Next Gen DRSSTC

[Hydron]

Got a UD3 all boxed up ready to go in my suitcase, will be re-united with my big coil in NZ next week




Will report back on how it goes - last time I used it with this coil was 2019 and it was a bit of a rush and without any case etc, plus lead-time was broken at that point (I still can't quite figure out how that part works other than the initial value offsetting the expected first period loaded into the PWM generator).

Am having to use serial-over-fibre control at this point as I don't have a fibrenet board built into this - it does actually have an ESP32 module inside (and an RF window on the end-plate for it) but apparently this is currently not working? I'll have a look at the ESP32 code but SW is not my strong suit sadly. The ESP32 is certainly the cheapest way to get galvanically isolated comms!

[zix]

That is a very nice looking enclosure!

BTW, we updated the documentation on the UD3 Wiki.  See https://github.com/Netzpfuscher/UD3/wiki

[Netzpfuscher]

Looks cool 

The ESP32 project is not working anymore. Since then we changed so many things, we needed to include a timecode in the frames and changed the packet sequence number from 8 to 32 bit. But with the Fibernet we simplified the network code. It's as simple a encapsulating the protocol frames into a UDP packet. I think I can port it to the ESP32 in short time, but need to find a bit of time to do it.

[Hydron]

Found the updates to the wiki earlier today - thanks! Would have been most handy about 3 weeks ago when I was bringing it up, but I managed with the basic info that was there at the time.

Looking through the info, I have a couple of questions about the Vsense/Isense connections:
- when using the "battery voltage" connection to look at mains voltage, other than a bridge rectifier does this need a capacitor to smooth out the resulting 100/120Hz signal, or is this filtered out in SW?
- where should the DC current sensor be placed? between the mains rectifier and the big bus capacitors, or between these capacitors and the rest of the bridge?

[Netzpfuscher]

I thought I had put a RMS filter in the voltage too, but it isn't. So actually you need a capacitor. Only the current signal has a RMS filter. I think I will add RMS to the voltage too.

The DC current sensor needs to be between rectifier and bus capacitor.


Edit:
Added RMS for the voltage. So theoretically no capacitor necessary.

https://github.com/Netzpfuscher/UD3/commit/a4ba5f52c30d6a502a41842188e9969980612b06

[TMaxElectronics]

Netzpfuscher and me finally got some high power run time on our UD3s, playing some SIDs at our show at the Extraschicht:

We had some issues with the synth filter though, so both small coils play the same tunes, we initially planned to split the notes between the three coils we had, with the other problem being a slight... lets just say "issue" with the third, 3.5m coil

[futurist]

Magic smoke caught in IGBT goop

[Mads Barnkob]

Very nice how and the SIDs are just extremely lovely, wonderful chip sound

The IGBT explosion, might deserve its own thread about the coil in question? Certainly must be a driver issue or pounding it too hard? Water cooling and everything and you still manage to blow of EVERY single die, that is impressive in itself!

[TMaxElectronics]

Magic smoke caught in IGBT goop

Hmm maybe we could pump it back in

The IGBT explosion, might deserve its own thread about the coil in question?

The big coil is from the guy that works at the HV Lab we test our coils at, not mine so idk if I'm qualified to make a post about it. I only helped installing and setting up the UD3

I suspect the problem that killed the IGBTs was just poor HF earth routing. We tested it inside for a long time and had no problems, but we had a very good earth there. When we went to the museum we only earthed the coil with quite a long cable, which caused the UD3 to crash and reset everytime we had a groundstrike and ultimately die (we had to replace the psoc for it on site ). After that the resonant circuit struggled to even reach 1000A in 400us (whereas it easily went to 2000A in half that time before), so something was likely already damaged by that point. Interestingly enough we didn't see excessive heat dissipation in the bridge, which makes me suspect an issue with switch timing but thats just a gut feeling. We didn't have the equipment there to verify everything and just risked a full power run... which didn't go well I wasn't at the controls at the time but I feel like the bang happened almost immediately after we started a song too, when the IGBTs were still cold.

The current limit was set at 1500A with the current peaking at a little under 2000A which is actually within the datasheet ratings of the IGBTs
We used two of those: https://www.infineon.com/dgdl/Infineon-FS450R17OE4-DS-v03_00-EN.pdf?fileId=db3a304336797ff90136c09b45906f3b
all three half bridges in each module were paralleled so I wouldn't be too sure about current balancing and thus the actual maximum current rating in that configuration, but I'm almost certain that the current we drove wasn't pushing the IGBTs.

[pache11]

Anyone have any CY8CKIT-059?  Looks to be about as rare as hen's teeth now.

[Mike]

Anyone have any CY8CKIT-059?  Looks to be about as rare as hen's teeth now.

You can get them from AliExpress, but the price is eye-watering. There's also the "legitimacy" question. I'd expect at least the PSOC itself is probably genuine, even if it's recycled and soldered on a new board.

So far as the more traditional distributors, it's not looking good for the next 12 months at least.

[Mads Barnkob]

Anyone have any CY8CKIT-059?  Looks to be about as rare as hen's teeth now.

You can get them from AliExpress, but the price is eye-watering. There's also the "legitimacy" question. I'd expect at least the PSOC itself is probably genuine, even if it's recycled and soldered on a new board.

So far as the more traditional distributors, it's not looking good for the next 12 months at least.

Is there a thing as virtual ICs? Like there is Virtual machines for computers. As times are right now, very few designs can be replicated by others or even by yourself a couple of years later

[dbach]

I have quite a few Psoc 5 dev boards that im not using (3 or 4 of them)

happy to sell them to someone here.

[pache11]

Are the boards CY8CKIT-059?  Look specifically for a build.   Quote me a price.  Thanks for a quick response!

[dannybeckett]

Couple questions guys. First, I am interested in aquiring a Psoc 5 dev board or anything I can desolder to build the UD3, so if there are any out there I'm more than willing to pay for one

Second, does the UD3 support freewheeling mode?

example:
http://classictesla.com/pslawinski/udplus.html

**post edited for clarity**

[Max]

Second, does this driver support freewheeling mode?

http://classictesla.com/pslawinski/udplus.html

I am a bit confused about this question. The page title literally is "UD+ (UD Plus) Freewheeling DRSSTC Driver"; the first half of the page is pretty much just about the freewheeling functionality.
So yes, it does indeed support freewheeling. If it helps, I can also confirm that the claims are true. I got a UD+ in my DRSSTC and indeed it does what the page says it would do.


Kind regards,
Max

[malte0811]

I think the question was whether the UD3 (the topic of this thread) supports freewheeling, and the link to the UD+ was meant as a reference for what freewheeling is supposed to be. The answer is yes, the UD3 uses freewheeling instead of stopping the pulse on overcurrent.

[Max]

That does make a whole lot more sense. Sorry for misunderstanding.

Kind regards,
Max

[dannybeckett]

Not at all, my bad for the confusing post Max! I was indeed referring to the functionality of UD3 and how it handles overcurrent situations.

Great to know that it freewheels in this situation This controller appears to have every single use-case covered! I'll look forward to constructing and playing with it.

Dan

[dannybeckett]

Quick question, does the UD3 support phase shift QCW control? Reading the docs and it's not jumped out at me yet