UD2.7 Gate Signals

[coilinator]

Hi Folks,

first of all I want to to give you some backgrounds because this is my first post here: My name is Andi, 25y, from Germany. I do Coiling for some years now and also builded a few DRSSTCs, VTTCs and so on. But you know there is allways something new

So the curent typ should be the QCW. I build up the needed buck converter and also the the drivers. I use the UD2.7 with phase lead, but without the FODXXX N P  smd bridges, instead I put in an UCC and some P and N smd fets, it should be the same as steve ward´s driver from 2008 I think.
I managed to receive a nice clean signal with no ringing (thanks to tvs diodes) on the gates of the dual fullbrdinge (8x60n65smd - something about 40nf gatecapa. in total o.O) And because the QCW has a super high switching f of about 400khz and more, its not that easy to drive them quick enough !

Today i scoped the gates while the coil is running. I see some super bad things, like spikes, ringing and so on but what makes me realy mad is something like here:
https://highvoltageforum.net/index.php?topic=569.0

the second pic, if you look at the ending of the signal, it will not instantly return to zero but makes such a kurve and then goes zero. Looks like som discharging kurve maybe the dc coupling cap in the GDT prim line ? but the two n fets are switched if there were no signals on the ucc so the cap will be discharged. My signal is nice lined, just the ending is similiar to the pic.

so does any body know where this is comming from ? I wrote also that these n p brdinges are very sensitive to  to many gate voltage so that they have many cross conductions ?

Iam familiar to ringing at the end but i think this is no ringing at all..


andi

[T3sl4co1l]

I don't know what else to tell you... the answers are in the very thread you linked

Tim

[coilinator]

Hi,

thanks Tim for the quik reply ! I also uses Loneoceans schematic.

Ahh, so hey, this is the circuit..? http://loneoceans.com/labs/ud27/UD27Cschematic.png
Hah, not bad at all, and there you have it, just a component value change will do.  C21 and C30 are 2.2uF, in parallel with an R+C of less, not more!  No wonder it's bouncing.  Increase C24 and C31 to 4.7-10uF, and probably decrease R19 and R24 accordingly.  Still won't fix the saturation, for which more turns are needed.

  So you mean I just have to change these Komponent values and the strange swing at the end will be gone ? That would be super cool 

Unfortunately I also have some trouble with the comparator, also the LT1016. Also the invertet Output on pin 7 shows a strange looking hook at the start and then the Amplitude falls to about 4V, so its still good for the AND gate but its a bit strange. The are two other things: on the 3rd pin should be a rectangular signal which comes from the diode-clamp.
Its just about 500mV in amplitude but with about 1V DC offset.

The other thing is, because I choose two way to different n and p fets for the gdt drive, there is cross conduction and  the whole thing puts out noise all over the board 

The first thing may be the signal generator, it just puts out 10V pp max but the feedback CT should do about 50V.

But whats about these conduction noise ? Does anybody had such problems too ? Can I do something against, some filters or have I just to look for new fets that matches better  ?

reguards,  Andi

[T3sl4co1l]

About the comparator, only thing that comes to mind is lack of supply bypass, or ground bounce, which also depends on layout.

Tim

[coilinator]

Hi

yesterday I viewed every thing twice because there has to be an error somewhere. And while grabbing around on the board I feel something warm on my finger but not the 5V V-reg or something like that it was the comparator it selfe 
Its not driving any load at all so I can´t imagine that this is its normaly state... It makes sense because the inverted output is doing this hook thing and it seems that the IC is switching, according to the feedback but not at the right time, maybe its because it can not detect the right voltage level any more, dont know.

So to be on the save side, a couple of new TL3116 are on the way- also run out of other ideas 

If the comparator do not thinks right, i´ll ending up having the phase lead induktor´s ferritcore drilled out because it´s doing nothing when turning it 

I also come up with a new GDT drive stage ! There for I also bought some FDD8424H smd drivers which originaly build in, in Loneocean´s UD2.7  Because they contain a pair of good matched P and N fets, the problems about crossconductions should be eliminated.

The new comparator is now the more significant part so the "ringing end of burst" thing has to wait meanwhile.

I´ll give you a update such as the parts arrives ^.^


reguards, Andi

[profdc9]

I don't know if it would help but I made a through-hole version of UD2.7 that might be easier to work on...

https://github.com/profdc9/DRSSTC-PCB-Pack/tree/master/ud27c%20thru%20hole

I also made a slightly more featureful one I call UD2.9

https://github.com/profdc9/DRSSTC-PCB-Pack/tree/master/ud29

Dan

Hi

yesterday I viewed every thing twice because there has to be an error somewhere. And while grabbing around on the board I feel something warm on my finger but not the 5V V-reg or something like that it was the comparator it selfe 
Its not driving any load at all so I can´t imagine that this is its normaly state... It makes sense because the inverted output is doing this hook thing and it seems that the IC is switching, according to the feedback but not at the right time, maybe its because it can not detect the right voltage level any more, dont know.

So to be on the save side, a couple of new TL3116 are on the way- also run out of other ideas 

If the comparator do not thinks right, i´ll ending up having the phase lead induktor´s ferritcore drilled out because it´s doing nothing when turning it 

I also come up with a new GDT drive stage ! There for I also bought some FDD8424H smd drivers which originaly build in, in Loneocean´s UD2.7  Because they contain a pair of good matched P and N fets, the problems about crossconductions should be eliminated.

The new comparator is now the more significant part so the "ringing end of burst" thing has to wait meanwhile.

I´ll give you a update such as the parts arrives ^.^


reguards, Andi

[coilinator]

hi

thx Dan for posting the versions - nice work ! my board is also a "through-hole" board design, because I like this "super tiny oh crap its falen down where is it..." things not that much - or maybe cleaning the workbench would do ? 

No but serious it is just a little bit more simple like you said, e.g. changing an IC, just pull instead of switching on hotair.
But of course if every thing is running its a nice thing to have super compact board...

Unfortunately I´ve not taken pictures from the comparator output but it looks a little bit like a gate wave from, at the beginning it has this little nose/hook and then its strait again, so no real ringing at all. This goes into the AND-gate but it seems that the AND not cares about this signal-form because the ampliutude of 4V is more than enough.

But we will see, think the parts will arrive til Wednesday or so 


Andi


 

[coilinator]

Good Morning Guys,

sorry for double posting, but there are some super gratifying news about the driver (and hope there was enough time between the last post  ) !!!

I managed to put in the new P-N  smd moduls, FDD8424H wich are very close matched complementary FETs in one caseing.
This improved the switching times by a lot, there are almost no crossconducting left, so no spikes !

I noticed a strange bang sound if if wound up the buck voltage poti. I thought it is the OCD limiter but the indication LED stays off so it hast to be something different. Yesterday I checked the LED - and ques what- the LED was not able to light  I was completely confused- but this problem was fixed by adding new LED... now i clearly see  that the OCD trips if the current exceeds the 70A that are set.

Also the phase lead funktion seems to work now, but I added about 80uH instead of say 10, Gao wrote to be good for to247 IGBTs... but this should not be a problem ^^

I know that such a QCW coil draw much few amps than regular DRSSTC will do. So I have to set the OCD little bit higher then...

But there is still some thing left to be considered: At the end of the burst, on the gate signal, there is still some little nose of about 6V, just when the signal is turned off. Is this the GDT ringing ? I think this can be solved by adding resistor to the DC cap of the GDT primary but I have not tried it yet.

I will also open extra topic for the QCW when this little problem is eliminated  The coil puts out about 80cm but this cracking sound afraids me and also the output is liminated by the OCD but hopefully I will break the one meter barrier !

reguards, Andi

[profdc9]

You might want to consider the UD2.9 version I posted earlier.  You can see the results here:

https://highvoltageforum.net/index.php?topic=346.0

It is a "skip pulse" version so that, unlike the UD2.7 where the spark stops growing when the overcurrent is tripped, the skip pulse will resume exciting the primary once the current falls below the threshold.  It will keep going on as long as you keep the interrupter pulse on.  It might be one way to grow the spark bigger.

[coilinator]

Hi,

thx for the post ! I don´t know much about pulse-skipping but like the name says the controller will skip pulses to bring down the Primary current so that the coil has not to be switched off completly ? That sounds also interessting but I think my current results are very promising ! I scewed up the ocd to 120A and the cracking noise disappeared ! The bridge stays cool, also the Buck IGBT-Brick heats just a little bit (not more than 30-40°C I would say). I can achive more than 1.20m Streamer from the 20cm coil (about 6 times the secondary length) and it is not tuned perfect at all- but seems that I catched the upper-pole quite good ^^.

There is just a little bit of "splitting" in the streamer but this may settle by varying the ramp(up) a little bit !
Also the nose at the end of the gate signal is still there,  have to do something against it as soon as possible.  So what you guys think will be the better way to "snubber" these ringing, adding large electrolytic capacitor parallel to the DC coupling cap or do the same thing with a resistor, say 20Ohms ? I know that mr. Ward allready suggested the resistor-variant here. 

But until now it seems to work just fine ! I may buy a larger tunable induktor with the 80uH because now I soldered in some nontunable one but you have to be able to adjust the phase lead while tuning the coil...so this is not the best solution 


reguards, Andi

[coilinator]

Hi,

today I scoped the gate signal and the "nose" is still there 



Hope the pic is not that bad... what do you mean, is it a kind of overshoot, ringing ?  My coupling capacitor is about 4.7uF . I added also 47uF bipolar electrolytic, parallel to it but with no other result. Signal is scoped at 10V/div and 2us.
It look like the signal will go on and just turns off little bit to late ?


Andi