Sig's Universal SSTC Logic Controller Board and Deluxe SSTC Kit Sale Thread

[Sigurthr]

Excellent, glad it got there safe and sound. Did you like the rigid support I used? haha! It was all I had on hand that I could think of. My fear was that some idiot postal worker would sit on it and crack it.

The population cost was a major factor in my decision to design my own board, as every other board I could find out there required a much larger BOM and parts cost. You should see what it costs to populate a PLL driver >.>, or worse a UD driver, it hurts the wallet.

 

[Speedy78]

I believe thats the metal base from a radiodhack project box or something similar lol.

 

[Things]

I've had some interesting effects from driving a coil with a much longer pulse width than it was intended, you can get some really weird fiery arcs. Luckily with my coil though the bridge was charged via a small transformer acting as a stepup, so running it like that just pops a fuse before it blows the IGBT's.

 

[Sigurthr]

I've had some interesting effects from driving a coil with a much longer pulse width than it was intended, you can get some really weird fiery arcs. Luckily with my coil though the bridge was charged via a small transformer acting as a stepup, so running it like that just pops a fuse before it blows the IGBT's.

What you're actually referencing is the current modulator status-quo; Generally, if it isn't a simple single 555 oscillator then two 555 timers are used, one to get pulse width in monostable mode and one to trigger the pulses in astable (oscillator) mode. The problem is that 555s cannot do <50% duty cycle, and cannot control pulse width and frequency independently. So, generally you wind up with a range of 50-99% duty cycle, and then they feed the signal through a logic inverter, resulting in 1%-50% duty cycle. However since you cannot control BPS/PW independently there are certain combinations that are unavailable. Additionally, for DRSSTC use, you can't set hard limiters on pulse width duration, and certainly can't set conditional limiters. This is why OCD monitors are so important for DR coils, to prevent bridge failures.

Such are the reasons why I moved to the microcontroller based interrupter; I can set it so that conditionally no ON time is greater than a certain pulse width in nS/uS/mS, and independently control frequency and pulse width with no hardware changes.

On a regular old singly resonant SSTC I've been able to emulate the arc structures over a range from thin wispy straight sparks to classic spark gap style to DRSSTC style to fiery flames.

 

[Sigurthr]

Update: Boards are out of Fab and on their way to me. These will be available very soon!

Also, I've set up a website to house all my stuff/projects/boards, etc. Feel free to check it out. I will offer boards for sale there, but they will be cheaper here!

Sigurthr Enterprises

 

[Sigurthr]

Sig's ATTiny85 Interrupter has been finalized and I'll be able to offer pre-programmed ATTiny85 MCU chips for sale for those interested. Pricing to be determined. Only one chip available for sale right now, but I'll order in more chips soon if there is interest.

Data files:
Application schematic: https://app.box.com/s/qilqivytxbots3hxummh NOT UPDATED, SEE Post #1
Code file: https://app.box.com/s/93dt6fso6wj3tvx13m7b

http://imagizer.imageshack.us/v2/640x480q90/713/3wi9.jpg

 

[Sigurthr]

UPDATE: BOARDS NOW AVAILABLE FOR SALE!

Direct Purchase w/ PayPal link available at www.SigurthrEnterprises.com

OR​

Private Message Me Here for manual PayPal transaction (I'll need your PayPal/email address to send you an invoice.)

Another Action Vid of the USSTCC:

​

 

[Speedy78]

On the BOM for the interrupter the part numbers for the fiber transmitter and receiver have an extra "-" in them. They should be FB129-ND and FB123-ND. Took me a while to find them, but then again I'm stupid so ya have to take that into consideration lol

 

[Sigurthr]

Oops, my mistake, sorry Speedy!

 

[The Lightning Stalker]

Have you been able to clock the Attinies higher yet? To change the fuse bytes using the Arduino IDE, modify the appropriate line in boards.txt and then burn the bootloader. At
16MHz, it should be possible to get down to 62.5ns.

I got some of the SPL LL85 the other day and had to get my Arduino to do this in order to
test them. It also gave me a chance to exercise the UCC37322. The rise time on those is
nothing spectacular, (on the order of 10 - 20ns) but it's a bot letter than driving directly from
the Arduino or even a 74HC14. The better (recommended) MOSFET drivers are over $20
apiece, so that is going to be out of my questoin.

Here is the code I've been using to pulse the Arduino pins at 62.5ns. Maybe it will give you an idea of what can be done. avr-objdump -d ???.o will decompile the object file into
assembly so you can see what the chip is actually doing.

// ns_Pulser.ino
// Apply 62.5ns (@ 16MHz) pulses

int laser = 2;

void setup() {
  // setup:
  pinMode(laser, OUTPUT);
  noInterrupts();          // disable interrupts just in case
}

void loop() {
  // main loop: 
  PIND = B00000100;        // writing to the pin register toggles
  PIND = B00000100;        // output on AVR microcontrollers
  delayMicroseconds(100);
}

As you can see, it bypasses the digitalWrite command and goes straight to the pin
registers. The digitalWrite function is notorious for using extra clock cycles. Time
sensitive sketches always use direct I/O to get the fastest most accurate timing. You can
also go to the nuclear option and use inline asm.

 

[Sigurthr]

I have to get my ATTiny85's in first. Used and sold off my stock. I am going to see about getting in some 20MHz versions and see how quick I can push them. My last batch I had running at 8MHz no problem.

As far as I can tell a major part of the problem is inaccuracy in the delay commands, both the uS and mS versions. Bypassing the digitalWrite helps a lot, but even when accessing the pins directly there is what seems like a nonlinear discrepancy between actual timing and programmed timing. I managed to get one of the 8MHz ones to do an accurate 350uS ON pulse but could shorten it no further. I seem to remember it actually backing off (as in shortening the delay in code actually lengthened the pulse!) when I tried it.

I have a sketch I found online that uses a custom library to get selectable 30/50/70% duty cycle pulse trains at up to 1MHz in 10^x increments. I can't decipher the code though, it is beyond my programming skill. I tested it on the scope and it is pretty damned accurate too. However, it is only able to do such a limited duty cycle range, and the pulse widths are heavily tied to frequency, so you couldn't do 500nS ON and 1mS off, despite it being able to do 500nS ON / 500nS OFF. I'll post it up sometime, but this is all better suited to the Arduino Thread, and not the sale thread here =).

 

[Sigurthr]

Still have more available, including one board + parts kit!

New videos showing the advanced expandability as well:







All three videos shot sequentially with no cool down time in between, running straight from mains on only passive cooling to the half-bridge.

 

[Speedy78]

Now how does one make the coil do that?

 

[Sigurthr]

Simple! Use the jumper for expansion pins 2-3-4 for CW mode, or put your interrupter to CW mode. Take output of 4046 and insert it into the antenna input port. I powered the 4046 using pins 1 and 2 of the expansion port. No feedback from the coil or driver are necessary.

The magic is in setting the PLL, but that isn't really hard, you just have to know the f0 of your coil and select a resistor and capacitor to suit. Assembly is simple. I'm still working on the design to see how far I can push it though, but I can give you working info on how to duplicate the effect for your coil(s).

 

[Speedy78]

Hmm. It would be hard to unsolder the ct and then solder in the leads for the chip. I assume one of the other pins on it is input for the audio?

Edit: Stupid question but where does the fiber plug into the driver or do I need to make something that hooks up to the expansion port that has the reciever?

 

[Sigurthr]

Hmm. It would be hard to unsolder the ct and then solder in the leads for the chip.

There's a reason I put rising cage clamp connectors at all the I/O ports.

I assume one of the other pins on it is input for the audio?

The PLL chips gets the audio input.

Edit: Stupid question but where does the fiber plug into the driver or do I need to make something that hooks up to the expansion port that has the reciever?

That would be the FB-129 (or was it the 123? I always mix the two of them up). It gets wired right to the 4 pin expansion port (through a rising cage clamp connector).

I mean you can leave out the connectors and direct solder everything but for saving like $3 worth on connectors you're buying yourself a whole lot of headache and limited expandability.