What Oscillator is in This PS?

There do seem to be some interesting errs in his published schematics. Maybe he's helping to convey the joy of gaining ones own grasp of the 'Concept" by having requiring the implementer to gain an understanding of the circuit in order to make it work.

That's OK. I enjoy such challenges.

I seem to have been able with the help here to get a 9V choked version I decided to simulate well enough. I have added some filter stages to the output. Doubt they were really necessary, the output was closte to 1% regulated.

Since I have had trouble simulating either transformer version well, transformers are tricky for me to model accurately anyway., I guess I just have to try the experimental approach with them.

Fred

Pity. You missed a mathematical tour-de-force, a whole new Field Theory.

snipped-for-privacy@4ax.com

I especially liked adding the thermal losses of the output stage to the *input signal power* to calculate gain. That makes the power gain of all class A amps below unity.

John

cycle

regulate

Since you pointed out a few different versions I went back and simulated various combinations: 3V & 9V supplies, with and without a Q2 base resistor, and 1-100uF timing capacitor values.

All of the combinations worked pretty well--they yield a not-blindingly-efficient little 500v converter that runs hard at first to charge up the output cap, then bangs the transformer a few times a second to maintain the 500v output into a low-current load.

It's cute.

I do not see the super long, power-wasting on-time of your simulations.

My simulations didn't use the 1n4007--I don't have that model, and it's a marginal choice anyhow--so maybe that's why I haven't seen the parasitic oscillations you report.

Are you still getting those?

I don't see the mechanism on the leading edge of the flyback pulse--the oscillator's already cut off, and extra feedback would just reinforce that. Are they in fact false triggers of the oscillator occurring at the *falling* transition, after flyback?

Is D5(*) upsetting the oscillator's front end on the falling edge? This design might be at risk of that.

If so, you could replace D5 with the c-b junction of an MPSA42 to cut the reverse recovery time (and undesired feedback capacitance).

(*) The 1n4007 from Q2(c) to R2 in your posted schematic.

I simulated using a 1:5 transformer with a 120mH primary, with a d.c.r. of 20 ohms.

HTH, James Arthur

cycle

caps

regulate

as

R1

I'm still working with the 10mH choke 22 DCR. I switched to 1n6529 diode model. Some of the erratic behavior disappeared. I switched to voltage tripler, and added RC filter. I had to adjust C2 and the base resistor for best performance.

It is amazing, and counter intuitive, how increasing the base resistor increases the energy stored in the choke. Maximizing the saturation conduction period, I think was mentioned. In any case I can see how the current in the choke ramps up farther with increases in the base resistor value :)

In the simulation I seem to be able to deliver 450V at 9uA with less than 1 volt ripple at an efficiency of about 43% without the feedback loop. I had trouble with that too. That will be further work. I can live without it though.

I got to say, this is one of the neatest circuits I ever seen. It's like a heartbeat.

Best regards,

Fred

cycle

caps

regulate

as

R1

I forgot to mention, I got it pulsing around 18cps, and the erratic behavior I find very similar to false triggering caused by gate oscillations in MOSFETs and IGBTs.

cycle

caps

regulate

as

R1

Also, the circuit does seem sensitive to the load. I see it having to be tuned for one particular tube, or tube of very similar nature.

[snip]

Something we already didn't know ?:-)

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

| 1962 | I love to cook with wine Sometimes I even put it in the food

Dang, all these years I've been buying 5 GHz MMICS that claimed to have 20 dB of gain. I've been cheated! JF has demonstrated that MMICS have no gain.

John

Wha--ut?

Are you putting it forth, that it is the thermal losses that tip the scales?

Here's a plot of this baby delivering 10uA at 748V with less than 1V ripple.

Say there James, do you have a fondness for Merlot?

Fred

Oh yea, I forgot, it's at 45% efficiency.

--
Hoo-boy!!! 

You sure get bent out of shape when someone shows you up as being wrong,
huh?

Seems the last time I saw you this way was when you said it was
impossible to run a relay with a DC coil on half-wave rectified AC and
keep it from buzzing.  Remember? 

I do; I proved you wrong by running five or ten of them that way and
they all worked perfectly.

It\'s about relays this time too, huh?  

Maybe that oughta be telling you something.

In any case, when did I ever demonstrate _anything_ WRT MMICs? 

   
JF

Your equation for power gain adds the thermal power loss of the output stage to the input power, the sum being the denominator of the gain equation. That's why your calculated gains are low by a huge factor. They are wrong for other reasons, too.

John

There is NO SUCH thing as Power Gain... the power to the load is ALWAYS less than the sum of the input powers

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

     Liberals are so cute.  Dumb as a box of rocks, but cute.

High-L, low duty cycle HV inverters are interesting. The first one I know of was a WWII sniperscope supply that ran on one D cell, for months as I recall. It used a watch spring thing to close a contact about once a second, and used a cold-cathode rectifier tube.

John

--
Yup, guilty as charged.

I've been looking at spec sheets for diodes, some of them leak more current than this PS puts out. In the multiplier at least they are sharing the voltage.

Fred

Literally a "flyback"? Reminds me of the old inverters for tube radios in cars where every once in a while you had to buy a new buzzer cartridge.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Just state the power in PMPO and then you can claim 1kW out of a 10W wall wart.

--
SCNR, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Yup, it had a transformer, and was a flyback. I'd imagine it was fairly efficient. I haven't seen the schematic since I was a kid, so I don't remember the details. It was in a biggish book I got from the US Printing Office, about sniperscope IR viewers; cost 50 cents or something.

The car radio vibrator supplies were forward converters, as I recall. Some of the car radios were quite good, and I used them as home radios, with a 6.3 volt filament transformer instead of the vibrator. Sounded great, but the radio transformer got really hot at 60 Hz.

John