EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: RRRoamer on October 06, 2018, 05:59:51 pm
-
Folks,
I'm missing something on this test circuit. Earth ground comes in on pin 2 of J3 along with neutral (pin1) and line (pin 3) and I have verified that all three pins are the correct "signals" through the IEC connector, switch, etc. One note: this is line driven. There is no upstream transformer in this circuit. It is using rectified mains power. Ok, note actually true during testing as I am using a autotransformer to keep the magic smoke in.
The problem is that when earth is connected, I am getting a great deal of current flow through, even with the PFC disconnected (and Qswitch1 and Rsense1 removed from the PCB just to be sure). It is showing an effective impedence of around 2 Ohms as when I dial up the variac to get 12VDC on the capacitor bank, I am getting around 5 amps through the fuse holder inline with the hot (pin 3 J3).
When I disconnect the earth ground from DC ground, the current draw drops right down into the 1 to 2 milliamp range even with the variac turned all the way up and the DC voltage sitting at 167VDC, which is what I expect from the circuit at this point.
I've checked Cin1 to make sure it wasn't shorted (it wasn't) and a few other simple tests to figure out what is going on. I've also measured the neutral to earth voltage in my electronics room and it is only 0.7VAC. When I ohmed out connector J3, all legs were in the megaohm range. I didn't find any low impedence paths between line and earth or line and neutral.
I have a few more things I want to try in the morning, but I wanted to see if anyone here can point out the (probably) obvious thing I am overlooking.
Any ideas where I am screwing up?
-
Negative side of bridge rectifier connected to circuit 'ground', now consider that mains ground and neutral are bonded at the incommer in most supply installations, and what happens when the L input swings negative WRT ground.
There should be NO GROUND anywhere near that circuit (Well possibly via a class Y cap to somewhere for EMC, but....).
I am a little worried by someone building a PFC input stage and thinking that bringing ground anywhere near it is a good idea.
Incidentally both line to J2 must be fed from a floating supply as well.
Regards, Dan.
-
Thanks Dan.
That does confirm what I have been thinking and researching. I'll disconnect that AND get an isolation transformer as well. I wouldn't be able to connect the scope without it.
-
dmills is right. A way if seeing this would be to show all points that are at the same potential. I quickly drew in red these points on the input circuit of your drawing that are basically connected together. You can see that the lower left diode has ground on both sides so it isn't doing anything in the circuit at all. The upper left diode is right across the line so on negative half-cycles it is shorting out the input.
My concern is that if you are having trouble with this circuit, which is apparently going to be high voltage and high power, perhaps this isn't a wise project for you to attempt. Here is the application note using the L4984 chip that has some good info.
https://www.st.com/content/ccc/resource/technical/document/application_note/46/55/97/5d/a0/21/43/62/DM00064513.pdf/files/DM00064513.pdf/jcr:content/translations/en.DM00064513.pdf (https://www.st.com/content/ccc/resource/technical/document/application_note/46/55/97/5d/a0/21/43/62/DM00064513.pdf/files/DM00064513.pdf/jcr:content/translations/en.DM00064513.pdf)
You will notice in the application note that they show 3 resistors in series every place the resistor string is placed across line level or higher voltage. Resistors have voltage ratings too so just one with 3 times the value to replace the three is a very bad idea.
-
ArthurDent,
As soon as I read dmills post I slapped my head. It's been a LONG time since I was in school for this (ended up as ME who plays with electronics and MOSTLY low voltage DC microcontroller stuff at that). I don't know what I was thinking when I decided to connect earth to DC ground.
Oh wait! I wasn't!
As for the resistors, they will be dropping nearly 300VDC in use. I chose to use single high voltage resistors (400VDC rated) instead of 3 regular 150VDC rated resistors. I had to order the parts in anyway and I have never liked using series/parallel parts to achieve the rating I require unless it is just necessary (parallel electrolytic caps for example).
When/If I ever build the final part (200VAC servo drive), I'll probably step up to the 1210 variant as they are rated for 600VDC to get a bit more margin on it.
And I actually agree with you on the "wise project" comment. But, that is one of the reasons I am pursuing it: to get out of my comfort zone in electronics and to scrape off some of that rust. I DO know this thing can kill me dead in an instant. And I do treat it accordingly. That's one reason I stopped and started this thread when I saw something that I didn't expect and my brain failed to call up why I was being stupid.
-
As for the resistors, they will be dropping nearly 300VDC in use. I chose to use single high voltage resistors (400VDC rated) instead of 3 regular 150VDC rated resistors. I had to order the parts in anyway and I have never liked using series/parallel parts to achieve the rating I require unless it is just necessary (parallel electrolytic caps for example).
A few of reasons to use several series resistors:
1. Increase the voltage rating, increased wattage rating
2. Increased creepage/clearance distances
3. Failure of a single resistor still leaves 2 resistors in circuit
4. Can often use regular resistors instead of special larger package or surge rated ones
5. Increased area of thermal dissipation