Hello again,

When you said there is no point it sounded like you were referring to the general overload operation so it sounded very strange. That makes you sound smarter now

But what kind of 'power' electronics did you work in.

When you now state different applications you didnt make that clear before this.
Yes a resistive load is different. In the previous post i quoted, the load is much different than a simple resistive load. My brother in law long time ago had the same problem when he tried using a gas generator to run a circular saw. The thing could not get the saw running because it was rated for the current the saw normally draws.

An oversized transformer is not an issue when reliability is the main concern. We did a project for a company i remember they ran their main frame computer i think it was in Newark NJ (this was a pretty long time ago). They did not order an inverter that was oversized. The converter, AND their computer power supply, blew up ... twice.
So again we have a different situation that requires a certain design.

But we had a saying in the lab, "Fools load", as a play on words for "Full load". That's because we ALL knew the general idea that you do not order a converter that will have to operate at the full load, you get an oversized one. This doesnt mean it is not tested for loads greater than that, including a short circuit, it just means the reliability has to be there and to get that it is not a good idea to order a converter that will have to operate at the full load on a continuous basis, which many do anyway and end up regretting it.

I wouldnt mind hearing your take on the same ideas but with a UPS system. For a UPS system you dont order a unit that will give you 5 minutes of run time if you need 5 minutes of run time, you order a unit that has at least twice the capacity.

Note in all of the above we didnt even take the manufacturer's honesty into account yet.

It's nice to meet someone who worked in power but it would be interesting to find out what kind of power devices and maybe max power and what you did there.

Thanks for the reply.

My journey into power electronics started at Marconi in 1983, as a student, aged 18, involved in testing thyristors for the HVDC link between Britain and France. After a brief foray into IC design (because it looked trendier) I returned to power electronics in Professional Audio and Lighting - in those days it was phase-fired triacs switching big halogen lamps, sometimes low-voltage transformer driven, so I know all about inrush currents! Then moving lights with stepper motors. Designing audio amplifiers is an exercise in overload protection circuit design. That industry was destroyed by Chinese imports in the early 2000s. Now I'm in emergency and off-grid power.
It's mainly 3-hour discharge systems for evacuation, not 5-minute UPS, but I imagine the lack of accurate battery data for that length of discharge is a real problem.
I deal a lot with Victron products - it is interesting that their 5kW inverters are rated 4kW continuous, and the transformers look as though they are no more than 3kW, but they have built-in thermistors to monitor the temperature. They also believe in using multiple (as in several dozen) TO220 transistors rather than the larger devices.
Generators have their own problems: Inrush (as you mentioned), but they don't like being run on low loads either, they really don't like leading power factors or highly distorted current waveforms

 

My journey into power electronics started at Marconi in 1983, as a student, aged 18, involved in testing thyristors for the HVDC link between Britain and France. After a brief foray into IC design (because it looked trendier) I returned to power electronics in Professional Audio and Lighting - in those days it was phase-fired triacs switching big halogen lamps, sometimes low-voltage transformer driven, so I know all about inrush currents! Then moving lights with stepper motors. Designing audio amplifiers is an exercise in overload protection circuit design. That industry was destroyed by Chinese imports in the early 2000s. Now I'm in emergency and off-grid power.
It's mainly 3-hour discharge systems for evacuation, not 5-minute UPS, but I imagine the lack of accurate battery data for that length of discharge is a real problem.
I deal a lot with Victron products - it is interesting that their 5kW inverters are rated 4kW continuous, and the transformers look as though they are no more than 3kW, but they have built-in thermistors to monitor the temperature. They also believe in using multiple (as in several dozen) TO220 transistors rather than the larger devices.
Generators have their own problems: Inrush (as you mentioned), but they don't like being run on low loads either, they really don't like leading power factors or highly distorted current waveforms

Hi,

Oh that sounds very interesting.

I worked in various fields really, both testing and then later design work. My field work took me to various places in the US and Canada. I worked with Sandia Labs on solar line tied converters and similar with Perdue University where a professor was doing a study on the effects of harmonics from line tied converters fed from solar panels. Had the opportunity to work with a million-dollar solar panel which took up the whole side of the building. Also worked with Israeli Aircraft and other military projects.

Areas included computer and automation design and programming in several different languages, measurement systems, control system theory and design (which i loved), and of course converters for 120vac, 240vac, and also 400Hz converters for aircraft.
The converters were pure sine converters where we developed our own novel method for generating the patterns as well as phase locking. There were also some high current DC to DC converters. I was lucky enough to work at a place that had it's own transformer/inductor winding department where we could order up just about any custom design.
During that time i saw the interesting transformation from big, high current metal bipolar transistors that cost perhaps $200 each to very large plastic body high current 100 amp bipolars and then the transition to high current MOSFETs all of which came with a much lower price tag.

I did a lot of other stuff too so i had a very wide range of experiences and have my own technical library. I especially liked the study of electrical circuit analysis and tried to analyze just about any circuit i came across. To date i probably analyzed a million circuits either on paper (in the older days) or on the computer using special math software and math software i had to develop myself for specific applications.

These days i am retired and do mostly stuff for students and friends and myself. Right now i am working on telemetry for my automobile which started out as a necessity and turned into a hobby.

It's been interesting. We'll have to talk more too at a later day.

 

The thread starter has already left the discussion.

This discussion reminds me of someone farting in a crowded elevator and then getting out on the next floor, and leaving the remaining riders to deal with the consequences.

 

The thread starter has already left the discussion.

This discussion reminds me of someone farting in a crowded elevator and then getting out on the next floor, and leaving the remaining riders to deal with the consequences.

I was wondering if he wanted the 20kW inverter to run an AC motor to power the DC generator. . .

 

Probably another over unity project

 

Some remarks on the 20 kW. First, seems the voltage stabilization is important between other factors. Thus the DC-DC cascade must making a stable 220 AC and rectifier (working on large capacitor array) revert it to 314 V DC. Then last but no least - how to make 20 kW 50 HZ transformerless. Such transformer costs within order of 100++ kiloUSD. So, first should be buck-boost, water cooled radiator (alu thickplate with drilled channels propped with conical thread - checked, normal thread works too bad)), igbt "bricks" and inductor as most weighty emlement. The cores, sure 0.36 T mighty 1x1x4 inch sized India ferrites (Cosmoferrites) - they are well to mount together in each wished configuration and rather cheap. Last time we bought a plethora of large boxes it cost less than 300 USD or between 1 and 2 USD per piece. The rectifier ought to be based on thyristors what cost less than diodes in that current category and allow to drive half-opened states to avoid short-circuit when charge-up the fully discharged capacitor array. Avoid the inductance of wires between this battery and output cascade, how to is well explained in my five year old article about 120 kW 36 kV power supply design experience, just look for it at google. But output, I guess, may be created as 50 Hz meander (2153 or similar) and then summing to it 33% 150 Hz and 20% 250 Hz and (if formfactor is extra important) 14% 350 Hz (floating cascades on the 50 Hz baseline). Other method is discrimination by time, using higher frequency (let it be 350 Hz) meander, what some cycles summing, some "not awaiking" thus in result making the same signal form. This case yet the output filter is necessary, but it is valuable even at first case as well. Again, that ferrite core and capacitor array. Dont forget about capacitor tan(delta) factor, because the typical 0.01 makes thermal output from 20 kW the 200 W what may harm the capacitors.
However, knowing a normal average houshold here at me, I wonder if one uses 20 kW more than 15 minutes a day. Then sure is more energo-economic to have one small Chinese 3 kW pseudosin inverter for "normal" activities and then a "grand" DIY box for extreme dances.

Some thoughts about "bricks" - I made a querry at Octopart for igbt+ 100A + 1200V and got 249 answers from F3L200R12W2H3B11BPSA1 (121 USD) to
IXYS
IXDN55N120D1
$ 29.572

 

Yes it appears that the original poster had gone away.

 

Hi guys, I am indeed back and not electrocuted (yet)!

A few things before we start. I like wooden telephone poles, they are, as someone mentioned, quaint and nostalgic. Bell Labs/AT&T used to test them off parts of the Jersey coast - I bet some of them are still there, stilted and rotting in the bay. Also, you're all correct, I have no experience with high-power circuits; my experience with electronics has always been with solid state, low power DC circuits. I have no business working with this high-power stuff. I do know enough to know to hide behind a wall when we switch it on. My buddy is more familiar with these things than I, so I'll trust him to keep us out of harms way.

Someone once told me there are no such things as stupid questions, but here, I have a few anyway.

I ended up buying this board off AliExpress: https://www.aliexpress.com/item/100...o.order_list.order_list_main.4.21ef1802P9FCz5

I've put it together and hooked up all the peripherals but I obviously want to test it before I go out and buy an 80kg transformer for the final output stage.

1/1.5. In the very vauge hand drawn schematic at the bottom of that URL, it's asked for a 47uH inductor to run in series with the low voltage side of the output transformer. Do I need to add a capacitor in parallel to complete the filer? Also, as there is no flyback circuit indicated on the schematic, would that suggest that the board has it's own built-in protection mechanism?

2: Is there a way to test the board without the feedback hooked up? I can't take readings without the transformer (i think) because the feedback circuit is expecting 220Vac but only 24Vac should come off the board.

If I test it at a lower power with a smaller transformer, will it still drive the MOSFETs?

Remember, for all I know I'm not even asking the right questions. I'm just looking for some more guidance. And a bit of a break here, kids. Don't be too hard on me. I would have stayed in school but life took a wild turn for me in the 90's and set me on a completely different path. Just tryin' to have some fun here in Cambodia,the land of no grounding rods (dont worry, we drove our own). Sorry if I offend any engineers' sensibility. Much love!

 

Oh, and no, we wont be running it at full capacity. And will probably drive it with 48VDC (fyi). Please forget about the 20kw I mentioned in my first post (as you all knew we would). Cheers

 

Looking at the schematic in that link, I wonder how the LED can work with only one connecting wire?
I see that the module comes with 150A worth of 'insurance'. That might come in handy!
The write-up mentions pipes and tubes. Is the module intended to be water-cooled?

 

Why would a "flyback diode" be part of a sine wave circuit??? That is taking the diode religion way too far. There is no even theoretical reason for any spike because there is no fast reduction of current..
Next, that series inductor is not a filter element, but rather a current limiting element to cover a design oversight problem.
To have an AC voltage feedback signal you can use a very low power size 220 to 24 volt transformer. A ten VA device (24 volts, 1/2 amp) will provide the AC 220 volts feedback signal. Just use proper care with the 220 volts wires.. You may not need a 48 volt source to power it, either. And there is something strange about a 24 volt output with a 48 volt supply. It would be useful to see the actual circuit of the inverter output section.

 

The inverter kit has no detailed specs and does not even show the output waveform.

 

I read the listing again and it seems that there may be a serious gap in the technical background of the writer, calling out a number of "tubes", when, I am guessing, they are referring to transistors of some kind.
And note the repeated statements if NO TECHNICAL SUPPORT!!!
I would take that as a warning that the package has been created by an individual with no more expertise in design than my friend's cat. And I am guessing that the TS will have no more success making it work than the cat would have. It would be very good for the TS to prove otherwise. Even with getting help.

 

Of coure the field tubes they mention are the MOSFETs and the Electrolysis are, of course, the electrolytic capacitors. Unless this is part of a hair growth for men circuit, I believe that's how it was translated from Chinese.

Thanks for the input!

 

My point was that the published description was created by one withoutmuch technical background, and so there may be some major factual technical errors. I recall reading the ad for a "100 watt amplifier system that was powered by a 12 volt 100 milliamp wall wart supply.

 

My point was that the published description was created by one withoutmuch technical background, and so there may be some major factual technical errors. I recall reading the ad for a "100 watt amplifier system that was powered by a 12 volt 100 milliamp wall wart supply.

I agree although sometimes it is just a language barrier.
I have read ads that read that the CPU is Intel, AMD, and Apple-something all in one, which of course is impossible. I have also read that the max current is 10 amps, then later 8 amps. Just people that are not used to technical writing, and also we sometimes have secretaries (male or female) that do not have any technical training at all editing and typing the actual add, as dictated by someone who is more knowledgeable in the field. They may even substitute a zero for the letter 'o' or vice versa. It gets very crazy sometimes.