Sam's TMS Coils

Version 1.10 (30-Oct-18)

Copyright © 1994-2019
Samuel M. Goldwasser
--- All Rights Reserved ---

For contact info, please see the
Sci.Electronics.Repair FAQ Email Links Page.

Reproduction of this document in whole or in part is permitted if both of the following conditions are satisfied:
  1. This notice is included in its entirety at the beginning.
  2. There is no charge except to cover the costs of copying.

Note: For the captions, "Toroid" is a bundle of wire wound with a roughly circular cross section, "Pancake" is a single layer of wire wound in a tight spiral, "Stacked" refers to multiple pancakes on top of one-another, "15T" denotes 15 turns, "2x5T" denotes 10 turns total in 2 pancakes of 5 turns each stacked, "D" is mean diameter.

Unless otherwise noted, the following air-core coils ran without problems, meaning that the pulser did not self destruct, at least during brief testing. :) However, only Sam #13 (the last one constructed to date) produces any cerebral response, and only with a minimum of 70 or 80 Joules of stored energy. Some or all of the other air-core coils may also work at a higher energy level but have not been tested with more than around the 26 J of the Ver. 1 pulser. However, at this energy, most of the air-core coils were truly excellent for peripheral nerve stimulation, and could launch aluminum rings across the room. ;-) The two iron-core coils were, to put it mildly, total duds in every respect, except perhaps for use as ugly paperweights. ;( :)

Sam #1, #2

Sam #1 was the initial test coil. Its design was based on the desire to have an inductance that would result in a biphasic period of 200 to 300 µs with the 25.6 µF energy storage capacitors of the Ver. 1 pulser, and have the physical appearance of a figure-8 TMS coil. ;-) The windings used #14 AWG stranded wire with PVC and outer plastic insulation. Its inductance was around 67 µH. Sam #1 functioned perfectly with a biphasic period of around 261 µs. It just proved useless for TMS, at least at the energy of the Ver. 1 pulser. However, Sam #1 did provide a truly fabulous neck massage and abdominal workout via peripheral nerve stimulation at 10 pps. ;-) Of all the coils constructed to date, it was probably the best in this regard.


#1: Figure-8 Twin Toroid 21T, 3" D                         #2: Figure-8 Twin Pancake 15T, 2.5" D

Sam #2 was also a twin coil design but the individual coils were pancakes instead of toroids. It was wound with #14 magnet wire laminated between plastic sheets. The twin pancake coils were side-by-side. Its inductance was around 34 µH, about half of that of Sam #1 so the biphasic period was around 164 µs. It was used to record the waveforms for the biphasic, polyphasic, and monophasic pulse types thus confirming the validity of the electronic design. This was done using a sense coil and integrator resulting in the scope photos in the pulser document.

Sam #3, #4, #5, #6

These were very simple single-loop coils. Sam #4 had the lowest inductance of any that had been built - around 3 µH resulting in a biphasic period of only 65 µs. Based on calculations, its peak dB/dt would be similar to that of Sam #13. The others were around 10 µH. All were wound with #14 AWG stranded insulated wire as above.


#3: Toroid 10T, 2.5" D     #4: Toroid 6T, 2" D     #5: Stacked Pancake 2x5T, 2.5" D           #6: Stacked Pancake 2x5T, 2.38" D

Sam #7

The one below used the yoke from some cheap electromagnet and as should have been expected did not work worth a darn at all due to eddy current losses in the central part of the yoke and almsot certainly core saturation as well. All three pulse types triggered the free-wheeling SCR and collapsed to the monophasic shape for anything but the smallest voltage.


#7: 12T Toroid Insulated and Installed in Electromagnet Core

Sam #8

The one below was made using laminations from an unidentified power transformer with the center leg cut off. But it was a total disaster due to eddy-current losses and/or core saturation. It blew the free-wheeling SCR on the first shot even though though it wasn't even being used for the monophasic pulse.


#8: 12T on Laminated 2x3" C-Core

Sam #9, #10

These are both small figure-8 coils with pancakes side-by-side. The photos show Coil #9 during three stages of construction, first pancake pair curing, then both curing, final coil with cable awaiting attachment, soldered, and insulated. Plastic sheets (~0.01" thick) are placed between the pair of pancakes, as well as top and bottom, with Epoxy holding everything together and providing additional insulation. Finally, the entire assembly is covered with electrical tape. Coil #10 is virtually identical, except with 2 sets of 6 turns instead of 10 turns. Sam #9 and #10 are wound with #14 magnet wire and have inducatances of around 10 µH and 4 µH, respectively.


#9: Figure-8 Twin Stacked Pancake each 2x5T, 1.2" D; Bare Windings Curing, Ready for Final Assembly, Before Outer Insulation, Complete

Calculations for Sam #10 predict that it should actually have the highest dB/dt of all the coils shown here. It is of similar construction with an average pancake diameter of only 3 cm, 6 turns per pancake, and a total inductance of around only 3.8 µH, the peak dB/dt should be almost 16 kT/s at 2 cm, and over 65 kT/s on its surface.

Sam #11, #12

Sam #11 and #12 were single stacked pancakes of 2x5T, 1.2" D and 3x5T, 2.5" D, with inductances of 9 µH and 22 µH respectively.

Sam #13

Sam #13 is modeled after the Medtronic/Magventure B70 but is slightly smaller with two sets of ~8 turns instead of 9 or 10 turns using #10 AWG magnet wire instead of #8. So, it might be called a B56. :-) The reduced size and number of turns was an attempt to more closely match the lower voltage of this pulser by reducing its inductance while maintaining the advantages of the overlapped pancake architecture. Total inductance is approximately 10.8 µH for the coil with an additional 0.2 µH for the 5 foot cable.


#13: Figure-8 Twin Pancake 8T Overlapped, 2.25" D; Diagram, Before Final Insulation, and Complete

As with the ones above, a jig is used to fix the coils during initial construction and gluing. The coils are insulated where they overlap by a piece of 0.02" Fiberglas-Epoxy PCB unplated material. After forming the wings, they were coated first in Epoxy and then RTV Silicone (Loctite 37467 Black RTV 598) with a pair of 0.02" Fiberglas-Epoxy PCB unplated sheets providing the primary high voltage insulation front and rear. Edges have Epoxy, RTV, two layers of 0.001" Kapton/Polyamide tape. Then the entire structure was covered in at least one layer of vinyl electrical tape. The cable is 3 #14 AWG flexible insulated wires for each direction inside a rubber hose for additional insulation. The equivalent wire is between AWG #10 and 9. The resistance is approximately 0.01 ohms for the coil and 0.008 ohms for the cable.

Testing with a Gauss probe at low current results in slightly more than 1 G/A at 2 cm (~0.8") from the coil surface near the center ("sweet spot").

Sam #13 is the only coil so far to have evoked a consistent cerebral neural response but this is most likely due to the fact that it is also the only one to have been used with the Ver 2. pulser and its higher energy of up to more than 175 J. All the others were only used at a maximum energy of around 27 J. Some some of those would probably have some effect, though Sam #13 is believed to be the best.

Sam #14 (Future)

Coil #14 is currently in the design phase. It is anticipated to have dimensions similar to those of the Medtronic B70 using dual 10 turn overlapped pancake windings made of 0.25" (3.2 cm) thick-wall copper tubing with cooling provided by a recirculating oil-to-air heat exchanger, so hopefully run time will be infinite with the Ver. 2 pulser. (Should an upgrade to Ver. 4 with its higher pulse rate and energy be completed, additional cooling would be required.) Due to the skin effect, the impedance of the copper tubing should be similar to that of #8 AWG copper magnet wire. Insulation between turns will be via Epoxy and Kapton tape. To increase dB/dt at the maximum voltage of 1,450 V, the two windings will be wired in parallel instead of series, resulting in an estimated inductance between 3 and 4 µH for a full biphasic cycle of 150-160 µs with the 167 µF energy storage capacitor or around 200 µs if the third available capacitor were to be added. However, a larger primary SCR may also be required. :)

#14: Figure-8 Twin Pancake 10T Overlapped, 2.67" D; Diagram Only

-- end V1.10 --