@teletypeguy You linked to my comment - thus, you’re preaching to the Choir (I’m happy with whatever a User wants for his/her symbols/footprints…etc). It’s @paulvdh that likes only what he likes…
1 Like
Hello @teletypeguy I made some changes to the circuit based on the advice you gave me.
I apologize if the schematic is not clear.
Is the leg grounded properly now? What about the pin out for the audiojacks?
I also changed the positioning of the filter based on a design I found online. Does it respect the high impedance signal captation now?
Should I add another OP AMP to bring the signal voltage up to 1-2V?
Do you have any other advice on this circuit?
Thank you for your reply.
1 Like
Thanks to everyone who took a moment to reply to this thread.
I went looking for a TL0721 data sheet, but couldn’t find the component.
This Op-amp seems to be a TL071 or TL072.
It also seems you have two dual op-amp packages, each with an unused op-amp. This seems a little strange. Also it is generally not good practice to leave unused op-amp inputs not tied to anything.
This is a good read.
I also think all your battery connections are reversed.
You need not have made these symbols as the pin connections are the standard 2 op-amp plus power symbol as used for , say, a LM358… see Kicad library.
Hi Giovanne: Yes you fixed the Leg connection and restored the high-impedance input for Chest-1 
But R8 should connect to the output of U1, and you are using a dual opamp and only using half of it. The TL072 (that’s a zero) is an old jfet-input dual opamp but is still a good performer at reasonable cost and is used in many audio circuits and guitar pedals. You could also use the single version, the TL071 – in the 8-pin package it has pins for adjusting dc offset. I don’t know if the filter stage will help much and didn’t run a spice sim to see where it breaks, but you could include it, and it can also be configured as a simple gain stage with different parts.
The big noise source is powerline freq (50 or 60 Hz), and to help keep it down a bit you can use shielded cables that are grounded at the jack but open at the body electrode to create an electrostatic shield on each cable. You can get pre-made patch cables with 2.5mm or 3.5mm plugs, cut them in half, strip the shield back a bit and cut it off and connect the center wire to your electrode. If you want to use just one cable I would use a four-conductor 3.5mm TRRS design, so you can ground the shield (like mentioned above) and have the three inner wires for your electrodes. Not sure where you are connecting the output, but you may want a series cap to ac-couple to the destination.
FWIW, I sketched it up more how I would draw it. Separate power supply stuff showing the divider for providing the system ground (can also add caps across the resistors). The 10meg input is easy to spot and shows that the stage is biased to ground. Added a simple gain stage to boost it a bit, which is ac-coupled since enough dc offset in the first stage could drive the second stage to one of the rails. You could add another dual opamp and put in a 4 or 5-pole filter. You might want to proto this up on a perfboard before committing to a pcb. It will be much noisier as a proto but you can adjust the circuit quickly with a soldering iron and fine-tune the basic design. It will only get better on a pcb with generous ground plane fill everywhere.
1 Like
Good catch on reversed power on the opamp pins – that’s why positive pins should be at the top of a symbol and negative are at the bottom, so it is obvious.
Hi @teletypeguy, I changed the circuit to match yours.
I added a sallen-and-key filter as well as a series capacitor to the output (what should the value be?).
Now my main concern is the audio jack connector compability. The input jack for the leads is a TRS so I wouldn´t be able to use a TRRS unless I use an adaptor which I´m not really sure how to use.
Same with the output jack, here in Brazil I found the TRS easier to find to buy for proto. Maybe for the PCB stage I could use the TRRS and TS because I´d probably get it from JLCPCB. The issue is that I won´t find input electrodes with 4 cables to use a TRRS unless I use a DB15 connected to a CONTEC ECG Cable and only use the 3 electrodes I need.
Would that change my schematic? Would I be able to use any pin I wanted to connect my RA, LA and LEG electrodes?
In order to use a TRS I connected the LEG to the sleeve and grounded it.
What value resistor I should use for the RIGHT-LEG_GND resistor?
Any advice about the circuit?
For safety could I add diodes before the inputs of the TL072 1?
Nice! A couple of observations:
- on the combo jack, remove the ground (or you short R2 since the labels connect).
- if you set C2 to 10 uF (with R5 at 100k) you get high-pass-filter (HPF) at 0.16 Hz – you need decent low-freq response for ecg signal.
f = 1/(2xPIxRxC) - C8 should connect between R11 and R12.
- if you leave U2B unused, connect output to neg-in and connect pos-in to gnd.
- Your gnd symbols don’t need clutter of “GND” unless it is special symbol (eg ISO-GND). I use triangle symbol for AGND and three-bar for digital GND, but up to you.
- I would mirror J1 and J4 vertically to put gnd at bottom (just because I am a bit anal).
- output cap value depends on impedance of destination (what is it?). This cap will roll off low end so not arbitrary. f = 1/(2xpixRxC)
- global labels are not needed here and kinda ugly – you can make custom V power symbols.
- the batt is external and does not need labels (my symbol is decorative).
- TRS jacks will be fine with LEG (body-gnd) on shield.
- the R in LEG is a future-proofing part (always try to make circuit flexible, and probably should add series R on chest-1 in as well). Just a jumper or solder blob for now. I would plop an 0603 or something here.
- sounds like you need to define what electrode cables you want to use.
- I don’t want to be giving much safety advice but you could add back-to-back schottkys between each lead and gnd – this clamps maximum delta-v to about +/-0.5V. Look at what could fail: if U1 fails and pin 3 is shorted to a power rail…
- again – only battery power, everywhere!
Don’t mean to be picky – just giving you some ideas.
1 Like
Thanks for the reply @teletypeguy
- if you set C2 to 10 uF (with R5 at 100k) you get high-pass-filter (HPF) at 0.16 Hz – you need decent low-freq response for ecg signal.
f = 1/(2xPIxRxC)
Is 10uF a good choice along with 0.16Hz low-freq response for ECG signal?
- output cap value depends on impedance of destination (what is it?). This cap will roll off low end so not arbitrary. f = 1/(2xpixRxC)
I found online that 32 Ohms are the usual impedance for headphones. Could I adopt this value for MAC and Windows computers audio jack port?
WIth that in hand I´ll have f=? (would it be the freq for the ECG signal?), R= 32 Ohm and C unknown?
- the R in LEG is a future-proofing part (always try to make circuit flexible, and probably should add series R on chest-1 in as well). Just a jumper or solder blob for now. I would plop an 0603 or something here.
Can I add potentiometers instead of resistors on the proto to make it more flexible? And change it to set values once I decide which are best?
- I don’t want to be giving much safety advice but you could add back-to-back schottkys between each lead and gnd – this clamps maximum delta-v to about +/-0.5V. Look at what could fail: if U1 fails and pin 3 is shorted to a power rail…
Did I add the diodes correctly?
I left some of the symbols advice unchanged for now.
When I go to buy the components, if there isn´t the specified values I need how close can I go? Should I recalculate the component values?
I would say that ac-coupling to about 0.16Hz high-pass is plenty for an egc. That works well with nice values 10uF and 100k (C2, R6): f = 1/ (2 * 3.14 * 100k * 10u).
Headphones? You won’t be plugging this into headphones will you? Hmm. If you are plugging into an audio-line-input on a laptop or something, it will have a input impedance of maybe 100k down to perhaps 10k (~50k typ). A 10uF output cap into 10k ohms will be a rolloff at about 1.6 Hz, which is kinda high but maybe ok for an ecg waveform. I would place a 1210 pad for up to a 100uF and just put a short in to start.
For limiting voltage on the channel inputs (not needed for leg), what I meant by back-to-back diodes is this clamping:
1 Like
Hear Hear Old Chap !
You mean “here here” or should that really be “ear ear” 
Typically JP designations are for jumper pins. Think of the 0.1" spaced square pins that you slide a shorting block over to configure a device. Can also be solder pads close together where you flow a bridge between them. Your JP1 would be typical for the first style of this.
Your JP2 is a TRS (Tip -Ring - Sleeve) jack, similar to a headphone jack. The L and R shown are for the left and right channels of unbalanced audio, for example a headphone jack. I would re-draw that and label the “L” to “T” so that T and R stand for Tip and Ring. The tip is at the end of the plug and the ring immediately next to it. The main body is the sleeve. While these are commonly used for stereo audio, they have many other uses and calling them out on the device specifically for audio isn’t in my opinion a good idea.
1 Like
Many decades ago in the Houses of parliament if a member had stood up and made a very popular statement, one that his colleagues and party members considered of great importance they could force the point by demanding at great volume that the opposition “Hear Him Hear Him” or “Let Him be Heard” nowadays it’s just a simple endorsement of whatever has been said that you wish to backup and has become “Hear Hear” 
1 Like
Your right, my brain-fade. I was too involved in the word play for those outstanding appendages on your head 
1 Like
Hi @teletypeguy and @BlackCoffee I built it and made it work.
This is the signal I got with a simple 59Hz low-pass RC filter:
This is the breadboard view without the jack connector as they´re not available on the software I used.
Thanks a lot for your help. Once I´m done with soldering this prototype I´ll test the sallen-and-key filter.
2 Likes
Would this circled part be a two terminal switch?
Why would there be a switch on this part of the circuit?
I apologize for the quality of the picture.
Is it the same as SW1 in my schematic?
Thank you!
Pin 1 of U2A on the Kicad version is missing the junction that is on the original.
I can’t read the text for the part in question.