Author Topic: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge  (Read 399 times)

Offline davekni

  • High Voltage Enthusiast
  • *
  • Posts: 45
  • Karma: +2/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« on: October 10, 2019, 06:31:02 AM »
Here's my first attempt at documenting a project.  It is my first DRSSTC - just complete enough to run, but with much enhancement work planned for the next couple years.  A few videos of operation, at the Portland (Oregon, USA) Mini-Maker Faire, and a couple in my back yard two weeks earlier :
   
   
   
   
   

Don't have a good assembled picture of the coil, but here's the parts:
   

MMC, with lots of room for planned expansion:
   

H-Bridge with 0.1mm x 100mm copper foil output leads.  E80-core current transformer is hidden under bump in output foil at right end.  Purple clip in the upper left quadrant is a crude optical probe (LED and resistor into plastic fiber) monitoring one of the high-side gate drive signals.  The pairs of TO220 FETs are the gate drives on floating 18.5V supplies, one PFET and one NFET per H-Bridge switch, driving 10 paralleled STGW60H65DRF IGBTs (hiding under the aluminum bar clamps):
   
The bulk cap array is under the H-Bridge, so not visible.  96 x "470" uF 450V.

Control circuit - in an aluminum-foil lined cardboard box at this point:
   

4-turn primary coil made of 200 strands of 27AWG wire.  Inner turn gets warm at 3kW.  Will need finer litz and/or cooling air directed that way to get 10kW eventually:
   

Bulk-cap circuit:
   

Main oscillator simulation schematic, using voltage-controlled-voltage-sources for gate drive, and FETs  for the H-Bridge because they simulate faster.  The center lower part is mostly a comparator made of discrete TO92 FETs, from M3 on the left through M10 and M11 on the right.  Nodes "vn" and "vp" are the inputs and nodes "v2" and "v3" are the true and inverted outputs.  R2, R3, R4, R7, R11, and C10 are the relevant feedback around the comparator.  No one would want to copy my comparator itself.
     

Current limit is also a bit unusual.  It is fed with a voltage-transformer (L1 and L2) from the oscillator current transformer output.  (My current transformer is 40:1 first-stage, then two 25:2 second stages.  However, one of the second stage outputs is for scoping only.)  The current limit circuit includes not only an immediate shutdown at 3500A peak (2600A in this initial version), but also simulation of the IGBT thermal transient response (C1 and C2 and associated resistors).  That way repeated long bursts will get aborted even if they don't reach 2600A.  (So far, the current limit trips only in my initial primary-only testing, not in real use so far.)
     

Finally, here's a hint of where I hope to go.  This is a sketch of a small part of what will be an array of ~500 TRIACs to switch in more primary tank capacitance as each arc grows.  My goal is to make it behave more like a QCW coil, but hopefully even better with a reverse-chirp in primary frequency to match the secondary frequency change with arc growth:
     

Before this fancy reverse-chirp attempt, I need to improve my MIDI interrupter setup, at least for next year's Maker Faire.  Also need to redo the primary tank circuit connections - going to use spade connectors instead of many paralleled 0.025" square pin connectors - that come apart too easily.  Going to explore detection of power arcs and termination of the drive enable pulse once a power arc forms.  That should make overall operation more efficient.

Hopefully I have the jpg attachments figured out now.
« Last Edit: October 11, 2019, 04:28:42 AM by davekni »
David Knierim

Offline johnf

  • High Voltage Enthusiast
  • *
  • Posts: 24
  • Karma: +0/-0
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #1 on: October 10, 2019, 08:44:38 AM »
oops
same pic repeated!!!

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 265
  • Karma: +8/-0
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #2 on: October 11, 2019, 10:06:32 PM »
Woah, was worth waiting for the new pics. Can't say I'll be copying the comparator bit, but I do like the idea of the OCD having more sophisticated behaviour than a simple on/off threshold.
Are you doing anything in particular to try and force the IGBTs to share current fairly? Or just relying on a low-inductance bridge configuration?

Offline davekni

  • High Voltage Enthusiast
  • *
  • Posts: 45
  • Karma: +2/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #3 on: October 12, 2019, 05:52:51 AM »
The IGBTs have a positive Vce-temperature coefficient, which helps with current sharing.  The only explicit step I took besides tight packing was to feed Vbus from one end of each half-bridge, and take the outputs from the other ends.  That way the Vbus voltage drop roughly matches the output voltage drop, keeping Vce roughly constant across the array.  I haven't attempted any specific measurements of current sharing.  Would like to check it before increasing max current from 2600A to 3500A, but not sure how.  (3500A still has some margin for sharing errors.  I tested one individual IGBT to 500A, both transistor and diode, at 90C initial temperature.  Rated peak current is 240A - the best I could find in a fast TO247 part.)

The reason for the discrete comparator (besides my habit of designing that way at times) is that it tolerates input voltages to 10V beyond supply rails.  That avoids diode-clamping the current-sense signal, which would make a non-linear load impedance on my phase-lead filter.

The phase-lead filter is a bit fancier than normal, designed for roughly constant time lead (not phase lead) from 50-80kHz.  That covers the range I'm planning with my eventual reverse-chirp primary resonant frequency.  The filter also has the advantage of not amplifying high-frequency noise as much as a single inductor or capacitor circuit.  (I'm currently running about 83kHz primary.  It works better with the arc-length constraint of the Faraday cage with the primary closer to the secondary's 92.2kHz resonance.)

Forgot to include a bit of basic information:
    Primary:  6.72uH (6.95uH including inductance of MMC and interconnect), 4 turns of 200 strands of 27AWG wire.
        222mm diameter of inner turn (diameter at center of wire).  308mm diameter at center of outside turn.
    Secondary:  68.0mH, 75 ohms DC, ~1800 turns of 24AWG wire, 157.5mm diameter, 1100mm high (not counting loose-wound ends).
    MMC:  10s16p of 330nF 1200VDC 630VAC Chinese induction cooker capacitors, 528nF total, half on each end of primary coil.
    Top Load:  194mm minor diameter, 632mm major diameter (826mm outside diameter).

Ignoring the comparator implementation, the circuitry around it is somewhat similar to the UD2.7 design.  A couple differences:  First, the bias point of the current-feedback is centered with no signal present, rather than a diode-drop low.  Second, I have positive capacitively-coupled feedback (as does UD2.7), but also weak resistive negative feedback.  That combination causes the circuit to oscillate with no current-sense present.  I've designed that oscillation frequency to roughly match the primary resonant frequency.  This allows synchronous oscillation to start even with very-low Vbus (5V works), independently of the initial state of the primary tank circuit.  No need for a resistor across the H-Bridge output.  (I do have bleed-down resistors with indicator LEDs across the MMC sections.)
David Knierim

Offline ScottE

  • High Voltage Enthusiast
  • *
  • Posts: 3
  • Karma: +0/-0
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #4 on: October 12, 2019, 07:38:47 AM »
I am planning on building a coil with litz wire on the primary and secondary coils. The secondary wire is made up of 100 strands of 0.07 mm wire (about 21 awg). The primary wire is made up of 4000 strands of 0.07 mm wire (about 6 awg) and has outer teflon covering. It should work up to 250 kHz.

I see that you are in Portland and I am quite sure I met you at Westercon when it was in Portland.
I know that you met Jon Hannis from Control H (PDX Hackerspace) at the maker fair. If you go Control H and talk to Jon he can show you samples the litz wire I have.
I am in LA but I will be building the coil at Control H in year or so after I move up to Portland.

Offline davekni

  • High Voltage Enthusiast
  • *
  • Posts: 45
  • Karma: +2/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #5 on: October 12, 2019, 06:57:55 PM »
Should be a great primary with 4000-strand litz!  With such fine wire, the inner turn may not get much warmer than the outer turns.  Would be fun to see thermal images of your primary after a run.  Mine clearly shows higher temperature on the first turn.

Litz will have minimal affect on the secondary.  Typical secondary Q is already high, so relatively little power loss.  Even if you got infinite secondary Q (zero power loss), it's not that much savings.  No harm in using litz, though.

I'd love to see your coil once your up here.  Haven't been to Westercon, but I think Don Anderson was there with his coils.  He's the one who got me interested in Tesla coils.  We share demo space at Maker Faire here.
David Knierim

Offline Weston

  • High Voltage Enthusiast
  • *
  • Posts: 9
  • Karma: +1/-0
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #6 on: October 13, 2019, 12:41:07 AM »
Cool coil! Its always nice to see people coming up with their own controller instead of directly copying the UCD design. I have a FPGA based controller for a QCW coil and I was considering implementing a similar limit on the total IGBT heating, its nice to see that someone else is already doing it.

Why did you choose to use so many paralleled TO-247 IGBTs? Surplus IGBT bricks are pretty cheap nowdays so I assume they would be cost competitive in addition to simplifying the bridge layout.

How much modeling have you done on the the concept of switching in additional capacitance with TRIACs? I would be worried about dv/dt induced turn on in the TRIACs beyond the general added complexity. QCW based coils typically get around the detuning issue by tuning the secondary above the primary and then having the controller run on the upper pole, so the added capacitance from the arc pulls the coil better into tune. 

Offline davekni

  • High Voltage Enthusiast
  • *
  • Posts: 45
  • Karma: +2/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #7 on: October 13, 2019, 06:36:13 AM »
The paralleled TO247 IGBTs were for fast switching, both because small IGBTs are rated faster, and the paralleling gives lower net lead inductance.  I would have had to parallel at least two sets of bricks to get 3500A peak.  Now that I'm reading what others are doing, it's clear that the bricks can be pushed to much higher frequencies in ZVS mode than I'd realized.  (The lower inductance does have an advantage still in peak voltage.  These 600V IGBTs still have sufficient headroom at 450Vbus, as there is very little overshoot during switching.  However, if I fry the H-Bridge, I may rebuild with bricks.  It was a lot of work to build a 40-IGBT bridge!

The thermal modeling seemed sensible for pushing devices to their limit.  The zener/resistor array at the bottom of the schematic page makes a current roughly proportional to the square of the primary current-sense transformer output, so proportional to forward-conduction power loss in the IGBTs.  That current charges C1 and C2, which are coupled with resistors modeling the transient thermal impedance from the IGBT data sheet.  I had no idea if anyone else was doing that - hadn't found much information before stumbling into this forum.  (Peak current causes all zeners to conduct, so behaves as infinite power, causing an immediate (next zero-current, rising or falling) end to that burst.  It also leaves the model in its hot state, delaying when the next burst is allowed.

The TRIAC switching will be a challenge!  I haven't found good modeling tools, so am relying mostly on testing individual TRIACs and gate-drive transformers etc.  Had to use NiZn ferrites for high electrical resistivity to avoid excess capacitive coupling.  MMC is split in half, one on each side of H-Bridge output to keep peak voltage to ground lower.  Finally, it's going to be inefficient in MMC capacitor use - many extra ones.  The plan is to build a huge MMC with extra voltage capability, then short out individual small groups of caps at the voltage zero-crossing.  Snubberless TRIACs have one quadrant where they won't turn on.  That allows the gate drive pulse to start early, before voltage zero-crossing, avoiding excessively-tight timing requirements.  Placing the TRIACs directly across capacitors prevents stray inductive spikes from triggering the TRIACs.  I did choose TRIACs with fast enough DVDT rating to handle the voltage slew rate of individual MMC sections.

I have some previous experience in failed TRIAC array switching.  My very first Tesla coil was an attempt to replace the spark gap with a TRIAC array.  Made three sparks before a failure cascaded through the array.  Turned that project into a rotary-gap coil.

I'll need to play with my basic coupled-resonator simulation some more.  (Anyone have a good SPICE model for a growing arc?  I've made a couple guesses, but not likely very accurate.)  In simulation, it seemed that starting on the upper pole didn't perform as well.  Does the ramped voltage make that upper pole work better?  I gather that most DRSSTC designs oscillate at the lower pole, as mine does.  Either way, I expect there's a limit to how far detuning can go, and I'm hoping to add to whatever that limit may be.
David Knierim

Offline T3sl4co1l

  • High Voltage Experimenter
  • **
  • Posts: 79
  • Karma: +8/-0
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #8 on: October 14, 2019, 07:10:39 AM »
I don't see that TRIACs can be used for tuning here.  They turn on automatically when exposed to high dV/dt.

A saturable inductor may be more feasible?

Capacitors can be tuned in the same way as well, but you can only do it with lower-Q ceramics, and you'd need a lot of them to handle that much reactive power.

Saturable inductors tend to have a low Q for the same reason.

Alternately, a vacuum variable capacitor (with servo control), with a transformer to match the impedance.  Exquisitely priced. :o

Perhaps a frequency tracking control, and an inverter with enough excess VA capacity to handle the change in SWR, would be easier.

Tim

Offline davekni

  • High Voltage Enthusiast
  • *
  • Posts: 45
  • Karma: +2/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #9 on: October 14, 2019, 07:01:34 PM »
I purposely selected TRIACs with high dV/dT rating, 1000V/us (BTB16-800BW from ST).  That's well beyond the 400V/us they'll see.  Each TRIAC will be shorting one MMC section, which will have 800V peak at 80kHz, thus 400V/us.

The TRIAC specification that is most risky is "critical rate of rise of on-state current".  These parts, as with almost all similar parts I could find, are specified at 50A/us maximum.  In my particular odd application, the rate-of-rise is controlled primarily by how fast the TRIAC manages to turn on.  (Ideally, it would instantly conduct all the capacitor current at the voltage zero-crossing.)  I made a small 3-cap MMC in a resonant circuit with one TRIAC shorting one stage.  It ended up with ~100A/us, about twice spec.  I've done some other single-shot tests running these to 250A/us, but not repeatedly nor with high die temperature.  It's been fun learning about "critical rate of rise of on-state current", as I had no idea previously what that spec. meant or why that would be limited.  (Turns out that TRIAC gates are in one corner of the die.  It takes time for the primary current to spread across the die.  It starts in the gate corner and spreads.  Too high current immediately overheats the gate corner of the die.  I can see this in the forward drop voltage/time behavior.  Starts at ~20V forward drop, which goes down to normal ~1V over 1-2us.)
David Knierim

Offline Weston

  • High Voltage Enthusiast
  • *
  • Posts: 9
  • Karma: +1/-0
    • View Profile
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #10 on: October 14, 2019, 08:46:54 PM »
In what way did the upper pole tuning did not preform well in simulation? The tuning relies on the fact that the capacitance from the discharge pulls the secondary into tune, so you need some sort of model relating power and spark length/capacitance. I recall there being some good posts on 4hv about the tuning method and models for arc impedance. The easiest model is assuming the arc has a capacitance equal to an equivalent length of wire, which is how I tuned my QCW coil.

One thing to note is that when the secondary is tuned above the primary the lower pole has higher gain, so a self oscillatory controller will lock onto the lower pole.

I am still unclear about what advantage the triac based system would have. It seems like a lot of work with a high potential to blow up if anything goes wrong.


Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1142
  • Karma: +17/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #11 on: October 15, 2019, 12:38:24 PM »
Impressive design and + for effort going your own way!

Regarding spark load models, Uspring shared his version 2 dynamical model ( https://highvoltageforum.net/index.php?topic=670.0 ) this summer, based on the new information from Hydron's spark load measurements ( https://highvoltageforum.net/index.php?topic=117.0 )

Usprings version 1 model, but without pictures as 4hv is broken: https://4hv.org/e107_plugins/forum/forum_viewtopic.php?153922
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

High Voltage Forum

Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
« Reply #11 on: October 15, 2019, 12:38:24 PM »

 


* Recent Topics and Posts

post Re: making a useable induction forge.
[Transformer (iron core)]
rustedone
Today at 05:55:24 PM
post Re: making a useable induction forge.
[Transformer (iron core)]
rustedone
Today at 05:49:02 PM
post Re: Analog Panel Meter Shunt Calculation and Custom Scale Design
[Beginners]
klugesmith
Today at 05:36:46 PM
post Re: making a useable induction forge.
[Transformer (iron core)]
T3sl4co1l
Today at 05:07:22 PM
post making a useable induction forge.
[Transformer (iron core)]
rustedone
Today at 04:20:15 PM
post Re: 220ac to 440dc 5a boost convertor
[Voltage Multipliers]
rustedone
Today at 04:03:58 PM
post Analog Panel Meter Shunt Calculation and Custom Scale Design
[Beginners]
Mads Barnkob
Today at 07:56:20 AM
post Re: Mysterious Chinese HV capacitors
[Capacitor banks]
MRMILSTAR
Today at 05:40:55 AM
post Re: Mysterious Chinese HV capacitors
[Capacitor banks]
davekni
Today at 03:33:52 AM
post WTS [US]: pole transformer, SRSG, PTs, NSTs, capacitors, variacs, [MORE]
[Sell / Buy / Trade]
acobaugh
October 15, 2019, 09:36:55 PM
post Mysterious Chinese HV capacitors
[Capacitor banks]
MRMILSTAR
October 15, 2019, 09:12:09 PM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
Mads Barnkob
October 15, 2019, 12:38:24 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
davekni
October 14, 2019, 10:02:07 PM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
Weston
October 14, 2019, 08:46:54 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
MRMILSTAR
October 14, 2019, 08:28:19 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
davekni
October 14, 2019, 07:27:22 PM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
davekni
October 14, 2019, 07:01:34 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
MRMILSTAR
October 14, 2019, 04:23:56 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
klugesmith
October 14, 2019, 07:27:49 AM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
T3sl4co1l
October 14, 2019, 07:10:39 AM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
MRMILSTAR
October 14, 2019, 05:43:57 AM
post Re: so far video
[Spark gap Tesla coils]
davekni
October 14, 2019, 05:40:27 AM
post Re: Drsstc 4 (new direction)
[Dual Resonant Solid State Tesla coils]
davekni
October 14, 2019, 05:23:48 AM
post Re: Drsstc 4 (new direction)
[Dual Resonant Solid State Tesla coils]
oneKone
October 14, 2019, 04:37:49 AM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
davekni
October 14, 2019, 03:47:03 AM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
klugesmith
October 13, 2019, 11:05:25 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
davekni
October 13, 2019, 09:29:25 PM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
MRMILSTAR
October 13, 2019, 04:43:47 PM
post Re: My first DRSSTC on bricks
[Dual Resonant Solid State Tesla coils]
davekni
October 13, 2019, 07:04:20 AM
post Re: How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
davekni
October 13, 2019, 07:00:57 AM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
davekni
October 13, 2019, 06:36:13 AM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
Weston
October 13, 2019, 12:41:07 AM
post Re: My first DRSSTC on bricks
[Dual Resonant Solid State Tesla coils]
Hydron
October 13, 2019, 12:17:39 AM
post Re: My first DRSSTC on bricks
[Dual Resonant Solid State Tesla coils]
Laci
October 12, 2019, 11:21:39 PM
post How to measure voltage output of AC flyback transformer
[Transformer (ferrite core)]
MRMILSTAR
October 12, 2019, 08:37:36 PM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
davekni
October 12, 2019, 06:57:55 PM
post Re: so far video
[Spark gap Tesla coils]
thedoc298
October 12, 2019, 05:28:01 PM
post Re: 220ac to 440dc 5a boost convertor
[Voltage Multipliers]
rustedone
October 12, 2019, 04:58:28 PM
post Re: My first DRSSTC on bricks
[Dual Resonant Solid State Tesla coils]
Netzpfuscher
October 12, 2019, 03:53:12 PM
post Re: My first DRSSTC on bricks
[Dual Resonant Solid State Tesla coils]
Laci
October 12, 2019, 02:12:59 PM
post Re: 220ac to 440dc 5a boost convertor
[Voltage Multipliers]
rustedone
October 12, 2019, 10:43:48 AM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
ScottE
October 12, 2019, 07:38:47 AM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic circuits]
davekni
October 12, 2019, 06:16:12 AM
post Re: DRSSTC with litz-wire primary and 40 x TO247 IGBT H-Bridge
[Dual Resonant Solid State Tesla coils]
davekni
October 12, 2019, 05:52:51 AM
post Re: so far video
[Spark gap Tesla coils]
MRMILSTAR
October 12, 2019, 05:13:28 AM
post Re: Problems with my first DRSSTC
[Dual Resonant Solid State Tesla coils]
davekni
October 12, 2019, 04:41:19 AM
post Re: so far video
[Spark gap Tesla coils]
thedoc298
October 12, 2019, 02:58:19 AM
post Re: so far video
[Spark gap Tesla coils]
thedoc298
October 12, 2019, 02:56:17 AM
post Re: so far video
[Spark gap Tesla coils]
MRMILSTAR
October 12, 2019, 02:38:43 AM
post so far video
[Spark gap Tesla coils]
thedoc298
October 12, 2019, 12:26:44 AM

Sitemap 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30