Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:36pm 11 May 2023
Copy link to clipboard
Print this post
Just started the big loads, on top of the full house load running the washing machine at the moment and one of the fridge freezers is also in run cycle, along with all workshop loads. Then we start the following loads.
Dust extractor @ 523A DC input for 2.5 seconds.
3Hp Compressor @ 540A
No Smoke - Troiod hardly made a sound, a tiny little brief hum.
220ohm across the C-Sense, preset turned right back, I will now advance preset a bit at a time until I get the Compressor to trip, just to check it has some adjustment in that range.
Been running along at around 800watts for over half an hour and the heat bars (where I have sensors screwed in), are still at 20 deg ambient temperature, the Toriod is at 25 deg, the inverter had been idling for a few hours before I switched it over to power the home.
And the slight ripple in the waveform is almost gone, it's just appearing slightly depending on the load variation, sometimes as little as 50 watts change. . Edited 2023-05-12 08:40 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 11:12pm 11 May 2023
Copy link to clipboard
Print this post
This is the waveform at 2.8kW load looks really good to me, again, for someone just reading ONE post, the flat topping is NOT the inverter.
Now that Nasty Heat gun on LOW.
It's blurred as I held the camera with one hand and clicked as soon as I switched the heat gun through LOW to off. This is what I got.
. Edited 2023-05-12 09:13 by KeepISIt's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 11:59pm 11 May 2023
Copy link to clipboard
Print this post
Just for interest sometime try reversing the A & N to the heatgun to see what happens when you use the low setting. As the EG8010 only samples once every 360 degrees you cant know whether the inverter is regulating from the driven half wave or the idling halfwave. I am guessing it will be slightly better one way than the other.
By the way - I looked at those humungous current values you posted and I'm sorry to say you have just voided your warranty
With regard to the flat topping that is usually caused through load rectification followed by a capacitive or battery charging type load as the max energy transfer is taken near the peak, if there is no power factor correction circuitry. Apart from that I dont know - usually our mains supply is even a bit flat topped as well... Edited 2023-05-12 10:13 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 12:35am 12 May 2023
Copy link to clipboard
Print this post
That is correct, the Mains supply is exactly the same flat topped waveform.
I decided to let it run a bit with that waveform and it does not seem to worry it, just buzzes back at you.
I'm rewiring the control cabling at the moment as this inverter looks like it's fine.
I'll try reversing the sense and see if anything changes as soon as I'm finished the changes.
Something to note, the coffee machine and milk frother always made the shed LED fluros flicker, however, 30 minutes ago when I made my daily hit, they didn't flicker at all with your design, and there was zero noise from the toriod and no indication a load was rapidly pulsing up to 2.3kW AC power levels.
Yes, understood, Warranty has been officially stamped null and void by the designer. .It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 01:12am 12 May 2023
Copy link to clipboard
Print this post
I think the secret behind the lack of flickering lights etc is due to the fact I am feeding back a totally unfiltered half sinusoid waveform for the Vfb feedback. This helps ensure the fastest possible reaction to mains events.
Thanks for building it and driving it harder than I could have ever done for quite a while to come due to my energy storage source limitations. This is what it was designed to do but its still nice to see it doing it.
What do you reckon was the total out of pocket costs for all the bits to make up the electronics not including the enclosure just the PCB electronic bits, including controller card. BTW I just received the new controller card with all the extra components including full wave active rectifier etc all done in copper tracks now. I tacked it onto another job and bought 5 PCBs for an extra ~$5 outlay. I will build one up and prove it works as intended after our holiday. Edited 2023-05-12 11:13 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 01:31am 12 May 2023
Copy link to clipboard
Print this post
That new controller sounds nice, next time you order some, kindly put me down for one if you remember.
I'll have to tally up total price, but in reality the biggest cost was the big Caps, really only need 4, the China board only had 11,000uf total, 4 of these is 40,000uf and really high current rating. The only reason I would use eight is if there was any heating. If I have 4 spare I'm happy as I doubt they will get cheaper, or even be available if thing go the way it looks, at least to some of us who, "looked up", a long time ago.
The semi conductors and other bits were surprisingly low cost, even considering the big amount I bought for the future.
Almost finished the hook up, just bringing out a LED and wiring a Reset Switch. .It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 02:03am 12 May 2023
Copy link to clipboard
Print this post
Sometime could you please do an incremental efficiency test preferably with resistive loads ?
For instance a 100 - 250W incandescent lamp load and then somewhere between 2.5 & 5 kW of resistive bar heater/oven element etc load .
Note the DC current and voltage (voltage at the PCB +/- terminals) for both loads and then subtract the smaller power from the larger and do the same for the mains power output for smaller and larger load.
Now divide the (incremental) results for input power change by output power change to get the true incremental efficiency without idling stuff like magnetising currents etc which could/should yield a figure above 95% or more.
The caps I used were from digikey and I think were 8,200u at 80V, I provided for the 4 each side, the cost was ~ $6-7 each from memory. I only ever fitted two to each side and there was no detectable heat rise, however I did not thrash my inverter by comparison to yours, I believe they were sufficient despite the fact I do like your big tower screw in caps.
I assume you put a (1K?) resistor from Q4 collector to drive an LED to ground for latch active ?
It would be easy to install a small connector into the adjacent unused holes for a neat terminal finish.
Another thought only 8 Fets, 4 upper and 4 lower are used for the switching part of each halfwave cycle, the other four upper and lower do the switching for the other half cycle - a sobering thought ?
I always considered that to make a stonking amount of power (basically what you are doing) you could use 2 complete power boards each one with 8 upper and 8 lower for each half cycle, both sides of each power PCB driven at once ie in phase in parallel and the the 2 x outputs per PCB joined together for one transformer connection to each board. Edited 2023-05-12 13:27 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:35am 12 May 2023
Copy link to clipboard
Print this post
Had a frustrating problem where a 24v relay controlled from your solenoid output would just emit a tone, it would not pull in. This after I made everything neat.
If the relay did manage to pull in, the inverter would cycle AC up and down. I traced it back to the Inverter OFF control line on pin 5 of J6.
I had put the cables from J6, pins 1 to 6, into a nice sleeve, only about 230mms long with nothing else near the loom. There appears to be coupling into the OFF pin from the other cables in the loom, it's stopping the enable of the solenoid, and if it does enable, it's causing the SWPMEM signal to turn off and on randomly.
With the OFF line at J6 disconnected it runs fine no AC up down, if I fit a 50K from pin 5 to pin 4 - ground, it will turn on and off, but it still causes the SPWMEN line to cycle randomly (I assume the reason for the part slow start and stop).
I just ran a separate lead from the on off switch to Pin 5, everything works again. It looks like Q7 is picking up noise when in loom with the other connections on J6.
Now the better news, you are spot on, the distortion moves to the trailing side of the sinewave when VFB is reversed and the Heat gun is set to destruct. .It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:39am 12 May 2023
Copy link to clipboard
Print this post
So what is the load spread across FETS on a standard inverter compared to this one? . Edited 2023-05-12 14:40 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:52am 12 May 2023
Copy link to clipboard
Print this post
A .1uf across pins 4 and 5 of the plug, problem solved. .It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 05:53am 12 May 2023
Copy link to clipboard
Print this post
Sorry not much time to talk atm across R49 would be better sometime not urgent didnt think the noise could ever partially turn on Q7, just goes to show...
Most other inverters passing these sorts of current have more than 4 upper and lower Fets I reckonIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 06:32am 12 May 2023
Copy link to clipboard
Print this post
Worried it may take to long to discharge across R49, but then again, a much smaller value cap would be used there.
Well these FETS are much higher rated, hope they hold up, time will tell.
Catch up when you have time in a few months.
BTW in case you missed it: You were spot on, the distortion moves to the trailing side of the sinewave when VFB is reversed and the Heat gun is set to destruct. . Edited 2023-05-12 16:33 by KeepISIt's all too hard. Mike.
noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 512
Posted: 07:54am 12 May 2023
Copy link to clipboard
Print this post
mine has 7 hy4008s per leg of h bridge... at full noise (9kw-9.5kw) the fans ramp up, never see over 50°c(even on a 35°c day) but its working.. haven't had a blow up unless i stuff something up, they are tough.. all sorts of loads ( household) ..
once they set up, VERY tough..I think it works !!
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 08:06am 12 May 2023
Copy link to clipboard
Print this post
The one I just took out also had 7 FETS per leg and totem pole, however, I'm running 4 x HY5608W per leg, these are a bit beefier.
HY5608W 360A, 500W
HY4008W 200A, 397W .It's all too hard. Mike.
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024
Posted: 10:24am 12 May 2023
Copy link to clipboard
Print this post
Thats some Excellent results Mike, well done. Good Efficiency at 300-500w, thats about what it takes to run the house overnight. Cheers Aaron Off The Grid
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:45pm 12 May 2023
Copy link to clipboard
Print this post
Thanks Aaron, I plan on doing the efficiency tests as per wiseguys suggestion to get better accuracy, but first thing I noticed was a power difference on the logging SW and definitely a few % better then the standard board.
It looks like I have good range on the current adjustment now, I backed it off from High testing (540A DC surge) to about half way late last night. This morning I turned on a small Switch mode supply, the LED lights flicked as the inverter current tripped and latched off. This was good as I got to verify that the Reset button is working, pressed reset and the inverter soft started and when 230Vac output had stabilized, the automatic transfer controller switched from Mains power back to the Inverter. A seamless transition with no loss of AC power over that dual transfer cycle. .It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 11:31pm 12 May 2023
Copy link to clipboard
Print this post
Of course this was meant to be the output power divided by the input power to work out the efficiency %
The cap would only discharge after the inverter is set to stop & after the solenoid opens at switch off and removes all power? The time constant for 100n & 10K is ~ 1mS (1/1000) second, on pins 4 & 5 when the solenoid removes all power the discharge is via 100K + 10K or 11 milliseconds - I dont think I have slipped any decimal places....
I was hoping that the waveform distortion might be slightly better by choosing the other half of the waveform to drive the gun. One day if you haven't looked already could you take a look inside the heat gun, is it possible an SCR is being used for the low setting instead of a diode ? That is just a wild theory from looking at the distortion shape. That little oscillation at around the zero crossing looks to be the same as the one you get when the solenoid turns off ~ 1kHz ? Very curious how it looks like the FETs aren't doing anything (conducting) at the onset of the oscillation ? It actually looks like the A&N were swapped at the small voltage sense transformer rather than between the ac outlet and the gun, the waveforms look too identical - which is not a bad thing either ? Edited 2023-05-13 09:38 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 01:16am 13 May 2023
Copy link to clipboard
Print this post
I have two heat sink sensors screwed into the Heat transfer bars, one on the Positive side and the other the Negative.
I haven't mounted a sensor on the heatsink as I can give that a feel at anytime
Inverter has been running at just over 2kW for the past hour, the temperature are:
Ambient temperature 22 deg.
Two of the Heat transfer bars 26.9 deg. (All 4 feel exactly the same temperature)
Toriod 32.4 deg.
Heat transfer appears to be perfect to the Heat sink at this stage, BIG Heat sink is around the same temperature "all over" as the Heat Bars or perhaps just a tad cooler.
The Heat gun has a "Diode" in series with the heater element for Low power.
Changing the A and N over at the Power lead has the same effect as changing the VFB connections over. Just moves it between leading and lagging, waveform appears to be about the same visually.
EDIT 1
Just started the Dust Extractor 520A, sine wave lost it's flat top, did not distort, or ripple, just looked like a normal sinewave as it powered up that load for 2.5 seconds, unless you were sitting in front of the inverter you would have heard nothing, it was virtually silent but for the faintest of a hum sound.
EDIT 2
Got the Infrared temperature gun out checking temps, The DC power connections to the Inverter power stage and connections to the Toriod are all around 26 deg, Power usage has gone up to a constant 2.9kW and temps by 1 deg, the huge back Heatsink is around 1 deg lower than the heat transfer bars according to the Infrared gun, being careful not to target any shiny areas.
The Toriod temperature has turned on the Fans for about 5 seconds a few times in the past 20 minutes as the inverter cycles around 3kW load, the toriod sensor is adjusted to keeps the Toriod below 35 deg. . Edited 2023-05-13 11:56 by KeepISIt's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 02:02am 13 May 2023
Copy link to clipboard
Print this post
Ok all looking good! The incremental efficiency will give the best indication of overall performance.
You have made one of the most powerful inverters on the Forum and yet escaped the drama of exploding FETs and a big toroidal wind exercise. Kudos to you for the lateral thinking of how to apply the 3 Toroidals you had to hand & how it has turned out.If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 02:25am 13 May 2023
Copy link to clipboard
Print this post
Thanks Mike, I'm trying not to mention the lack of any blown FETS in the builds, you never know who might be listening and planning something
Yes, I really want to do the incremental test, BUT, in the meantime, the Inverter has an AC power meter that also logs total kWh throughput, Freq, PF,A ,V ,W ,etc, the DC input also has a meter doing exactly that along with running hours, Bat%, A, V, W, etc.
With the Heat gun on high and a big fan, the meters show 1.92kW in and 1.83kW out and holding constant power power levels in and out on the meters, so 95.3% quick and dirty . Edited 2023-05-13 12:26 by KeepISIt's all too hard. Mike.
Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Print this post
