Must I have a buffer?

[biomed12]

Hello dear members,

This design, which is attached, is part of my portable prototype. Particularly, 9V input is to 7805 and its output(5V) is input to ltc1983. LTC1983's output is -5V. My aim is getting either +5v and -5V for opamps from single power source. Their regulating capacitors are connected cascade, naturally. So, I wonder, will any problem arise in this desing? Should I add a buffer between them?

[Some Anon Guy]

Capacitors in parallel are just fine. It is the same as having 1 capacitor equal to the total capacitance of both combined. (kind of...)
Capacitors of multiple values are commonly used as bypass capacitors because different values and internal resistances will "smooth" ripples at different frequencies...
 

[solutions]

What do you mean "add a buffer between them"?

Can't tell you if you'll have problems if we can't see the rest of the circuit (we designed most of it here for you so no big deal to put it up, I'd think).

No idea how much power you need - in any case, you are cutting your battery life in half with this approach.

[Wizpic]

I would use an lt1305(step up converter) and a couple of 14500 3.7v 2500mah in parallel to give a 5AH supply it will run longer than a 9v battery. If it's a AC power pack I'd use a step down converter to 5v. I would say your -5V would be ok if you're only using them to power a couple of op-amps as the LTC1983 can supply 100ma.
I've used the LT1305 in a project I did and get a very good run time. I've also used these cheap and cheerful
http://www.ebay.co.uk/itm/DC-DC-Boost-Converter-Step-Up-Module-1-5V-to-5V-500mA-Power-Module-New-TM/381788902587?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D38661%26meid%3D24e2571c90a446e7ab3bb44d0a66908c%26pid%3D100005%26rk%3D1%26rkt%3D6%26sd%3D262779837203
They still work very well cheaper than building your own 

[biomed12]

I would use an lt1305(step up converter) and a couple of 14500 3.7v 2500mah in parallel to give a 5AH supply it will run longer than a 9v battery. If it's a AC power pack I'd use a step down converter to 5v. I would say your -5V would be ok if you're only using them to power a couple of op-amps as the LTC1983 can supply 100ma.
I've used the LT1305 in a project I did and get a very good run time. I've also used these cheap and cheerful
http://www.ebay.co.uk/itm/DC-DC-Boost-Converter-Step-Up-Module-1-5V-to-5V-500mA-Power-Module-New-TM/381788902587?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D38661%26meid%3D24e2571c90a446e7ab3bb44d0a66908c%26pid%3D100005%26rk%3D1%26rkt%3D6%26sd%3D262779837203
They still work very well cheaper than building your own 

Power consumption is not very important now. I will get just about 10ma from -5V source for opamps. I read in an OPA series opamp's datasheet that, V+ and V- just consume 3-4 mA. Maybe also probably, I can use a 9V 500ma DC Adapter for supply design. All of the power consumption of the circuit is about 150mA. In fact, another reason to use a DC Power Adapter is that.

Just I had an doubt if 2 ICs will affects other and doesn't provide a proper voltage because of cascade connection. 

Thanks.

[Signal]

This design, which is attached, is part of my portable prototype.

I would use an lt1305(step up converter)

I think the point is worth to be explained some more. In assumption that "portable" means battery powered a 7805 is not a good choice.
The main reason - the 7805 is a linear type regulator - it's theoretical maximum efficiency is 5V/9V=55%. And 7805 has comparatively big ground current 5-8 mA. Anyway I consider 78xx obsolete.
About input capacitor - rule of thumb - it's capacitance should be greater then output one. Some regulators are critical to this some not. Of course there are factors as type of supply (output impedance/capacitance) and distance between supply and regulator.

V+ and V- just consume 3-4 mA.

Add to this value a load current. Assume that average "positive" and "negative" load current is 2mA (maybe less actually).
Then roughly: 4mA+2mA +2*(4mA+2mA) = 18mA from positive rail (7805 output). For 6mA output current an efficiency of LTC1983 is about  55% (TPS60403 has 97% @6mA load).
Considering efficiency of LTC1983 a consumpion of OPAMP from 7805 output will be 4mA+2mA + (2*(4mA+2mA)/0.55) = 24mA.

Maybe also probably, I can use a 9V 500ma DC Adapter for supply design. All of the power consumption of the circuit is about 150mA. In fact, another reason to use a DC Power Adapter is that.

Then I'd use switched type regulated DC adapter. Then I'd choose from
A) 5V DC adapter with eliminated 7805 with just good filtering/conditionning
or B) 6V DC adapter followed by modern LDO regulator to 5V.

Do not forget about power dissipation. For assumed 150mA and 9V 7805 will take 0.6W - it could be hot. Choose proper package/radiator.

Just I had an doubt if 2 ICs will affects other and doesn't provide a proper voltage because of cascade connection. 

It is a proper connection - to take LTC1983 input just from 7805 using it's output capacitor as input capacitor for 1983. They will work just fine without bad influence on each other.
BUT. Situation becomes more complicate concerning different loads and theirs requirements to supply and how different loads affect supply (make noise).
Then there are possible solutions for power supply:
1) just your circuit with refined capacitance values;
2) filter between 7805 and 1983 with separated capacitance for 7805 output and 1983 input;
3) filters between +5V /-5V rails and opamp;
4) separate 9V/5V regulator exclusively for OPAMP supply;
...
So it depends.
Mentioned filters could be resistors or ferrite beads followed by capacitor for example. It also depends on PCB layout. For example long upcoupled 5V trace (between 7805 and 1983 for example) is a good candidate for ferrite bead (think about AC-shorted loop "CAP1-power_trace-CAP2-ground-CAP1" and "CAP1-powertrace-ferritebead-CAP2-ground-CAP1" variant).

[biomed12]

Dear Signal,

Thanks for your detailed descriptions and suggestions. In fact, I dont have wide knowledge about power electronics, this is why I am confused.

Then I'd use switched type regulated DC adapter. Then I'd choose from
A) 5V DC adapter with eliminated 7805 with just good filtering/conditionning
or B) 6V DC adapter followed by modern LDO regulator to 5V.

Actually, the reason of using regulator is I don't trust bad conditioned chinese adapters. For example I have one which should give 5V but it provides 5.6V. I have 2 MCU and fixed 5V is so critical for me, as you know. Traditionally, I buy 7805 for my circuits. Which part do you recommend particularly?

For assumed 150mA and 9V 7805 will take 0.6W - it could be hot

Yes, it was a bit hot. I thought that I should have used 2 7805 paralelly connected at that time. I think, it will be hotter. Because, I will add ethernet controller to my circuit and it will add to the total consumption about 50mA.

2) filter between 7805 and 1983 with separated capacitance for 7805 output and 1983 input;
3) filters between +5V /-5V rails and opamp;

Do you mean RC filter? In fact, I dont have much background about this topic. Do you have any recommend to me, how can I get a good practical knowledges like you? I have "the art of electronics" book. Is is enough or do you recommend any other sources?

[zac]

A linear regulator like the 7805 is inefficient, but among the least noisy power supply arrangements.  There are also lots of low dropout linear regulators available that may allow a more efficient design if the input voltage is close to the output voltage.  The 7805 has a typical dropout voltage (input-output differential) of about 2.0V compared to 0.5V on some low dropout types.    On the other hand, the LT1983 charge pump will create lots of power supply noise so would seem to negate the low noise benefit of a linear regulator. 

[Gallymimu]

The short answer is:

Yes, the circuit will work fine for the purpose you specified and asked about.  Good job.

Nothing wrong with what you did and as expressed by everyone else there are dozens of ways to get the same outcome.  (I've used your exact design myself many times for simple compact power solutions.)

[solutions]

Since you don't know what you are doing, do not even THINK about attaching this system of yours to a human, preferably not to an animal you don't plan to immediately eat either. That battery does not excuse an electrocution via Ethernet

[Wizpic]

Signal,
I used the LT1303 which is a step up converter 1.5to4.2v input to give a 5V output rated at 500ma, using 14500 batteries these are the size of AA batteries, my system draws 89Milli-amp giving me 20hrs continuous usage off one signal cell 40hrs doubled up even though I don't run it continuously. But I now use those chep step up converters from e-bay these work just as good and these may be even cheaper than a 7805 regulator ready built

[biomed12]

Since you don't know what you are doing, do not even THINK about attaching this system of yours to a human, preferably not to an animal you don't plan to immediately eat either. That battery does not excuse an electrocution via Ethernet

I will attach the prototype to me actually I am not scared of low voltages like 5V and low power values.

[Signal]

Signal,
I used the LT1303 which is a step up converter 1.5to4.2v input to give a 5V output rated at 500ma, using 14500 batteries these are the size of AA batteries, my system draws 89Milli-amp giving me 20hrs continuous usage off one signal cell 40hrs doubled up even though I don't run it continuously. But I now use those chep step up converters from e-bay these work just as good and these may be even cheaper than a 7805 regulator ready built

I agree, the use of step-up converter is a modern approach (though not exclusive) for battery powered applications for reasons.
Two notes.
1) "1.5to4.2v input" is not correct. 1.5 is not guaranteed nor stated for LT1303.
LT1303 "Supply Voltage As Low As 1.8V" is justified with intended use "2-Cell and 3-Cell to 5V Conversion".
2) I used MCP1640 in similar situation (1 cell alcaline AA to 3.7V out). I think it is better (because perhaps newer) in all aspects than LT1303.

[Gallymimu]

I agree, the use of step-up converter is a modern approach (though not exclusive) for battery powered applications for reasons.
Two notes.
1) "1.5to4.2v input" is not correct. 1.5 is not guaranteed nor stated for LT1303.
LT1303 "Supply Voltage As Low As 1.8V" is justified with intended use "2-Cell and 3-Cell to 5V Conversion".
2) I used MCP1640 in similar situation (1 cell alcaline AA to 3.7V out). I think it is better (because perhaps newer) in all aspects than LT1303.

ACTUALLY for VERY low power applications switching converters can be WORSE as they have much higher quiescent current than a linear regulator.  TI, in fact, has some dual linear/switched buck converters to address this issue i.e. linear in sleep mode, buck while awake.

-edit: corrected my statement-

[Wizpic]

Yes the Lt1303 is 1.8v input the ones of e-bay work down to 1.5v which I have tested but I have a cut off of 2.95v but never reached that voltage yet

[Signal]

Actually, the reason of using regulator is I don't trust bad conditioned chinese adapters. For example I have one which should give 5V but it provides 5.6V. I have 2 MCU and fixed 5V is so critical for me, as you know.

Yes, it is common practice to incorporate conditioning of unknown input in one inseparable block with a consumer of this input.
The worst part of unknown adapter is not the country of origin (most of electronics is from china and most of it is great). The worst part of unknown adapter is that it is unknown and requires investigation before use. It could be classic transformer-bridge-capacitor with high output impedance (so unstable output) and pulsed output under load (without oscilloscope the measured average 5.6V means nothing). It could be switching type with short over-regulation pulse to 6V at startup (not a big problem after adding some more capacitance). On the other hand it could be fine, precise regulated with acceptable 5% tolerance working on your lab desk but customer will try to replace this adapter by another with similar plug.

Anyway, read datasheet of all used components. Upper limit of power voltage for some PICs could be 5.6V operational and up to 7V absolute maximum (as I recall).
As a variant, 3/3.3V parts (controllers) require additional 5V to 3V regulator, while OPAMP could have much higher than 5V upper limit, then coarse regulation of input 5V could be not a problem.

Traditionally, I buy 7805 for my circuits. Which part do you recommend particularly?

I can not recommend anything without requirements. For example, 7805 is a right horse for task expressed as "to take about 9V from any picked up wall AC/DC adapter and with minimal attention and least work to get precise 5V". 7805 is cheap (or free if scattered on shelf), is least critical to output capacitance, very few parts were prone to zero output current fail. It just works. It is not your idea to buy 7805, you do it because all do it, and it is because of long story of success.

Yes, it was a bit hot. I thought that I should have used 2 7805 paralelly connected at that time. I think, it will be hotter. Because, I will add ethernet controller to my circuit and it will add to the total consumption about 50mA.

Hey, don't you remember that voltage sources do not like to be paralleled? Remember why? This is a good example of the fact that theoretical knowledge works well in practice.
There are several possible solution for higher dissipated power.
1) add radiator - screw TO220 to aluminium. Anyway pick the part with appropriate parameters - maximum current, thermal resistance etc.
2) reduce consumption of target application (just to complete list)
3) increase efficiency of voltage conversion by reducing input voltage (that requires more stable source and LDO type linear regulator).
4) increase efficiency of voltage conversion by using switched type regulator. In common there is a threshold of output current when switching regulators become more efficient than linear. Nowadays this threshold becomes much less than former 100mA. And of course in all cases where you sense hot a switching regulator will be much cooler.
Types are boost(step up), buck (step down), step up/down (SEPIC). I have just looked in my usual store for cheapest buck converter and found MC34063 0.15euro (without external inductor, diode...). I did not use it - just for price scale.

I do not recommend you personally to use switching regulator now. It requires some study that is time. And spending time for this part is not efficient at curent stage of your project and your education. When you want to drink tea you do not want to cut wood for stove, don't you?

Do you mean RC filter?

Yes, rather RC than LC.

Posted on: December 29, 2016, 08:05:11 20:05 - Automerged

ACTUALLY for VERY low power applications switching converters can be WORSE as they have much higher quiescent current than a switching regulator.  TI, in fact, has some dual linear/switched buck converters to address this issue i.e. linear in sleep mode, buck while awake.

I was very impressed and happy when one day found these long awaited dual type regulators. But they are targeted at different application - efficient high power conversion with efficient retention of powered stand-by state. As a side effect this regulator could be well suited for loads that vary a lot that causes a compromise selection of inductor for usual buck regulators.

But discussed LT1303 and MCP1640 (both are quite offtopic here) as representatives of "non exclusive" approach are targeted at different problem - to get higher voltage from minimum amount of cells and to eat this cell completely.

[Sideshow Bob]

@Biomed12
Is it any specific reason why you are asking this or these question(s). Is it a question from your professor or are you having problems with noise in your ECG setup as an other example. If you struggle it is better to tell why and not holding things back. Remember we are not sitting at your work bench. As a tip then testing you must use proper medical grade electrodes. To ensure the best results. Also some electrodes dry out very quick and become 100% useless. A hairy chest is also not a good thing for ECG pickup

[CocaCola]

Also some electrodes dry out very quick and become 100% useless. A hairy chest is also not a good thing for ECG pickup

Tensive Conductive Adhesive by Parker Labs will solve both those problems 

[Signal]

In fact, I dont have much background about this topic. Do you have any recommend to me, how can I get a good practical knowledges like you? I have "the art of electronics" book. Is is enough or do you recommend any other sources?

There is no magic pill. Beginner becomes experienced obviously after getting experience that implies time (years) spending in area of interest. That way consists of steps. One at a time. Even for genius. You just solve your current task (that you able to handle) one after another. Nobody expects that after finishing your courses you are an expert in whole area. But quality of limited educated knowledge is clearly seen and is a sign of potential to learn and do more. I am  saying it now because it seems that you try to overextend the program of your education. That intention is not bad but risky. Anyway sorry for that lyrical digression and of course I could be wrong again.

My background was formed many years ago in different environment with sources in different language. I do not know good educational books in your native language (that I consider preferable) or in English. But of course I can recommend what you can do. In short - one book is not enough even if you already own it ;)

 - At first, do not hurry. Take your time to study a thing. Do not scatter your interest among many things at the same time.

 - Focus on tasks that inspire you, not on instruments. If you will be asked you will say "I wrote a book" instead of "I know MS Word".

 - Do not waste your time seeking for best instrument from set of competitors. They all are good in spite of opinions of holy wars participants. They all are produced in millions and make their work. Select one that is sufficient for current task and is easier for you personally to use right now and go forward. Dozen of % of difference of some parameters means nothing to you really. Your time is the most valuable parameter.
So if you are familiar with PIC do not tune yourself to ARM just because you heard that eclipse is better than netbeans or whatever. Do it if you want to know yet another processor/platform/environment and this is your current task (he-he). That is the most contradictory statement in this list but maybe you can get it ;)

 - Simpler instruments are here for a reason. They are easier to deal with and it was easier to make them. Read errata on each component (controller) you plan to use. It helps to understand what is behind complexity. Sometimes it should be done even before you read datasheet. There could be showstoppers for your task. That way you will not be like most (if not all) participants of holy wars. Obviously it requires a time - see the point number one.

 - You are a professional. Just RTFM! (datasheet or whatever). Each. Completely. Understanding each stated parameter. period. Oh, again, it needs time.

 - Now in epoch of internet self-education becomes much simpler (not easier) than ever. All you need is a desire (and time, see number one). Of course, it is easier and with much more reliable result especially for start to take courses with mentor (because it takes mentor's time, besides other things)

 - Having a mentor do not focus on what he missed, focus on what he passes to you.

 - Valuable sources of knowledge require time (again!) to be composed. It is more expected in articles and books than in forum's posts. Use this as criteria while looking for information. Of course, forums are also important, here we are sharing;). Forum says your circuit is OK. Are you happy? It is not enough. Now you should find the understanding why.

 - Educational or third party literature is not a replacement of original sources.

 - Good application examples are very useful. Starting point - original application notes from parts manufacturers. But only a starting point.

 - Two main sources: academic and from manufacturer. No one could be ignored.

 - Continuous learning is a condition for successful career. Till the very end. So do not be too fast. To be consistent is enough.

And finally, I have spent my time just to say some things that you can consider useless, while expecting different kind of help. Do not blame me for this. I did my best ;)

[solutions]

I will attach the prototype to me actually I am not scared of low voltages like 5V and low power values.

And that is EXACTLY what will kill either you or your volunteer.

You said "wall adapter"...

[CocaCola]

And that is EXACTLY what will kill either you or your volunteer.

You said "wall adapter"...

That and it's not the volts that kill, we commonly get shocked with 10s of thousands of volts aka 'static', stun guns, electric fences and what not but as little as 20Ma of current can cause heart issues, while around the 100Ma mark is considered the edge of lethality...  When you start reducing that resistance with conductive agents and probes, things can go bad real quick, especially if the circuit malfunctions while it's connected to a mains line that has every bit enough juice to stop a heart...

[vern]

If you want to build a medical device that will use electrodes to attach to a body you will have to make sure that there can never be a dangerous current.
Most simple method of course is to put a series resistor with a high voltage rating in every electrode wire to limit currents to about 10uA.
This resistor has to be inside the medical device at the start of the electrodes to make sure that it works even if the electrode wire breaks or the insulation is damaged.
That makes sure that even a wall cube failure does no harm.

[Sideshow Bob]

If you want to build a medical device that will use electrodes to attach to a body you will have to make sure that there can never be a dangerous current.
Most simple method of course is to put a series resistor with a high voltage rating in every electrode wire to limit currents to about 10uA.
This resistor has to be inside the medical device at the start of the electrodes to make sure that it works even if the electrode wire breaks or the insulation is damaged.
That makes sure that even a wall cube failure does no harm.

It is more complex than that. And some equipment like electrosurgery or nerve stimulators may output currents that may cause heart failure. But they are safe to use as long as the current do not take a path through the heart. And then you have the defibrillator designed to put a "lethal" current through the heart.
Here is an open link that may give you a crash course in medical safety rules
http://www.imt.liu.se/edu/singlesubjectcourses/2015/2.tuesday/2/A%20PRACTICAL%20GUIDE%20TO%20IEC60601%20%20Standard-1.pdf

[vern]

sideshow Bob,
thanks for the information, good paper, very useful.
My advise was only for a beginner to make sure he is not electrocuting anyone (including himself)
I hope no one here in this forum wants to build a defibrillator!

[biomed12]

Vern,

You are right, I have been building a medical device prototype for 4-5 months, particularly a digital ECG. I know some basic rules. For example, "don't attach your prototype with another medical device which has a connection with human body if your medical prototype has some electrical standarts". I am using two battery for intsrumentation now and I dont care about this topic for now. After my dc converters has been delivered from digikey, I will migrate all of the design to a fabricating PCB or DIY PCB. I had much workout for this final project and I think, It will be the best project of my department ever. I had completed that steps: ethernet communication, usb communication, digital filter layer... The last and the most complex layet is instrumentation layer of the design. As you know, biosignals are completely bulls*it. For instance, ecg signal has just 1mV amplitude. I have a traditional instrumentation layer and I dont like its performance yet because, I have some noises on ecg signals that I dont know their source. I need a bit fft maybe... I will have a 24-days internship term which will start on next week. In this term, I will design an instrumentation circuit which will be the best of me during first 2 weeks. After all of these jobs, I will focus another details like safety.

Thanks.