Turns on Before Surgery, Won't Turn on After

[Kevan Hashemi]

Some facts:

(A) At ION/UCL we have two transmitters that turned on before surgery, but would not turn on after surgery.

(B) ION reports that several transmitters were running when they checked a few weeks after delivery. Last month, we had two transmitters out of twelve running upon arrival in Paris, after two days journey from Boston.

(C) We check that all transmitters are turned off before and after packing them up in a box.

Our best guess as to what is happening is as follows.

(1) During shipping, the transmitters are subject to a magnetic field so strong that it turns some of them on.

(2) Upon arrival, if we do not check them immediately and turn off any that are running, those that are turned on run down their batteries and turn off. In the weeks that follow, their batteries recover a little.

(3) When it comes to surgery day, the transmitters will turn on briefly, but later they will not run after surgery because what little charge the battery had left is gone.

To stop failures like this we recommend three steps.

(A) We will pack the transmitters in a box that is more like a cube, so as to increase the minimum distance between the transmitters and the box wall. Every extra centimeter between the transmitter and the source of the magnetic field will decrease the probability that the field can activate the transmitter's magnetic switch.

(B) All recipients check their transmitters upon arrival and turn off any that are running.

(C) Before surgery, instead of merely detecting reception, check the average value of the signals, from which we can deduce the battery voltage using: V_bat = 1.8 * 65536 / V_ave. If the battery voltage is less than 2.6V, the transmitter has less than 10% of its battery left.

In the long run, we would like to provide a simple "detector box" you can keep in your office, and in which you can store transmitters. If one of them is running, a light will turn on. A box that is inexpensive enough that you won't hesitate to buy one. We will also consider using a less sensitive magnetic switch.

[Kevan Hashemi]

We dissect two transmitters we brought back from ION/UCL. We hear from ION that these turned on before surgery but not after surgery. Both have damaged radio-frequency oscillators. We investigate whether it is possible to damage the radio-frequency oscillator by applying various electrical stimuli to the read and blue leads of an encapsulated transmitter. We apply 20-V pulses through 50 Ohms. We apply sparks with the help of a plasma ball. We connect our battery tab spot welder to the leads. The transmitter keeps running. Our best guess is that whatever damaged the radio-frequency oscillator took place before we encapsulated the transmitter, at our own facility, so that this failure is a "manufacturing defect".

[Kevan Hashemi]

We strip insulation off the antenna of an A3028 transmitter and apply 10-V pulses between the antenna and blue reference lead. The transmitter turns off when we apply pulses, but is otherwise unharmed. We apply sparks from our plasma ball directly to the antenna and the transmitter turns off, will not turn on again, and consumes enough current to drain a CR1225 coin cell in one hour.

We read some old notes on damaged transmitters, in which we report damaged transmitters running for several hours before they fail. We examine our transmitter assembly procedure and note that we told our staff the transmitters were vulnerable to static electricity until they were encapsulated. We should have said "until they are encapsulated in epoxy and coated in silicone". Our hypothesis is that transmitters were being damaged by static electricity during the epoxy curing process in mid-winter, and the damage did not become apparent until after implantation.

The A3028 transmitter has been replaced by the A3048 and A3049. Both new circuits contain antenna matching networks that we plan to adapt to act as static protection networks. We will wear anti-static boots while handling transmitters until after they are coated in silicone. We will include in our final quality control a twenty-four hour burn-in of all transmitters to reveal any delayed symptoms of static damage. With any luck these precautions will prevent us from shipping static-damaged transmitters in the future.

[Kevan Hashemi]

Update on our study of transmitters failing during surgery.

(1) We cause permanent damage to a transmitter when we discharge a spark from our plasma ball onto the radio frequency oscillator's output, which is connected directly to the antenna on the A3028 assembly.

(2) We are deploying a protection circuit in the new A3048 and A3049 assemblies. This circuit allows us to spark the antenna as much as we like with our plasma ball without damaging the transmitter.

(3) We have observed the problem described in the topic title only at once location at UCL. At no other location have we observed transmitters to work before surgery but not after surgery. We have no idea why this one site is suffering these failures.

Our hope is that the A3048 and A3049 assemblies will put a stop to these failures at UCL, and anywhere else they might occur.