Reply by ●February 6, 20192019-02-06
dakupoto@gmail.com wrote

>I agree that an emitter current limiting resistor acts as a current source, but connecting the emitter directly
>to ground makes me queasy. So, I have a small resistance, 10 Ohm with a bypass capacitor. This is a side project, and I am
>trying to understand the workings of an air-core transformer.


The way I go about transformers is like this:

First the no-load situation, get the primary inductance.

If a switcher look at the maximum on-time of the switch.
 _______
|       | 
|+      )L
U       ) 
|-      |
|       / switch
|       |
--------

i
^   /
|  /
| /
--------->t

The current will then linear rise as U.t/L,
where t is the time the switch his on.
so in a 1 Henry coil after 1 second the current will be 1 Ampere.

L should be so big as to not create a huge no-load current (maybe 10% or less of full load).

So from L you can now calculate the number of primary turns,

Air-coils do not go into saturation, but if you keep the switch on too long the current will not rise linear
as it will be limited by the R of the coil.

And that is really, all there is too it, for a switcher that is.

The fun starts when the switch opens..... and a large flyback voltage appears across L if there is no load.
Reply by ●February 6, 20192019-02-06
On Tuesday, February 5, 2019 at 4:33:38 AM UTC-5, 69883925...@nospam.org wrote:

> dakupoto@gmail.com wrote
> 
> >On Monday, February 4, 2019 at 3:58:36 PM UTC-5, whit3rd wrote:
> >> On Monday, February 4, 2019 at 3:50:29 AM UTC-8, daku...@gmail.com wrote:
> >> 
> >> > The base of each 2N3055 is driven by a relaxation
> >> > oscillator at 150 Kilohertz. What is so fast about that ?
> >> 
> >> If those transistors go into saturation, the rise time when they turn off
> >> is very poor, by switchmode standards.  A bit of capacitance across
> >> the 10 ohm resistors would help.   So would a different transistor, with
> >> lower Rbb.
> >
> >Thanks, but I already tried out your suggestions 
> >before I made my initial post. For example, using a 0.01uF capacitor across the 10 Ohm and BC548(instead 
> >of 2N3055) boosts the current to 125 mA and the 
> >emitter voltage to 500 mV. The emitter output voltage consists of nice square wave
> >and the collector current is also a nuce square wave.  
> 
> The 2N3055 was not even up to audio frequencies with a beta of 20 to 70.. Ft of < 2.5 MHz
> The BC548 is low power with a beta of 200 (-B and -C version even higher), and an Ft of 100 MHz. 
> 
> For the current in an inductor i = t / L 
> where t is the time in seconds the transistor is 100% on (as a switch).
> A 3055 will likely not be 100% on any time soon in less than 1 / 150.000 seconds ;-)
> 
> A current limiting resistor in the emitter will create a current source,
> You do not want that if you use the transistor as a switch,
> unless it is a very low value and perhaps used to detect the current to dynamically trigger a cutoff.
> A small current transformer in the collector for feedback is better.
> 
> 
> Example of current sense feedback to a micro to dynamically current limit Ic every cycle:
>            +
>            |
>            L 
>            |
>       _____|_ 
> feedback |   }||
>         [ ]  }|( 10:1 current transformer
>       ___|__ }||
>            |  
>            c
>       -- b
>            e
>            |
>           ///
> 
> What are you trying to do, circuit diagram?
> 
> 
> I have used IRLZ34N MOSFETS as switch at that frequency.
> There are many suitable MOSFETS.


I agree that an emitter current limiting resistor acts as a current source, but connecting the emitter directly
to ground makes me queasy. So, I have a small resistance, 10 Ohm with a bypass capacitor. This is a side project, and I am trying to understand the workings of an air-core transformer.
Reply by ●February 5, 20192019-02-05
On Mon, 4 Feb 2019 03:50:24 -0800 (PST), dakupoto@gmail.com wrote:


>On Monday, February 4, 2019 at 5:25:03 AM UTC-5, Clifford Heath wrote:
>> On 4/2/19 6:03 pm, dakupoto@gmail.com wrote:
>> > Could some electronics guru here please clarify the
>> > following.
>> > 
>> > I did a SPICE simulation of a coil driver. A simple
>> > astable multivibrator @ 150 KHz provides the signals
>> > @ the bases of two 2N3055 BJTs.
>> 
>> I don't think the 2N3055 is specified to be fast enough for that, though 
>> most recently manufactured ones are much faster than spec. I'm sure Phil 
>> A will chime in if I'm wrong.
>
>The base of each 2N3055 is driven by a relaxation
>oscillator at 150 Kilohertz. What is so fast about that ?


For original 2N3055 the fT was just 800 kHz at small current, so the
current gain for a 150 kHz sine wave would be only 5. The edges of a
non-sine waveform would be severely rounded, hence requiring
considerable base current.  Modern 2N3055 are perhaps 3 times faster,
but still very slow.
Reply by ●February 5, 20192019-02-05
dakupoto@gmail.com wrote


>On Monday, February 4, 2019 at 3:58:36 PM UTC-5, whit3rd wrote:
>> On Monday, February 4, 2019 at 3:50:29 AM UTC-8, daku...@gmail.com wrote:
>> 
>> > The base of each 2N3055 is driven by a relaxation
>> > oscillator at 150 Kilohertz. What is so fast about that ?
>> 
>> If those transistors go into saturation, the rise time when they turn off
>> is very poor, by switchmode standards.  A bit of capacitance across
>> the 10 ohm resistors would help.   So would a different transistor, with
>> lower Rbb.
>
>Thanks, but I already tried out your suggestions 
>before I made my initial post. For example, using a 0.01uF capacitor across the 10 Ohm and BC548(instead 
>of 2N3055) boosts the current to 125 mA and the 
>emitter voltage to 500 mV. The emitter output voltage consists of nice square wave
>and the collector current is also a nuce square wave.  


The 2N3055 was not even up to audio frequencies with a beta of 20 to 70.. Ft of < 2.5 MHz
The BC548 is low power with a beta of 200 (-B and -C version even higher), and an Ft of 100 MHz. 

For the current in an inductor i = t / L 
where t is the time in seconds the transistor is 100% on (as a switch).
A 3055 will likely not be 100% on any time soon in less than 1 / 150.000 seconds ;-)

A current limiting resistor in the emitter will create a current source,
You do not want that if you use the transistor as a switch,
unless it is a very low value and perhaps used to detect the current to dynamically trigger a cutoff.
A small current transformer in the collector for feedback is better.


Example of current sense feedback to a micro to dynamically current limit Ic every cycle:
           +
           |
           L 
           |
      _____|_ 
feedback |   }||
        [ ]  }|( 10:1 current transformer
      ___|__ }||
           |  
           c
      -- b
           e
           |
          ///

What are you trying to do, circuit diagram?


I have used IRLZ34N MOSFETS as switch at that frequency.
There are many suitable MOSFETS.
Reply by ●February 5, 20192019-02-05
On Monday, February 4, 2019 at 3:58:36 PM UTC-5, whit3rd wrote:

> On Monday, February 4, 2019 at 3:50:29 AM UTC-8, daku...@gmail.com wrote:
> 
> > The base of each 2N3055 is driven by a relaxation
> > oscillator at 150 Kilohertz. What is so fast about that ?
> 
> If those transistors go into saturation, the rise time when they turn off
> is very poor, by switchmode standards.  A bit of capacitance across
> the 10 ohm resistors would help.   So would a different transistor, with
> lower Rbb.


Thanks, but I already tried out your suggestions 
before I made my initial post. For example, using a 0.01uF capacitor across the 10 Ohm and BC548(instead 
of 2N3055) boosts the current to 125 mA and the 
emitter voltage to 500 mV. The emitter output voltage consists of nice square wave
and the collector current is also a nuce square wave.
Reply by George Herold ●February 4, 20192019-02-04
On Monday, February 4, 2019 at 2:03:31 AM UTC-5, daku...@gmail.com wrote:

> Could some electronics guru here please clarify the
> following. 
> 
> I did a SPICE simulation of a coil driver. A simple
> astable multivibrator @ 150 KHz provides the signals
> @ the bases of two 2N3055 BJTs. The collectors of 
> each 2N3055 is tied to a coil. The coil is bifilar,
> with the common node connected to a 15 V DC source.
> Each inductor of the coil has value of 0.075 uH. The
> emitter of each 2N3055 is connected to a 5 Ohm
> resistor to ground, as a current limiter. The voltage 
> output at the emitter consists of triangular spikes of 
> of mahnitude 125mV abd the curent through the same 
> node consists odf triangular spikes of magnitude 30mA 
> 
> Couls someobe please tell me if the results are as 
> expected ? I have never worked on coil drivers 
> and such in the past.
> 
> All hints, suggestions are welcome. Thanks in 
> advance.


A schematic would always help.  What are you trying to do with the 
coils?  At 30 mA why do you need such a beefy transistor?  

George H.
Reply by whit3rd ●February 4, 20192019-02-04
On Monday, February 4, 2019 at 3:50:29 AM UTC-8, daku...@gmail.com wrote:


> The base of each 2N3055 is driven by a relaxation
> oscillator at 150 Kilohertz. What is so fast about that ?


If those transistors go into saturation, the rise time when they turn off
is very poor, by switchmode standards.  A bit of capacitance across
the 10 ohm resistors would help.   So would a different transistor, with
lower Rbb.
Reply by Phil Hobbs ●February 4, 20192019-02-04
On 2/4/19 2:03 AM, dakupoto@gmail.com wrote:

> Could some electronics guru here please clarify the
> following.
> 
> I did a SPICE simulation of a coil driver. A simple
> astable multivibrator @ 150 KHz provides the signals
> @ the bases of two 2N3055 BJTs. The collectors of
> each 2N3055 is tied to a coil. The coil is bifilar,
> with the common node connected to a 15 V DC source.
> Each inductor of the coil has value of 0.075 uH. The
> emitter of each 2N3055 is connected to a 5 Ohm
> resistor to ground, as a current limiter. The voltage
> output at the emitter consists of triangular spikes of
> of mahnitude 125mV abd the curent through the same
> node consists odf triangular spikes of magnitude 30mA
> 
> Couls someobe please tell me if the results are as
> expected ? I have never worked on coil drivers
> and such in the past.
> 
> All hints, suggestions are welcome. Thanks in
> advance.
>    
>   
> 


If it's LTspice, post the .asc file.

Recently-made 2N3055s aren't as slow as the originals, but they're still 
slow.  I don't know what speed your model runs at.

30 mA in 3 us from 15V corresponds to

3 us * 15V / 30 mA = 1500 uH

so it isn't inductive.

I suspect that you're dragging down your driver stage.  Try using a 
voltage source to drive one of the bases and see.

When you fix it, watch out for the magic pixel smoke. ;)

Cheers

Phil Hobbs



-- 
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com
Reply by ●February 4, 20192019-02-04
On Monday, February 4, 2019 at 5:25:03 AM UTC-5, Clifford Heath wrote:

> On 4/2/19 6:03 pm, dakupoto@gmail.com wrote:
> > Could some electronics guru here please clarify the
> > following.
> > 
> > I did a SPICE simulation of a coil driver. A simple
> > astable multivibrator @ 150 KHz provides the signals
> > @ the bases of two 2N3055 BJTs.
> 
> I don't think the 2N3055 is specified to be fast enough for that, though 
> most recently manufactured ones are much faster than spec. I'm sure Phil 
> A will chime in if I'm wrong.


The base of each 2N3055 is driven by a relaxation
oscillator at 150 Kilohertz. What is so fast about that ?
Reply by ●February 4, 20192019-02-04
On Monday, February 4, 2019 at 3:11:06 AM UTC-5, klaus.k...@gmail.com wrote:

> I think your numbers are wrong
> 
> 75nH is very low and at a on time of 3us at 15V will give you a peak current of 666A
> 
> Please check your simulation and of possible share a screenshot,  lot easier to help then
> 
> Cheers 
> 
> Klaus


Could you please be a bit more specific ? I would be 
happy to send a screenshot of the emiiter output 
current and/or voltage, but I would then need a email address or something, as this news group does not 
allow attachments.