Quick, cheap, low power 3 phase supply

[cupoftea]

Hi,
We need a 3 phase 115vac , pure sine, delta power source to shovel into a unit...only a few watts......is the attached the easiest way to do this?
LTspice and jpeg as atatched.

 

[FvM]

Too many circuit flaws, e.g. massively exceeded transistor voltage and power rating.
Relation of power supply voltage and achieved voltage swing is mysterious.

 

[D.A.(Tony)Stewart]

Isn't it massive overkill to use 3 phase AC power source for anything < 100 watts?

Why using +/-200V DC to AC to DC just to get 120V at 2W?

This is another case of designing something without specs for cost & performance and then specifying for an "Easy " solution with a bad example.

Considering the frequency of design questions from cupoftea, I would expect he has learnt this by now.

 

[cupoftea]

Thanks, its just to give a basic test input to 3 phase equipment.

 

[D.A.(Tony)Stewart]

Thanks, its just to give a basic test input to 3 phase equipment.

For 1 unit the cost of parts does not matter, but you need Vceo >=500V transistors and you need to specify output impedance Z(f) and load regulation error. This design looks rather high Z.

 

[Easy peasy]

if it's only a few watts, a split 36V DC psu, 3 x small class AB audio amps, a signal source, x 3, phase locked, and three small transformers to go from:

12Vac rms ( audio amp output ) say, line to star point, up to 115Vac, line to line ( i.e. star - delta )

the code for the signal could be done on a PiC or any small uP with enough D/A for 3 outputs

If you use a 5V uP, and sit it at +/- 2v5 you will get just the right level shifted signal to go straight into the amplifiers.

easy peasy

 

[danadakk]

Simple single chip approach to gen the 3 phase using a SOC. Lots
of other resources available for other tasks onchip, see right hand
column.

$15 board would suffice :

Regards, Dana.

 

[BradtheRad]

Self-oscillating chaser (3-stage, half-bridges) creates 3-phase sinewaves to drive a delta arrangement. The load is the delta formation of 2.7 kOhm resistors.

LC arrangements were devised because they yield a sine waveform. However there ought to be a substitute for the 1 Henry inductors. As it stands my simulation wastes current unnecessarily. Anyway it illustrates the concept.

 

[cupoftea]

Thanks all,
@ Easy Peasy...and presumably we should do the "2 phase is 3 phase" version of your audio amp based supply?
(LTspice and jpeg as attached)

Audio amps into transformers seems like the way to go, as you kindly say.

 

[cupoftea]

Hi,

As such, I believe what we need to produce a quick 3 phase delta supply for our test purposes, is three Audio amplifiers using TDA7292 chip....

TDA7292 audio amp driver
https://4donline.ihs.com/images/Vip...5-1.pdf?hkey=6D3A4C79FDBF58556ACFDE234799DDF0

(So we would use fig22 page 17 and put the transformer primary where the speaker is on that schem..we would only use one channel as they cant be paralleled))

We need an Arduino to help give us the three phase shifted input sine waves. We will get the reference sines from filtered square waves. We will then feed each sine into the TDA7292 audio amplifier. The output of that will go into the CTFCS200-18U mains transformer (but input into “secondary”).

CTFCS200-18U transformer

https://www.farnell.com/datasheets/3764833.pdf

We then simply connect the three transformer outputs in delta.

Do you agree , this is a good modus operandi?
(We’d rather buy off the shelf audio amp, but there are none for cheap)

 

[betwixt]

It may not be efficient but it should work.
Suggestion - use a phase shift oscillator (3 x caps, 3 resistors in a single transistor feedback loop) and take the phases from each stage of the RC network. It's much simpler than generating phased square waves and filtering.

Brian.

 

[cupoftea]

Thanks, the attached phase shift osc is doing 60Hz now, but need to adjust it for amplitudes. Also, will need two of them for the 120degs (LTspice and jpeg)

 

[betwixt]

Check the signals at the output of the IC and the top of the two capacitors.

Brian.

 

[danadakk]

Result of post #7. No T and V issues with passives due to onchip timing accuracy
and onchip Vref for DACs.

Took 4 lines of code.




Regards, Dana.

 

[cupoftea]

Thanks Danadakk, ill look into the SOC situation.
At the moment, the analog way looks like only giving one stable sine wave, which needs filtering...hence i am still thinking of the "square wave in to LC tuned to 60Hz".

..3 times, and phase shifted 120degs

LTspice and jpeg attached

 

[danadakk]

If T variation matters your electrolytics look like this :

https://www.tdk-electronics.tdk.com/download/185386/e724fb43668a157bc547c65b0cff75f8/pdf-generaltechnicalinformation.pdf

You could raise your R's and lower the C's until you get into film caps.....
Or the digital approach.,


Regards, Dana.

 

[FvM]

Nothing has be said about using electrolytic caps in the circuit. You surely won't.
RC impedance level could be higher though.

 

[danadakk]

Nothing has be said about using electrolytic caps in the circuit. You surely won't.
RC impedance level could be higher though.

Circuit shows 47 uF, so I assumed electrolytic or tants....

But could be ceramics, like MLCC, class 1 or film but we are just guessing not knowing
OP goal for frequency accuracy over T & V.....

Understanding Ceramic Capacitor Temp-Coefficients

A common question when looking at ceramic capacitors is what do the temperature coefficient numbers/letters mean? These numbers will generally break down to a temperature range and the variation in capacitance over that specific range. The first thing you need to understand with what standard...

forum.digikey.com

Regards, Dana.

--- Updated Oct 30, 2023 ---

Thanks Danadakk, ill look into the SOC situation.
At the moment, the analog way looks like only giving one stable sine wave, which needs filtering...hence i am still thinking of the "square wave in to LC tuned to 60Hz".

..3 times, and phase shifted 120degs

LTspice and jpeg attached

If you want stability using L:C than focus on the L to get high Q.

LC tank Q factor

I am working on parallel LC tank. It is Rp // Lp // C . Rp = Rs( 1 + Q^2 ) Since inductor dominates capacitor under high frequency, Rs is parasitic resistance of inductor. It seems to me Q fac...

electronics.stackexchange.com

Q Factor and Bandwidth of a Resonant Circuit | Resonance | Electronics Textbook

Read about Q Factor and Bandwidth of a Resonant Circuit (Resonance) in our free Electronics Textbook

www.allaboutcircuits.com

Of course yiouw ant a high slope to phase vs freq to achieve stability -

LC Circuit: Parallel And Series Circuits, Equations & Transfer Function | Electrical4U

A SIMPLE explanation of an LC Circuit. Learn what an LC Circuit is, series & parallel LC Circuits, and the equations & transfer function for an LC Circuit. LC circuit analysis involves ...

www.electrical4u.com

Regards, Dana.

 

[FvM]

Circuit shows 47 uF, so I assumed electrolytic or tants....

Which circuit? Post #15 shows 4.7 uF, I would expect film capacitor in the first place.

Capacitor value his however unnecessarily high. It's not clear at all why component values have been selected so.

I feel that none of the analog generator implementations being shown along this thread is well considered. But specs aren't really clear, frequency variable yes/no?, expected magnitude stability?

Regarding recently suggested LC filters, they are feasible but not very practical for 50/60 Hz. Better use active RC filter. I don't see a particular advantage in bandpass usage, low pass works as well.

 

[D.A.(Tony)Stewart]

V240 has a gain a bit more than 2 to restore the Vpp amplitude of the input lag filter.
V120 is the inverted sum of V0 and V240.
V0 & V180 are both unity inverters.

This result works for 60 Hz