Thinking Cap v0.1

http://www.grindhousewetware.com/projects

If there are any questions or concerns please let us know.  
Tagged:

Comments

  • rdbrdb
    edited June 2012
    I've checked it out, and I'm a bit worried about your design since not only can it not actually induce a proper tDCS effect, it offers very little in safety features.  You seem to use PWM to adjust the current intensity, which produces a square wave, and not a constant voltage, so it is technically not tDCS, and I'm worried about side effects that this alternating current would have on the brain.  At least it should be lowpassed by adding a capacitor near the limiting resistor, but that's not quite sufficient since the target impedance can vary, which would change the cutoff frequency, plus there are other safety implications of just adding a capacitor like that.

    Even then, assuming a very optimistic transcranial+electrode impedance of 2.5 kΩ, and given that you're using a 5 volt PWM signal, it can't really produce currents in excess of 5/(2500+2500)=1 mA.  In reality, the impedance will be higher, and therefore the maximum current attainable will be significantly lower than the advertised 2 mA.

    Also worrying is the fact that you don't have any kind of feedback mechanism, no way to measure and accurately control the amount of current.  At least hook the other side of the resistor to one of the ADC pins (through a 50k resistor or so) so that you can measure the actual amount of current flowing through the resistor!

    Have you guys see the design I made and showed on the other thread?  It follows up the PWM with a lowpass filter and a transconductance op-amp setup to tightly and precisely regulate the amount of current, has a backup current limiter that doesn't significantly increase impedance like your resistor does, and it has several more safety features including ESD protection.
    Feel free to adapt concepts from my design!

    Sorry if I'm stepping on anyone's toes here, I just don't want anyone to get hurt as a result of using a design with potentially overlooked design flaws.  That is not to say that GHWW's stimulation device couldn't be useful to someone - it's just not tDCS, and it's probably not very safe.

    I spit out my coffee when I saw the pricetag: $100.  Isn't that a *bit* overpriced?  The hardware doesn't cost more than $5 altogether...

    PS.  Not that I would care, but since you don't have proper certification it is illegal to sell the device like that unless you specifically state that the device is for educational use only or something like that, and that it should not be used in vivo.
  • From Tim Cannon: 

    "rcb,

    Thanks for your feedback it sounds like you are quite knowledgeable on the subject, so we appreciate you critiquing our little device. As you know most tDCs devices take rudimentary electronics skills, and to build one with features begins to require so math, programming etc. Most devices in the market are priced over $300 we are just trying to stay competitive. We personally spend a lot of money investigating all forms of biohacking. We make this available for people with those skills to make it themselves and improve it as they would like. For those who would rather not learn the skills and build it themselves they will help us make some of our research money back, and help us fund the research of other projects. We release all products every 6 months so many of the features you describe are in that product back log for the next release.

    As for possible amounts of mA coming out of the device, I am think I will trust my multimeter and it said 2.12mA so I am not sure what the deal is there.

    Hat $4
    Chip $4
    Clock Circuit $2
    12 pack trodes $4
    Shipping at least $5

    So yes we do have a mark up, but that mark up is for people who would rather pay $100 than learn those skills. Hell we even provide a recipe."
  • @Lukas please don't take that as personal attack on you or your work, or your buddies work. But as electrical engineer i can assure you that rdb is correct in all points.

    The biggest problem is that the device does not deliver constant current at all, but pulses instead. Your multimeter will not tell you. You need a scope to meassure that.

    The limited operating voltage, the "safety" resistor and the lack of controll/regulation of the current make the circuit very prone to changes in the electrode-impedance (which can, and therefor will happen, maybe not in your experiments, but it is a variable the user has influence on so it needs to me taken care of).

    Other than the resistor, there are no further safeguards. Keep in mind, this is a medical device, consider failure to be lethal. Altho there is little risk of failure in the circuit itself, there is always the user. And in case the user connects a wrong power supply there is nothing that stops your circuit from frying a brain.

    I do understand your desire to keep the cost and complexity low. But that must not interfere with the intended operation and safety. And as you can see in rdb's circuit, a current safequard from just a jfet+resistor is easy enough to add.
    A propper PWM-controlled CC-source can be made with a lowpass connected to a BJT's base and a resistor between emitter and gnd.
    That makes 6 discrete components, the atmega (no need for an external clock), and the power supply.

    As electrical engineer i really advise you to un-release the circuit and redo it. Or ask someone with experience to do it for you. Right now, a potentiometer makes a better tDCS, sorry but that's a sad physical fact.

    i won't comment on the pricetag other than there's just no excuse.

    I hope you had a fun learning experience while working on this project.
  • rdbrdb
    edited June 2012
    I would like to add that the circuit I made is barely $0.50 more expensive than the one you've made, and that I think that you should not make compromises when it's about people's safety.  Make no mistake that something as simple as an electrostatic discharge, or accidentally hooking up the battery in reverse, could kill or severely injure someone using your device.

    The circuit I linked to doesn't use concepts that aren't taught within the first semester of an electronics course, but I'd be happy to explain anything that isn't understood.  I think it is important that anyone who builds something that's going to be hooked up to a body using electrodes understands exactly what they are doing and what could possibly go wrong, especially in a DIY setting.

    Your multimeter doesn't have the same impedance that your electrodes and skin do (current meters typically have one approaching zero, whereas your transcranial impedance is in the kΩ range), and only measures direct current.  Your device does not produce a direct current, so you cannot rely on your multimeter readings.
    The math, however, doesn't lie.  Your device cannot produce 2 mA of transcranial current.

    (By the way, you could use a simple $0.50 ATtiny instead of the expensive ATmega microcontroller.  An ATtiny is a lot cheaper, smaller and consumes less power; and even the cheapest ones are nonetheless perfectly capable of running your program.  You don't need a clock circuit as Thomas pointed out.)

    I'm sorry for the large amount of criticism coming from me.  I don't want to seem like a condescending jerk, but I just don't want people to get the impression that this is a safe tDCS device, or even a tDCS device at all.  More importantly, I don't want anyone to get hurt.
  • I definitely don't interpret this as a personal attack.  I appreciated the feedback that we are getting from you guys - this is what the community is for. 
  • It seems as though it may be a good idea to have the community informed on the projects as they are being done by grindhouse. Like a journal-like update topic for the community to give feedback on during the process to change any possible problems before a release.

    I personally don't mind the high price tag as you guys have every right to get money back on money spent for research. Not to mention payment for all the time spent on the project. rdb makes some very important points though which unjustifies the price and marketing of the device.
  • edited June 2012
    Hey guys, mostly in agreement, the feedback you gave is valuable and we will implement the safety suggestions in the next release. 

    I do question why the signal coming from that pin(post ramp up) would be square wave if we use digitalWrite(HIGH) which I believe automatically detects that you want to switch the pin mode and use it as a digital pin which isn't a PWM signal at that point is it? 

    But while I am here I want to correct one really bad misconception here.... safety is NOT our highest concern, people whose highest concern is safety don't put magnets in their fingers, or dampen sponges and zap there brains.

    @ThomasEgi this thing is a medical device in the same way that heroine is medicine. No heroine dealer in the world is trying to make his product less safe, just making it more safe isn't the highest priority, the highest priority is user experience.

    @rdb the criticism is good, keep it coming but lets try to stick with facts, we don't care what you think of the price point(not being mean, we just don't care), you are not our target market. Apple doesn't care what a java developer thinks about iphone, the target market is rich old women and scarf wearing hipster douchebags. ideally your target market can't easily reproduce your design, I wouldn't market a prebuilt PC to you guys either, no price i could offer would be worth it for both me and you because of our skill level. in this case you are leveraging curiosity with people who don't want to learn our skill set. I am not charging for the device, the device is free. I am charging to build it and to know how to build it. you want to offer it to people at cost be my guest. 
  • rdbrdb
    edited June 2012
    At 100% duty cycle it is indeed a constant signal.  Anything below that is not tDCS and potentially an additional danger.  But you must understand that even at 5V, you can't possibly come close to 2 mA of output current.  Are you not taking the impedance of the electrodes and your head into account at all?  You won't achieve 2 mA unless your head is made of metal.  The math for a constant current goes I = 5 / (2500 + Z), where Z is the impedance of your electrode loop, which is going to be several kiloohms, typically this can be up to 5 kΩ, limiting the current to just 0.67 mA.  (If it is not, then there is definitely a short circuit somewhere!)
    This impedance can change all the time, though, as Thomas pointed out, so due to the lack of regulation the resulting current could vary.  (At which point a bunch of other components are introduced into the impedance calculation).  Since there is no measuring circuit built-in (just pull the electrode lead to an ADC input with a high-resistance resistor), there is no way to even tell what kind of current is currently flowing.

    My main concern isn't the price, it's that you're offering this as an end-user product with no disclaimer whatsoever.  The customers you are targeting will have no idea what kind of risk they are taking.  These are exactly the kind of customers that would wire up a battery the wrong way, or something stupid like that.  As Thomas pointed out, you should assume that failure is lethal.

    Don't get me wrong, I'm not trying to discourage you in your endeavour.  I think it's great that you want to lower the threshold for people to tinker with such things.  I just think that you should be aware of the kind of trouble you or your customers could be in for if something were to go wrong (and likely, something will go wrong due to the lack of safety features).  You are liable for any injury other people might incur due to use of your device, besides the fact that it is illegal for you to sell it as such without proper certification (which you won't get without adding said safety features).  I don't really care about the latter, but you might when you suddenly get a big-money lawsuit against you.

    So really, I don't mind you selling an arduino hooked up to a resistor, some sponges and a hat for $100.  But I think it is important to be responsible and honest with your customers, put up a big disclaimer and tell them exactly what the risks are of using this device, that they are to understand the circuit before they use it, that they could die using it, that you don't give a damn about whether or not they die using it, and that you aren't to be held liable when they do die.  Doing otherwise is in my opinion misleading, dishonest, not to mention illegal.

    That said, I do not see why one wouldn't toss in some additional cheap components and make it safer.  I disagree with the idea that grinding is inherently unsafe, and I am personally committed to lowering the threshold for people to mod their bodies by making it as safe as possible.  It is the unnecessary risks that grinders often take that give us a bad image in mainstream society's view.

    Anyway, I know I may sound a bit harsh, but please don't see any of this as a personal attack or as an attempt to degrade you or your work.  I admire the cause, I am simply worried that this hasn't been thought out as well as it should have.

    I also agree with Shishou.  One of the great things about this community is that we can collaborate on projects like these, and discuss the changes until an iteration is reached that everyone can agree on.
  • IanIan
    edited June 2012
    I'm the main designer behind the thinking cap, and as such the design is mainly my responsibility.

    Both @rdb and @ThomasEgi seem to think that the cap is delivering square pulses throughout the entire stimulation period.  It isn't.  PWM is only used for the first ten seconds of the stimulation (or some other amount of time, which can be set by the user).  The rest of the stimulation is a completely constant current (well, constant to within what it is possible to be constant, anyway).  Also note that this should mean that the reading from any DC multimeter should be accurate enough.

    So, what about the ramp-up period's effects being too much of an unknown to be considered safe?  Well, square pulses are actually used in another form of brain stimulation called Cranial Electrotherapy Stimulation.  While the jury is still out on the efficacy of CES, its safety is as well-established as that of tDCS or tMS.  Granted, CES isn't ramping the current up like I am, but I don't see why that should make a difference.  Now, the long-term effects of CES are unknown, but AFAIK so are the long-term effects of tDCS.

    Now for the big one:  safety.  Does the thinking cap need more safety features?  Absolutely.  I'm already trying to think of ways to make it safer, such as using a current-limiter instead of a resistor, and your suggestions to this end (such as the feedback mechanism) are greatly appreciated.  Having said that, I think I can stand by the safety of this device to within what would be expected from a v0.1 release.  This machine simply doesn't have enough moving parts for many, if any, things to go wrong.  Even the scenario of a user connecting the cap to the wrong power source is very unlikely, as there simply is no potential "wrong power source" to connect to.

    Again, I do want to emphasize that newer models will have more safety features than this one.

    I do know of @rdb's proposed circuit in the other thread, but I unfortunately haven't been able to really take a good look at it.  Your description of it, though, does make me interested in looking at it, and even using it.

    Believe me, the current model of the thinking cap is still far from what I envision for it.

    ~Ian

    P.S.  I agree that the device is somewhat overpriced; that one wasn't my decision.  I don't know anything about finances, though, so you'd have to ask @Lukas or someone else why it has the price it does.
  • i highly recommend you look at rdb's design. he educated himself rather well, and 2 more electrical engineers looked over the circuit. not only does it handle regular operation as desired , but also remains safe in many failure cases. such as ESD, powerloss and software-failure.

    btw, mind to draw a rough shematic of your meassuring circuit? cause if you meassured the wrong way, the result may look right but might be wrong anyway.
  • @rdc all biohacking products ours included should come with the following disclaimer.

    "Buyer beware, doing this is stupid, the guys who made this.... are stupid. They have crazy ideas, some of which may have come from psychedelic mushrooms. There is no way to determine how this might effect you, frankly you might be wasting your money. If you do this incorrectly... and even if you do it correctly you could come down with a serious case of deadness. Consider this device hazardous consider something safer like juggling swords whilst blindfolded."
  • @TimmyCNinja:  Lol.  Like I always say, "I'm stupid.  Don't copy my stupid ass."

    @ThomasEgi:  I certainly intend to look over it.  From a brief glance, I don't have the components to test it properly right now, but once I get back to the United States I'll have more time and components available.
  • Dear biohack.me,

    After speaking further with the designer we realize that our initial price point which we were basing off of currently available units was off. We apologize and as you can tell are new to business.  Thank you for your honest and strait forward opinion.  Expect an upgraded version of the Thinking Cap within the coming weeks.  

    Thank you for your patience,

    Lucas 
  • rdbrdb
    edited June 2012
    @Ian: I'm glad to hear that.  Just as a clarification, the primary feedback mechanism is primarily to maintain a constant current independent of the head impedance.  Your current design is more of a voltage source than a current source, and as such would require the head+electrode impedance to remain constant for it to produce a constant current.  The resulting current will always depend on this impedance, and it will always be much lower than 2 mA as I've demonstrated above even assuming 100% duty cycle.
    The op-amp in my design always tries to adjust the output to keep the voltage potential on the + and - inputs equal, which means that it will always maintain a voltage over the 390 Ω resistor equal to the one set by the PWM+lowpass, so the current flowing through the resistor will always be set_voltage / 390, where set_voltage is in the range from 0 to 1.2 V (I've voltage divided 5v to 1.2v so that the internal reference of the ATmega can be used for the ADCs).

    I didn't know about CES, interesting.  If that has proven to be safe, then I withdraw my objection about that being a potential danger - assuming your PWM frequency is within the ranges tolerated by CES.  My other concerns still stand, of course.

    @TimmyCNinja: Good to hear.  I couldn't find that disclaimer on the product page though - are you only telling your customers that after they have purchased the products?
Sign In or Register to comment.