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: 1685
Posted: 10:36am 08 Jun 2024
Copy link to clipboard
Print this post
The AC voltage regulation from no load to 5kw results in exactly the same voltage on a true RMS meter on my inverter. The PWM % hardly changes.
But my secondary is wound to 280V AC and with primary, will easily hold regulation down to around < 40v - can't quite remember but it's low and I would never get below 52V at over 15kW.
Yes a read to a grounded A7 ADC input is included
Mike did not see your reply until after I posted!! . Edited 2024-06-08 20:46 by KeepISIt's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 10:58am 08 Jun 2024
Copy link to clipboard
Print this post
Dex my only focus was to diagnose where your problem lies in as few a steps as possible. If unplugging the current sensor has an effect or not will tell me (and KeepIS) lots, I will still pass you back to KeepIS but at least it may eliminate a major step that will otherwise have you chasing your tail.
Something may be affecting the VFB voltage seen at the Nano (artificially making it higher than it should be) which causes the code to reduce the voltage to be in balance again. To be honest I am unsure but suspect the % modulation should change lower due to the artificially increased VFB. where as it should have gone to a higher % for losses due to system current/voltage drops or transformer/choke influences. I love servo loops..... Edited 2024-06-08 21:11 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
-dex- Regular Member Joined: 11/01/2024 Location: PolandPosts: 99
Posted: 01:02pm 08 Jun 2024
Copy link to clipboard
Print this post
I disconnected the current sensor cable and then loaded the inverter. There is still the same AC voltage drop, as on the photos below.
On the photos you can see additional power meter, it is at the inverter output, the inverter LCD shows the remaining parameters you asked about.
Please do not rely on the AC voltage from the inverter's LCD, it has probably not been precisely calibrated.
The mod with the link on pin A7 to GND has already been applied previously. Then this solved the problem of zeroing the current on the display. It did nothing to stabilize the voltage. To be sure, I checked again, A7 is shorted to ground.
Edited 2024-06-08 23:13 by -dex-
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 02:02pm 08 Jun 2024
Copy link to clipboard
Print this post
Thanks for that Dex it is very revealing. It tells me the servo loop from the DC input through inverter transformers choke etc to Ac output& fed back around the loop to the voltage feedback is happy. I deduce this because the two modulation percentages are almost the same. If there were impedance problems anywhere the % would increase to make up the shortfall.
I suspect that if you measure with an AC RMS voltmeter at the two pins feeding back the AC output right at the Nano control board it will have exactly the same voltage reading of 230.5 give or take half a volt for the instance of no load and ~8A load.
It appears there is impedance between the 230V transformer output and your multi function meter. It can be roughly calculated by using the difference between the loaded and unloaded AC voltage which is ~ 16V and the current difference is ~ 7.8A that tells me there is a voltage drop between the AC output winding and when it finally is measured by the meter. The resistance you are looking for is ~16/7.8 or around 2 ohms.
Maybe you should grab a true RMS meter and start hunting. Hint, at 8A load, measure (ACV) between the same active wire immediately at the closest point to the toroidal with the other end of the meter to the active out of the multifunction meter, note it - then do the same at the other end of the 230V mains wiring (neutral?) from tranformer to meter neutral out. Add the two noted AC voltages together and I suspect the sum will be damn close to 16VAC. Good luck and be careful probing !! Note the AC output probably passes also through the Mains contactor/relay assuming you installed one, it obviously needs to be closed to measure the voltage drops. Welcome to ohms law. (I hope this is all correct I hate humble pie)
Do you have an AC line filter in the 230VAC circuit too? Edited 2024-06-09 00:33 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1685
Posted: 10:04pm 08 Jun 2024
Copy link to clipboard
Print this post
Dex just out of interest, you might also try it with a purely Resistive load - I have big old 2.2kW electric jug I use.
This way LCD ACV voltage should be within a volt or so the the true RMS Mains monitor. Not going to make any difference to your fault finding, but at least you will be looking at roughly the same numbers on the meters and any handheld meter that you are suing.
BTW That AC monitor model appears to be very accurate against my expensive True RMS meter. .It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 11:02pm 08 Jun 2024
Copy link to clipboard
Print this post
After more thought, that would mean 120W odd watts is going awol (missing) which is a fair bit of energy. If the two voltage readings at no load and under load are different at the connector for mains feedback on the controller board, that is still possible and yet the inverter/servo loop is happy.
The only other explanation I can see is somehow noise is getting into the chain between the precision rectifier input and the VFB at the Nano A0 (pin 19). What I mean is that A0 sees the same feedback voltage for both of the unloaded and loaded conditions that is why the modulation drive % is essentially unchanged.
If this is the case I know you have discovered a huge amount of electrical noise, emanating from the inverter cabinet and into you radio gear, so maybe it is time to look at some noise mitigation techniques. To improve this issue as far as affecting the the controller goes, it can be a combination of noise reduction at the source and also how the noise (or the effect of the noise) induced into the control card can be reduced. An aluminium backing plate ~ 6-10mm from the rear of the controller card is a good first step. The plate does not have to be earthed but the next step is to put a capacitor of at least 10nF from the PCB ground-plane pads to the screw solder pads just near J1. This makes the plate a Faraday screen to outside noise. It can be grounded no problem but the extra capacitor coupling the ground plane to the plate is a very important part to fit.
The last comment I will make is that if you (or anyone with issues) asks for help, I suggest it is best practice to work with that person and despite for instance 2 or more people wanting to help you, all having different ideas on how to solve it and wanting to go on different paths, you need to decide & announce whose help and advice you want to follow for a time to be fair to all the helper/s. I would be relieved for you to choose to follow for instance KeepIS's advice and I can sit back and observe and maybe even learn something. But when too many cooks get involved pulling you in different directions all asynchronously the landscape (ie noise problem) will keep changing in directions that are hard to keep up with and follow for anyone trying to understand or help solve the same problem. Others may have "stuff" they could or should be doing and therefore multiple people are not all spending time and resources designing the same wheel.
I would be one of the first to say ok I need help here if no progress is being made. But what seems like no progress at first may not be the case as every step that does not have the desired outcome is eliminating sections of circuitry or wiring, and I would like to think homing ever closer to the real cure. Edit: I am not saying any advice you have received to date is wrong or on the wrong path but more to focus on one path at a time is usually less confusing. And making one change at a time not multiple changes as then we can better detect what step actually fixed it or started us in the right direction. Edited 2024-06-09 11:33 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1685
Posted: 02:18am 09 Jun 2024
Copy link to clipboard
Print this post
It's all too hard. Mike.
-dex- Regular Member Joined: 11/01/2024 Location: PolandPosts: 99
Posted: 06:03am 09 Jun 2024
Copy link to clipboard
Print this post
Good morning!
So I'll refer to the first answer. There are basically 3 components on the AC source - input pin route on the controller board:
1. EMI filter 2. Fuse 3. Ferrite beads
Everything is connected as in the diagram.
Both meters on din rails and the "inverter's meter" wires are connected at one common point.
RV1 behaves correctly, I can change the input voltage with it. But why does the system do nothing to change the SPWM % when the AC voltage drops?
I could observe large changes in % SPWM in case of significant changes in the DC voltage on the power rail.
Edited 2024-06-09 17:16 by -dex-
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 07:40am 09 Jun 2024
Copy link to clipboard
Print this post
Hi Dex I was waiting for you to do the first test,
"I suspect that if you measure with an AC RMS voltmeter at the two pins (J2) feeding back the AC output right at the Nano control board it will have exactly the same voltage reading of 230.5 give or take half a volt for the instance of no load and ~8A load."
Please measure the AC volts at the connector with an RMS meter for no load and with an 8 Amp load 8 is not critical 6-12 amps will do higher is better. What are the 2 AC voltages ?
Thanks for the pictures and wiring diagram, do you have the control card just some millimetres above the main Power Fet PCB ?
Re your question
"RV1 behaves correctly, I can change the input voltage with it. But why does the system do nothing to change the SPWM % when the AC voltage drops?"
That is what these tests will hopefully determine. The system only has one AC signal source which is then converted to a DC level to determine what % to use, does the AC voltage change at the connector or stay the same? Edited 2024-06-09 18:10 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2155
Posted: 08:13am 09 Jun 2024
Copy link to clipboard
Print this post
Another possible thing to try is to move the secondary winding earth point to the output side of the EMI filter.
As the metering and feedback are taken from there it might minimize any interference. It may also do nothing, but there is only one way to know for sure.
-dex- Regular Member Joined: 11/01/2024 Location: PolandPosts: 99
Posted: 08:56am 09 Jun 2024
Copy link to clipboard
Print this post
Here they are.
No load
~8A resistive load
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 08:59am 09 Jun 2024
Copy link to clipboard
Print this post
Excellent - we are making great progress!
Thanks for the pictures and wiring diagram, do you have the control card mounted just some millimetres above the main Power Fet PCB ? Edited 2024-06-09 19:01 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
-dex- Regular Member Joined: 11/01/2024 Location: PolandPosts: 99
Posted: 09:15am 09 Jun 2024
Copy link to clipboard
Print this post
Yes, about 4cm above power fet pcb.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 09:22am 09 Jun 2024
Copy link to clipboard
Print this post
I believe that is the problem it is noise pick up under U1 & U2. Rather than you mount a whole metal plate, can you fabricate a small copper foil shield that covers the underside of U1 & U2 and the D1, D2 & D3 about 40mm square at a guess and soldered to 3 or 4 earth points 5-6mm clearance is fine. It does not have to be real pretty just avoid the mains connections and check clearance between other points and the shield and that nothing squashes it against other soldered points when you install it.
I would like to make an observation that it would have been better mounted to the side of the main switching parts or at least have a metal shield the same size as the board if you are going to mount it so close to the noise - I have never tried doing that.
The only way that the servo loop can be happy with less volts at the feedback point is if injected noise is adding to the feedback volts causing the voltage at A0 to be the same for the cases of 230V & 214V. More load = more noise - reduce output voltage a bit to be in balance again. Edited 2024-06-09 19:36 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
-dex- Regular Member Joined: 11/01/2024 Location: PolandPosts: 99
Posted: 09:51am 09 Jun 2024
Copy link to clipboard
Print this post
I can add an aluminum plate underneath the entire controller board.
Please clarify what kind of grounding you mean.
Casing is earthed, so the heatsink and controller holder also have casing = earth potential - red arrows.
Should the aluminum plate have the "-" potential of the battery or the earth potential?
And there's one more thing that makes me wonder. If the noise signals the controller to reduce the voltage at the output, I should notice it in % SPWM, and according to what the LCD shows, it remains unchanged.
Edited 2024-06-09 19:57 by -dex-
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 10:00am 09 Jun 2024
Copy link to clipboard
Print this post
I would be happy for either or neither it does not have to be earthed. I would put it to ground if I had to chose. I believe that KeepIS put it to B-.
The plate does not have to be earthed to ground or B- but it must be connected to the PCB groundplane area (through the 10n capacitor you need to fit). Next step is to put the capacitor of at least 10nF - preferably a multilayer ceramic, from the PCB ground-plane pads to the screw solder pads just near J1. The screw should have a metal spacer to the plate once the plate and the ground plane and sensitive nodes are at the same "noise" potential I believe it should cure your problem. Edited 2024-06-09 20:02 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
-dex- Regular Member Joined: 11/01/2024 Location: PolandPosts: 99
Posted: 10:42am 09 Jun 2024
Copy link to clipboard
Print this post
Before I do anything, I want to be sure that I understand what I read, and I'm not sure of that.
Is this what it's supposed to look like?
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 11:10am 09 Jun 2024
Copy link to clipboard
Print this post
It is easier to write than draw the picture. Make a plate of aluminium same size as control PCB
Screw aluminium plate where pcb is now, use spacers to screw the plate to the existing spacers (I hope you have 4 spare spacers female to male).
Now solder 1 and only 10n capacitor on PCB pads just next to J1.
Now screw control PCB onto the spacers sticking up from the aluminium plate.
Put wires/plugs back to control card and retest.
I will draw a picture but not possible for about 1 - 1.5 hour. Edited 2024-06-09 21:10 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 11:24am 09 Jun 2024
Copy link to clipboard
Print this post
I got my 4 year old assistant to draw this.........
What can I say...... drawn with a fingernail on a touch pad....If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 03:48pm 09 Jun 2024
Copy link to clipboard
Print this post
Dex after a lot more thought & calculations I am going to make a prediction that the plate may make a little difference but I suspect not nearly, enough so it is unlikely to fix the problems that we are seeing. I calculated your turns ration as about 1:8.55 what is the real ratio?
There is something going on here that I cannot explain - yet. The fact that with the output at 214V the input was down to 53.3V from 53.6V, the % was down to 71.6 from 72.4 I calculated that the output voltage should be only down by about 4.1V to 226.4 not what we are seeing 214 ie down 16V.
I agree with your comment that if it was significant noise in the feedback signal we should see the % drop further than what we see.
The paradox is that we have a servo system. If the output droops, the drive % increases, if the output climbs, the % drive reduces. Yet we are seeing the output droop fairly significantly and the servo loop % change we should see is ~1.8% change but we see ~7.17%.
I suggest tests are carried out as below just one at a time then re-test the droop.
1) What is the load we are looking at in the above pictures you posted, is it purely resistive ? If not try a resistive load as Phil suggested and re-test.
2)Could you also remove the AC feedback wires from the clip on ferrite suppressors and retest. (leave them off for now)
3) Could you please remove the ground connection marked in a triangle with ! on it ie the node of the transformer leg and the mains filter leg and retest. (leave it off for now).
4) Oh damn we got to 4)...can you please ask anyone else here if they can make sense of what is going on ?
5) Is the software % readout figure lying to us? Is it possible the AC voltage feedback wires to the controller are actually coming from the input to the mains filter not the output (ie right at the transformer) - small wiring error?
I'm calling it a night Dex hope you get somewhere and I see a happy post in the morning. What you are seeing is not normal. Maybe a few more pictures of the inverter construction from a few different angles might help too? Edited 2024-06-10 01:54 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers 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
