Advanced Wargame Systems
Creating innovative designs and modifications for AEGs since 2007, from the first AEG braking MOSFETs to the powerful Scorpion Hi-Realism controllers!
Supercapacitors
Aerogel supercapacitors will be included in the upcoming v4 controllers to act as a backup "battery", so in some worst case situations where the line voltage drops below 2V (ie: using a weak battery to startup the motor WHILE the spring is fully compressed), the controller will switch over to the supercapacitor as a secondary power source until the motor gains enough momentum, allowing the line voltage to recover above 2V. If the line voltage does not recover within a preprogrammed time limit or the thermal sensor detects a sudden spike in the MOSFET temperature, the appropriate cutoff is activated.
 Final v4 controller layout |
We've long considered adding a supercapacitor but haven't done so until now due to their low voltage ratings of 2.5-3.3V. 5V supercapacitors are too large and expensive, and we couldn't reduce the microcontroller supply to 3.3V otherwise it wouldn't be able to drive the UV LED in the Scorpion modded hopups, which have a forward voltage drop of ~3.2V. But this all changes with the new power regulation design in the v4 controller. By having the linear regulator output 1.8-2.5V (Vin > Vout), we now have a place to put a 2.5V supercapacitor.
 Linear & Boost regulators |
The supercapacitor is 0.22F (220,000uF), fully charged within 5 seconds (although 3 is enough), and can supply the boost regulator for up to 30 seconds, allowing it to output a steady 5V the whole time even though the supercapacitor's voltage is decreasing linearly! This is perfect for the microcontroller's ADC (analogue digital converter) since its reference voltage, which is typically VCC, will remain at a stable 5V no matter how how severe the battery voltage drop is.
Due to this design change, I threw out the 200 perfectly good PCBs I ordered based on the design in my previous post below. I hate it when I come up with new ideas too soon!
Site Changes:
Added "DIY" (do it yourself) to our product configuration options. If you select this, we'll allow YOU to solder on the wires and do the heat shrinking (which are still included with the order). A small discount will be applied and as thanks for saving us time. :)
- Frank (PsyphyerVII)
June 25th, 2009
Smaller Components & New Experimental Design
I have just returned from my visit to Taiwan. Had very limited internet connection, so for those of you that emailed while I was away, expect a response soon. All the controllers have been put back in stock, but if you see "out of stock", then it likely means we ran out of a certain component and are reordering from Digi-key so check back every 2-3 days. Now for the updates:
While looking over the new controller layout I posted two weeks ago, a thought occured me: If I'm going to have the controllers fully assembled by high-precision component placement machinery, why not use even smaller components? There is no increase in the assembly cost, the components are actually cheaper, and more room will be available on the controller. So begins another redesign, this time replacing most of the 0603-sized resistors/capacitors with thier 0402-sized counterparts, which has a 55% smaller footprint and is smaller than an ant! The size of the ICs have also been decreased to the point thier pins are within just 0.015 inch of each other, which would require extremely steady hands if placed manually.
 Revised v4 layout with smaller components |
By downsizing the resistors which originally occupied about 1/3 of the board, we can now fit them in a circular fashion around the microcontroller, saving tons of space and reducing the complexity of the trace paths (above image). With the extra room, we managed to implement a unique and very important feature that is lacking in every MOSFET controller currently on the market. Most of you have probably heard of, or even experienced, AEG MOSFETs frying even though the MOSFET's current and voltage rating are well above those used by the AEG, so why do they burn out?
The answer lies within the line voltage drop when the motor operates under heavy load; basically the more current you draw out of the battery (or if your battery is not rated for the current the your AEG draws), the greater the instantaneous voltage drop. Same reason why most house or car lights appear to dim when a power hungry device, such as an AC, powers up. Once the current is cutoff, the voltage jumps back up to the true voltage level.
Since ALL power MOSFETs require a minimum driving voltage of 4V or 10V, too large of a voltage drop can pose a problem. Underdriving a MOSFET causes it to become resistive and will easily overheat. Using a non-logic-level MOSFET, which requires a 10V gate drive, runs a very high risk of frying the MOSFET (which happens to many who attempt to make thier own MOSFET units without reading the datasheets closely). Logic-level MOSFETs require only a 4V gate drive, but even that doesn't always make it immune to extremely high current-consuming AEGs, especially those with hi-speed setups or too tightly shimmed gears. We've had a customer in the past whose high-speed AEG activated our controller's <5V cutoff activate right before it could fire the first shot, no doubt caused by this very problem.
So we made a new change that has never been tried before in any AEG MOSFET system: the minimum required voltage is now down to just 2V! This is achieved by using a series of linear and boost regulators. Battery voltages ranging from 2-18V gets converted by the linear regulator to 1.8-5V (Vin > Vout), which is then feed into the boost regulator and boosted to 5V with an output current capacity of 120mA, which is more than the controller needs. Voltages <2V will be considered a short circuit and thus will activate the cutoff. If it's not a short circuit, then that person should seriously consider upgrading their battery to one with a higher rated discharge or using a higher torque setup! This new low-voltage requirement feature will also implemented into the next v4 controller version.
Another small improvement was the addition of a jumper resistor located between the pins of the right-side MOSFET in the above image. Since we're going to be using a more powerful MOSFET, it makes sense to increase the trace widths to handle more current. The trace where the jumper is placed over might become a bottleneck under high current, so the addition of the jumper should increase the overall current paths.
The PCB layout design files have been sent to the fabrication house for production and should arrive in about 8 days.
Product update:
- Minimum voltage requirement reduced from 5.5V to 2V to withstand severe voltage dips
- Frank (PsyphyerVII)
June 9th, 2009
Too Many Orders!
Most items have been temporarily marked as "out of stock" since we have recieved too many orders, they're coming in at at the same rate than we can make them! Since I'll be going overseas soon, I cannot have too many order piling up, therefore any further orders will have to wait until I come back from Taiwan. I have to modify several hopups for the Scorpion orders, install controllers in AEGs that cutomers have sent in, and finish most the orders, all within a short amount of time. :(
On the bright side, just the profit we made over the past week was enough to cover 1/3 the costs of having the upcoming controllers FULLY assembled by machine, not just partially like the current v3 controllers (which I now regret)! Not bad considering this is the only place I ever announce AWS product updates. Guess word spreads fast. :)
- Frank (PsyphyerVII)
May 13th, 2009
Visiting Taiwan Soon
Now that the university finals are over (Senior CompE major), we no longer have to constantly place our products as "out of stock" due to endless project/all-nighters/exams, we can now place more attention on AWS! Although I had little time to work on orders during the semesters, as long as I had my laptop in front of me, the controller design continued to evolve. The lack of time was one of the main reasons for having the current v3 boards partially assembled by machine. I greatly appreciate those of you that had the patience and willingness to wait months at a time while I was prototyping various designs! :)
Product stock update:
All the controllers are in stock, BUT I'll be leaving for Taiwan on May 15th so there is a small time frame for those of you that wants to order the Pulse and the Scorpion (but you must send your hopups quick!). Once I have too many orders or need to prepare for departure they'll be set to out of stock for 3 weeks until I return. The Sentry controller and Microfet will remain in stock as Rijvi (Lighthazard) will be handling those while I'm away.
Once I get back, things will go full speed from there on! Now that I get emails direct to my phone, you can rest assured I'll be reading your emails within minutes after you send it ;). ALL emails and IMs will be responded to right away when we're online and not AFK. Order production will go much faster and once we get enough money it'll all be invested in designing new products (especially the new controller) and having them FULLY assembled this time, not just partially like the current controllers. Very expensive process, but it allows us to handle even more orders since we'll only be programming the chip and soldering on the wires, nothing more. Maybe we'll offer a DIY option for the user to solder on the wires and heatshrink the controller themselves at a discount with instant shipping (since we do nothing). Also, nothing will be allowed to remain out of stock for more than a week unless we state otherwise (such as my trip).
Site Changes:
New tutorial section is up (top nav menu)! It shows step by step photos of how to install our products, so you won't have to worry about not following the manual correctly ;). We'll add more tutorials for various AEG and product combinations as time goes on.
Added "wire length" to our product configuration options. If the default 12" wires [total] is too short, then you may select up to 48" wires! There will be small price increases for longer wires to make up for the higher shipping costs.
- Frank (PsyphyerVII)
May 11th, 2009
More Ideas & Improvements - v4?
We recieved quite some feedback when the Sentry controllers was placed back in stock over the past weeks, and three good suggestions came up:
- 1) Possibility of being able to change certain configs (burst count, li-poly threshold, etc) on the fly without having to reorder or send in for reprogramming.
- 2) Increase the MOSFET current capacity, such as using the 450A IRF1324S MOSFETs instead of the current 80A FDB8441 MOSFETs.
- 3) Error flashes are a bit hard to count and to remember what error each flashing sequence represents. Perhaps use motor vibrations?
So this is our attempt at a new version which will attempt to meet those requests:
 Proposed v4 controller |
These are the changes we have applied:
Complete Flexibility
The "control panel" currently on the sentry/pulse/scorpion controllers consists of 3 DIP switches and one side-actuated potentiometer. The switches can disable/enable burst, lipoly protection(pulse & scorpion), and realism mode(scorpion only) and the potentiometer controls the RoF/motor speed. Normally I would try to put as many switches a possible so the users gets more flexibility with the controllers, but at that time it seemed a 3 DIP switch was the biggest I could fit on the board without increasing the width. But now that changes after stumbling across this: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=563-1019-2-ND, a half-pitch 8 DIP switch, so small that it fits on the current controller and is just 1/2 the height of the current DIP switch, reducing the overall dimension of the controller! Below is the current 3 DIP setup compared with what I have planned for the 8 DIP setup:
|
|
|
|
|
|
|
|
Switches 1 & 2 - Realism Mode
|
Switches 3 & 4 - Burst Fire
|
Switches 5 & 6 - Lipoly Mode
|
Switches 7 & 8 - Speed Control
|
Massive Current Rating
I usually find it pointless to have a MOSFET rated for anything above 100A (continuous) since every AEG I've worked with has never drawn even close to that much. The fact that AEG fuse ratings are usually 10-30A and barely any AEG batteries are rated for >50A shows just how unlikely it is to be drawing 100A. But I guess I can't ignore the possibility that somewhere, someone does have a >100A setup (or maybe too tight shim job), and since people like higher current ratings anyways, I shall upgrade the MOSFET. :)
The MOSFET some people have suggested is the IRF1324S: http://www.irf.com/product-info/datasheets/data/irf1324s-7ppbf.pdf
This seems like a great choice, very high 450A current rating, and at a very great price of $1.50 in bulk. BUT one thing that caught my attention was the fact it's not a logic level MOSFET according to the manufacturer, meaning the gate MUST be driven at >10V before it can operate at its claimed ratings, whereas microcontrollers can only output 5V. At 5V gate voltage, the IRF1324S can only handle a maximum current of ~100A (datasheet page3 - figure3), this means it can never get anywhere close to its 1640A pulsed rating nor its 430A continuous rating! We can use voltage charge-pumps but their charge-up times are too slow for a decent PWM frequency for speed control. A MOSFET driver/buffer will solve the problem, but only if your battery is >10V and it (just by itself) will typically consume, compared to our controllers, 30x more current, which risks damaging any li-poly that is left connected for about a month or two. The IRF1324 is clearly better than the 80A MOSFET we're currently using, but still presents some downsides no matter how it is interfaced with a microcontroller, so we had to look for a simpler alternative, preferably a high power logic level MOSFET.
A quick search on IRF turned up the perfect candidate, the IRLS3034: http://www.irf.com/product-info/datasheets/data/irls3034pbf.pdf
Several features of this MOSFET makes it more suitable than the IRF1324S, besides the lower current rating of 350A, for our application. The IRF3034 is logic level according to IRF, and at just 5V gate voltage it can pass current >1000A (datasheet page3 - figure3)! This means it can handle its 1370A pulsed rating and 350A continuous rating without any delay-causing or power-draining interface, just a simple 5V direct from the microcontroller will do. But it doesn't end there, the Power rating is a very high 375W and the resistance is <0.002ohms, making the IRLS3034 the clear winner for our new design. The top of the datasheet even mentions "DC Motor Drive" as one of its main applications (but not on the IRF1324S datasheet), which has to mean something! ;)
So for the next release, the new controllers will contain a IRLS3034 MOSFET with 1370A peak current & 350A continuous current rating!
Simpler Error Indication
Having a single red LED rapidly flash 1-3 times, pausing for 0.5 second, then repeating, might be a bit hard to count or to understand exactly what error it is. So instead we replaced the red LED with a duo-color LED (red & green). Turning on both the red and green would give yellow to use as a third color. Perhaps the color can be used to indicate severity, such as green flash for startup success (when battery connected), yellow flashes for low battery, red flashes for overheat. The LED is located right below the microcontroller in the picture above.
And no, we cannot use any motor vibrations as indication since most of our realism customers opposes having any non-realistic effects coming out of their AEGs. But if you do, you can email support or my email directly (psyphyervii@hotmail.com) to have a custom coded controller. Don't worry, I enjoy challenges! ;)
- Frank (PsyphyerVII)
May 9th, 2009
v3 Controllers Available!
After assembling a few of the new circuit boards by hand, we quickly realized that applying just the right amount of solder paste on each pad and placing the tiny components on was VERY time consuming and tedious. We took a couple of weeks to research the pricing and reputation of various PCB assembly companies and drawing the documentations required for assembly. We settled with MyroPCB (who also makes our PCBs), who offered to assemble all the tiny components onto the PCB for a very low price, whereas the large/easy-to-handle components will still done by hand to reduce costs. We went ahead and shipped out the all PCBs and components, and it was all assembled and shipped back within a week.
 Partially assembled PCBs |
 Close-up view |
After soldering on the larger components and testing the controllers, everything worked perfectly, which means the new controllers are now available!
 New v3 Sentry controller |
 Modular design |
Product update:
- Small tweak: microcontroller internal clock increased from 8MHz to 16MHz.
Product stock update:
- Sentry controllers are now in stock.
- UberFET controllers are now in stock.
- Pulse/Scorpion will be set to "in stock" after we finish the pre-orders.
- Frank (PsyphyerVII)
March 23th, 2009
v2 Controllers Sold-Out & Obsolete
We've been getting a lot of e-mails as to when we'll be back in stock. Since the green-board v2 controllers are out of stock and discontinued, we must wait for the new v3 controllers. Although we already recieved new PCBs, we've decided to send them to an assembly house. We are in the final stages of code tweaks and testing the controllers. Rest assured, we will inform you all on the front page when we are ready to begin sales!
- Rijvi (LightHazard)
March 15th, 2009
v3 Controller PCBs Arrived
The new batch of PCBs have arrived! Hopefully I dont come up with new ideas too soon or else these will be going in the trash too..
 PCBs for the v3 controllers |
The quality of the PCBs are top-notch, no visible flaws on any of the boards. Now to assemble one and give it a test run over the next couple of days.
- Frank (PsyphyerVII)
Feburary 14th, 2009
v3 Controller Progress
Its been few months since my last update and I have received tons of email regarding the progress of the new controllers. Since my last post below, the prototype v3 controllers have undergone multiple revisions, which each taking weeks to design, order parts, program microcontroller, test on breadboard, fabricate PCBs, assemble parts, test in AEG, have new idea, discard PCBs, repeat process, etc. resulting in delays (including college). We have thrown away many perfectly good boards in the process (including the ones in the previous post). But now we have settled on the following design:
 Finalized v3 board layout |
The biggest and most important change was to switch to the much more advanced PIC18LF2321 microcontrollers in the super tiny QFN28 package/casing, which requires an expensive QFN28-to-DIP adapter to program o_o.
 Old (top) & new (bottom) microcontrollers [yes, that's a BB] |
We are now in the process ordering the PCBs and updating the product pages and coding a new customization panel for each product page so you can have your controller customized the way you want it (burst amount, lipoly threshold, wiring, etc).
Product updates:
[Sentry/Pulse/Scorpion]
- The pre-soldered wires are now replaced with tough thinner-jacketed 16AWG wires. The thinner-jacket makes the 16AWG wires the same size as 18AWG wires, making it compatible with gearboxes that have limited wiring space such as the V3 gearbox motor cage.
- More customization options are now available. NOTE: since there will be many possible variants, ALL ORDERS WILL BE BUILT-TO-ORDER from now on. Since your order will need to be programmed and assembled first, do not expect us to ship right away! Failure to understand this will likely result in your order being cancelled.
- If the user accidentally shorts the trigger (solder bridge, exposed wiring) during installation, an "anti-heart-attack" feature will prevent the AEG from suddenly firing when the battery is plugged in.
- Many have asked for us to add an lipoly undervoltage protection, and so we did! It uses an algorithm that can compensate for voltage drops during motor operation, allowing it to compute the actual battery voltage. I think we have added pretty much everything that has been asked for on the Scorpion! :)
- In my previous post I mentioned switching the P-Ch braking MOSFET to an N-Ch MOSFET driven by a high-side MOSFET driver. That turned out to be a bad idea since MOSFET drivers are very slow, making them unsuitable for the PWM signals needed for controling the braking intensity. Thus we reverted back to our original P-Ch MOSFET braking design with the MOSFETs' tabs placed facing each other (with temperature sensor right in between).
- Now uses the PIC18LF2321 microcontroller. More I/O ports for sensors and features, and yet only 1/4 the footprint of the old PIC16F88 microcontrollers in SOIC18 package/casing! More room -> more components -> no compromise.
- Since the new microcontroller have two PWM ports, the multiplexer mentioned in my previous post is no longer needed, freeing even up more room.
- All this new room allows us to add an on-board control panel for the adjusting of options and motor speed. No more messy wirings! Refer to the manual for more details on the control panel.
[Pulse/Scorpion]
- The shot detection is sensed by an angle toggle switch which will need to be installed near the sector gear. The switch is activated by the sector gear's tappet cam per cycle/shot. Refer to the manual for more details.
[Scorpion]
- New ammo detection algorithm, more reliable and able determine if BBs have jammed in the hopup (or barrel, when the BBs back up into the hopup).
- New simplified sensor hopup design. The microphone (old sound-based shot detection in v2 controllers) is removed and the reflective sensor (black rectangular object in image) is replaced with a simple phototransistor to sense the tracer LED light reflecting off the BBs.
New product:
The UberFET is an advanced version of the standard AEG MOSFETs. The pad normally occupied by the P-Ch braking MOSFET is occupied by a secondary N-Ch MOSFET and wired in parallel with the primary N-Ch trigger MOSFET, allowing for double the current capacity. A microcontroller gives the UberFET the same protection as it does for the Sentry/Pulse/Scorpion controllers, making it a very powerful dual-MOSFET controller! Starting price is $28.
Price changes:
Sentry cost increased from $28 to $38 due to more expensive components.
Pulse cost increased from $55 to $58 due to more expensive components.
Scorpion cost REDUCED from $115 to $78 since hopups are easier to modify now.
The prices might soon get lower once we begin to have the PCBs partially assembled by machine, assuming the assembly costs aren't too high.
Documentation:
New installation manual is up [link]
Updated flash presentation [link]
- Frank (PsyphyerVII)
January 27th, 2009
Christmas Sale!
We're having a Christmas and New Year's special! Use the discount code Q5293D on your next order during checkout and get 5% off! Please note that this offer only lasts while the controllers are still in stock, so act quickly!
- Rijvi (LightHazard)
December 24th, 2008
Prototype v3 PCBs Arrived
The PCBs have arrived! Took a bit long since we decided to switch over to a PCB fabricator in Hong Kong, which required some research on thier capabilities. The quality turned out to be much better than I had expected!
 First batch of [prototype] v3 PCBs |
I'll post more after I've finished programming the microcontroller, assembled the components, and given the controller some test runs. The current v2 controllers will remain on sale with Rijvi (lighthazard) taking over most of the orders while I test the v3 controllers.
- Frank (PsyphyerVII)
November 7th, 2008
Prototype 3v Controller Preview
Due to the success of the v2 controllers, we have begun working on a new version of the controller with even more improvements and features. Preview of the upcoming v3 Sentry/Pulse/Scorpion controller boards:
 Board layout rendered in Eagle3D |
Changes:
- Circuit width reduced by 25%, now it is smaller than a typical AEG fuse holder.
- User-adjustable motor speed control via onboard potentiometer.
- Soft startup to reduce large current spikes during motor startup.
- P-channel braking MOSFET replaced with N-channel MOSFET & high-side MOSFET driver [needs testing].
- Microcontroller PWM pin multiplexed to both MOSFETs for full braking and speed control.
- Resettable fuse removed, replaced with current/temperature sensor. Protects circuit without the added resistance of a fuse.
- [Scorpion & Pulse] New high-reliability shot/gear position detection via an angle-toggle switch positioned next to the sector gear. Replaces the sound-based shot detection methods in previous models.
-
- Easy to install, long operating life, and fits in most gearboxes, even the "filled" V2 reinforced gearboxes.
- Provides burst-fire with absolutely NO adjustments needed (timer/sound/current-based burst controllers are not always reliable, requiring frequent adjustments/reprogramming).
Not changed:
- The price! My goal is to design the ultimate AEG circuit, I don't care too much about profit...yet. Most goes towards research and development anyways. :)
This circuit could be the most feature-packed AEG controller board on the market! There is currently no other AEG controller (other than the Systema Revolution gearbox) that has full-cycle completion or offers next-gen realism features to various AEGs (Systema PTW is currently limited to the MP5 and M4/16 models).
- Frank (PsyphyerVII)
October 14th, 2008terms