 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. |
Forum Index : Electronics : Hopefully? Another 48vdc-240vac Toriod Inverter build.
  Page 18 of 26   |
|||||
| Author | Message | ||||
KeepIS Guru  Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
I tested the Inverter very early this morning, it was very quite in the workshop at that time, I wanted to check just how quite it was running the Aircon on Heat after the changes I made to stop that loud deafening racket it made yesterday. I kept waiting to hear something as it ramped up to 2kw, but there was absolutely nothing to indicate that the inverter was even running, but the Meters didn't lie, and the AirCon was now blowing nice hot air. I went as far as opening the cabinet door and placing my ear against the Toriod, there was not a sound. The Heat gun on low is now down to a quite Hum, instead of a nasty bark, amazing difference  . Edited 2023-05-22 13:25 by KeepIS It's all too hard. Mike. |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 852 |
Am I correct in understanding that dropping back to just the one 10 stack choke ... got rid of the racket occurring when running your aircon on heat? Cheers, Roger |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Yes, that's correct, I have posted a few screen captures of the first few cycles of 500 odd peak DC amps input being dropped onto the Inverter. Just to clear up the heatgun info, when I carried out previous tests with the heatgun on low, there was almost no difference to the waveform or noise, what I had not done correctly for that type of load was to clamp a really heavy cable across the chokes I was testing when bypassing them, so I did not get the full effect. The inverter-drive reverse cycle aircon when producing FULL Heat output was around 20 times as loud as the Heatgun ever was, I mean it was loud!! Quickly bypassing the chokes made a huge difference to it, but once the wiring was bolted down to correctly bypass the choke, the difference was even greater, as in "complete silence" when running the Aircon on Full heat output, that's when I tried the Heatgun again, and it was now down to just a low hum. So the AirCon is presenting a different problematic load when compared to the bad load the heatgun presents, but the choke change made one load completely silent, and reduced the other by around 60%. I hope that makes sense to you, I need a coffee. . Edited 2023-05-22 17:01 by KeepIS It's all too hard. Mike. |
||||
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Thanks, That would be great, I can see there is a chance of low battery voltage during heavy starting loads.  Looks like you have your inverter working well with all your different loads, the 10 stack choke is the go.  I will get on to the other stuff soon and post the results somewhere. Cheers Aaron Off The Grid |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
You have essentially saved me a lot of effort with your discovery search for the ultimate choke. Despite your initial results with the heavy loads and the use of ferrite E cores and second smaller toroid chokes I was having some trouble reconciling and understanding your results. Mind you though, my results were really Nano based so who was I to query what appeared to be excellent results from your setup with the EG8010. Recently though I had already started to get better results when I reduced a couple of turns on my ring core choke stack that was then less than 50uH and I was only running ~ 3-4kW. It appears the sweet spot for the right choke is not minimum idling current but to reduce the choke value a bit and sacrifice a bit of apparent efficiency low down for best performance at high loads. Whether this can be further simplified with a rule of thumb or semi simple calculation is still wanting for more research I think. With your setup the only change I might make would be to have two 5 ring chokes, one for each transformer leg, each having the 2 x choke windings giving the same total inductance of the 10 stack cores. The only real advantage in doing this is I believe there is less noise likely to be coupled from the primary to the secondary. The other advantage is that each primary winding at the choke primary connection shows a half 50 Hz sinusoid, the other side of the choke is the HF PWM "mess". Happy to hear that the inverter & heat gun noises have been tamed now. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| analog8484 Senior Member  Joined: 11/11/2021 Location: United StatesPosts: 108 |
Thanks for the clarification. So, doesn't reducing the total choke value change the resonance frequency from the original tuning (1.5x line frequency?) assuming you haven't changed the AC output capacitor value? It's a bit surprising that a small load can do that but good to know. That's good. I was wondering about it because the PWM waveform/frequency on the legs can be different depending on the modulation scheme. Glad to see the choke placement doesn't make much of a difference. Sorry for that. Will do. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
No, the choke or chokes on the primary winding have no effect on transformer resonance, that's purely a Capacitor and transformer secondary relationship, obviously that include transformer construction etc. In my testing, nothing I did to the primary with Chokes or connections to the Inverter made the slightest difference to the measured 75hz resonance. This was verified with the two toriodal transformers that I was testing, one a single core Aerosharp and the other a 3 stack toriod that's in my inverter. . It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Thank Mike, no need to sacrifice idling current, the idle current is exactly the same. I arrived at 37uH as the minimum inductance sweet point for idling current in my early inverter testing, then I set about making it's saturation as high as I could within the space I had allocated in the inverter cabinet. Once I established the main chokes minimum inductance for the inverter, I went about trying various choke values / combinations to get best average load waveforms. I think I was just chasing my tail doing that, as "the best" waveform changes constantly with numerous loads combinations. We must be mind melding, I have two sets of 5 cores out, I'll be making two 5 stack chokes at 18uh each, I've just got to squeeze that chore in sometime over the next few busy days. Happy travels. . Edited 2023-05-23 10:11 by KeepIS It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Actually that's not a problem, if you remember, once started, those small DC/AC->DC isolated supplys don't drop out until very low voltages, it's just starting some of the mongrels that is problematic. I'll post when I have a look at the value of R I used in my inverter. . Edited 2023-05-23 10:18 by KeepIS It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Mike, I might not have made this clear in my last post to you. You are correct in that I did sacrifice the idling current "slightly" as I went down to 37uH, but I was as you inferred, purposely sacrificing the lowest idling current in favor of the lowest inductance required for the choke to achieve acceptable efficiency at low power. The reason I did that was obviously to increase the saturation point of the the choke for the chosen number of cores, and the number of turns I could fit. It appears that ultimately, and under the worst kind of loads, that this may have been the best approach. A value of 37uH has so far proven itself under huge startup currents as well. There is an increase in peak DC current with 300A to 540A inputs over the first few cycles when using one choke and less inductance, as I suggested it would be, but it's not a lot more, and the AC waveform is better over those first few cycles with 24kW DC input load conditions. My canary in the mine, the UPS, is completely silent now under "every" startup load condition. Once again, I hope all that prattle made sense. . Edited 2023-05-23 18:13 by KeepIS It's all too hard. Mike. |
||||
| nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458 |
The transformer is like many mH's and the chokes are uH so this has little to no effect. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108 |
Good to know. So, it's still not obvious how the chokes interact with loads like your air-con to create the disturbances you observed. |
||||
| tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438 |
Here is an alternate part: 22V - 290V input isolated output: 5V - 15V. One unit I tested starts to operate at 25vdc. Hope this helps. I intend to use it on my PV Array monitoring circuit. Edited 2023-05-24 08:16 by tinyt |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Thanks for the link, unfortunately it's currently unavailable, I'll keep an eye on it though. The units I have are so far proving 100% reliable on startup and low voltage dropout now. . It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Apology for my long ramblings, but it's more a recap for me for when I look back in the future: It is obvious in my mind, but putting that into the correct "exact" technical description that won't be picked apart for every error in terminology is a game that I'm really not going to play, life and time is to short. However, the choke, transformer, and secondary capacitor form a filter that turns PWM into a nice smooth sine wave. The capacitor has other effects on what the PWM bridge sees looking in to the transformer, and of course these helps reduce idle / no load losses. So the choke reduces PWM current peaks and FET switching losses. The choke also buffers the FETs during on / off transitions under high current, less heat in the FETs. Then you have the load variation on the Secondary from a very Stiff powerful Toriod transformer and the effect of this load disruption is seen in the primary, normally wound with a dozen or so windings as thick as your thumb, the primary is driven by FETS switching hundreds of Amps as 50 volts across that winding. The only thing standing between the FETS and the Toriod is that choke. The choke will also change inductance with the load current. Are the goal posts moving? How stiff you wind the primary and secondary, how big the Toroidal Core is, DC coupling losses, wiring losses, and the list goes on, all can effect the exact value of inductance required, and all can effect how severe the Toriod reacts to various complex load conditions. I ask myself, why wouldn't the Toroiod growl under any load other than a straight resistive load. I know there is is a small window of values between too much and too little inductance for a given inverter build, and I could see a resonance in the Toriod output with the AirCon, but only on full heat mode. I had just needed the right load to finally allow me to select the correct value for my inverter. This sort of condition has not been apparent under 96% of the Loads the inverter had been subject to in this installation. . It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Recap on the chokes testing: While there has been a known problem with some Heatguns on LOW on the forums for a long time, nothing had made a racket like that Reverse cycle Inverter Aircon ramping up to MAX heat output. It's silent on its Freeze setting. The other thing I want to emphasize again is this, I had tried reducing the effect of the Heatgun by bypassing chokes, my setup gave me various values to easily to test against. I had simply held a short 4G cable across big terminals while monitoring the difference in AC out and DC input waveforms on a DSO, whilst also monitoring AC/DC currents on meters and audibly listening to the Buzz / HUM from the Toriod. Choke values I could test with were around 69uH, 37uh, 53uH, 32uH and 16uH, but the mistake I had made with testing the Heatgun was not unbolting each choke and then bolting a 2G cable across the connections. When I was holding a short 4G cable across the terminals, the noise and waveform had only changed slightly. The AirCon is where I noted a substantial difference. Once I removed the chokes and clamped a 2G cable across the terminals, only then did I realize my error of holding a short 4G cable across the chokes, the 10% of residual noise had completely disappeared. So today, I set about retesting the other offending loads. A week ago, my wife had used a newish Hair curler for the first time in months, it made a fast pulsing rattling in the toriod, quite loud, a bit worse than the Heatgun, and it also made the LED shed lights flicker like crazy. I didn't bother to investigate this as it's not something that's used very often, and is not really needed in any case, as I had no luck in reducing the Heatgun noise by much on three inverters, why bother with this. This morning we got that Hair curler out and fired it up, not the slightest flicker from the LED lights, and only the "faintest" higher pitch buzz from the big toriod, I actually had to stand right next to the inverter to hear anything at all, this is an amazing difference. Finally, I had bought a new unused Heatgun a while back (garage sale for $15) with variable heat control, it was also nasty like the other Heatgun on low, almost zero difference in fact, so in the past I could see no reason to use it over the old Heatgun. I just finished trying that new Heatgun again, it's completely silent on high and the variable heat control does not cause any noise at all. On low setting there is only a faint hum now. I wonder if they are still using a diode on the Low setting and vary the heating duty cycle down from there? I can see no other reason for it to sound exactly like Half wave conduction through a diode on low power, interestingly, that slight hum on low disappeared when I turned the variable heat back slightly, it looks like this inverter is happy with SCR phase control on both high and low power. Result: I can now use every crappy device we have without any concern for this inverter. . Edited 2023-05-24 11:47 by KeepIS It's all too hard. Mike. |
||||
| Godoh Guru Joined: 26/09/2020 Location: AustraliaPosts: 458 |
Hi Mike, I have three 8010 based inverters at my place that I put together. The only thing I cannot quite get sorted is the overcurrent protection. How did you go about setting the over load or over current settings. So far I have not managed to get hold of any information on how to set these boards up properly. Other than just turn the pots and hope for the best cheers Pete |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
FYI: My Wife has taken quite some time off to help with my recovery. Her hobby in woodworking in various forms, from big to small timber, the workshop is next to my play area, and the gear being used is running start stop all day every day. The inverter may as well be on a construction site, big drop saw, routers, sanders, big grinder and sharpener, thicknesser, huge VFD Lathe, external shop Vac, the bigger gear is in the 2.5Hp to 3Hp range with most of it starting and stopping countless times an hour. At times 3 to 4 devices running at once as it's easier to just keep them running for and hour as it causes less strain / heat constantly starting and stopping big loaded inducting motors, especially the dust extractor. Everything else on the property is also connected to this Inverter, including all my electronic gear in my man cave / workshop. The inverter is silent, not a blink from the LED lights in my area, it just amazes me how good both the inverter driver units have run, but this inverter board and controller of wiseguys design is just running incredibly well. You simply forget that your not on mains AC. The Inverter is sitting not far behind me in this fairly sound proof workshop area, I have to look over to check it's still running as it's absolutely silent, and still no fans starting, but then the only thing they would start up for is the Toriod above 35 deg, and even the toriod is just sitting at 28 deg. . It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Pete, I didn't set this on the China board as it's a lot more difficult due to the current sense point, and if it blew then so what in this case. On Wiseguys design and other builds on here it's pretty easy, but the first few steps may be the way to do it. Likely this may be what you are trying to avoid? so here goes. Set adjustment low until it trips around 1kW. Mark Adjustment point. Repeat above for 2kW and 3kW (roughly). Increases until it's tripping at the biggest load you will likely "START", adjust slightly more until not tripping. "Start" as you know for me is a big induction motor startup. This is that part that you can't do on my china board. Once I established that, I took the board out and fed my signal generator at 50Hz into the IFB terminals, I found that I could easily adjust the input level to the Controller and get the over current Led and circuit to reliably trip at the same point. I noted the Generator output value for each trip point, just to give me some reference points, I have not tried to see if these will accurately tabulate to a simple way to set the current in future, just too many thing going on at the moment. . Edited 2023-05-24 12:44 by KeepIS It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Quick and dirty mod for DC-DC supply to lower startup voltage and drop out voltage. This is very successful on some of these modules, makes a few volts difference to startup. Check max output before and after mod, some of these are not close tolerance in there stated output voltage and do moves a bit as input voltages increases. Soldered a 220K resistor to the SMD that's mounted closest to the N terminal and the end closest to the transformer pins. The other end of the 220K goes to the pin on the SMD Rectifier or the positive terminal of the big CAP. Note the startup voltage is also determined by the load current, 47 to 100 ohm should test it out. Cap moved back for photo, it folds back in place.  . It's all too hard. Mike. |
||||
| Page 18 of 26 |
|||||
 Print this page |
