[NEONIXIE-L] Transformers
Matthew Smith
Hi Folks
Whilst I am more than happy to power my Nixies from John Taylor's little
supplies, as they don't run up to 450+V, I've got to bite the bullet and
build my own supply/supplies for Dekatron purposes.
I am trying to do so with what I have to hand. This includes a few
LT1171 SMPS controllers - for which the most wonderful application note
was written (actually for the LT1070) and a box of small, medium and
large-ish bobbins and cores. Oh, and some spools of varying gauges of
transformer wire.
So, I have the makings of a power supply, but have some questions that
Application Note 19 does not answer:
* Flyback secondaries are in anti-phase to the primary. I know that I
can just connect from the other end of the winding, but what effect
would winding in the other direction (say clockwise vs counterclockwise
have?) Just asking this out of curiosity.
* I've seen flyback converters with multiple secondaries, but can you
have a single, multi-tapped secondary? I need about 3 or 4 different
voltages in addition to my B+ and logic voltages - looking at the best
way to get them. Resistive dividers sound like a good way of generating
heat (a potential problem for me due to the hot climate) and wasting energy.
* My cores are ungapped, a pair of 'E' sections. I've seen TV lineout
transformers with simple shims between the two 'U' sections. Could I
convert my non-gapped cores to gapped ones simply by putting, say, some
Kapton tape on each leg of one of the sections? [I made some inductors
up using these same cores. Glued them together with the thinnest layer
of epoxy - you would not BELIEVE the drop in inductance!]
This is all scary analogue stuff where I can't fix up any problems in
software, so thought I'd ask the experts first ;-)
BTW - this PSU will be running of 12V DC. Also considering an offline
version as I've got some TOP202s in my "spares box." Same issues apply,
as I'd still be building flyback tranformers.
Cheers
M
--
Matthew Smith
Smiffytech - Technology Consulting & Web Application Development
Business: http://www.smiffytech.com/
Blog/personal: http://www.smiffysplace.com/
LinkedIn: http://www.linkedin.com/in/smiffy
Skype: msmiffy
Twitter: @smiffy
Mike Harrison
On Tue, 09 Feb 2010 20:54:01 +1030, you wrote:
>Hi Folks
>
>Whilst I am more than happy to power my Nixies from John Taylor's little
>supplies, as they don't run up to 450+V, I've got to bite the bullet and
> build my own supply/supplies for Dekatron purposes.
>
>I am trying to do so with what I have to hand. This includes a few
>LT1171 SMPS controllers - for which the most wonderful application note
>was written (actually for the LT1070) and a box of small, medium and
>large-ish bobbins and cores. Oh, and some spools of varying gauges of
>transformer wire.
>
>So, I have the makings of a power supply, but have some questions that
>Application Note 19 does not answer:
>
>* Flyback secondaries are in anti-phase to the primary. I know that I
>can just connect from the other end of the winding, but what effect
>would winding in the other direction (say clockwise vs counterclockwise
>have?) Just asking this out of curiosity.
Winding in the other direction is the same as swapping the leads. Layering order isn't significant
for this type of application.
>* I've seen flyback converters with multiple secondaries, but can you
>have a single, multi-tapped secondary? I need about 3 or 4 different
>voltages in addition to my B+ and logic voltages - looking at the best
>way to get them. Resistive dividers sound like a good way of generating
>heat (a potential problem for me due to the hot climate) and wasting energy.
You may get some regulation issues if different parts of the winding have different loadings.
For low-power apps it probably works - it's not normally done on conventional step-down flybacks as
the optimum thickness of wire will be different for different windings.
>* My cores are ungapped, a pair of 'E' sections. I've seen TV lineout
>transformers with simple shims between the two 'U' sections. Could I
>convert my non-gapped cores to gapped ones simply by putting, say, some
>Kapton tape on each leg of one of the sections? [I made some inductors
>up using these same cores. Glued them together with the thinnest layer
>of epoxy - you would not BELIEVE the drop in inductance!]
Yes.
David Forbes
I'll jump in here, as I became an expert while making CRT power supplies.
At 8:54 PM +1030 2/9/10, Matthew Smith wrote:
>* Flyback secondaries are in anti-phase to the primary. I know that I
>can just connect from the other end of the winding, but what effect
>would winding in the other direction (say clockwise vs counterclockwise
>have?) Just asking this out of curiosity.
The start end labeling assumes winding all windings in the same
direction. Why would you wind in two different directions? It's
harder to do. If you do, then the start of the backwards winding is
called the end.
>* I've seen flyback converters with multiple secondaries, but can you
>have a single, multi-tapped secondary? I need about 3 or 4 different
>voltages in addition to my B+ and logic voltages - looking at the best
>way to get them. Resistive dividers sound like a good way of generating
>heat (a potential problem for me due to the hot climate) and wasting energy.
Yes, you can have as many secondaries as you like. I put five
secondaries on my scope clock transformer. Only one of them will be
tightly regulated, but that is not a problem for "digital" loads such
as Dekatrons.
>* My cores are ungapped, a pair of 'E' sections. I've seen TV lineout
>transformers with simple shims between the two 'U' sections. Could I
>convert my non-gapped cores to gapped ones simply by putting, say, some
>Kapton tape on each leg of one of the sections? [I made some inductors
>up using these same cores. Glued them together with the thinnest layer
>of epoxy - you would not BELIEVE the drop in inductance!]
Kapton tape is a fine gap maker. The formula for gap assumes only the
center leg has a gap ground into it, so if you use tape, make your
gap half as thick as the calculations say. But I've found that the
gap size is not at all critical. Having it at all is critical.
>This is all scary analogue stuff where I can't fix up any problems in
>software, so thought I'd ask the experts first ;-)
But you can always rewind the transformer if it doesn't work quite right.
>BTW - this PSU will be running of 12V DC. Also considering an offline
>version as I've got some TOP202s in my "spares box." Same issues apply,
>as I'd still be building flyback tranformers.
I'd stick with 12V if I were you. I made a bunch of flyback supplies
with mains input, but I wouldn't do it again.
The LM2586 is a fine chip also.
>Cheers
>
>M
--
--David Forbes, Tucson, AZ
http://www.cathodecorner.com/
James
--- In NEONIXIE-L@yahoogroups.com, Matthew Smith <matt@...> wrote:
>
> Hi Folks
>
> Whilst I am more than happy to power my Nixies from John Taylor's little
> supplies, as they don't run up to 450+V, I've got to bite the bullet and
> build my own supply/supplies for Dekatron purposes.
>
>
I had a chat with Bill Buzbee at the Maker Faire last year about a dekatron spinner he had there. He said he uses CCFL inverters to power them and it works well. I haven't tried it myself, but they're easy to find cheap and all you need to do is add a fast rectifier and a small filter capacitor and you have several hundred volts.
Matthew Smith
Mike/David, many thanks for your replies.
Quoth Mike Harrison at 2010-02-09 21:13...
...
> You may get some regulation issues if different parts of
> the winding have different loadings. For low-power apps
> it probably works - it's not normally done on conventional
> step-down flybacks as the optimum thickness of wire will
> be different for different windings.
OK, might try that. I don't think that regulation will be a real issue
with the various supplies for the dekatrons. Logic will be on a
separate supply anyway.
Quoth David Forbes at 2010-02-10 02:12...
...
> The start end labeling assumes winding all windings in the same
> direction. Why would you wind in two different directions?
I really was just asking that out of curiosity - wanted to know if my
understanding of phasing was correct or not. Haven't done anything like
that since college and that was over 20 years of dead braincells ago.
> Yes, you can have as many secondaries as you like. I put five
> secondaries on my scope clock transformer. Only one of them will be
> tightly regulated, but that is not a problem for "digital" loads such
> as Dekatrons.
Ties in with what I understood from Application Note 19. Feedback for
regulation comes off the most critical secondary, other secondaries get
something approximating to the relevant turns ratio.
> Kapton tape is a fine gap maker. The formula for gap assumes only the
> center leg has a gap ground into it, so if you use tape, make your
> gap half as thick as the calculations say. But I've found that the
> gap size is not at all critical. Having it at all is critical.
Ah, that's good! Gapping sounded too much like a Black Art. "A bit of
tape is good enough" sounds like one less variable with which to be
concerned.
> I'd stick with 12V if I were you. I made a bunch of flyback supplies
> with mains input, but I wouldn't do it again.
I think I will. I'm not unduly worried about working with 240V - it's
240V coming from a 15A breaker that tends to worry me. Plenty of time
for something to catch fire before the breaker finally trips. The
LT1171 has onboard current limiting - it should never pull more than the
rated current. Pop in a fuse between there and the DC supply and the
risk of conflagration should be minimised.
Now all I need to do is to build a turns counter so I can actually start
to wind the damn things. Sadly, I have no dead tape recorders to
pillage so I might turn the problem on its head and rotate the bobbin
with a stepper motor.
Cheers
M
--
Matthew Smith
Smiffytech - Technology Consulting & Web Application Development
Business: http://www.smiffytech.com/
Blog/personal: http://www.smiffysplace.com/
LinkedIn: http://www.linkedin.com/in/smiffy
Skype: msmiffy
Twitter: @smiffy
- New Members 10
- New Photos 17
Matthew Smith
Quoth David Forbes at 2010-02-10 02:12...
...
> The LM2586 is a fine chip also.
And $5 AUD to nearly $20 AUD for an LT1070. Listed by Farnell, but not
in stock anywhere for some reason. Not listed as end-of-line by National
so I wonder why.
Will certainly consider when I've burned out, er, used all my LT1171s.
--
Matthew Smith
Smiffytech - Technology Consulting & Web Application Development
Business: http://www.smiffytech.com/
Blog/personal: http://www.smiffysplace.com/
LinkedIn: http://www.linkedin.com/in/smiffy
Skype: msmiffy
Twitter: @smiffy
- New Members 10
- New Photos 17
Charles MacDonald
Matthew Smith wrote:
> I think I will. I'm not unduly worried about working with 240V - it's
> 240V coming from a 15A breaker that tends to worry me. Plenty of time
> for something to catch fire before the breaker finally trips.
I have a habit of ALWAYS puting a 3AG fuse and Holder right where the
power comes in on anything, (including modifing older gear) a guess as
to the expected draw selects 1/4, 1/2, 1 or 2 Amp as the right fuse to
use.
--
Charles MacDonald Stittsville Ontario
cmacd@zeusprune.ca Just Beyond the Fringe
http://www.TelecomOttawa.net/~cmacd/
No Microsoft Products were used in sending this e-mail.
- New Members 9
- New Photos 17
xaxiak
> * My cores are ungapped, a pair of 'E' sections.
Inductance will be high with no gap but the core will saturate at a very low current. Not good for energy storage. Adding some gap lowers the inductance and allows a much higher current. I use bits of paper. Its a good idea to monitor the inductor current waveform by placing a low value resistor between the switching device and ground rail. The waveform should look linear like a sawtooth. Core saturation is indicated by a sudden steepening.
- New Members 9
- New Photos 17
David
[edited by A.J. - please trim quoted material]
--- In NEONIXIE-L@yahoogroups.com, Mike Harrison <mike@...> wrote:
> On Tue, 09 Feb 2010 20:54:01 +1030, you wrote:
<snip>
> >* My cores are ungapped, a pair of 'E' sections. I've seen TV lineout
> >transformers with simple shims between the two 'U' sections. Could I
> >convert my non-gapped cores to gapped ones simply by putting, say, some
> >Kapton tape on each leg of one of the sections? [I made some inductors
> >up using these same cores. Glued them together with the thinnest layer
> >of epoxy - you would not BELIEVE the drop in inductance!]
>
> Yes.
You are correct in that the inductance drops when using a gapped core, however, the inductance is much more constant with varying current levels in the windings. If you have the space to use a gapped core (with corresponding larger windings), you'll find that the inductor (and consequently, the entire PSU) is easier to design.
When you introduce a large magnetizing current into the inductor, its inductance changes significantly with respect to its value as measured with a common inductance meter.
Cheers,
Dave M
- New Members 9
- New Photos 17
Matthew Smith
One last query on this thread before I start figuring out how I'm going
to build my coil winder. (Don't REALLY want to build a turns counter, so
may be simple crank and nothing else.)
I measured up the core/bobbin I was going to use and reckon that the 20
turns of 26AWG and 400 turns of 30AWG will fit comfortably.
However, having decided on extra secondaries rather than taking the 450V
down with resistive dividers to get the "extra" voltages, I have no room
for any more without going up to the next core/bobbin size.
The next size up is significantly bigger. Other than the fact that it
will take up more board room, are there any issues I should watch out
for with larger cores? I can only see advantages, such as greater
headroom when it comes to power, room for more insulation between
windings, etc. But then I might be missing something important.
Cheers
M
--
Matthew Smith
Smiffytech - Technology Consulting & Web Application Development
Business: http://www.smiffytech.com/
Blog/personal: http://www.smiffysplace.com/
LinkedIn: http://www.linkedin.com/in/smiffy
Skype: msmiffy
Twitter: @smiffy
- New Members 6
- New Photos 15
David Forbes
At 3:26 PM +1030 2/12/10, Matthew Smith wrote:
>One last query on this thread before I start figuring out how I'm going
>to build my coil winder. (Don't REALLY want to build a turns counter, so
>may be simple crank and nothing else.)
The Morris Coil Winder has a 100:1 gear reduction driving a clock
hand on a round dial calibrated in 100 radial marks.
>I measured up the core/bobbin I was going to use and reckon that the 20
>turns of 26AWG and 400 turns of 30AWG will fit comfortably.
>
>However, having decided on extra secondaries rather than taking the 450V
>down with resistive dividers to get the "extra" voltages, I have no room
>for any more without going up to the next core/bobbin size.
You can put taps in your HV winding to get a lower voltage for free.
If 400 turns = 450V, then 100 turns = 112V and so forth. Calculate
how many turns each lower voltage requires, and tie the secondary to
a pin after that many turns, then keep winding.
The only reason I used so many windings in my scope clock transformer
is that they all needed to be isolated from each other to run the
cathode circuits vs the anode circuits.
>The next size up is significantly bigger. Other than the fact that it
>will take up more board room, are there any issues I should watch out
>for with larger cores? I can only see advantages, such as greater
>headroom when it comes to power, room for more insulation between
>windings, etc. But then I might be missing something important.
Bigger cores are almost always better, in terms of power handling and
cooling etc. But a bigger core is bigger, so it won't fit in a small
space. That probably doesn't matter if you're running a bunch of
dekatrons, but it sure matters in an iPhone or a nixie watch.
--
--David Forbes, Tucson, AZ
http://www.cathodecorner.com/
John Rehwinkel
> One last query on this thread before I start figuring out how I'm going
> to build my coil winder. (Don't REALLY want to build a turns counter, so
> may be simple crank and nothing else.)
You can also calculate the length of wire needed and use that instead of a turns counter.
> However, having decided on extra secondaries rather than taking the 450V
> down with resistive dividers to get the "extra" voltages, I have no room
> for any more without going up to the next core/bobbin size.
You could use a tapped secondary to provide the intermediate voltages without requiring extra windings, assuming you don't need the other voltages to float. You can get negative voltages easily enough by using diodes in different directions, and/or choosing a tap that's not at one end for the secondary ground.
- John
- New Members 7
- New Photos 15
John Rehwinkel
> I'd stick with 12V if I were you. I made a bunch of flyback supplies
> with mains input, but I wouldn't do it again.
Yeah, I see that the redesigned scope clock uses 12V in and a more "traditional" switching power supply than the TopSwitch-based line powered one.
> The LM2586 is a fine chip also.
It must be, as it's the one you used in the aforementioned design! I've used some of those Simple Switcher chips, they're solid, well behaved, and easy to use.
- John
- New Members 4
- New Files 4
Matthew Smith
Quoth John Rehwinkel at 24/02/10 11:09...
>> The LM2586 is a fine chip also.
>
> It must be, as it's the one you used in the aforementioned
> design! I've used some of those Simple Switcher chips,
> they're solid, well behaved, and easy to use.
I'm trying to fathom out what the difference is between this and the
LM2576 (comparing the -ADJ version of each) other than the 2576 working
on a fixed 52kHz and the 2586 being adjustable between 100 and 200kHz.
Can anyone elaborate on this?
I have a rail each of the 5V and 12V 2567's that I picked up for very
little, a year or so ago.
Oh - the 2576 is half the price of 2586 - that's the other difference I
spotted.
Cheers
M
--
Matthew Smith
Smiffytech - Technology Consulting & Web Application Development
Business: http://www.smiffytech.com/
Blog/personal: http://www.smiffysplace.com/
LinkedIn: http://www.linkedin.com/in/smiffy
Skype: msmiffy
Twitter: @smiffy
David Forbes
At 2:00 PM +1030 2/25/10, Matthew Smith wrote:
>Quoth John Rehwinkel at 24/02/10 11:09...
>>> The LM2586 is a fine chip also.
>>
>> It must be, as it's the one you used in the aforementioned
> > design! I've used some of those Simple Switcher chips,
> > they're solid, well behaved, and easy to use.
>
>I'm trying to fathom out what the difference is between this and the
>LM2576 (comparing the -ADJ version of each) other than the 2576 working
>on a fixed 52kHz and the 2586 being adjustable between 100 and 200kHz.
>
>Can anyone elaborate on this?
>
The LM2576 is a buck regulator, with the switch in series with the
output. It makes only lower output voltage, so it's useless for
powering tubes.
The LM2586 is a flyback regulator, with the switch grounded. It can
run a transformer, so it will make many volts. That's why it's useful
to us.
Quite a difference.
--
--David Forbes, Tucson, AZ
http://www.cathodecorner.com/
Matthew Smith
Quoth David Forbes at 25/02/10 14:26...
> The LM2576 is a buck regulator, with the switch in series with the
> output. It makes only lower output voltage, so it's useless for
> powering tubes.
>
> The LM2586 is a flyback regulator, with the switch grounded. It can
> run a transformer, so it will make many volts. That's why it's useful
> to us.
Ah, thanks. I was getting confused by the "Isolated Flyback Converter
using minimum number of components" line under "Applications" on the
datasheet.
--
Matthew Smith
Smiffytech - Technology Consulting & Web Application Development
Business: http://www.smiffytech.com/
Blog/personal: http://www.smiffysplace.com/
LinkedIn: http://www.linkedin.com/in/smiffy
Skype: msmiffy
Twitter: @smiffy
Mike Harrison
Item 261105444127
kay486
Mike Harrison
>Nice find! Do you know how much do they usualy sell for?
