ECG EKG lead protection - TVS Diodes

Thread Starter

mmarc

Joined Mar 31, 2022
12
ECG leads generally see voltages in the range of -5mV to 5mV and a frequency range of .05Hz - 150Hz. We are peripherally stimulating the patients with up to 200v at a frequency of 1kHz elsewhere on the body for about 100mSec and it disrupts our front-end ECG chip. I need to protect the ECG leads from the stimulation. Although the system recovers, it takes far too long. How could we do this?
 

prairiemystic

Joined Jun 5, 2018
246
If you're stimulating on a human body (somewhere) at up to 200V while there are EKG electrodes on a patient, you have a serious safety issue. Electrodes on the chest will be part of the return path for the stimulation current if you use clamp diodes, and around the heart it's not a sane idea. This device sounds unsafe and I don't hear of any isolation between channels or if the "peripherally stimulating" is EMG or what.
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
If you're stimulating on a human body (somewhere) at up to 200V while there are EKG electrodes on a patient, you have a serious safety issue. Electrodes on the chest will be part of the return path for the stimulation current if you use clamp diodes, and around the heart it's not a sane idea. This device sounds unsafe and I don't hear of any isolation between channels or if the "peripherally stimulating" is EMG or what.
The peripheral stimulation path is from the finger to the thumb on the right hand. The constant current is biphasic for 10mSec at 10mA max, hardly dangerous. The HV circuit is obviously isolated as is the sensor and the whole thing is operated on one 3.3v Lion cell that has discharge protection. There are also lead off and other protections built-in. Safety is not the issue. Marc
 

DickCappels

Joined Aug 21, 2008
8,710
I would put a transistor (a 2N2222 is a good choice) across C55 with the collector grounded, and a second transistor across C57 with the collector grounded. Then using a low value resistor like 1K in series with each base drive the bases with 5V and hold the transistors on for the duration of the 200V stimulation. Note, the collector and emitter of the transistors have swapped roles because using the emitter as the collector gets you a lower resistance connection to ground.
 

prairiemystic

Joined Jun 5, 2018
246
That would work, I think you need a transistor with reverse-hFE for -ve currents. I'd first try a couple BAV199 clamp diodes, one at each input to keep it below say +/-0.6V seen by the preamp. It will still saturate it though. Your choice before or after the "energy rated" 47k resistors.
The MAX30003 has a (manual) fast recovery mode where in F/W you can activate it, then generate the stim pulse, the deactivate it. Something to try in firmware for muting things.
If this is still too slow in recovery time, there's really two methods to blanking stimulus artifact.
One, as mentioned is to quickly/briefly short out the preamplifier's inputs, I would try using a dual (AC) mosfet configuration or CMOS switch, on during the stimulus portion, as well as have protective clamp diodes. This can be slow because the electrode gel and skin can store charge from the stim, making recovery slow.
Two, is to quickly/briefly simultaneously open-circuit input electrodes and mute (short out) the preamplifier's inputs. You'd need a another mosfet or CMOS switch. If the clamp diodes work limiting to say a 1V surge, then many IC's can work. You'd need a pair of SPDT CMOS switch like TI little logic SN74AUC1G66 or 2G53 kind of IC gated by firmware.

The safety issue is mainly any possible current path between the stimulator and heart electrodes, under any circumstances. Using a common battery means the isolation is safety critical. Your HV power supply/stim can malfunction.
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
I would put a transistor (a 2N2222 is a good choice) across C55 with the collector grounded, and a second transistor across C57 grounded. Then using a low value resistor like 1K in series with each base drive the bases with 5V and hold the transistors on for the duration of the 200V stimulation. Note, the collector and emitter of the transistors have swapped roles because using the emitter as the collector gets you a lower resistance connection to earth.
Like this?
 

Attachments

Thread Starter

mmarc

Joined Mar 31, 2022
12
That would work, I think you need a transistor with reverse-hFE for -ve currents. I'd first try a couple BAV199 clamp diodes, one at each input to keep it below say +/-0.6V seen by the preamp. It will still saturate it though. Your choice before or after the "energy rated" 47k resistors.
The MAX30003 has a (manual) fast recovery mode where in F/W you can activate it, then generate the stim pulse, the deactivate it. Something to try in firmware for muting things.
If this is still too slow in recovery time, there's really two methods to blanking stimulus artifact.
One, as mentioned is to quickly/briefly short out the preamplifier's inputs, I would try using a dual (AC) mosfet configuration or CMOS switch, on during the stimulus portion, as well as have protective clamp diodes. This can be slow because the electrode gel and skin can store charge from the stim, making recovery slow.
Two, is to quickly/briefly simultaneously open-circuit input electrodes and mute (short out) the preamplifier's inputs. You'd need a another mosfet or CMOS switch. If the clamp diodes work limiting to say a 1V surge, then many IC's can work. You'd need a pair of SPDT CMOS switch like TI little logic SN74AUC1G66 or 2G53 kind of IC gated by firmware.

The safety issue is mainly any possible current path between the stimulator and heart electrodes, under any circumstances. Using a common battery means the isolation is safety critical. Your HV power supply/stim can malfunction.
Thanks so much. We have the fast recovery mode on, but it is much faster for the EKG than for the Peak Detection Trigger. We want to trigger at about 80ms after the peak, as we stimulate approx. when the systolic pressure pulse hits the baroceptors. The recovery is far too slow - >1 minute. The short out, as far as we can tell, also triggers the recovery mode, which is the problem.

On safety, if the HV completely fails, the output is limited to 10mA to the hand - the EKG sensor, the max, is completely isolated.
 

DickCappels

Joined Aug 21, 2008
8,710
Like this?
Yes like that, and do it for C55 too.

I just realized that those capacitors are very small, only 10 pf, and a 2N2222 will have almost that much capacitance between the base and collector, which means you might want to revisit the values in that filter after you see this work. There are much lower ouput capacitance RF transistors that you can try if you can find them, but start with a good "health" transistor like the 2N2222 first.
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
Yes like that, and do it for C55 too.

I just realized that those capacitors are very small, only 10 pf, and a 2N2222 will have almost that much capacitance between the base and collector, which means you might want to revisit the values in that filter after you see this work. There are much lower ouput capacitance RF transistors that you can try if you can find them, but start with a good "health" transistor like the 2N2222 first.
Thank you for your help! - I still have a problem. My stim source is more complicated and created in hardware off of a trigger, not from a uP. See lower picture. How do I get the 5v to drive the transistor? It varies in amplitude from +- 2-70v. Lower picture 9ms on and 15ms off for 105impuls stim waveform.jpegms. To test the concept, I can test at one amplitude level.
 

DickCappels

Joined Aug 21, 2008
8,710
A few questions: About the On/OFF Stimuls waveform, is that a digital control signal and if so what are the DC levels during the stimulus burst (high and low levels) and during the time between bursts. Is that the signal you want to trigger the front end blanking?

If the signal you want to use to trigger the blanking is the actual 200 volt peak (or is that peak-to-peak) stimulus burst being applied to the patient the blanking will probably be a little less helpful because some amount, still unknown, will sneak through before the blanking can start -this might only be a few microseconds. Just so you know.

I think the rest of your illustration is very clear.

To summarize my question: What is the range of voltages where you want to sense when blanking should start, and what is the DC level of the signal before blanking is to start and when blanking should start?
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
A few questions: About the On/OFF Stimuls waveform, is that a digital control signal and if so what are the DC levels during the stimulus burst (high and low levels) and during the time between bursts. Is that the signal you want to trigger the front end blanking?

If the signal you want to use to trigger the blanking is the actual 200 volt peak (or is that peak-to-peak) stimulus burst being applied to the patient the blanking will probably be a little less helpful because some amount, still unknown, will sneak through before the blanking can start -this might only be a few microseconds. Just so you know.

I think the rest of your illustration is very clear.

To summarize my question: What is the range of voltages where you want to sense when blanking should start, and what is the DC level of the signal before blanking is to start and when blanking should start?
This is a University project at the University of Marburg, Medical Psychologie, for chronic pain. We stimulate the hand (finger to thumb) with conductive gel and round the finger electrodes.. The amplifier is a constant current with amp levels set in Epots. The biphasic on-off waveform (the picture) is delivered at 200uA - hand resistance varies but is around 3400. We slowly increase the amp levels and ask the patients for feedback and capture sensitivity, threshold, and tolerance levels. We then stimulate at precise points in the cardiac cycle and predict when the heart will beat again. We identify the ECG peak (by the AFE ECG chip) and then time the pictured pulse train, which is all at one amplitude however, the subsequent stimulation is about 5sec later, will be at a different amplitude. The problem is that the stim resets the AFE so I cannot get a trigger, so I have to protect the EKG leads. The body has a DC offset voltage of typically +-12mV. The EKG signal is +-5mV, measured on the right and left shoulder. The stim voltage goes up to +-100V, but is usually +-2-20V. The blanking should occur when we stimulate (during the pulse train or the 1kHz pulses in the pulse train, perhaps a little longer due to body capacitance/inductance effects. All this is too small to measure in the scope - I hope that I have answered the questions.
 

DickCappels

Joined Aug 21, 2008
8,710
Are the 1 kHz pulses only accessible from the stimulus signal applied to the subject's fingers?

Where would you get power for this circuit?

Sorry about the questions but I am working on understanding exactly what you need and how it can be connected to what you have.
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
Are the 1 kHz pulses only accessible from the stimulus signal applied to the subject's fingers?

Where would you get power for this circuit?

Sorry about the questions but I am working on understanding exactly what you need and how it can be connected to what you have.
Should we talk? Yes. The circuit is constant current, so it can supply the power. The power comes from a 3.3v Lion cell that gets boosted. (Email address removed by moderator -you can communicate here via private message.)
 
Last edited by a moderator:

DickCappels

Joined Aug 21, 2008
8,710
mmarc wrote:
"The biphasic on-off waveform (the picture) is delivered at 200uA - hand resistance varies but is around 3400."

That is only around 680 millivolts. I set the threshold to 310 millivolts. You can adjust this threshold by changing voltage divider R3 and R4.

C2 should be a good quality multi-layer ceramic capacitor mounted very close to U2.
C1 would best be a tantalum capacitor but an aluminum electrolytic would probably work too.

1649176453024.png

We did not talk about your 3.3 volt battery. Make sure it is ok for supplying about 6 milliamps to the circuit. This current may be reduced to less than 3 milliamps by increasing the resistance of R7, R8, and R9 but that might be at the expense of reduced attenuation of the stimulus signal getting into your preamp.

D1 may be redundant but better to be safe than sorry.
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
mmarc wrote:
"The biphasic on-off waveform (the picture) is delivered at 200uA - hand resistance varies but is around 3400."

That is only around 680 millivolts. I set the threshold to 310 millivolts. You can adjust this threshold by changing voltage divider R3 and R4.

C2 should be a good quality multi-layer ceramic capacitor mounted very close to U2.
C1 would best be a tantalum capacitor but an aluminum electrolytic would probably work too.

View attachment 264367

We did not talk about your 3.3 volt battery. Make sure it is ok for supplying about 6 milliamps to the circuit. This current may be reduced to less than 3 milliamps by increasing the resistance of R7, R8, and R9 but that might be at the expense of reduced attenuation of the stimulus signal getting into your preamp.

D1 may be redundant but better to be safe than sorry.
wow, thanks. We order the parts breadboard and test.
 

Thread Starter

mmarc

Joined Mar 31, 2022
12
hi, I need to send you the stim circuit, but we change poles to create the biphasic 1kHz signal. Is there an email to send this? Marc
 
Top