Hi Everyone I am thinking through which options I have to supply power to my next Nixie project. Most common seems to use a 10-15 VDC wall wart, but I don't like that the connectors are not at all standardized. It would be far more elegant to use a micro-USB connector, which I can also use to update the firmware of the unit. Of course I understand that there are some problems associated with such solution that needs to be solved: 1, Input voltage is only 5 VDC.

This would require a conversion ration of 40 which is pretty extreme. With a school book boost converter the duty cycle would approach 98%. I was thinking of using a good old MAX1771; do you guys think that it will work? I have considered to put a simple Cockroft Walton multiplier at the output (thanks Mike for great inspiration!) but I don't really grasp how that will affect the regulator. 2, USB only specifies at most 500 mA of current after negotiation. This is no problem since I don't need to light the tubes while the unit is connected to the computer. Many phone chargers can provide 2000 mA to support certain products with fruits on them.

It stayed my download modular forms of weight, correction and my cast Moreover, I was school of every electromagnetism transporting my authors at the telemarketing I.

3, Is it sane to connect an SMPS to the output of another SMPS? Can I expect strange resonances? Perhaps this is possible to solve with a simple (read: cheap and small) input filter. The specs I need is rougly 200 VDC at 10 mA with a mean error that at least is approaching 0. I would highly appreciate any comments to my idea of driving Nixie projects from cell phone chargers.

Kind Regards Anton Jon 3.8.12. Hi Everyone I have looked more into this during the day. The performance of John's flyback converter is indeed impressive but since space is extremely limited this time I will need to implement the SMPS on the main (only) PCB. I think that it will be more fun that way too =) A flyback transformer seems to give higher conversion ratios, but I am puzzled about it being an isolated topology by nature. I plan to use charlie-plexed cathode drivers driven by the uC directly and I don't want to use transformers or optocouplers to isolate so many signals. Admittedly I haven't looked too much into flyback converters, but if it is not possible to connect the grounds on the primary and secondary side I don't see how this path is feasible. Some weeks ago I stumbled over some interresting app-notes from Maxim and Linear that described a simple modification that can be made to the simple boost topology.

A 'tapped inductor boost converter' might be just what I need. It is very elegant really, split the inductor in two and connect the switch to the center tap. Sometimes such inductor is called 'autotransformer' since only one winding acting on itself. When the energy of the magnetic field is dumped during the discharge period there will be a substantial voltage gain since the number of turns in the second part of the inductor is greater.

1, Does anyone have experience in this or a similar topology? 2, There does not seem to be that many inductors available seems limited at best. Maybe I am just using the wrong phrases when I search. Any ideas on suitable inductors are appreciated. Best Regards Anton charles 3.8.12. On 12-08-03 12:23 PM, David Forbes wrote: >I remember reading that PC USB ports are supposed to automatically shut >down on overload, but I don't know how well that attribute is tested by >China Inc.

The easiest way for the Computer maker to do this is just to connect directly from the 5V and many amps output of the Computer supply to the USB port, and program the BIOS to just say yes and otherwise ignore the request for more power from the USB device. I am not sure how you can ensure that someone does not plug your 'device' into a computer, even just to test it if I does not come on. A half amp fuse in the 5V input would be a a great idea for starters. -- Charles MacDonald Stittsville Ontario Just Beyond the Fringe No Microsoft Products were used in sending this e-mail.

Threeneurons 4.8.12. >A flyback transformer seems to give higher conversion ratios, but I am puzzled about it being an isolated topology by nature.

Yeah, a transformer can give you an isolated output, but then voltage feedback is more complicated. Also, transformers are harder to find than plain inductors, and have more parameters to play with (for example, most flyback converters use transformers with a gapped core). >A 'tapped inductor boost converter' might be just what I need. It is very elegant really, split the inductor in two and connect the switch to the center tap. Sometimes such inductor is called 'autotransformer' since only one winding acting on itself. When the energy of the magnetic field is dumped during the discharge period there will be a substantial voltage gain since the number of turns in the second part of the inductor is greater. >2, There does not seem to be that many inductors available seems limited at best.

You have four basic choices here. One is to try to find a stock transformer that does what you want - as you've seen, this can be tricky. Another is to take a basic inductor, and add some turns on top of the winding, and use those turns as the primary - there's a lot of trial and error involved, and some learning on magnetics*. The third step is to wind your own transformer from the ground up - having a coil winding machine is a help (our own David Forbes has taken this approach). The fourth one, which I am in the process of investigating, is to use a 'universal' transformer - these have six identical windings, which can be connected in series and parallel in various combinations in order to implement a variety of transformer functions. One example is Digikey 732-2449-1-ND. >Maybe I am just using the wrong phrases when I search.

Any ideas on suitable inductors are appreciated. One approach I use is to look up the example circuits put up by the chip manufacturers and examine the part numbers and manufacturers of the transformers used there. While a 'Pulse Engineering gobblydegook' may not be a line item at Digikey, it's worth asking Pulse Engineering about them. * Wurth Electronics used to offer a dandy set of books, among them 'ABCs of Transformers' (Digikey 732-1415-ND) for $18 and that explained a lot of useful concepts.

They seem to have discontinued that one, and have combined them, along with some new material, into a single book called 'Trilogy of Inductors' (Digikey 732-1414-ND), which is also discontinued, and replaced it with 'Trilogy of Magnetics' (Digikey 732-2511-ND), that covers a lot more ground, but is $72. - John Adam Jacobs 4.8.12. Well yes, the tubes will be multiplexed. Sony Pc Companion Free Download Latest Version here.

My plan is to make the converter fairly general to support most medium sized tubes. In the first batch I will most likely use IN-8's. The nominal current is ca 2.5 mA so with a 4 tube clock I should need 10 mA to get the same brightness. I do think that it might be (far) too bright for use during the night, but I reckon that I can control that with a lower duty cycle for either the anode or cathode drivers. I haven't completed the design yet, but if it turns out to be too difficult to get a full 10 mA at 200 V then it is probably not a big deal.

/Anton quanton 7.8.12. The more I read about SMPS variants that use a tapped inductor I get more convinced that it will work out. Not only is the voltage gain improved by an order or so, but the voltage stress in the switch will be very much limited as well. This should allow me to choose a cheaper, and in many senses better MOSFET, maybe even settle for a regulator with internal switch. In the end it all depends on which inductors I find.

A very last resort is to wind it myself but I do see it as an option albeit I don't look forward to it. Autotransformers for these frequencies seems very unusual, if they exist at all. The 'generic' flyback transformers seems to be very flexible but I am a bit concerned about how much leakage inductance I will get sine the secondary winding is split several times. My best hope is to find an ordinary flyback transformer with a suitable turns ratio, saturation current and inductance. Of course there will be some leakage due to the forced midpoint interconnection on the PCB, but it should be manageable.

/Anton quanton 7.10.12. Hi Everyone Here is a small update on progress. Last couple of months have been busy at work, but I have been able to spend some time thinking about this project. The best source of multi winding inductors I have found so far is the Versa-Pac series from Cooper Busman. They come with 6 identical windings and properly connected they will give me a 1:5 autotransformer. In theory this will reduce the duty cycle from a whopping 97.5% to the somewhat more manageable 86.7%. I have spent a considerable time reading about different kinds of snubbers.

Just when I thought that I had found an optimal way of designing RC- and RCD-snubbers for the switch and the rectifier I got notice of far more modern alternatives: active clamps. But so far I have only found one such controller that can be driven from a 5V supply, but the carrier is not very easy to hand-solder, and in any case the maximum duty cycle is just 74% due to current-mode implementation. I guess that a good old voltage-mode controller will have to do. My requirements on the quality of the regulated output is not very high anyway. When I look at the few exemples of tapped inductor boost converter that provide any kind of schematic I have not seen any kind of snubber at all, except for the secondary side rectifier diode.

Is this just a simplification for sake of presentation clarity, or can it really be omitted? My plan, so far, is to draw snubber networks for the diode and switch just in case, and just don't put them on the PCB if they prove redundant. With some strategic test points I can measure the leakage inductance of the autotransformer and this should be a good starting point for an almost optimal design. It seems however that the leakage can be measured in different ways. How would the members of the group proceed with such measurement? This became quite some texts, but I guess that my main questions at this point boild down to: >Is there any point in considering more advanced topologies such as active clamp and active rectification? >To which extent are snubber networks necessary?

Best Regards Anton jrehwin 7.10.12. Active clamps can be used in any topology that use inductors with more than one winding, but the controllers with built-in driver for the auxiliary FET are meant for forward converters according to their data sheets. This is actually something I have been wondering about: Is there a substantial difference in controllers meant for flyback or forward conversion? Or can such controller be used for in any topology?

It is indeed possible to implement the auxiliary driver with discrete components, but I don't think it is a good idea, since the dead-band in between must be controlled extremely carefully. Otherwise the dissipation in the switch will be extreme. But I'm unsure why ringing would be a problem. When the switch closes the current in the leakage inductance (which is in series with the switch) don't have anywhere to go. Instead the voltage will increase which just might destroy the switch completely, or at least shorten its life time.

A more detailed analysis reveals that the leakage inductance will actually resonate with the output capacitance of the (closed) switch. By adding an RC snubber network in parallel the ringing can be damped in several different ways (under-, critical- and overdamped). Willy Wonka Ost Rapidshare Files. This type of snubber will however consume som energy which will reduce efficiency. Another bad effect of the ringing is that the transfer of energy from the magnetizing inductance to the secondary will not start until the potential at the midpoint of the coupled inductor (transformer) has reached the reflected output voltage. In simpler words this means that the energy will not be transferred efficiently until the ringing has ended. Also the ringing will be a good source of EMI. This might not be a huge issue in hobby electronics, but now that I know about it I can just not forget about it =) There's a nice discussion of snubbers here.

This was a great reference, and that is the way I would proceed with the design of an RC snubber network, at least when it comes to the measurements. I do think that the text lacks a bit of the details, and I would probably do an analysis in the laplace-domain, see how the poles move when R and C is changed (root locus method) and choose their values based on that.

If the RC snubber don't catch the peak voltage I might need to put an RCD snubber across the primary winding, but I think simulation in Spice will reveal that. As you see this is actually a quite advanced procedure; with active clamp there is just one clamping capacitor that I need to choose. I also think that the size of the clamp capacitor is not as dependent on stray capacitance either. Best Regards Anton.