FPGA TC driver board / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

FPGA TC driver board

previous 1 2 3 4 next

Move Thread

LAN_403

Intra

Mon Jun 25 2012, 12:35PM

Registered Member #2694 Joined: Mon Feb 22 2010, 11:52PM
Location: Russia, Volgograd (Stalingrad).
Posts: 97

Guys!
I think FPGA driver is already done.
Last year BSVi doing this work on his forum.
Here you can see some tests of this driver

Steve Conner

Tue Jun 26 2012, 10:26AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Well, maybe so, but it's not a lot of use to us if it's all in Russian. We have to do it again in English.

Steve Ward

Tue Jun 26 2012, 10:49PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Wow, you guys have been busy. Not even sure where to begin, but to put in my 2 cents.

If i were making it, i would have probably 2 to 4 super fast ADCs, and maybe 2 to 4 slow ones. As Marko points out, this doesnt have to only be for tesla coils, it could do all sorts of power electronics controls.

To figure out the switch timing, id look at using a voltage transformer off the bridge output, compared with the current output of the bridge. Should be pretty straight forward once you have both of those signals, to compensate the timing however you want, automagically within the driver board. Whether its a delay line or digital filter (i prefer this way, btw) i think is up to the user, the hardware should be able to support either algorithm. And why do i prefer the IIR filter? Well because the operational frequency of the tesla coil will shift somewhat (easily 20%) and the lead-network method naturally compensates this and keeps up, but a delay line would need to be adjusted throughout the bang. Essentially, the IIR filter could run open loop (like the LR predictor thing), while a delay line would need to close the loop with some timing feedback from the bridge output. Oh, and its been way too long since ive done any DSP, so making a filter sounds kinda fun.

I hate toslink, mechanically. We use them for isolating our stage audio, and the cables suck, and the connectors suck harder. Maybe fine for your home theater system, but total rubbish compared to a proper ST fiber in my opinion. Id totally be willing to compromise and just put more than 1 comm link on the thing.

As to the gate drivers, well obviously i like mine. Yes, it runs out of steam below 10khz or something, but you could easily use a bigger boot-strap cap and make it go to 1k, i just dont see the need. High side drivers would be really awesome, but does add quite a bit of stuff to the board to have decent drivers capable of driving bricks and such. Not sure how to resolve this one... depending on the coil i was building i might want direct drive, or i might want GDT drive. The requirements for each method vary somewhat, mainly, i really like to have negative bias on my gate drive for direct drive setups. I will say this though, my "hacked" gate driver (as Marko says

) has been the *only* gate driver that has never let me down. Its never spontaneously failed like many commercial gate drivers have once they have to deal with some junk coupled back into them. Not only that, but they are super fast and lots of current output. If someone's got something better, i promise my ego wont get in the way, but id like to see it tested before investing too much into it.

BSVi

Sat Jun 30 2012, 02:54PM

Registered Member #1637 Joined: Sat Aug 16 2008, 04:47AM
Location: Kiev, Ukraine
Posts: 83

Ok, I have FPGA driver done already :) It's commercial project (*lots* of time, thought and money is there) and soon It will be availibe for sale.

It has phase corrector, dead time generator, lots of different interruptors (including midi with ajustable pw->frequency dependency).

I took videos in russian only :) My english speech is really ugly.

Here is phase corrector demonstration

And MIDI/Internal interrupter demo.

Fortunately, PC-side software is in english, so I have something to show:

Now this driver is in beta-testing stage.

As for hardware side, it uses xc3s50an part with picoblaze core inside. The gate drivers are IXDN609 or IXDN630 with 24v peak gate voltage (48Vpp). These drivers allows for CW usage.

The driver uses either optical (avago versalink) or copper wire method to transfer interrupter signal or midi data.

Steve Ward

Sat Jun 30 2012, 06:16PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

BSVi, that's very impressive! So much control, its great! When can we get one?

BSVi

Sun Jul 01 2012, 11:04AM

Registered Member #1637 Joined: Sat Aug 16 2008, 04:47AM
Location: Kiev, Ukraine
Posts: 83

Thanks :) I think, new batch should be avalible in month or two.

My current plans are:

* I have to respin my boards as I'm not quite satisfied with my current manufacturer's quality.

* I have to write documentation in both english and russian.

* I have to make test rig to test those controllers. It's pretty hard to test them manually.

If you aren't concerned about PCB's appearence, I can ship it in a week. Here what's current PCB looks like:

Solder mask is week and not uniform, it's exfoliates as leads are soldered. The edges of the board are not straight Ðž_Ð¾.

And, yes. MIDI could be transfered via bluetooth (module on top) with no problems at all.

Also, I have no idea how to transfer money from the USA, as paypal doesnt work in the Ukraine. I have to figure out how to do that.

BTW: If anyone, who interested in tesla coils is enjoying euro2012 in Kiev right now, PM me. It's should be really cool to meet someone form 4hv live.

BSVi

Sun Jul 08 2012, 07:43PM

Registered Member #1637 Joined: Sat Aug 16 2008, 04:47AM
Location: Kiev, Ukraine
Posts: 83

Just have implemented something like your phase-shifting QCW mode in my controller. No PC-side software ready yet, but I'll write it one day.

Plenty of logical resources still availible, but for BRAMs, but I have no idea what they can be used for except for code and phase-shift table.

I'm thinking about adding features to this controller before I respin boards, but the more features, the more pricy controller is. What price point for controller like that you thik is acceptible?

I'm writing datasheet now. Its in russian, but you can use google translate to read it:

I have also published schematic:

Samples of controller is shipped now to beta-testers.

stay tuned :)

1341776635 1637 FT140387 Phase Shift Bridge

mikeselectricstuff

Thu Jul 26 2012, 11:39PM

Registered Member #311 Joined: Sun Mar 12 2006, 08:28PM
Location:
Posts: 253

Steve Conner wrote ...

The FPGA I use for my work projects is the XC3S100E-TQG144, with the XCF04S config PROM. It's available from Farnell for Â£10. I made an Eagle footprint for it.

More recent parts (S3A, S6) can use cheap SPI flash to hold the config data, which is about a tenth the cost of the Xilinx part.

BTW Chinese clones of Xilinx' JTAG programmer are available or about Â£25 on ebay.

Goodchild

Sat Jul 28 2012, 08:08AM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

Holy cow this is some impressive work going on here.... I have only worked with FPGAs and CPLDs a little so I can't say I have much working knowledge of the stuff.

I do have a question though, why use an FPGA rather than discreet logic? To me FPGA seems much more complex, however I can see configurability and control being a plus. What I wonder is if it's worth it for all the extra time and effort put into the software/hardware algorithms.

Maybe I just like creating discreet logic circuits mixed with analog as I see it more of a jigsaw puzzle every time I do it for something new.

I look foreword to reading more of this thread as it comes!!!

Steve Conner

Sat Jul 28 2012, 11:38AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Mike: The FPGA I'm using also supports the SPI PROMs. I weighed the cost saving of the 3rd party SPI PROM against the extra development time needed for me to figure out how to program it, and the Xilinx part won. The instrument it's used in sells for about Â£20k.

Eric: You have a good point

IMO, Tesla coiling is a way for electronic hobbyists to do something adventurous and push their limits. Learning about FPGAs could be a part of that. Also, the big boys all use FPGAs to control their electric trains and so on.

Anyway, all I want out of it for the time being is a multiplexed fibre data link.

previous 1 2 3 4 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob