So, I have been using a selfmade estim box for a while. It’s quite bulky due to the audio transformers which can handle much more power than needed.
Inspired by a post I saw of someone using a single EE14 1300:8 transformer per channel, I ordered a pack of 5 transformers off Aliexpress and started experimenting. The single transformer output was too weak, and not having a series resistor made it feel very ‘buzzy’.
So I decided to double up the transformers, to reach twice the output voltage. Putting the 8 ohm sides in parallel and the 1300 ohm sides in series, we get an effective turns ratio of 1:25, comparable to the big transformers. Perfect!
I am driving the outputs with a cheap ZK-502MT bluetooth amp that does not have output filters. Note that these transformer can handle frequencies well above the audio range. When I measured the output waveform of the transformers with my oscilloscope, there were some very short spikes visible, so I added a 100 nF capacitor to smooth those out. (The capacitor combined with the transformers resistance creates an RC lowpass filter.) Take care to use a capacitor with sufficient voltage rating.
Here is the circuit (for one channel, repeat for stereo):
Transformers are labeled as 8:1300 to indicate that the ‘primary’ is used as output. The primary side is typically indicated with a dot or bump.
I power the amp with 12V DC, coming from a USB powerbank and USB-C PD trigger board.
I am very satisfied with the result. My new stim-box is now compact enough to be portable. The feeling is more ‘direct’ than the big transformers, and I can reliably achieve a HFO.
Edit: Updated circuit, moved the filter capaciter to the low voltage side.
The filter capacitor could be omitted if your amp has built-in filters.
RC filter with 3.9 Ohm and 2 uF has a cutoff frequency around 20 kHz.
I’m a bit concerned about the voltage rating on the caps. With an effective 25:1 turns ratio, the open-circuit voltage can reach up to 300v. Have you considered placing the caps before the transformer? This is the approach I take in an upcoming design.
1300:8 is the impedance ratio, meaning the turns ratio is sqrt(1300/8) = 12.7.
Assuming a full bridge and 12 V supply, the amp output is 12 V peak, transformed to 153 V peak (minus losses).
Fair point, I measured the unloaded output at around 70-100 Vac, but don’t remember if that was at max volume. There is certainly some power loss in the series resistor.
.Placing the capacitor before the transformers is possible too; it will form a RC lowpass with the 3.9 ohm series resistor. You will need a larger capacitance though.
I often wonder about the purpose of the series resistor. I think, the usual scenario happening is the electrode contact suddenly getting worse, not getting better.
With a low impedance source, and the load (body) impedance rising, the voltage would stay constant. So, if for example half of a pad peels off, the current density (current per area) of the other half keeps the same as before.
With a high impedance source, and the body impedance rising, the voltage will increase, and so will the current density of the remaining pad, which might cause pain.
So why does every circuit I have seen use a series resistor?
The amp is a voltage source, designed for maximum power into a typical 4 - 8 Ohm load. Adding the series resistor turns the amp+resistor combo into an (imperfect) current source feeding into the transformer(s).
That’s what I’m talking about, the series resistor turns a voltage source in a (more or less) current source. I half a pad comes off, a current source drives the same current through half the contact area. I wonder why that’s the desired behavior.
The commonly used amplifiers have a minimum output impedance of about 4 ohm (exact value can be found in the datasheet). Without the series resistor, the amp will very likely blow itself up very soon after the transformer hits saturation. I managed to blow up one chinese tpa3116-based amp when testing small xicon 42TU200 transformers without a series resistor.
Total current flowing through the electrode is a much better proxy for subjective intensity than current density. Simple things like getting hard of moving around can change the skin resistance by about 20%. I would focus more on those scenario’s than the occasional pad peeling off.
Ok, I re-measured output voltage with all volumes at 100%.
Sine wave 1000 Hz, no load: 96 V.
Sine wave 750 Hz, no load: 89 V.
At full volume I hear a distorted sine emanating from the circuit, the transformers being driven into saturation? I rarely go above 50% volume, not noticing any sound.
The easiest way to see if the transformer hits saturation, is to hook up the scope.and measure the waveform before and after.
I just hooked up my stereostim box with 10:1 turns ratio. With an input voltage of ~12v, it reaches an output amplitude of 125v before evidence of clipping occurs, as expected with this turns ratio.
The higher voltage with the higher frequency waveform suggests saturation occurs.
Discovered I cannot trust my multimeter
The oscilloscope shows 250 V peak-to-peak sine wave at 1 kHz with volume at 75%.
Further volume increase only adds distortion, no increase in peak voltage.
My usual volume setting corresponds to 50V peak-to-peak. This is unloaded, so in practice will be a bit lower. The capacitors seem to have survived the test on full power. Nevertheless, they should be upgraded.
