1	$16.67
3.3V curr limit	6	0.2	3.3	86%	4.6	25000	0.1	$11.51
20v supply cur limit	6	0.2	20	100%	24.0	25000	0.1	$60.00
16v supply cur limit	6	0.2	16	100%	19.2	25000	0.1	$48.00
2.5VAC xfmr	6	0.2	2.5	80%	3.8	25000	0.1	$9.38

[Jon Jackson]

@gregebert  What is "perfect solution" with respect to the 1.1V supply?  Is that your fancy FPGA soft-start setup ???

Jon J.

--
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/e71909e8-7f1e-4d7f-8dc7-7a389b3ccffe%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[gregebert]

Perfect solution is a supply that is 100% efficient and gives 1.1V @ 1200mA (power for 6 NIMO tubes), and magically limits the current to 200mA.

Nothing will ever be this good; it's just a reference point for cost-comparison.

The FPGA solution I described is an added feature for the 2.5V filament transformer. The only advantage it could provide is reducing the peak current from 260mA to about 200mA. My gut feeling is that the 260mA surge current is probably low-enough that it wont cause any degradation of the NIMO tube.

[Dekatron42]

I contacted Ametherm regarding inrush current limiters as I’ve used their products to limit inrush current on toroidal transformers and they told me they had limiters used for heat lamps and similar applications and they believed those would work. I’ll try to order some and see how they behave.

Otherwise I’ll probably go for the relay solution shorting out series resistors.

/Martin

[Nick]

On Tuesday, 23 October 2018 18:04:32 UTC+4, GastonP wrote:

You don't need to put all of the tubes if you use limiting resistors (or some other protection device) in series with the filaments. Just the one you calculated for normal use will do.

This case (very expensive, unobtanium tubes that will be gone forever if zapped) fully justifies a slow turn on regulated DC power supply instead of an "economical" solution.

Gastón

I have some stock of IEE 6000-31-0000 if spares are needed.

Nick