Optimum way to drive a transformer Primary for minimum heat loss

Thread Starter

cipherman

Joined Jul 10, 2015
9
Hi everybody, I am the newbee on the block and have a simple question that I think might have a complex answer. I would like to know..."How can I drive the primary of a standard transformer the most efficient way?" A Push/Pull driver would be my choice, but there are Center Tap options and Center Tap Output options that could apply.

The objective is to build a 3kW, 48vdc to 220vac/50Hz Inverter with minimum Heat Generation. Also, to confound the project with additional complexity,..."Is it necessary to have a Feed-Back winding in the Secondary (Output) to allow for Output Voltage adjustment by feeding it back to the Input Circuitry?"

I realize that this might be an old question for some (most) of you, but I consider it a worthy in order to finally put this question to bed.

If you had everything you needed (no superconductors please) to build an optimum Inverter...how would you do it.

I hold everybody in the highest esteem and look forward to your valuable comments.
 

AnalogKid

Joined Aug 1, 2013
10,987
If you mean linear drive, then a pure sinewave is most efficient because it has no harmonics. But for minimum heat generation you would need a high frequency switcher. In either case you're looking at 80 A of primary current and 80% efficiency for the switcher, much less for linear. One guy working alone, I'd figure 6 months minimum if he's experienced in high power power conversion, probably more like a year. Resonant techniques can add 10% to the efficiency, but double the development time.

ak
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
Hi AK:
If I use High Frequency techniques, I would still have to get back to 50Hz. I dont favor a High Frequency/Transformer-less design because my experience has shown that they are too fragile when driving high inductance motors. I would prefer a straight transfer from Primary to Secondary in a transformer, not easy to break this as the electronics are galvanicly isolated from the load.

If I digitally generate a Pure Sine Wave (close enough) using analog techniques and high current followers/trackers, I can use the DC side as my working voltage rails. It does, however, still generate heat loss across the Follower Transistors. I could use a PWM circuit and pass it through a Low Pass Flter to generate the Sine Wave and then apply this to the transformer primary.

All very interesting stuff (I could go on for hours), so I must limit this discussion to the following question. "What is the finest approach to solving thos problem"...dont be afraid to be innovative, I have considerable resources at my disposal.
 

cmartinez

Joined Jan 17, 2007
8,220
Hi AK:
If I use High Frequency techniques, I would still have to get back to 50Hz. I dont favor a High Frequency/Transformer-less design because my experience has shown that they are too fragile when driving high inductance motors. I would prefer a straight transfer from Primary to Secondary in a transformer, not easy to break this as the electronics are galvanicly isolated from the load.

If I digitally generate a Pure Sine Wave (close enough) using analog techniques and high current followers/trackers, I can use the DC side as my working voltage rails. It does, however, still generate heat loss across the Follower Transistors. I could use a PWM circuit and pass it through a Low Pass Flter to generate the Sine Wave and then apply this to the transformer primary.

All very interesting stuff (I could go on for hours), so I must limit this discussion to the following question. "What is the finest approach to solving thos problem"...dont be afraid to be innovative, I have considerable resources at my disposal.
You might be a newbie but you seem to know far more about the problem you're tackling than me, at least... I'll be watching this thread closely. In the meantime, I have a question. Is there a specific duty factor for this transformer of yours you'd be working with? Or would the transformer be used intermittently (with standby periods) for your application?
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
The applications would be for 2hp water pumps, so they would run for about 1-hour and then rest for 2-hours.
As you know, there are many different ways to skin this cat. I have designed hundreds of different systems in by day and have been a consultant for NASA on the space shuttle. I get success because I dont consider that I know everything, humble is the way to approach any project, I think.

I am curious to see what the member s of this Forum think. Remember, this is a very familiar requirement in electronics and power systems. Wouldn't it be interesting to find the "Very Best Approach" to this problem.

I actually think I could prototype the driver in 3-weeks from the GO Time and finish the project in 2-months, including all the heat sink and enclosure design...I have a lot of time on my hands and nothing better to do...give it your best shot.

I do hope others Jump in here.
 

cmartinez

Joined Jan 17, 2007
8,220
The applications would be for 2hp water pumps, so they would run for about 1-hour and then rest for 2-hours.
As you know, there are many different ways to skin this cat. I have designed hundreds of different systems in by day and have been a consultant for NASA on the space shuttle. I get success because I dont consider that I know everything, humble is the way to approach any project, I think.

I am curious to see what the member s of this Forum think. Remember, this is a very familiar requirement in electronics and power systems. Wouldn't it be interesting to find the "Very Best Approach" to this problem.

I actually think I could prototype the driver in 3-weeks from the GO Time and finish the project in 2-months, including all the heat sink and enclosure design...I have a lot of time on my hands and nothing better to do...give it your best shot.

I do hope others Jump in here.
Maybe this stuff is too basic for you, but @ronv helped me design a 12VDC to 120VAC power supply a few months ago... you might want to take a look at how he did it. He did not use a sinewave, since high efficiency was not a requirement, but maybe you'd still like to see our discussion.
Take a look at the LTspice file that I've attached.
 

Attachments

Last edited:

Thread Starter

cipherman

Joined Jul 10, 2015
9
Thank you gentlemen...I will review.
No matter what the help, I will keep,everybody informed if I get info outside this group...we will all learn something.
 

AnalogKid

Joined Aug 1, 2013
10,987
Both of the techniques I mentioned have fully isolated transformers. The linear one is pretty straightforward in concept - grow a sinewave, amplify it, drive it through a really big 50/60 Hz transformer. The switcher is basically like any other switcher, except that the reference voltage is a sinewave rather than DC. Sinewave, high frequency PWM, switching drivers, high frequency high power transformer, filter on the output to recover the sinewave.

Can't you just buy this? Multi-kW AC power sources are switch mode these days. They do exactly what you want, except that they start with rectified/filtered AC line voltage. Somewhere between the automotive market and the test market there's got to be a 48V input flavor. Telecom, probably. Or an industrial-sized UPS with a 48 V battery bank.

ak
 

nsaspook

Joined Aug 27, 2009
13,082
Can't you just buy this? Multi-kW AC power sources are switch mode these days. They do exactly what you want, except that they start with rectified/filtered AC line voltage. Somewhere between the automotive market and the test market there's got to be a 48V input flavor. Telecom, probably. Or an industrial-sized UPS with a 48 V battery bank.

ak
These are very common in the off-grid solar industry.
http://www.theinverterstore.com/7000-watt-power-inverter-48-volt-240vac.html
http://www.wholesalesolar.com/products.folder/inverter-folder/Magnum_MS4448.html
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
I am sure that I could buy a product from a source with no idea of the quality, no idea how it worked, impossible to service myself, long repair turn-around time and a million other bad things to contend with...If I were to just buy it, we would not be having fun on this forum.

Almost everything in made-in-China. It is not easy getting ownership of a manufacturing problem from China...they hate to part with money. Also it takes a long time to get a repair turn-around.
 

Lestraveled

Joined May 19, 2014
1,946
You could buy at least 5 inverters, (1 actual with 4 spares) for the money and effort you would spend designing and building one custom unit. Why don't you buy 48V pumps? What is missing in this equation??
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
It is uncustomary for me reply to such things, but this one I must reply to: "whats missing in this equation...". I have a very large order for a good inverter that I can service myself in my own electronics company. The initial order after testing the prototype is 500/month. Considering I use 300/month of the Chinese failure products, this is a good monthly quantity.

If I were to buy 4-inverters and keep 3-spares for the inevitable failures, the revenue from the initial sale would be dissolved on the freight shipping the next 3 units sent.

I believe that knowledge is power and NOT building this would be a mistake...maybe we should adopt the attitude of actually learning nothing and buy everything from a foreign country...assuming we have space for all the failures we replace.
 

Lestraveled

Joined May 19, 2014
1,946
Thank you, that was what was missing in the equation. The scope, volume, and your intent. This information will improve the quality of the posts to your issue.
 

nsaspook

Joined Aug 27, 2009
13,082
Companies like Magnum Energy make inverters in the USA if that's the main objection. I hope you have the resources for a successful product development cycle.
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
OK...what a pity. I started out asking a question that I thought would generate some interest and interesting threads. It started out well with useful comments from "AnalogKid" and then it quickly degenerated with the following posts.

Comments from: Lestraveled had nothing to contribute to the threads developing in this forum
Then to my surprise a "Senior Member" nasaspook contributed nothing to the "forum question "

LET IT BE KNOWN:
I remove this Forum Question as it is apparent that I am asking the wrong people. A note to everybody, kindly re frame from commenting further. If there is a way I can remove this topic, will somebody please advise...thank you.
 

Lestraveled

Joined May 19, 2014
1,946
Yes, it is a pity, but not the way you are presenting it. You come on to this forum. You start by giving limited information. You say you were a consultant for NASA. (I really have worked with NASA engineers.) People press you for more information and then you say "what a pity", why didn't all of you fall for my B.S. HaHaHa , Yea, too bad, Go back to NASA.
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
My question was "HOW WOULD YOU DO IT". How is this limiting information...there is NO information given...none. It was a question for YOU to answer....dont you understand English now?
 

Thread Starter

cipherman

Joined Jul 10, 2015
9
OK...I have disabled emails so we wont be receiving any more childish remarks from this group...go back to NASA, not a good idea there are 10 managers for every engineer...bye everybody.
 

Lestraveled

Joined May 19, 2014
1,946
This is not that kind of a forum. We assist people here that are trying to really do things. You have to show that you that you have tried something and really need help to make it work. If you would like to tone down your swagger you might get some assistance here. Do you even know how to solder??
 
Top