How “Telecard" works

Currently, our obviously underestimated estimates, by means of "Telecard" they give not less than 100 000 telemedicine consultations per year in Russia and Ukraine. Even if we assume that only one of hundreds transmitted electrocardiograms is really vital, and the probability of a correct decision on the provision of emergency assistance is 1/2, then not less than five hundred people a year do not give up the ghost or become disabled as a result of a timely remote diagnosis. Naturally, loss of the state decreases also - according to official data, about 15,000 USD per one person - a total of 15000х500=7.5 million cabbage annually.

It would seem that at the state level - use and enjoy and expand your network, because the existing level of teleconsultations - only a pale shadow of Soviet successes in this area, which with a much more primitive and difficult to use equipment provided (only in Russian Federation), more than 800 thousand consultations a year! (From the order of Ministry of Health of the Russian Federation in functional diagnosis from 11/20/93 № 283: "Compared to 1991 [in 1992], the number of remote diagnostic rooms lowered from 354 to 286, and the number of ECG studies conducted in them - from 887.7 to 857.1 thousand " ) . Of course, there are currently good results (film about the urgent case of "Telecard" usage), but the Soviet level is still very far.

But no, as soon as the planned procurement and tender is announced, immediately showered in abundance all kinds of "estimates", "reviews", examination findings, conclusions, publications in the media, etc. And immediately, "archaic technical solutions", "remix of Soviet-era “Volna””, "pseudo-digital TX", "functionally defective product", "yesterdays technology", "junk in the Internet age", etc. There are more straight-from-the-shoulder and stupider ones. No, of course, "sitting on the armor and slurping out of the pot", I calmly refer to expert opinions, but, first, all of this is unpleasant, and second, I am often asked to explain how the "Telecard" works, and if it really is junk, how they allow achieving such impressive results?

Here, the developer of medical products is at a distinct disadvantage. After all, if an unreasonable or abusive criticism, say, of a Kalashnikov machine gun can (in extreme cases) be answered by a short burst, and the incident will be settled (sometimes together with the opponent), in our situation the description of the device, it seems, is really necessary.

So, let's start from history. No matter how offensive for human dogmas, but the first and by a wide margin the best was the totalitarian and terrible USSR. The level of its scientific schools and engineering developments was very high. So, to 70’s of the last century, nothing doubting, in order to assess the prospects and potential benefits of remote ECG diagnosis, they took at the state level the decision – to make this piece. Especially, as the results of ECG transmission from the spacecraft have timely arrived. And for mundane tasks they selected a phone. Reasonable? Absolutely! After all, it is the most common, and then almost the only available civil communication channel.

And now estimate the level of developers: you cannot even mention any memory chips, saying nothing about microprocessors for civil engineering. That is, if you please - stream live signal in real time – there is no memory. And phones are all different, with carbon microphones, for example. And wheezing ancient switches of village authorities, and non-compliance on the phone lines with all conceivable standards on bandwidth, dynamic range, echo levels, nonlinear distortions (I'll tell that these problems remain, and new ones were added). That is, you know, not to work in the prosperous West with licked-up standardized telephone lines, it's something like drawing products behind tanks. And again, absolutely right decision - not to connect the device to a telephone line, but in acoustic way - so that you can use any phone and do not look for the outlet, which may be absent. Looking ahead again, I say that our firm did not succeed this at once, i.e. first series of "Telecard" were wired to a radio phone handset, that was very uncomfortable.

Thus, Soviet engineers made ​​the only correct decision at that time - serial registration of ECG channels (as in a single-channel electrocardiograph), transmission during registration, and selection of modulation: FM, frequency detector at the receiving node, amplifier and ink recorder at the output. ECG recording looked like this: "Tanya, baby, give lead V1». Tanya puts the electrode on the desired place and presses the button. Beep sounds, at the receiving station they turn on the recorder, record a part of the ECG. Stop, and then: "Tanya, baby, give the next." And so all 12 channels.

Time-consuming? Certainly. But the system performed its functions. There were even cases where in the absence of a house phone, the patient was taken to a payphone, dragged into the cabin and there the recording was made. And nothing, everything worked.

No, of course we can, grinning in IT mustache and scratching with the latest smartphone model, sniff at the sight of such work. Especially because doctors’ white coats at the receiving station were always in a spray of ink – at interference the pen trembled and ink was fanned around. And if you saw how they broke walls when mounting the receiving stations - because they were so huge that could not pass through usual doors...

But this equipment, with which the Union has equipped almost all the "heartland" in two (!!) years, recorded and transmitted 800,000 ECGs per year! Those people were able to work. And we now rejoice a hundred thousand. Maybe because there was no tenders, kickbacks, again? And developers were pros, and not like it is now - every dot-com, to put it mildly, promotes its own "dot".

How it ended? August came. And slowly, slowly, everything went down the drain. I.e. not the drain, of course. They drew all richest heritage, and became Israel ahead of the rest of the world in the field of teleconsultations. Overall - correctly. If we are not able to value ours, have no self-esteem, then we deserve this.

But the holy place is never empty, and sitting with nothing - not the most fruitful occupation. Especially - the situation with communication has changed dramatically. And the old equipment has long since gone to scrap.

ECG teleconsultations are needed. This is really useful and necessary thing, unlike frankly splendiferous videoconferencing of "front" telemedicine. And for this very remote ECG a device is needed. Device for "hinterland", for weakly trained personnel, for mass. Mass device. Easy to use and reliable as a Kalashnikov machine gun.

Now, dear reader, let's try together to design such a device. I bet you will not be bored. You will understand almost everything, and if you are an engineer, then - everything. And become a staunch supporter of "Telecard". Moreover, in the subsequent discussion of transtelephonic ECG you will regularly use such terms as "reliability of equipment cannot be replaced by any other virtue", "if something can be done in a wrong way, so be sure there will be a person who will do just that", "there’s no need to install a jet engine on a cart", "higher education is a good thing, but sometimes it needs a secondary one as a company... ".

Let's start with the preliminary design. And the first question that must be answered: what communication channels we shall use? For now, there are three main widely used channels: a wired phone, cell phone of GSM standard and digital wireless communications (GPRS, CDMA, 4G, etc.)

We will proceed from the actual conditions of the device operation. Where will the recording and transmission of ECG take place? Naturally, in the habitat of our patient. Because if he/she is, say, in the hospital, the ECG recording and transfer to the diagnostic center is hardly necessary; simply take off the shelf a usual electrocardiograph and record the ECG. That is, we initially assume that the patient, who needs an electrocardiogram to be recorded, is at home, at work or in some outpatient clinics, where there is no qualified cardiologist.

What is the likelihood that we will have to use only wired or only a cell phone? Or vice versa: there is no connection, except 4G? Or another situation: something extraordinary happened, cellular communication is "laid down" or blocked by security services, the switch of wired phones is out of order, and we are left only with the radio communication of the emergency services. Or we need to use only a satellite phone, as our patient is in remote areas, and other communication there has never been and never will.

If we correctly estimate these probabilities, then we design the device correctly, and it will be in demand. If we go on about the euphoric fantasies and ideas about "the rapid development of telecommunications", then we will make another "rubbish", i.e. thing that seems to be workable, but of no use to anyone.

Making a preliminary assessment. We believe the most common communication channels are GSM cell phone and landline phone, and arrive at the obvious solution: our designed unit must transmit ECG through normal and cell phones, the connection must be acoustic - it should work like the old Soviet "Volna". That is, place it on the microphone of the handset - and that’s all - the transfer must be ensured. Any handset, be it modern smartphone, or archaic TAP-57.

Is it possible to waive this requirement and simplify your life by making the transfer through, for example, GPRS channel of a cell phone? No, you cannot! The reason is obvious: cellular communication is not everywhere, oddly enough. For example, in Kharkov region, I can easily show dozens of villages, where mobile phones either do not work, or work only when you climb up a hill. And here's the situation: in rural dispensaries we have an ancient as mammoth dung, wired phone, and here we have a patient with suspected acute myocardial infarction, and cellular communication is only on the hilltop. And what we have to do - waddle with the patient uphill? Or record the ECG, and then with the device and cell phone up that same hill at a trot?

And even worse, when a cellular network is "piled up" - New Year's Eve, for example. There is not that MMS transfer, there is not always possible to call. A heart attack, alas, does not recognize the holidays and dislikes waiting, eats and eats the heart muscle irreversibly. So, if you please, dear developers, do so, that we, in this backward country, can use a phone, and that the quality of ECG transmission was "digital" and no worse than using a cutting-edge "devices." Is it difficult? Yes, it is very difficult! But it is necessary.

Now we estimate the prospects of modern digital communication channels. That is, assume that in the place where we need to record the ECG, Internet is available, but there are no phones. This is an extremely unlikely situation, but we assume that it is. Then our audio signal, which contains information about the recorded ECG, can be transferred by IP telephony. It is not available too? Do not worry, we record the acoustic signal through a standard computer microphone and transmit this file in any way.

This is non-optimal solution? Yes. But this situation is extremely rare and optimizing the transfer is necessary to sacrifice in favor of transmission method unification. As academician Moiseev said, "Nothing is optimal in the nature, except a miniskirt on maxi-ass - minimum cost with maximum effect."

If we want to maximize the use of available capacity of any communication, we need to convert the recorded and digitized ECG in an acoustic signal, which will pass wherever a human voice can be transmitted. And will transmit this same ECG "in digit."

What a task! It's necessary to design its own dedicated modem that is capable to break through the lines with poor signal-to-noise ratio, through a mobile phone vocoder (absit omen!), through any PBX with speech signal compression, through adaptive filters of modern telephones, etc., etc. Moreover, it must have a simplex transmission, be insensitive to short-term interruptions, and even transmit a digital signal through a carbon microphone (Holy Saints!). Plus - coding should be in interference-resistant codes with error correction. And the electrocardiogram must be transmitted in a reasonable time - 2-3 minutes maximum. Yeah, there remains only to commemorate in the corresponding expressions King David and all his meekness.

Eureka! And if we simplify our life, maybe go by the path of Israeli (and some Russian) developers? God bless it, this digital transmission, we will transmit the recorded ECG by the proven method of "Volna", i.e. transform into an analog signal again and feed it consistently, channel by channel, to an FM modulator? At the receiving end of the line we will place a modern computer, create a software frequency detector and here you are, dear doctors, the advanced technique. And if you do not like it, we'll explain that exactly in the same way work advanced devices of some terribly developed countries, and you’d better not raise the tail.

Or still GPRS? Declare it a promising and innovative technology, ignore the slowpokes who are going to develop their own modem, or better yet – we’ll take ready-made blocks, thank God they are for sale in bulk, use, for example, TCP/IP protocol, and forward, march! All, who didn’t appreciate our youthful logic we declare reptiles-reactionaries. We will not listen to doctors-obscurantists, and quietly close our eyes on the fact that the GPRS transmission has low priority in GSM networks, and can either be delayed or even lost.

But enough jokes. Our way, dear reader, is to the difficult but fascinating field of design of electronic equipment.

How can we transmit through a telephone line a digitized ECG? Shall we try to use close to optimum pseudo-noise wide-base signals having the autocorrelation function like a delta pulse? Let us use the direct digital synthesis, because the frequency range allows this. The redundancy is close to zero, the reception in the form of convolution of the parcel with its prototype promises theoretically optimal noise immunity, the bandwidth is fully used, in short - even the portraits of the great mathematicians are smiling from the wall. But wait! What pseudo-noise signals? We forgot about the vocoder. He, the dog of British design, only seeing a pseudo-noise signal, will immediately turn the digital stream in the message: "noise signal with amplitude N transmitted”. And we, at the receiving station, will have, instead of our pseudo-noise, a foreign pseudo-noise. Terminus, this theoretically optimal charm will not pass through a cell phone!

Then, angry at our mistake and having spent a fair amount of time and effort, we will try to use the standard solution in the form of energy-optimal discrete-phase modulation. Trouble again: whatever parameters of the shaping Nyquist filter we use, such signal spectrum (which the vocoder analyzes) is obtained, that it recognizes it as bandwidth-limited noise. Here is a plague on our heads...

And the whole group of amplitude modulations, including quadrature one? We cannot, we have the acoustic transmission and the signal level constantly "floats". Woe is me! And combined modulations? Impossible, the amplitude-frequency response of the communication channel is unknown and may change during transmission, it will be torment to decrypt the received signal. And if we adjust the AFR during transmission? No, we cannot, we need a receipt, i.e. switching to duplex, and because we have the requirements to transmit over a cell phone, we cannot do that - it will send us a receipt with a large and arbitrary delay.

Have arrived! So, Soviet engineers were right, only the classical frequency modulation from Armstrong? It also will not pass through the cell phone vocoder without distortion, alas.

Yes, Soviet engineers were right, but we will try to modify their mode of transmission. We make the frequency sending discrete in frequency and time. And in order to force the vocoder not to understand, what’s happening, we minimize the spectrum widening at phase transitions. That is one sending starts from the signal zero value, makes a few sine wave oscillations, and when it gets to zero (in integer number of periods), we change the frequency. And again start drawing a sinusoid from zero signal, but at other frequency. Actually it is MSK, i.e. the Minimum Shift Keying.

Well, we turn on? Works? Excellently works! Picture - sight for sore eyes, and through a cell phone, and through dirty and oblique lines and curves of the old, and through sophisticated digital wireless phones (with adaptive filters of long-existing tones), and via radio stations. It always goes. Noise immunity? When decrypting by a spectrum analyzer with sliding window - leaves nothing to be desired! So, the digital communications from our device through all possible telephone and radio lines has appeared. And the transmission rate? But with this, to put it mildly, sadly.

But let's pause for a moment. And answer the question: how is the “digit” transmitted when using such modulation. Because they do not believe. Serious young people, loaded with netbooks, smartphones, tablets and other "devices", immersed in the Internet from the age of puberty and advising chief physicians on technical issues - do not believe! Categorically refuse to understand this: "N- Noo, (and the eyes with a sly Leninist squint), no! You signal passes through a microphone, then - it is the analog one. Don’t throw dust in our eyes, your transmission is not digital, but pseudo-digital." The Chief Doctor nods: "And you explain to us, so that both I and my expert understand."

Okay, what the demand from the Chief? He’s responsible for the entire household of the hospital and a couple of hundred employees, for him all these technical terms are like the ancient Indian grammar. The only thing he learned in this thread: digital = good, analog = bad. That's all. And here some firm tries to offer him noodles, and his expert, a great specialist, (imagine, he has connected the Chief to Internet!) says by the way - crap! Deceiving! And the Chief believes, of course.

But our employees at the sight of such experts usually go berserk and start mumbling something softly through their teeth. We will not specify, what, but the words are very bad. Because the Soviet education has pushed up daisies, and higher education (education?) remained, but the secondary one does not appear in this case.

Okay, look at the picture.

It's very simple: each segment of the sinusoid (they are color-highlighted) is responsible for the transfer of the relevant figure. That is, we have received, for example, a length of the sinusoid with frequency 700 Hertz, so, they have sent us a number, say "5". The next segment had frequency 1200 Hz - they sent us zero. Then came, for example, the length of 1500 Hertz - they sent us four, etc.

– Well, it's understandable - the expert says. - But will this pass through a microphone? You’d have to use "bluetooth", you have antiquated technical solutions and technologies of yesterday. The Chief Doctor nods in agreement with the face expressing administrative concern.

– And why would it not pass? - our man says. – The microphone transmits these frequencies, isn’t it?

– A-r-r-r, kuldyk kuldyk-kuldyk(!), and once again kuldyk(!) – explodes the expert. - What antiquity, what archaic in the Internet era! Your developers - mossy stumps, Soviet-style school, trying to assemble a spaceship on a kitchen table! While 4G format firmly controls a virtual interactive self-swabbing toilet bowl, and Microsoft breaks brains of all mankind with new versions of their product... and you here(!) waist our time?! - And goes away proudly wrapped in a peacock's tail.

– Yes, folk, - says the Chief (he does not know whom to believe now), with products, you have some trouble, you would learn a little bit more ... and then come, good-bye!

Glitch came out, sorry. How do we explain in a way that it will be quite simple? And if so: let there be a rural wedding - it is easy to imagine. At the wedding, people drink and sing. We do not order a fight, although it will be. And we need to transmit from this wedding (where all the people pissed as a newt) a digital code, well, let's say, number 328. And transmit in noisy environment, i.e. noise, din, fights, etc. Then we take and enumerate songs that usually are sung in such cases. Let "Katyusha" have number 8, "Ty zh mene pidmanula" (Ukrainian folk song) be the second number, and "Kalina" be the third number in the list. Then, if we can influence the repertoire of singers, let them first sing "Kalina", then "Ty zh mene pidmanula" and finish with "Katyusha". They sing, sing well, even on the phone, we understand, what they are singing now. And now they are fighting - noise, clamor, broken crockery. But we still know that at this point they sang "Kalina", i.e. transmitted figure 3.

Thus, using this uncomplicated method, we have transmitted by a phone from the room where a drunken wedding roars, a digital message, in this case 328. But it could be any other. That is, transmission of digital information has been provided under condition of noise and interference.

Similarly behaves "Telecard". It takes a number of its memory and sings the proper song. Only this song is very simple, just one note. And the songs are sung one after the other without interruption. And they are transmitted very quickly.

And the receiving station listens to these songs and determines their number in the list. This very number will be the received digital message. Note that even if some part of the song is missing, it can still be identified. This is a noise-immune communication channel based on the discrete frequency modulation with Minimal Shift Keying.

– Oh, weren’t up to tell this to the Chief.

– Are you sure that he did not care? Moreover, he already knew, but these homegrown experts ...

– Let’s not talk about sad things. They do not concern technique.

And we are with you, dear reader, will continue to develop the device. How we manage to transmit a "figure" is already clear. How to receive it, in general, is also understandable: to measure the frequency of these incoming sendings and restore the numeric message that "Telecard" sent to us. Incidentally, this is the most complex node of the entire hardware and software system. Because the telephone line can do with our signal anything. Can change the level of the signal, and can arbitrarily change it, the more that our user is often literally waving the device in front of the phone, or vice versa. It is clear that the volume changes and sometimes very quickly. But it's still nonsense. The telephone line can make noise (not terrible, we have a comb AFR at the reception). It can buzz (not terrible, we have a WLN system (wide-band amp. – limiter – narrow-band amp.) at the input). May have regular monochromatic interference (not terrible, we adapt to the spectra of the received signal). May have non-linear distortions (unpleasant, but not scary, because we distinguish only the main harmonic sendings), may have a frequency shift (and this is worse, but to a value of ± 50 Hz we will stand), may have a short-term signal level drop to zero (unpleasant, but adjusted by noise-immune codes). Can have a pronounced echo (absolutely not scary), can mix into our signal another conversation (and we ignore it).

And what's scary? Scary is strong overquoting in mobile communication lines. This is when the cell load exceeds the nominal one. That is the provider saves on equipment and "cuts loot." As a result, the GSM system has an overload, and the phone does not transmit some components of speech and pauses periodically (yes you will undoubtedly have meet this at cheap and vile providers, when you understand the speaker every second word, basically - guess). But we have learned to fight this to some extent.

Terrible are also "self-made" PBX, as well as self-made GSM-gateways. Here are monsters. There russische-Erfinder can make such in-ven-tions, that rough words somehow can be transmitted along these lines and received (which is amazing), but it is practically impossible to understand the remaining words. But since all our subscribers wield the profanity, for them there is no problem, but "Telecard" may “lie”. Very rarely, but - happens.

Of course, the noise-immunity of the modulation used is an important thing, but still during the service errors in transmission will occur. That is, they were going to sing "Katyusha", and then Mykola arrives. And in his jalopy - subwoofer. We listen, listen and cannot understand: it is not "Kalina", nor "Katyusha", but only the hooting of prehistoric reptiles, interspersed with cries of either wounded hyenas, or rabid jackals. What would you do, what decision to take?

But for this, there are error-correcting codes: if we have lost "Katyusha", then by the rest of used repertoire we can accurately restore the message that we should have received. Naturally, this method requires transmission of additional verification characters and badly affects the speed. And yet, the correction of erroneous characters can only be done in the package, controlled by checkup characters. The larger is the package, the smaller number (in respect to information characters) of verification characters we require, the lesser we affect the transmission rate. Naturally, at any digital transmission the developers are seeking to increase the package. And rightly so, but not always.

For example, the package was irreversibly damaged. We cannot restore it. It needs to be retransmitted. How behaves the standard system: it waits for the receipt from the receiver, and if the packet is not received (damaged), the receiver informs the transmitter (not sends the reception receipt). The transmitter retransmits until it receives a receipt of correct reception. But the larger is the package, the greater is the number of lost characters, the longer we need to retransmit the package that has been lost. And time runs. And if we cannot receive several packages in a row? And time is running...

Clearly, in our case of urgent transfer of an electrocardiogram, we are forced to use a short package and simplest code of type SEC/DED (single error is corrected, multiple error is detected only). But this is not because the developer of "Telecard" is deeply senile, it is because the nature of interference can be non-Gaussian one. You can, of course, apply Reed-Solomon code with corrected sequence interleaving, and it would work perfectly. But at falling-out of a large chunk of the transmitted signal the loss would be much, much more serious than even in the absence of correction at all.

ИThat's why "Telecard" fully compensates for any errors that it can, and all that it can not - fixes basing on a priori knowledge of the characteristics of the ECG signal and knowledge of the a posteriori probability of a transmission error of each symbol. Naturally, all errors that are corrected at least with some error are marked on the received electrocardiogram. The electrocardiogram must reach its destination! Be it a carcass, be it even stuffed, even with badly readable fragments - must reach! Not tomorrow, not after the rain, it must be laid on the cardiologist’s table now.

It's like a machine gun: its main task - to shoot. And shoot it must - dirty, unpolished, clogged with residue, frozen or overheated – it must shoot. It’s master’s life depends on it. And in our case sometimes - the life of our patients. But it's so, you know, inappropriate pathos in the triumphant consumer society, where health and human life itself - also a marketable commodity (almost as to Friedman).

Well, here we have a digital signal broadcast through any channel for transmitting voice. The decoding of the received signal is also designed and works (here in the description I have omitted pretty much: from compensation of slow jitter of packages to selection of the data window in the spectral analysis of accepted implementations).

Now back to the transmission rate. It's small. Why? It's simple: our applied type of modulation is good, but there are drawbacks. In real development would not otherwise happen. We need sendings, discrete in frequency, starting from zero and going to zero at a chosen duration. Our duration - 10 ms, it cannot be less – the vocoder will not understand. In this duration fit with an integer period only frequencies that are multiples of hundred hertz. That is 100 .. 200 .. 300 etc. In the bandwidth permitted for usage they are few, from 300 to 3000. In reality - even less if we expect to meet a faulty telephone line. As a result, our digital channel has notoriously low speed, although very noise-immune. And if we try to transmit a digitized ECG "directly", then we will wait until the transfer is over, so to say, till that’ll be the day! Of course, we need a compander, i.e. link compressor/decompressor. Moreover, the compression must be undistorted, as we put the goal of full compliance of the ECG transmitted by telephone, with the electrocardiogram recorded on the best stationary electrocardiograph. Looking ahead, I can say that our compander and transmission path really provide their complete correspondence: you can put one ECG over the other and the lines will coincide in detail.

How to achieve this? Create a specialized compressor specifically for ECG signal (for other signals, it will be less effective). First, let’s perform the probabilistic analysis of instant increments of electrocardiogram and construct the distribution law.

It is obvious that the number of small increments is incomparably greater than that of the larger ones. It is clear, large increments only appear during QRS complex and small - all the rest of the time. So, that's something. We can apply the well-known since the days of Samuel Morse method: we encode small increments with a short codeword, and large ones – with long.

Now let us look at the energy spectrum of the PQRST complex and see the almost exponential decay of power of spectral components with increasing frequency. We move the sliding window into the interval of diastole and see that the high-frequency components have completely disappeared. Hence, we can reduce the sampling rate during diastole, because Kotelnikov’s condition is still satisfied (unless, of course, we have a quality amplifier and not rubbish from "LLC"). Okay, now we can digitize the ECG with normal frequency, and then, depending on the values ​​of the increments, do resampling, dynamic resampling, when in each section the corresponding ECG sampling frequency will be used. This is more serious, we will use this method and ask ourselves the question: why do we need a uniform sampling as to level in the region of large increments? Let us switch to the B-coding. Then there is the problem of the error of large increment transmission. And if we compensate it at the next step, i.e. record and add the discrepancy to the input signal at the next cycle? We obtain the adaptive differential pulse-code modulation (one of its variants). Excellent, but for the perfect "depiction" of wave tops we need high accuracy. Then we use the transaction method - simply put - we give change. That is, we will transfer this discrete reading of electrocardiogram as many times as necessary to reach the zero error. Let us optimize this method and adjust the curve of B-encoding. We will carry out research and find out that the number of old banknotes 1.3.5.10.25.50 rubles is much more convenient to use and easy to obtain the change than relatively modern ones, i.e. 1.2.5.10.20.50 hryvnias.

Thus, a consistent combination of these methods gives the desired result: ECG is transmitted without loss of quality, 10-second segment of 12 lead ECG for 80-140 seconds. Why is there such variation in the transmission time? And it's simple: the more complex is the ECG signal, the greater is the obtained digital stream, the longer we transmit. Disadvantages of the method? There are, of course (we remember that development is the art of compromise, and there are no methods without drawbacks). We need a good amplifier because if it makes a noise or hum, i.e. complicates the signal, our system will transmit ECG for unreasonably long time (maybe up to five minutes or even more).

Although, we do declare the full compliance with a stationary electrocardiograph. So let's make a better amplifier that it has, the recording quality is never too high.

And what can be improved in the ECG amplifier, because the circuit is common, all have almost the same? Nevertheless, it can be done.

Let us pay attention to the conditions under which work a usual electrocardiograph and "Telecard". If the usual device has long wires that go to the amplifier unit, and on the other side there is the power cord, then "Telecard" has nothing of that. That is, for the electrocardiograph its connections represent a quite decent dipole and the reception of RF signals (and even with a maximum at certain frequencies) is effectively guaranteed and in our device this level of noise is much lower. Due to its small size, "Telecard" is a little counterbalance to antenna wires, and most importantly, its capacity relative to the patient's body is much larger than the capacity to ground. As a result, we have a lower level of radio frequency interference, which have the nasty property to leak into the useful signal due to detection on the non-linearity of the input circuits, and because of the effect of "spectrum masking." Okay, but then why we need a standard compensator circuit (applying interference to the black electrode in opposite phase)? Not needed at all (if the notch filters are good and the dynamic range "is not completely exhausted"). Down with him! And with it goes the additional noise introduced by this circuit. A bit, but won.

Come on. We would make the electrode cable wire unshielded. Then at breakage or damage of the cable, it will be easy to solder it yourself, and if necessary, simply twist together and roll up with tape - the device will work. This is very important in actual operation. Can we do this? It turns out - can. But then certainly we will make symmetrical all bypass capacitors at the amplifier input, single "Mecca of grounding"; it is now permissible, because there is no longer a compensator on the black electrode. We compare our scheme of connections with the standard one: our scheme is better as to noise and RF interference, and at least not worse as to AC mains interference. Great, let's proceed to the bandwidth. Please note that we will never need to keep records with the upper limit of the band more than 75 Hertz. Let’s get rid of stereotypes and limit the bandwidth by analog filters by 6 dB at 100Hz. Now all sampling will say us "thank you", because we strictly fulfill Kotelnikov’s condition, and the transfer of high-frequency components through the half of the sampling frequency (that gives noise and dirt in the "self-made" ECG amplifiers) is virtually suppressed by a very decent amount (more than 18dB relative to the standard scheme).

We now have no debris at the input (remember, Garbage In - Garbage Out). And no need to excel with its cleaning, so it does not hit the compressor. It remains to suppress the mains interference of 50 Hz and its harmonics. Yes, yes, yes, harmonics must also be removed. Including even onesi. And do not allow any non-linearities of PFR (very important) at digital preprocessing of the signal.

Well, you just do not mess up the analog preamplifiers. Select the appropriate OA. And optimize the value of noise current and voltage applied to the expected output impedance of the signal source. And our firm makes it good. As a result, "Telecard" records electrocardiograms "cleaner" than other instruments, and in noisy environments our advantage is still increasing.

And it's not because we grab God for his beard, we just have a dedicated device, which is always better than a universal one under those conditions for which it was designed. Nobody catches with one universal rod a small roach and huge catfish?

You are a little tired, dear reader? Relax, while I tell you two cases of operating experience with "Telecard".

In one of the regional centers of Ukraine a commission from Kiev comes to check for the use of purchased by budget funds "Telecard" in order to detect improper use of these same international labor money. That us takers, crooks and embezzlers come to check the people who really work very well. It's in our Slavic traditions as a rule: smart and intelligent break their backs, and any mold depicts leaders. In the regional clinic - chickie! Preparing for the arrival of the leading brethren. They called us and demanded our representative.

- But why? You have everything working fine, I can even say - excellent.

- You do not understand, there’s a revision...

- Yes, you spit on them ...

- No, please help us.

- Well, we do.

And the entire director hurtles through chillness to support. He came, and the commission has already pushed the bushy eyebrows and watches. But with the transfer of ECG - trouble, they fight for half an hour, but instead of electrocardiogram records - not figure-understand what.

Reviewers already salivate in anticipation of the impact of their mission – well, caught by the hand the plunderers of public funds. With the region from which the transmission is performed, the connection is gorgeous, one can hear poor nurse crying out loud.

Our director picks up the phone.

- What is your name?

- Na-ata-asha-a (in tears)

- Natasha, let's not cry. You transmitted ECG yesterday?

- Yes-ah all day practicing.

- And everything was fine?

- Yes-ah, but today (sob)

- Natasha, do not cry. Tell me how you put the electrodes?

- How I wa-as tea-ach-ed and I have exhausted all of the alcoho-ol.

- Wait, what alcohol? With what then you wetted electrodes yesterday?

- Water, we had nothing else.

- And today?

- And today I was given alcohol, because there Komi-used B-Ia, and from Key-Ieva.

- So, then you water moistened electrodes?

- And I was told that alcohol is better, as there is the Komi-sion.

- Natasha, alcohol does not conduct electric current, moisten the electrodes with water and retransmit. And stop sobbing!

Next - excellent transmission and controlling gang went empty-handed, to the office of the Chief physician - to profit at least something.

And alcohol really does not conduct electrical current!

The second case was more useful for us. We went to Kiev, the mother of Russian cities, to a highly scientific institution in order to show "Telecard". Out of this institution came a middle-aged scientist of medium masterness (i.e. did not a master, but so-so – submaster-infante), but with a proud turn of the head, and took away our device. For review. And we have to sit in the corridor. But we waited not for long – he went out with the device.

- The conclusion is negative!

- Can we see the results of the tests?

- It does not work.

– ???

We take the device, check it, all right.

- It is OK.

- No, our "Schiller" records ECG, and your device - not.

- The electrodes are the same?

- Of course.

- Can we look at them? (I intruded)

- Come

And we find ourselves in a scientific sanctuary where embryos of future Nobel Prize winners may potentially club, although now there is nobody there, but the electrodes, really - are on the table. Without saying another word, I take the tester and measure the resistance. Holy Mother of God (and so on) – yes, they have the resistance in the cable over megohm! I look closely – yes, it's soured the whole, blue-green flakes of wire oxides dangling from the connector. They unclog drains with them, or what? But "Schiller" works, the record, however, is very poor, but readable.

- Here, learn, a European instrument, here's how to work!

My colleagues looked at me accusingly. Nothing to answer, the fact is that in the instrument at that time there was a control system of electrode resistance. And first it warned about impermissible resistance, and then switched off the device. Needless to say that we took this system and threw it to hell? Because if in the metropolitan churches of science electrode cables are used to clean toilets, then on the periphery, apparently, they will stun with them pigs before slaughter. And anyway, if our man has invented something himself (for example, to use shampoo instead of electrode gel and get a result of unacceptably high impedance), he will never give up, overcome all obstacles, all inconveniences, all the hardware and software locks, but will never refuse from his idea.

So, “Telecard” operation for the user:

1. We put electrodes on the patient.
2. Press a single button on the device (it adapts 12 seconds and 10 seconds of ECG records). At the end of the recording, the phone beeps and the ECG is stored.
3. If the receiving station is ready to receive the ECG, then click this single button once more. The instrument sings his song, bring it to the phone, with which we spoke to the central station operator, and wait.
4. After about two minutes the unit will beep and turn off, the voice communication with the central station is restored, get the advice of the cardiologist.
5. Decide what to do with the patient: urgent hospitalization, appropriate emergency aid on the spot or a sedative kind word. That’s all. We have spent for the procedure no more than five minutes. Unless, of course, we do not ask for ten times what drug in what dosage to administer.

And now about the criticism of the "functionally defective products" (guessed as to such stupidity Microsoft victims). In what, you ask, this disability, honey?

- A small screen is not present, the electrocardiogram at the bedside is not visible, ha-ha.

- Yes, it is not a problem to install a screen, but we just do not need it. If the staff who came to the patient, is able to adequately analyze the ECG of the patient, no telemedicine consultation is necessary. And if not able - then no small screen would help him. And yet, in order to correctly understand the difficult situation at the patient’s bedside, there, in addition to electrocardiograph should be a cardiologist, and they are not yet in queue for the medical assistant job (though it may be, we soon will bring the native medicine to such insanity too?)

- But we have seen a system, genuine, advanced, so there via "bluetooth" everything flies on the phone, from the phone flies on a cloud server, in general, real high-tech. And yet - ECG itself falls to the receiving station, and from there SMS falls on the doctor's phone, in short, there is the automatic reception and even the automatic conclusion, that's it! And you have one button on the device, just like in decrepit USSR.

- What is traveling and where that falls, you kidding me? Have you long ago been pleased to visit a village? And a provincial ambulance? Have you ever visited patients at their homes with a local doctor? How do you imagine a doctor's work in an ambulance car, seeing American TV series? Please, understand, the device for transtelephonic ECG works under the most difficult conditions: it is dropped, its wires are torn away, it’s left forgotten on thirty-degree frost, thrown under glass in an ambulance car, which is heated by the sun, taken from the cold into a warm house and they record an ECG with the frosty device. What the hell, bluetooth and smartphones? They only drastically reduce the reliability of the system and complicate its operation. We regularly replace torn with the root wires, smashed device bodies, forgotten and leaking batteries. And about one button: so our engineers worked for a few months to ensure that the device is driven by a single button. Because a rural nurse milked a cow in the evening, and then her grandfather Ivan has trouble with the heart, and straight from the cow, she went to see her grandfather and transmit his ECG to the district. She did not need to understand the buttons and screens, she will still forget them, but press on "Telecard" one button once and then again. And listen to what they will say from the district hospital. Got that?

- And yet, why you do not have automatic reception of ECG and automatic conclusion? It's so convenient, the system automatically analyzes the received electrocardiogram, and if it is normal, the doctor does not care if the nurse can understand herself, then sends his opinion, and only if, well, in very complicated case – she sends an SMS the doctor, asking him to look at the ECG.

- And never will! Neither one nor the other, categorical - no! With automatic conclusion everything is clear, it is, that you know, forbidden in developed countries. But such devices are allowed for export to the yellow and black and CIS countries. Just try to use such a device in the United States, you’ll quickly have your rectum turned inside out.

- Ugh, how rude! And you could not make an automatic reception? Not enough qualification? Copying half a century old Soviet developments?

- It is perhaps still the only sensible question from you. I’ll try to show on a simple example. Let the automatic reception be available. Received ECG are stored, the doctor at the receiving station periodically scans them and calls back to those who are in need. Now imagine yourself in the place of an ambulance car paramedic. He has three urgent calls and arrives at the first patient: the patient has either infarction, or severe intercostal neuralgia. ECG is necessary, it is recorded, transmitted and - silence. The phone is silent. And two other calls are waiting. And the phone is silent. The patient feels himself worse and he is also half seas over, some foam appeared, the pulse is faster, looks like arrhythmia. And the phone is silent. But his wife said that they boozed, and then nibbled some nonsense, and all his friends also spat foam. Your decision? And the phone is silent. Whether the ECG was not received, or the doctor left for a store, or they cannot get through? Or maybe, there is nothing on the ECG, therefore they do not call? And the phone is silent. The relatives look at you with the evil eye. Call yourself? Call, but ask yourself the question, what the heck this same automatic reception was made for? By the way, those two other calls are still waiting; they already called back the dispatcher and swore, one patient is really bad.

Impressed?

 - Well, why "Telecard" cannot operate as an electrocardiograph?

- But this is not true. It can, and very easily and conveniently. Using any computer, laptop, netbook, tablet or even a smartphone. Just install on it our free software. Hold "Telecard" to the microphone of the device you have selected, and after two minutes you’ll have the ECG on your screen, on the printer, you can either show it to the doctor, or send it to an empyrean server, or even further.

- You have not convinced us in anything.

- And I did not hope to.

- And that's what, you could not talk much, you’re unintelligent.

- I will answer with Timofeev-Resovskii’s words: "I'm not an intellectual - I have a profession."

In conclusion, I would like to give the illustration of the electrocardiogram transmitted by "Telecard" via a phone, let the doctors estimate. I think that the recording quality will satisfy even professionals from IT. Although, you never know, there are few buttons on the device, and no screen at all.

- Aha, - says the critic - you have created ideal conditions in your laboratory. You would still have connected a simulator instead of a person. But show us the real work, when the record is made at the patient’s home, and not by the staff of your company, and via only God knows what phone.

- Please - I answer - there is nothing easier, that's why we have more than a hundred thousand of these transmissions done every year. Now we ask, and we will receive the results. Directly "from the conveyor", to keep the purity of the experiment.

- Well, not bad for the "flow" of research?

- You know what, stop boasting with technological achievements, you have excellent records. If all ECG were such, the problems would not exist. You better tell me - how you deal with the staff?

- Yes, Doctor, it's a question "in the top ten." Staff - the alpha and omega of our troubles and misfortunes. However, and success - too. If literate people have gathered, and want to work, the results are amazing.

And of all the transmission problems - only one of ten has the technical nature. All the rest - the "human factor". In a purely empirical way, we have found that it is necessary to force, I emphasize - force administratively, everyone who received "Telecard", to make ten transmissions to the center. No matter, sick, healthy, can record himself, but must make ten transmissions. A remote diagnostic center immediately after system commissioning should receive a hundred ECGs. Then it will go. Otherwise - will freeze, and the instrument will be moved quietly into dusty storerooms and the computer will be taken to the accounting department. And before the system dies, we will receive the complaints.

– Hello, thiz iz Tredex? We trenzmit your ECG and it iz not received! (Soft South Russian dialect causes in this case not a smile, but the desire to curse).

– Yes, it is. How do you transmit ECG?

– Teik, record, button pressed, called Re-De-Ce, and it iz not teiken!!

– And what have they said in RDC (Remote Diagnostic Center)?

– Said it iz not teiken, no signal.

– Do they hear you OK?

– Yea-ah.

Well-well-well, the picture becomes clearer.

– And do you bring “Telecard” close to the phone?

– No-a, Mykolanikolayevich said zat we don’t need to, there is blutuz.

– Who the hell is this Mykolayovych? You've got the instruction and the figure in it!

– Dohtor.

– Young?

– From institut.

Yeah, one more advanced user on our heads. But nothing, it is treated. After 40 minutes of intense and nervous consultations over the telephone - HURRAH(!) - The first electrocardiogram is sent and received. Our employee wipes sweat from his forehead and promises to start looking for another job tomorrow, he says ... never mind, he will calm down in half an hour, and everything will be fine.

Significantly worse, when the system is sentenced to a dusty storage room in silence. Sometimes it is because people do not understand: a phone has a microphone (they do not read the user manual and traditionally ignore it). They put "Telecard" to the place where they hear the subscriber’s voice "so they can hear", and apply the reverse side, the "Telecard" speaker to herself. The computer at the receiving station, of course, swears and in Russian language writes a warning in red letters. And well, if in RDC the operator reads this message and forces the paramedic too work correctly with "Telecard", but not waves his hand: "Bad system!".

Our domestic staff - is not well-trained and well-paid German or French nurses. This is entropy in its purest form. And it is impossible to influence them in any way. And the equipment in such circumstances must exceed all conceivable standards as to the simplicity of use, reliability and fool-proofness, even the military ones. Because if a soldier stubbornly inserts the dispenser magazine into his machine gun upside down, his comrades will quickly correct his brains in most undemocratic and not tolerant way, and he will instantly realized his mistake. And in our case (if anything cannot be inserted upside down) the sentence is fast, ruthless and not appealable: "Does not work. Because chaffy!".

And you say - hi-tech, telecommunications, information technology...

By the way, the beloved administration is also not lagging behind. Can make such things that a combined battalion of drunken nurses admitted to medical work, will look like extra-class professionals. For example, you can choose a corporate network from the cheapest (in all senses) provider. It will be the communication for teenagers: quality – zero, reliability below zero, coverage - fragmentary. Or, for example, cheap discounted cell phones. For the urgent system (real case!). Or they have installed the receiving and transmitting equipment in a primary care unit, where and who now will transmit ECG, what’s the meaning, whom who advises? And now nothing can be changed - it is easier to purchase additional receiving stations for hospitals. And this, remark, is a purely administrative breakdown associated with the fact that neither with us, nor with their specialists nobody consulted before purchasing, and not even intended to. Screwed from all administrative soul.

A typical attitude to doctors? They install the "Telecard" receiving station in an intensive care unit. And - go ahead, doctors dear - treat ECG from the whole city unpaid. And your responsibilities in the intensive care unit were not canceled, arbeiten-arbeiten, work makes you free! The result? Doctors lift the telephone receiver and ... do not put it back. Impossible to call. Finita la comedia.

In fairness, the administration can be smart and professional. In one of the cities in Ukraine they have found the information themselves, they have bought for ambulance cars, work successfully and professionally, and while in Kharkov understood what was happening, they transmitted more ECG than the one-and-half million first capital.

So, dear reader, the medical equipment developer's life is hard, but not boring. However, our job is done. The device exists, is produced in batch and really works, and under domestic, very, very not greenhouse conditions. And the more it is used, the less there will be regrets about that: "Oh, if we'd figured out, and helped, maybe Ivan Ivanovich would be sitting at this table with us".