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

[diglet]

The issue with the unbalanced center channel kept bugging me, so I decided to perform more measurements. This time with an WYE resistor network. This is the expected output (voltage over the resistor) when rotating the position.



This happens on actual hardware:



It's clear that the hardware has an easier time driving the neutral electrode. It's a bit difficult to see because I used the cheapest multimeter on the planet to perform these measurements, but the full degrees of freedom are there.

I hypothesize the hardware does a modified version of the ab-transform in which the poles are not at a 120° angle. The first image was generated with the usual ab-transform, what happens if we use a modified ab transform with the poles at an angle of 150° relative to neutral?



That looks a lot like the measurements from the 220 ohm case. let's call this transform the hardware transform. Now all we have to do is calculate the inverse of this transformation matrix, multiply that with the waveform before we send it to the box, and the box will output a sweet, sweet balanced waveform. One-parameter calibration, no hardware modifications necessary. I will document the full details later.


I think we can do the same in hardware by simply adding a resistor to the neutral output. I tried adding a potmeter, but it didn't take long before it released it's magic smoke...


Edit: here is a measurement after calibration:

There remains a small L/R calibration issue, but otherwise the box generates the waveforms I expect.

[Stéropès]

Hello!

Thanks for all the info you share here!
I made my own stereo e-stim box based the parts list proposed by mantrid.

It worked very well... until the third session when the 10 ohm 10W (Edit: 5W in fact) resistors started to burn out!

I'm not sure I understood mantrid's calculations correctly. I thought that 5W was enough for a normal use. But in fact, if you look at the characteristics of the amplifier I used, at 12V it can deliver 20W per channel. Is it possible to reach these values without feeling any pain?
The transformers are also indicated for a power of 65 mW, which is much less than what I see in the calculations of mantrid. Is it ok to use them?

Is it just a question of dissipation, and I should have fixed the resistors to a metal plate?

Would 20W ceramic resistors be more appropriate or would it be better to use a metal heatsink with a higher capacity?

Here are some pictures of the resistors (you can see that I also made the mistake of connecting cables too close together which started to burn).

Capture d’écran du 2023-03-01 12-34-09.png (995.44 KiB) Viewed 2412 times

Capture d’écran du 2023-03-01 12-28-10.png (609.51 KiB) Viewed 2412 times

[diglet]

Are you sure those are 10W resistors? They look like they're about 17mm long, an appropriate length for a 3W resistor?


Last time I tested my box, a sustained total load of 12W was sufficient to push the 2x25W resistors in an enclosed box to 60°. Under normal use, temperature and power draw are lower.

I tested a 22mm long 5W square ceramic resistor under a load of 2W, and it took only a few minutes to hit 70°. I conclude that your resistors might be able to dissipate 10W indefinitely, but they will do so at a temperature hot enough to melt anything else nearby...

I think mantrid sized those resistors with good airflow and aggressive amplitude modulation in mind.

Edit: I found a datasheet of a 10w round resistor with a length of 38 mils, the datasheet specifies that a sustained load of 10W results in a surface temperature of 300° over ambient.

[Stéropès]

Are you sure those are 10W resistors? They look like they're about 17mm long, an appropriate length for a 3W resistor?

Good point!
They are 5W resistors. I don't know why, I was convinced that I bought 10W.

https://www.digikey.fr/fr/products/deta ... 0R/9926178

[edger477]

I made my own stereo e-stim box based the parts list proposed by mantrid.

Can you also post the schematic you used? I am assuming no parallel resistors, so amp power would actually be limited due to resistance connected being higher than 4 Ohms. For my box I used these resistors, screwed to the tin plate that holds everything in box, and with a bit of thermal paste between https://www.amazon.de/gp/product/B08FM6M76Y

[Stéropès]

Yes, there is no parallel resistor.
Only the high pass filter and a TVS diode.



Thanks for the suggestion. That's what I was going to do too, buy resistors with heatsink and put thermal paste.
But I wanted to make sure I understood where the problem was before making the modifications.

[Stéropès]

Is putting a resistor parallel to the transformer useful in my case?

The DC resistance at the primary side is only 1Ω. As I understand it, putting a parallel resistor of 22Ω or more will result of an equivalent resistor just a little less than 1Ω, out of the 4Ω - 8Ω scope of the amplifier. Right?

[diglet]

The purpose of the parallel resistor is to make the device act more like a current controlled device, as opposed to a voltage controlled device. I have not tested this but reportedly this improves the sensations.

[edger477]

Parallel resistor can only change sensations if amplifier's response is different at higher volume.

All parallel resistor does is only adds more load to amp, and that additional load is converted to heat. So to have same output on transformers you need to turn amp to higher volume. You do not want a parallel resistor to burn out while you are stimming, so if you will use one, better be enough power and properly cooled.

The only reason where it will improve sensations is on amps that need 4-8 Ohms load to work properly and with only serial resistor + transformer amp might not have enough load to function properly, causing it to be instable, have painful bursts etc. If amp works fine without parallel resistor then there is no need to have it (on the contrary you just produce more heat and spend more enerfy with it).

[Stéropès]

The purpose of the parallel resistor is to make the device act more like a current controlled device, as opposed to a voltage controlled device. I have not tested this but reportedly this improves the sensations.

Thanks, but I didn't understand that.

For me the goal is to have a load for the amplifier that is in any case close to that for which it was designed (4 - 8 ohm) to avoid unpleasant effects when increasing the volume.

Also, the side effect (and I think undesirable) that has been explained in this topic is the opposite, the resistance makes the output less current driven and more voltage driven:

This parallel resistor also makes the voltage on the output less dependent on the load connected (basically you have less voltage drop when you use electrodes with larger surface area), might be useful for some scenarios, i.e. if you make venus/tremblr receiver lined with conductive rubber, so your electrode surface area is variable.

Parallel resistor makes the output less current controlled and more voltage controlled. That can also be achieved by choosing a lower series resistor value of by choosing a smaller winding ratio (I would go the latter way). Whether the estim signal is current controlled or voltage controlled matters when the contact surface changes. With current control reduction of the contact surface can cause high voltages and high current densities which can feel stinging. With voltage control the opposite happens: The signal feels less intense if the contact surface is reduced. If you experience that problems, you should change you electrode setup.

As the winding ratio of my transmitters is 22,4:1, I think I can manage to put a resistor in parallel while keeping the whole set sufficiently dependent on the current.
So, can I achieve this goal with a only 1Ω DC resistance at the primary side of my transformers or a parallel resistor will count for nothing in that case?

Edit:

Parallel resistor can only change sensations if amplifier's response is different at higher volume.

All parallel resistor does is only adds more load to amp, and that additional load is converted to heat. So to have same output on transformers you need to turn amp to higher volume. You do not want a parallel resistor to burn out while you are stimming, so if you will use one, better be enough power and properly cooled.

The only reason where it will improve sensations is on amps that need 4-8 Ohms load to work properly and with only serial resistor + transformer amp might not have enough load to function properly, causing it to be instable, have painful bursts etc. If amp works fine without parallel resistor then there is no need to have it (on the contrary you just produce more heat and spend more enerfy with it).

How to know if amp works fine without parallel resistor?
I don't understand how the parallel resistor add load, the transformer and the resistor doesn't add like 2 resistors (1/Req=1/R+1/RtransformerDC)?

[mantrid]

Parallel resistor makes the output less current controlled and more voltage controlled. That can also be achieved by choosing a lower series resistor value of by choosing a smaller winding ratio (I would go the latter way). Whether the estim signal is current controlled or voltage controlled matters when the contact surface changes. With current control reduction of the contact surface can cause high voltages and high current densities which can feel stinging. With voltage control the opposite happens: The signal feels less intense if the contact surface is reduced. If you experience that problems, you should change you electrode setup.

As the winding ratio of my transmitters is 22,4:1, I think I can manage to put a resistor in parallel while keeping the whole set sufficiently dependent on the current.
So, can I achieve this goal with a only 1Ω DC resistance at the primary side of my transformers or a parallel resistor will count for nothing in that case?

In normal operation a parallel resistor just generates heat (and a 1 ohm resistor will generate a lot of it). If it feels better with a parallel resistor, you can achieve the same by reducing the serial resistance or the winding ratio..

An other possible effect is that the parallel resistor (which increases the load) drives the amplifier into clipping. That clipped signal feels softer.

How to know if amp works fine without parallel resistor?
I don't understand how the parallel resistor add load, the transformer and the resistor doesn't add like 2 resistors (1/Req=1/R+1/RtransformerDC)?

The impedance (=AC resistance) on the primary side of an ideal transformer is

  Z = R1 || ( Rp + Rc/n² + Rl/n²), where

R1 ... optional parallel resistor
Rp ... DC resistance of the primary winding
Rs ... DC resistance of the secondary winding
Rl ... Load at the electrodes
n  ... winding ratiop
|| ... parallel operator: a || b := 1/(1/a+1/b)

Edit: Formula above contained an error. Fixed this.

[Stéropès]

Thank you mantrid, I understand better how to compute the impedance of the transformer.

In my case, I will stay with the 10Ω serial resistor but with a heatsink and I will mount it with thermal paste on the metal frame.

I have also plane to open a new topic to centralize the different information you have detailed in different posts on the forum mantrid, edger477 and others. This would also be an opportunity to discuss in more detail other design evolution, such as the addition of fuses for example.

Does this seem relevant to you? Does it interest other people here?

[edger477]

I don't see purpose a fuse would play, because when you short these boxes for reasonable amount of time (i.e. <30 sec) there should be no damage done to box, so fuse would not prevent any damage, but it would end your session when you have to replace it.

[zebbg69]

I have also plane to open a new topic to centralize the different information you have detailed in different posts on the forum...

Does this seem relevant to you? Does it interest other people here?

I would like to hear more about the parallel resistor shifting from "voltage control" to "current control" or creating "smoothing." I have been puzzled by these claims for a long time and can only imagine that they are based on interactions with the amplifier's output stage, or perhaps assumptions about capacitance in the body resulting in a filter.

What mantrid just said about the extra current draw pushing the amp toward clipping is a hint in this direction, and I would like to hear more to understand this better. Thanks!

[mantrid]

I have also plane to open a new topic to centralize the different information you have detailed in different posts on the forum mantrid, edger477 and others. This would also be an opportunity to discuss in more detail other design evolution, such as the addition of fuses for example.

Does this seem relevant to you? Does it interest other people here?

A forum is not the best place for such a knowledge collection. After a while these information are lost because its almost impossible to dig them out. That's also the reason why the same topics are discussed over and over again.

The right place would be a Wiki. Unfortunately neither Milovana or nor similar sites (e.g. EroScripts or SexScripts) have one.