Bob Boyce TPU thread

[Jdo300]

Hey Bob,

I definitely understand what it's like being busy with life. I actually just moved all of my stuff to my apartment in Columbus in preparation for my fall Co-op with the Battelle Institute. So unfortunately, all of my lab equipment and supplies are stuck there (I wont be living there for about another week). However, in the mean time, I'm spending some time learning more about FPGAs and how to program them to make a super duper control circuit with all the bells and whistles we'll ever need.

I just have a couple of questions about driving the coils. I did order some samples of the driver chips you gave us links to (they just arrived yesterday) and I was wondering how you are driving your coils with your hex driver. Are you driving the coils with one wire using MOSFETs ar just using the driver chip output directly? This is the only thing I need to know with regard to the circuit since I will (hopefully) be integrating all the other functionality into the FPGA.

My second question is about the circuit diagram you made that shows the HV bias connections and things. Could you give us some basic component values to work with? At the moment, I am assuming standard values for the filter cap and choke (like 440 uF and 1H??) but I don't have the slightest clue of what to use for the DC blocking cap. Should it be a high value DC cap or a small valued cap?

God Bless,
Jason O

[eldarion]

Bob,

I have been experimenting with the attached setup, and had a few questions.  I am using a toroidal neodymium magnet for the core (for a fixed amount of magnetic bias), and have one winding of 22 gauge all the way around the core, and three (ugly!  ) windings of 24 gauge for primaries.  I am driving the primaries open-ended with fast MOSFET driver chips, at a frequency of about 62KHz, pulse width 100ns, primary 1 is at 0?, primary 2 is at 120?, and primary 3 is at 240?.

I realize my setup is way outside of what you recommend using, but I was wondering of you could shed some light on the following:
1.) The HVDC bias does absolutely nothing--when I connect it, the output waveform I have attached does not change at all.  It is connected as follows: 160VDC positive lead, high value resistor, then connected to one end of the secondary coil, then that end goes to a 0.22uF DC blocking capacitor, then the load, and finally back to ground.  Scope is connected directly across the 470-ohm load resistor, 5V per vertical division, 10us per horizontal division.

2.) As you can see, I am getting almost zero power output.  Is this because of my setup, or some other issue?

3.) What type of output waveform should I be getting?

This setup is so far out of spec that I am not expecting any overunity, I just thought that the effects would at least try to manifest themselves in some small way...

Thank you very much!  I hope we receive our cores soon, so we can actually build something that will work!

Eldarion

[Jdo300]

Hi Eldarion,

I can answer that question right off the bat. That neo magnet you have for the core is overpowering ALL of the fields from the coils that you are applying. The whole point of the low voltage DC bias is to turn the 2D rotating magnetic field into a 3D vortex field but this also implies that the rotating field must be strong enough to influence the bias fied. Most Neo magnets have a B-field strength of over 2000+ gauss so there is NO WAY you could setup a vortex in a field that strong without having field coils of comparable strength.

So in other words, that big magnet is choking your whole TPU. If I were you, I would replace that magnet with something like a ceramic 5 magnet, whose field strength is much easier to perturb than the Neo.

God Bless,
Jason O

[Jdo300]

Hi Everyone,

I've still been doing some digging around for the perfect control circuit and I found what I believe to be the ULTIMITE IC for the job. It is called the AD9959 which is a 4-Channel 500 MSPS DDS
with 10-Bit DACs. Basically, you have 4 independent, synchronized sine wave generators that can potentially go up to 100MHz all in a single IC!!! Each channel is programmable which means we can use any microcontroller (doesn't matter how fast) to set the frequencies, amplitude, and phase shift for each channel!

Check out the features listed on the datasheet:

FEATURES

• 4 synchronized DDS channels @ 500 MSPS
• Independent frequency/phase/amplitude control between channels
• Matched latencies for frequency/phase/amplitude changes
• Excellent channel-to-channel isolation (>65 dB)
• Linear frequency/phase/amplitude sweeping capability
• Up to 16 levels of frequency/phase/amplitude modulation (pin-selectable)
• 4 integrated 10-bit D/A converters (DACs)
• Individually programmable DAC full-scale currents
• 32-bit frequency tuning resolution
• 14-bit phase offset resolution
• 10-bit output amplitude scaling resolution
• Serial I/O Port (SPI) with enhanced data throughput

Here's the link to the datasheet:
http://www.analog.com/UploadedFiles/Data_Sheets/210775240AD9959_0.pdf

The output is a sine wave but it would be a simple matter to send the output into a buffer IC to convert it to a square wave. If we use a flip flop, circuit like the one Earl suggested, we can then chop the pulse width way down before sending it to the driver chips.

The only bad thing about this chip is the fact that it only comes in a surface mount package (56 lead LFCSP) which is ridiculously tiny. I"m currently looking around to see if there are any adapter boards for it. Otherwise, I may have to find someone who can solder one to a board for me 'cause I can't see well enough to do it on my own.

God Bless,
Jason O

[Jdo300]

By the way, I forgot to mention that these ICs are a bit pricey, $37.14. if you buy it from Analog's website. BUT, you can get up to two free sample chips so I ordered two of them. Here's a link to the page to order samples if you are interested:

http://www.analog.com/en/prod/0%2C2877%2CAD9959%2C00.html#price

God Bless,
Jason O