[Tutorial] Building your own DIY E-Stim Stereo Device

[JakofClubs]

I believe you need a "true RMS" meter to measure the audio voltage. They have came way down in price.

[tigrotto]

Leaving this here in case someone can help this guy on reddit since new subscriptions are halted because of the issues with the site.

[RacerXxX]

I managed to find the amplifier and the Speco transformer.
Now I was wondering if "10W 3.9 ohm resistors" would fit.
Later I was wondering if I could power it with a classic 12v led power supply. does it need direct or alternating current?

Thanks

[JakofClubs]

Later I was wondering if I could power it with a classic 12v led power supply. does it need direct or alternating current?

Thanks

Direct current. 12v DC, 5 amp output recommended. Double insulated "square within a square" logo for safety. The classic open frame LED power supplies I've seen aren't double insulated.

[RacerXxX]

Instead, do you think a 15w/1.3A power supply is enough?

[diglet]

I think there is a good chance it will be enough for most sessions. But if you hit the limits of your power supply, this will result in unpredictable behavior, so I would strongly recommend a better power supply.

[diglet]

Electronics question

I noticed when stimming in a three-phase setup, the center channel (when left and right are out-of-phase) appears to be more sensitive to volume changes. I'm not entirely sure what's going on, but I feel there might be a box issue.

I think it's obvious why the box is doing this -- the parallel resistor was chosen to limit the current over the left or right channel, if we use triphase, the resistors are only half as effective on the center channel. What box design changes can be made to limit the current over the center channel just like left/right?

[edger477]

Electronics question

I cannot understand this question. There is no "center channel", as there are only left and right channels. There is no volume control for center channel, so not sure how it can be more sensitive to volume control. Center electrode current completely and only depends on currents through L and R channels and electrode configuration - all the current that flows through that electrode comes from either L or R channel.

Maybe what you are saying is that when using triphase setup with +/- as common (out of phase?), then center electrode feels stronger when you increase volume to one (or both?) of the channels? As you used phrase "out of phase" so it could also mean you are using ++ or -- as common but volume change (to one or both channels?) is pronounced when the signals are out of phase (like funscript was at 100?)?

In ++/-- common, when L/R signals are out of phase, reducing volume on one of channels will redirect the current from other channel towards the common electrode (because it is flowing between L/R when the volumes are equal, and common will receive difference in volume).

In +- common, when L/R signals are out of phase, the common will be receiving sum of both currents from L/R so reducing any of them will also reduce the common.

I do not see how it could be related to resistors because the serial ones only make amplifier output less powerful so we have better volume control, and parallel ones prevent the amp from "stopping" due to inadequate load (with serial + transformer the load on amp might be too low for it to produce output consistently, most of them are rated for 4-8 or 4-16 Ohms), and also depending on transformer, it helps reduce spikes caused by transient states.

[diglet]

Maybe I don't understand the purpose of the resistors. I thought that the purpose of the parallel 22ohm resistor was to make the device act more like a constant current device. Without it, it acts more like a constant voltage device. This is important because the resistance of the skin is very nonlinear.

This all works well for a single phase setup. Let's assume for a moment that the resistors are tuned to limit the output at 20ma. But for a three-phase setup with a common neutral, the setup isn't symmetrical, so the center channel can see higher currents (20ma from each channel = 40ma).

What I sense is that, when I turn up the box volume, the sensation at the center channel increases much faster than the sensation at the left or right channel. This might be caused by the problem above, or it might be caused by differences in skin resistance/electrode distance/nerve sensitivity. I'm just guessing at this point.

I should *really* get an oscilloscope...

[edger477]

Oscilloscope will not help you much here, as even if it was 2 perfect current sources - if you make them parallel - the common will have sum of both electrodes, and vice versa.

I think this problem cannot be solved on electronics level. I think better approach would be to "normalize" the funscript to a level that is comfortable for certain electrode setup - essentially set max at 0.8 or whatever is the value where it does not sting (and also if it is stinging on 0 then move minimum to i.e. 0.2). This might be better approach than what I did with reducing the volume, because reducing the volume removes all sensation from the L/R electrodes when we are reducing it because of common, or all sensation from common when we are reducing it because of L/R stinging. I think we might not want the L and R to ever be fully out of phase or fully in phase because at these extremes either common has 0% of current other electrode's current, or 200% of their current.

[diglet]

I decided to take some measurements with the old fashioned multimeter. Test setup:



Test signal "R": 1vrms over output R-neutral, 0.5vrms over both L-neutral and L-R
Test signal "L": 1vrms over output L-neutral, 0.5vrms other output pairs.
Test signal "M": 1vrms over output L-R, 0.5vrms other output pairs.
(vrms is normalized to 1)

Test results without resistors, measuring directly at the box outputs:
Test signal R, vrms over output R-neutral: 26.4. vrms over the other outputs is half of that.
Test signal L, vrms over output L-neutral: 24.3
Test signal M, vrms over output R-L: 25.5
Conclusion: the test signals work as expected with no load. There might be a small l/r balance issue which could be corrected by turning the pots on the board.

Test results with 10k ohm resistors:
Test signal R, vrms over resistor R: 49.1
Test signal L, vrms over resistor L: 46.7
Test signal M, vrms over resistor M: 40.7
At higher/lower volume, the ratio stay the same.
What I expected to see: the same vrms over all resistors.

Test results with 220 ohm resistors:
Test signal R, vrms over resistor R: 10.9
Test signal L, vrms over resistor L: 10.7
Test signal M, vrms over resistor M: 4.6
At higher/lower volume, the ratio stay the same.
What I expected to see: the same vrms over all resistors.

With both test signal R and L, barely any voltage (~2.5vrms) is over M. It looks like the amp doesn't like generating a potential between R+ and L+ when loaded, but has no problems doing so if the resistance is very high.

I changed the volume between the tests because I didn't want to smoke any resistors.

conclusion: the box is clearly not symmetric in a three-phase setup. As the resistance drops, the voltage drops faster on the center channel than on left/right. I expected the opposite.

I think this problem cannot be solved on electronics level. I think better approach would be to "normalize" the funscript to a level that is comfortable for certain electrode setup - essentially set max at 0.8 or whatever is the value where it does not sting (and also if it is stinging on 0 then move minimum to i.e. 0.2). This might be better approach than what I did with reducing the volume, because reducing the volume removes all sensation from the L/R electrodes when we are reducing it because of common, or all sensation from common when we are reducing it because of L/R stinging. I think we might not want the L and R to ever be fully out of phase or fully in phase because at these extremes either common has 0% of current other electrode's current, or 200% of their current.

My software solves this exact problem . But if the box output (or nerve sensitivity) is nonlinear, configuring the software becomes very complicated...

[edger477]

With both test signal R and L, barely any voltage (~2.5vrms) is over M. It looks like the amp doesn't like generating a potential between R+ and L+ when loaded, but has no problems doing so if the resistance is very high.

I changed the volume between the tests because I didn't want to smoke any resistors.

conclusion: the box is clearly not symmetric in a three-phase setup. As the resistance drops, the voltage drops faster on the center channel than on left/right. I expected the opposite.

Did you connect to amp directly or to transformers? On 10k and even to some extent with ~1k you will have pretty expected behavior, but with loads higher than that (less resistance), the internal transformer resistances (which would be order of magnitude 100-200Ohms) need to be taken into account. I forgot how we model those resistances, as they are inside the signal source, so source is both resistance and voltage/current generator - probably can be represented as another resistor in series with source, but it gets more complicated when you connect them into triphase configuration. Also different volume would put transformers into different areas of their characteristics, so you can't really compare the measurements.

Also - box does not generate potential between R+ and L+ - the channels are totally independent thanks to transformers. All potential between R+ and L+ will occur when R- and L- are connected, and potential will be result of difference in input signal (from amp) to 2 transformers.

[mantrid]

the internal transformer resistances (which would be order of magnitude 100-200Ohms) need to be taken into account. I forgot how we model those resistances, as they are inside the signal source, so source is both resistance and voltage/current generator - probably can be represented as another resistor in series with source,

These calculations can be found in a spreadsheet I provided many years ago: series resistance R on the primary side is equivalent to series resistance R*n² on the secondary side, where n is the winding ratio. (In the spreadsheet series resistance on the secondary side has to be added to the load R_L.) DC-resistance (on the primary/secondary side) of the transformer is equivalent to series resistance (on the primary/secondary side).

but it gets more complicated when you connect them into triphase configuration. Also different volume would put transformers into different areas of their characteristics, so you can't really compare the measurements.

With current controlled output current at common is the sum of the currents at the electrodes. That's it. (And that does not mean the RMS current at common is the sum of RMS currents at the electrodes)

[Johankain]

I just wanted to show off my device - build it identical to the schematics in the first post. Had some problems fitting everything in the box ... for now only tested it on my leg ... it feels completly different than all other commercal devices i own.
I am somehow scared of connecting it to my junk
Will try it with a tease next week probably

I finished buiding the latest version with the new amp.
However I noticed that the two channels influence each other much more than on the version before.
What I mean is that when I switch on triphase one channel influences the other extemely.

E.g. on triphase the feeling on the left channel may vanishes or gets too stron when dialing up the right one.

I did not notice this in the previous unit?

EDIT: 2022/07/07 - I was testing the unit with different files ... after testing the two units with the same files I must say that this issue is not really present, I am sorry for writing stuff before properly testing the unit

Hi
I have ordered all the parts for the v2 of this device and will soon build the whole thing. I read the whole thread and noticed that you had possible issues and did not see them really resolved.

The way you have mounted the transformers leads to induction of the signal from one transformer to another. this happens because the windings are on the same plane and their magnetic fields affect each other. If you flip one of them to its side this does no longer happen. Same principle is kind of explained here http://www.troelsgravesen.dk/coils.htm

[edger477]

You would have to screw one of them to the side of the box for this, but I would argue that coils from this link are more prone to the effect due to not having the cores. Transformers we use mostly have core going through the center and around, so it will have much less magnetic field around because most of it will go through the iron core.