A little less Simple DIY TDCS circuit using CRDs | Insight

Regarding the transients at turn on, that is a circuit issue. So to remedy, it is easy to add an LC filter at the output see picture below for new schematic. The LC filter acts to dampen any transients. Bench testing shows the ramp up to be 500ms, which is plenty to dampen any turn on pulses, but unfortunately not enough to prevent any flash that occurs with certain montages. At mouser.com, the L can be 22R105C and the C can be UKL1E100KDDANA.

via Insight, a growth project driven by tDCS: A little less Simple DIY TDCS circuit using CRDs.

8 thoughts on “A little less Simple DIY TDCS circuit using CRDs | Insight

  1. My own design, which I’m planning to publish sometime on a blog, uses a very different approach. I included an audio taper potentiometer shunting the regulated line to ground.
    You set up with the current shunted 100% to ground, and apply the current by hand. The audio taper is to counterbalance the hyperbolic curve of current division as you change the shunt resistance.

  2. As a beta tester of this design, I agree the capacitor might present a problem.
    I have always put the electrodes on first except one time. I did get tiny shock out of the cathode that one time. Otherwise, I am quite happy with the way it works. The core issue as I see it is how to find a montage that really helps. I am looking for improving the quality of my rest. I I found something that I am liking. I do not run continously. I am trying 15 seconds on and 15 seconds off. Its seems just as effective and less buzzy in my case. Including ammeter and potentiometer might make this circuit a little tweekable.

  3. Seriously now. What is wrong with you people?
    Tell me how you rationalize assuming that the user’s head is a constant resistor? It’s NOT! Never assume!
    There will be at least 25% of the idiots using this thing who will first apply power, THEN put on the trodes, then plug them in, then get a VERY NASTY ELECTRIC SHOCK!
    But even if you discount those as Darwin award winners. This is a DIY thing. People will have trodes coming loose, cables getting unplugged, etc.
    And in that very moment, that “filter” ceases being a filter, and becomes a bloody life hazard!

    Question: What is the voltage across the terminals with the electrodes unplugged or unattached?
    Question: What does the capacitor do when the circuit is re-established?
    See? It’s a hazard! Don’t EVER build this!

    • Thanks for the feedback Sigmoid. If you’d like to write a guest post about which DIY designs have problems and which are solid devices I’d really appreciate it. John

      • Hi. I’m not Sigmoid but I agree with his observations regarding “The Open tDCS Project” circuit, It’s a bad circuit and you should remove it.
        These circuits adjust current to the one we want (1-2mA) using feedback. When the current is below our reference, the circuit ups the voltage across the electrodes to increase the current. The opposite when the current is above the reference.
        I see 2 problems with that circuit: If it’s turned on before applying the electrodes, the current will be zero at that moment and the circuit will try to compensate, increasing the voltage across the electrodes to the maximum. Then the user will put the electrodes and will receive the maximum current, which is not the desired result. To prevent this, there should be something limiting the maximum current to 2mA, maybe a resistor in series in one of the electrodes.
        The other problem is the capacitor in parallel with the electrodes. When the circuit is turned on without the electrodes in place, or when one electrode gets loose or a wire gets disconnected, the current will be zero, the circuit tries to compensate and the voltage across the electrodes will be maximum, and this will charge the capacitor. When the electrodes are put again or the wire reconnected, the user will receive maximum current and probably the circuit won’t be able to adjust the current fast because the capacitor is charged and will force the higher voltage for some time. The result will probably be a high current for an extended period of time.
        A better circuit won’t have this capacitor, will have a series resistor (maintaining the fuse) and will have a variable resistor (in place of R1-R4) to regulate the current (0-2mA). This resistor allows the user to slowly increase the current, allowing for a smooth start.

        • Do you mean the OpenStim circuit? There are 8 circuits in the post. If you click on the circuit and send me a copy the url of the diagram I’ll know exactly which you’re speaking of. Again, alas, at this point in time I myself don’t much about electronics. I’m still hoping someone emerges to lead the Open tDCS project but so far no takers.

          • I’m sorry for not cooperating with you on the project, I had my hands pretty much full, and I also wanted to do some tests beforehand.

            I’ll have a look at the upverter project you sent me, for potential safety problems.

            Anyway… I just did some tests using an oscilloscope on my CRD-based stim unit.

            The test scenario is as follows: The circuit contains a CRD and a 1k series resistor (which also acts as a shunt for current measurement). It also contains a shunt potentiometer to ground, for turning the device on and off, but that was left closed (at 50k resistance) for the entire test. The test load I used was a 4k resistor.
            (Note that with optimal electrode contact and salinity, the load resistance can be less than half this value, which means any transient spikes would in fact be much harsher!)

            Upon connecting the battery first, and the load second, the current does in fact overshoot (even with the series resistor to provide a buffer). The overshoot wasn’t very large (about 5%), but it still shows that while the CRD is a very reliable controller, it’s not magic.

            I think the lesson is that I was right in building in the ground shunt pot, as it allows me to safely connect everything, ie. keeping the electrodes cold, while also letting the current flow – which keeps the business end of the CRDs at a low voltage.

            So while a CRD in itself would overshoot, and contribute to phosphemes and a potential jolt if connected carelessly, I didn’t even go and check what a capacitor would do. It would make it worse – how much worse depends on its capacitance.

            So the answer to “what design should be removed”, is that any and all that feature a so-called soft start capacitor across the terminals.
            A choke in series with the anode would potentially tame the CRD further, however the size and weight of chokes might limit their usefulness.

    • Hi Sigmoid,

      Actually, the CRD’s are very good for varied resistance; when regulating current and the electrodes shift or dry and the resistance increases the CRD’s still regulate the current.

      Now the other point about turn on and turn off sequence is a valid point. I have been using a similar design and turning on and off at the battery pack with electrodes on. But if you switch after the capacitor with the power on, a charge will remain on the cap to the battery voltage. When the circuit is again flowing that charge will dissipate through the head and produce an on transient that is higher than the regulated current. It will be small duration, and the level will depend on the resistance of the head, and the 1k resistor in line will limit it. I think I will investigate the various on transients for the circuit with and without the C, and publish in the blog.

      Hence, to foolproof the design, the capacitor should not be included.

Leave a Reply

Your email address will not be published.