EEVblog #72 - Let's Design a Product

hi welcome to the EEV blog an electronics engineering video blog of interest to anyone involved in electronics design I'm your host Dave Jones hi Ivo suggested some time back if I could do a blog about designing a product from scratch from start to finish animal it sounds like a good idea but it's actually incredibly difficult because there's a lot of issues and concepts which go into a designing an entire product but I'm willing to give it a go so here we go I'm going to show you how I designed a product from start from concept to finish so the best example I could come up with for this I think is my microcurrent adapter you've seen it before and it was published in silicon chip magazine April 2009 and I think this is a really good example of how to design a simple product from start to finish because there's not much in the circuit it's a very very simplistic circuit but as you'll see there's a lot more to design in a nice little product than just the circuit so what's the first step well the first tip is to define the problem and define what product you're going to design to solve that problem in this case its burden voltage on multimeters now I won't go into why burden voltage is a problem you can see my previous one of my previous vlogs for that so I won't go into why I'm actually going to design this but let's just say it's a problem and we got to fix it so what do we have here let's take a look at the problem the problem is you've got a power supply unit you've got a multimeter which has a shunt resistor in there to measure the current when you're powering your particular circuit now the voltage across here this shunt resistor is too high in a lot of cases and that causes all sorts of problems so we want to lower that value of shunt resistor and still use our multimeter ok so what we want to do is design a little do that box that that works with our multimeter that plugs in in series with it and it lowers this value of shunt resistor here that lowers the burden voltage so here it is we need a box let's put - we need and we still need a current chakras because that's how you measure current you have a shunt resistor and and you measure the folders drop nothing's going to change there there's no secret to it so it's very simple but we just need to make it much lower than the one inside a typical multimeter how much lower well ah as always in electronics you talk you know an order of magnitude or ten times R is always a good thing so 10 times less in value that's not bad but we'll go into that later maybe a hundred times but so the value will be lower and because the value is lower is a kookabura tell we're in Australia now um because the value the shunt resistors lower the multimeters not going to be able to measure that value so you need an amplifier of gain that we haven't determined yet so that's it that's all we're going to have in our box is a shut resistor and an amplifier too simple and there'll be a battery to power it as well but that's our entire product but let's look at the detail which goes into actually designing the final product next up let's look at the basic specs we want to do now the main problem I wanted to overcome is basically on the typically on the micro amps and the milliamp ranges I didn't really care much about the amps range so what we want to do is have multiple ranges obviously one range isn't going to do the whole thing it's just not going to work as we'll probably deduce later so we need something with different ranges we basically need are three different value current shunt resistors now there's two basic ways to to do this either you have one set of input sockets like we do here okay and you switch in you you have a three ways well let's say a three-way switching you're switching different current shunt resistor values or you can have one ground terminal and then a different arm a different formula meter banana terminal for each current range but these banana terminals are quite expensive and it's just it takes up front panel board space and I wouldn't to make this thing as small as possible so really I think I thought it was better just to do the range switching based on a switch okay so we've decided that we're actually got which the inputs for the different ranges so we get rid of this idea down here and now we need to figure out what value these resistors need to be for our different ranges now because what we're basically trying to do here is use a multimeter to measure current we can't use the existing current range on the multimeter because it has its own current shunt resistor so what we do is we use the voltage range of the multimeter which just so happens to be the most accurate range as we'll come and do that's actually an important advantage now as it turns out usually the millivolt range on a digital multimeter is the most accurate and it's the easiest to use so that makes sense to use the 200 millivolt range on a 2,000 count multimeter so what we want is a different current ranges we want let's say we want to directly convert a current into voltage so we want 1 millivolt per milliamp for the milliamp range 1 millivolt per micro amp for the microwave range in one variable / nano and because I think it might be handy if we include some nano amps as well now let's not limit our options here we may actually want to measure amps as well just for good measure if it's easy to add to the design but we'll find that out later so you might have one knew of our program now let's look at what value can't shut resistors we need to do that directly 1 millivolt per ab using Ohm's law is 1 milli ohm 1 millivolt per milli amperes 101 millivolt from micro amperes 1k 1 millivolt per nano amp is 1 Meg there what we need for the current shunt resistors but really that's the same as what's in a regular multimeter we don't want that we want to decrease it at least an order of magnitude or several orders of magnitude so um you know we have to sort of make a decision there one order of magnitude 10 times is pretty good but I'd prefer a hundred times so I'm just going to pick that as an arbitrary figure I'd like my design to be a hundred times lower so in this case umm well o equals ten milli a milli ohms because we want it the value to be 1/100 of that value so it's 10 milli ohms likewise 1k is 10 ohms and 1 Meg is actually 10k ohms so they're the values for our three can't shut resistors if we use a times 100 amplifier now we have to take a cursory look at the amps range to see if it's easy to add to our design now I'm a standard for 1 millivolt per amp understand a multimeter it's a 1 million shot resistor that's already very low 1 millivolt per amp and all typical multimeter has a 10 amp current range so it's going to be only a 10 millivolt drop so it's not really a big deal so I don't really think that there's a problem to be solved there by having amps and when you go up and in very high currents like this you have real problems with your connection resistances as will actually see later even on the milliamp rein so really we you know abs is quite hard to add I mean we could change that instead of 1 millivolt per and we could change it to 100 millivolts per amp and gain a you know a 10 x advantage or something like that but you know I don't think there's a problem to be solved there so let's not bother I think we'll scrap that amps range because we don't want to use a 10 micro ohm resistor to get the same to get the same ratio ranges as our other ones so it's a it's a bad idea no amps it's gone all right so our basic design is taking shape we have three different current ranges milliamps micro amps and nano amps we have three different current shunt resistors which are a hundred times lower than a standard multimeter basically and to compensate for that we've got a times 100 amplifier in there as well and the good thing is is that there's a direct relationship between 1 millivolt and 1 milliamp 1 millivolt one microwave and so on so that our multimeter arm is going to read directly in amps per milli volts so if you're on the 200 millivolt range and it's reading 200 you're actually reading 200 mili-amps that's really nice you don't need any the user doesn't have to do any conversions or anything like that so that's a really nice design criteria which we meet easily by having by not having oddball value resistors we're going to choose them to give us a matching relationship between millivolts and milliamps so the design is pretty simple we've got three current shunt resistors and a times one hand one hundred amplifier it sounds pretty simple but now here's where the practical considerations come in now if you know your basic op amp Theory no op amp is perfect okay that's going to have an input offset voltage which is called v.o.s okay the input offset voltage can be you know in a typical general purpose op amp it might be millivolts okay but that value is going to be because we've got a times 100 gain if we've got a 1 millivolt offset on our op amp we're going to get 100 millivolts offset we're going to get 100 millivolts offset so even when we're not feeding in any current at all our output could read up to 100 millivolts now if we use in the millivolt range where or you know if found it's gonna read if we're on 1 millivolt per milliamp that means our output is going to read 100 millivolts the user will think oh we're feeding 100 milliamps feed me nothing like it so with the times 100 gain we also need basically a 100 fold reduction or so in that V in that input offset voltage so we need to find a very schmick op amp that can actually do that now we should look at how we're actually going to use this doodad with the multimeter let's take a bottom of the range three and a half digit 2000 count motivator okay we know we're using on the millivolt range okay so that's going to have a value of 200 point zero millivolts okay so it's resolution is naught point one milli volts now when we feed in no current into here we want it to measure nothing on the output so we need the offset voltage because we have a gain of 100 we knew the offset voltage to be a hundred times lower than this resolution the least significant digit so point one millivolts divided by 100 is one micro volts and that's what we need for L V for our input offset voltage of this op-amp so we need to look search for an op-amp that can do typically 1 microvolt offset so when we feed in no current into here we basically read zero on our multimeter now from my industry and knowledge I know 1 microvolt input offset voltage is incredibly low and you're going to need a very very schmick op-amp to do it and there's not many on the market that are actually going to meet that spec that we have here so but um I've done a blog on this before as it so happens and if you do your basic electronics theory you'll do different types so from op amps and one of them is called a chopper or an auto 0i PAMP and these have the characteristic of having incredibly low input offset voltages it's almost zero because they automatically there's an automatic process inside the amplifier that that nulls out the input offset voltage now even if you didn't know about chopper amplifier auto zero amplifiers what you would do in this case if you wanted 1 microvolt offset voltage you would go on the web and you would search either digi-key or mouse or one of the other component suppliers that gives you a parametric search for fir a particular op-amp so you might type in op EV and then you go into the op amp categories you go into precision op amps or whatever and then it'll have a whole list of tables of all the parameters and one of them will be the offset voltage and they'll be listed from the point 1 micro volts down all the way to you know a horrible one at you know 10 20 millivolts or something like that so what you want to do is you want to select the ones that are in the range 0.1 micro volts 2 to 1 micro volt or something like that and then you want to narrow your search down to that now you can also do the same thing on the individual manufacturers website so you might know if you're experienced industry you might know that max them do op-amps national semiconductor linear technology companies like that so you'll go to their individual websites but that can typically and do the same parametric search but that can take longer so I often use digi-key or Mouser or new I for one of the other component suppliers and search all the different brands parametrically so your parametric search spits out numbers like lmp2 o 1/5 o LM V 2 0 1 1 or or the max 42 38 42 39 so what you do is you open up the data sheets for all these different chips and you start comparing them in this case our main requirement really is that input offset voltage now one of the things you've got to take into account with practical designs is a temperature range things like input offset voltage and all sorts of other parameters vary over temperature so we don't want our product just working at room temperature we want it to work over a decent you know industrial temperature range because that's good design practice so you will search the data sheets and you will look for the typical input offset voltage over the entire temperature range not just at not just a 20 degrees Celsius at room temperature know you want to over the whole range in and you compare these different chips so I compare them and it turned out that the maximum 42-38 / 42:39 was was pretty much the best with a typical offset voltage of 2.5 microvolts over the entire temperature range it was a reasonable price it's only a couple of dollars so I decided to base my design around that and there are other things to look for as well like overload recovery with chopper amps and and and gain bandwidth product and stuff like that because but because this is a little current adapter and your average mole you made only has a couple of kilohertz bandwidth anyway then pretty much any of these are chopper amps is going to meet the bandwidth requirement so the max 42:39 looked um looked you know fairly ideal and because maximum have a nice free chips sample service that's that's an extra sweetener even though I have issues with maxim chips I put that and I decided to go with the maxim device this time so we've chosen our max 42:39 chip now if you look at the data sheet it has a vos a typical vos if not point 1 microvolt beauty at all but the main thing is is that it's offset voltage is still only 2.5 microvolts over the entire temperature range so even absolute worst-case if we feed no current into here there times 100 we're only going to get 2.5 it's only going to read plus minus 2.5 on the meter and well it's you know in practice is going to be better than that it's going to be you know 0.1 or something like that so I can live with that that's fine so let's give it a go okay so now have our basic circuit we've got three different current range resistors 10k 10 ohms and 10 milli ohms via an input switch which switches a range and we're also going to need a switch as well which is a dual gained switch to actually tap off the voltage from one of the resistors into the amplifier so we need a threat we need a double pole three-way again switch there now we've got our max 42:39 as configured as a times 100 amplifier here and that's it that should give us our basic functionality that we're after but there's one other can thing to consider as well now um do we use a standard single ended amplifier or do we implement a differential amplifier well bit because we're not I'm dealing with long lines or anything like that our current shunt resistor is going to be on the same border right next to the chip we don't really need a differential amplifier and because the whole thing is going to be battery-powered in a box there's no problems with our you know it mains reference or any other system reference like that so really we can get away with just a um single ended amplifier so all we need is a positive x 100 gain so that we configure the op-amp in a standard non inverting configuration of x 100 so what values do we choose for the non-inverting various well because we don't want to use too much current it's going to be battery-powered or make that one say 1k and this one has to be it has to have a game that is 100 so your basic formula for your non-inverting op-amp is r1 on r2 plus one so if this is 1k to give you a gain of 100 this needs to be 99 K so 99 K divided by 1 K is gain of 99 plus 1 gives you a gain of 100 Izzy and because we're only dealing with the i/o bandwidth of only several kilohertz these values can be reasonably high and you don't have to worry about straight capacitances and all that sort of stuff now there's one other thing I mentioned before about the large current ranges the amps range and the contact resistances on the components and the connectors and things like that can make it can swap the value of your low value shunt resistor well even on our milliamp range here this is our milliamp range we've got a value of 10 milli ohms now that's very that's that's actually very low um and your contact resistance of a typical switch might be in the order of Mille ohms so it's going to be it's you know it's going to swap that it's going to swap that in terms of accuracy so when you're measuring something you don't want to measure between ground here and this and the input connector because then you're actually measuring effectively a resistor in series with that that might be saying you know it might be 1 milli ohm so what 1 milli ohm so you're 1 milli ohm in series with 10 milli ohms you've got 11 million actresses just gone out the window so you don't want that so what you need is to tap it straight off the actual resistor like that so you want to actually connect it right onto the junction I'll show it like this but it's actually the junction right on the actual resistor now you can actually get special current shunt resistors that do exactly this they're actually got 4 terminals on them your regular resistor and then $0.02 terminals like that so that's what you actually want to tap off into the amplifier so it doesn't matter what the switch the value of the the resistance of your switch is now that your switch is going affect your bird the arm total burden voltage on the milliamp arrangement that's not a big deal because we're still in order more than order magnitude lower than a regular multimeter so it's just fine and of course when we're talking 10 ohms and 10k on the other range as well you know the contact resistance of the switch doesn't matter there's yet another thing you have to think about there's always something to think about in even basic designs like this in this case it's the operation of it we want to measure current um positive and negative so if the current actually is flowing in like that it's going to generate a positive voltage across the air and it's going to generate a positive output there but what if we've accidentally users hooked it up backwards or the current is um AC then it's going to be flowing in that direction in this case we're going to get out a negative voltage with respect to our output ground terminal so what we need we can't power this off air from a single supply we're going to need a dual supply so we're going to need a positive supply and a negative supply so that um if this is your import ground here okay this input terminal is effectively ground because it's battery it's just an internal ground it's not mains earth referenced or anything like that if we power this from positive and negative supplies then our op amp can actually get out can generate positive and negative output voltages and that's what we want okay so that's no big deal you need a positive and negative supply what's so hard about that well there's actually a bit of thought which needs to go into this as well and it's about the practicality of your whole design and things like that there's three different methods you can use to get a positive and negative supply from some from batteries from a little device the first one is to put two batteries in series like that in this case two doublea's or triple-a is 1.5 volts each and the center tap becomes the ground and then the either side becomes the positive one point 5 and the negative 1.5 or positive positive 3 if you put 2 in series and negative 3 if you put 2 in series etc now look that seems like an easy way to do it but the problem that is that the is that the current drain is actually going to be the current drain to be different for different cells so that means some a good design I wanted to have a low battery focus detecting this so how do you detect when your battery's low you've got two separate batteries during a drawing different currents so really there's there's practical considerations there which pretty much ruled that one out if you wanted and easy to design low voltage battery detection circuit so you think about the second one the second one is to generate is to have two batteries in series the same so you have three volts say or one single coin cell lithium or whatever and then you use a switch capacitor voltage inverter like the classic seven double six over voltage inverter and that actually just inverts your voltage from plus three to minus three so this point is ground here and it's you've got closer here and it generates minus three now the problem with that is that um these can generate noise they generate switching noise so you've actually got a filter that and take that into consideration as well and in a very low noise design like this one we're talking about you know we're talking about micro volts so really it's a bad idea to introduce a switching element into your design and it also costs money as well it's an extra bill of materials part you need a couple of capacitors and well it's just not the preferred method so number three is to get your same batteries a three volt battery and then you split it in half with two series resistors here now these need to be very high value so you might make them a hundred K or something like that so you're not draining your battery voltage and you tap off they're the same value so you tap off you're effectively tapping off 1.5 volts and then you need to buffer that with with a voltage follower op-amp so that it's low impedance and that becomes your ground now you're not actually shorting the output of there because it's not actually referenced to anything this becomes the ground the output of the op-amp becomes the ground of your circuit and then by that nature you've got plus 1.5 volts and minus 1.5 volts and bingo there's you've got a single low-cost cheap op-amp no switching noise and you've generated your plus/minus rails and then you've only got a single battery to worry about for your low voltage battery detection circuit so I choose number three with this third option you can actually buy a chip which actually just does this it's a it's a specific voltage supply split out chip and essentially integrates the the two resistors in the op amp in there for you but they're fairly exotic and they actually they're not that cheap they might be like a dollar each whereas they a you know a jellybean op amp with you know doesn't matter what the input offset voltage is in this case can be you know 10 or 20 cents so and and the resistors cost virtually nothing so we'll go with this option now because we actually have ranges based on nine milli amp micro ampere nano air by in order of a thousand that means I'm to switch up to the next range you have to go to a thousand millivolts output voltage on the device now that's within the plus minus 1.5 volt range of the device so that's perfect and a typical meter is going to be either it's going to be in a 10,000 count or lower if it's 20,000 count it just means you get an extra digit of resolution it doesn't mean that it goes to 2 volts or something like that so everything are everything seemed to fitted and we can easily power the device from one point plus minus 1.5 volts not a problem now it's a choice of two parts with the max on the max 42-38 / max 42:39 chose the max 42:39 because it's a higher bandwidth version but it's only limitation is that it needs a gain of 10 but we're going to use a gain of a hundred so there's no problem at all so we'll go with the higher bandwidth part okay this is terrific our circuits really coming together here it is we have our three range resistors we have a Joule gained switch we have our x 100 gain arm precision max 43 9 op-amp let's whack in a series resistor there just to provide some overload protection for the op-amp so the or diodes can clap let's add a hundred ohm series resistor on the output so that it ensures stability of the op-amp if it's driving a capacitive load we don't want to do something silly or if you shorted out we don't under on the op-amp so we choose 100 by 100 our nice round values no reason for them they're just nice and round they do the job and you want to try and keep values similar in a design to lower your bill of materials parts count as well even though they're the same cost you don't want to use a hundred ohms here and and 220 ohms here because well you know that's just silly but you've got to be careful with this output resistor because you remember we're driving a multimeter and it's got an input impedance so you don't want to get extra induced error there but most digital multimeters are going to be 10 mega ohms now if you put 10 mega ohms if you do the math 100 ohm in series with 10 mega ohms the error is negligible in fact you have to get down to about 100 K input impedance of the meter before the hundred ohms starts being a problem at about point one percent error so really it's you know it's a hundred ohm is a good value to protect it and not provide access error when you're driving a meter one of the big things are taking the count is the type of battery the battery voltage and also your compatibility with the main device you're using now the next 4239 can work anywhere single supply from 2.7 volts to 5.5 volts so we don't want though we're powering it from a split supply the op amp doesn't really realize that it just thinks it's a single supply so if your parent it from a 3 volt battery plus minus 1.5 volts that's fine that's within the range but the battery voltage is going to drop so in this case you don't want it to go essentially below that two point seven volt minimum operational range for the chip so if we decide to power a thing from three volts then we need a low-voltage battery detection circuit which cuts out at about two point seven volts now there's many ways to do low voltage battery detection circuits but they're all very essentially the same thing they have a voltage referee precision voltage reference Anakin once it drops below a certain level it's it switches on an output and you can turn on LED seen low voltage now I didn't want to muck around with things like that it turns out that you can get a whole slew for your digi-key or Mouse you can get a whole slew of these dedicated low voltage battery detection chips there are a tiny little tiny little three pin device and they just give you an output once they get a below a predetermined level and you can buy them in different voltages in this case I bought the TPS 38 oh nine el 30 which is actually a two point six five volt reference device so in the input voltage from the battery drops below two point six five volts which is close enough to an operational range of two point seven it well in this case it's got a negative output source which is the LED off so I'm going to have a feature on the micro current which has which the LED is on when it's above two point six five volts and the LED turns off when it's below so if the LEDs off that determines that it's you know you've got low battery whoo there you have it that's essentially our circuit and this is what essentially the final design of my micro current what was published in silicon chip and the product I actually sold so I'm but that's not the end of the story no because I design a product design is much more than just the circuit sure you can build this up on a bit of a vera border on a breadboard and you know it's it's gonna work it's gonna do the job well apart from the low noise stuff but let's not go into that um you can build it up and it's going to work but that doesn't make a good product there's lots of choices which will also back interact with this circuit so you might have to change the circuit later based on the physical construction and the price constraints and other constraints in you to actually get your practical product so let's take a look at that now although I designed the circuit first in this video that's not necessarily the order that I'm going to do it when I design a a practical project like this because functionality and the form factor can make or break your product as well as can the price price point is very important so I wanted this design to be small and low-cost like I'm I could put it in a standard large size jiffy box like this I mean in in Australia we've got these jiffy boxes and they're sold by Jay Caron and dick Smith and now tronics and and others and they're pretty much a standard box in your drill holes for your switches and your drill holes for your terminals and yeah there's things to mount your board inside there this little standoffs to mount your board and you screw them screwing your board menu wire things to the switches and well that's that's pretty amateurish and it can actually add to the cost and the complexity so I wanted something smaller and low-cost so I put a fair bit of effort into actually minimizing the parts cost at the circuit design stage as well as the assembly stage too so I pretty much decided that I didn't really want to put in one of these large sized Tiffany boxes although there's plenty of room to make it fit so I decided to get the smallest jiffy box which also happens to be the lowest cost is the ub5 jiffy box and that's what I actually turned into the product because these jiffy boxes you can get for like you know a dollar 20 or something like that in in reasonably small volume so they're very cheap and simple to design around the next decision you got to make is how you're going to do your front panel now it comes with a it comes with a lid like this which you can have our professionally silkscreen but that's an extra step and it cost money and then you've got to have it punched or drilled for your switches and your terminals and your LED and stuff like and it's just it's really annoying and it just adds to the cost and complexity of of doing a small project run for something like this in this case you might want to make 50 or 100 of them go into the effort to silkscreen and punch a panel can be a pain in the ass so I wanted to avoid that so what I did is I reverted to one of my standard techniques to make use of the PCB I've got to design a PCB anyway right so why not make it duplicate it as the front panel and mount all your components on the board so that's exactly what I did here is the microcurrent and it's everything's mounted and self-contained on that one boy there's no wiring at all and I made it fit the standoffs in here and it just so happens the standoffs sit below the surface by one point six millimeters you stick the board in there and it becomes the front panel and it and it looks very professional now side-effect of using your PCB as your front panel and your entire ear for your entire design is that well unless you want to see the circuitry on the top which you don't you've got to put it all underneath and you don't want to see the solder joints on the top so that means you're forced to use surface mount components so it was pretty obvious early on in my design decision that everything every chip I chose every part had to be a surface mount part to meet this design criteria all right so I've made the decision to put in a small UV five jiffy boxes just enough room for your two output terminals in which case these output terminals have to be these banana four millimeter banana post they have to be the standard 19 millimeter industry standard spacing and you've got to use these binding posts that double is banana Jack's so you can use standard multimeter banana plug probes or you can actually screw wires into the middle well now these are reasonably expensive so but pretty much I didn't have a choice I had to use those or I wanted to use those for the input but for the output I don't need that because it's just going to the multimeter so although it's always a good idea to are to use common parts in your designed to lower your bill of materials not at the expense of overall project cost in this case in this case the input banana binding post terminals they're expensive so I didn't want to use them on the output because really the outputs just perfectly connected your multimeter you don't want to wire it up you want to just use banana plugs that plug in like that which then goes into your multimeter simple so I used the low-cost super cheap ones they're about a quarter of the price these little tiny 4-millimeter Mountain Pose so that lowered my project cost considerably next up was wiring I didn't want a wire the damn things because that's an extra assembly step and that adds cost and complexity so what I did is I just put large pads on the on the back on the design them into the board and then I just used the existing screws to mount them on there same for the same for the banana terminals as well now in case of the banana / binding terminal say they can come loose so you've got to put two screws on there to hold it in place and and a bit of Loctite sometimes but that works and I can there's no wiring at all in this entire design it's nice next up was your battery consideration do I use double A's triple A's we've determined that I have three volts is a good voltage to actually choose for that but well what do you use double a is triple a is a lithium coin cell in this case I didn't want extra wiring and have the batteries rattling around in the box and things like that so I decided to go for a lithium coin cell battery which goes straight on the board like that I've got a surface mount one in your plug in as standard cr2032 battery so I made that choice pretty early on so when I chose the parts from my design the max 42:39 the the current draw of that chip was important same with the op amp for the voltage leader in the low voltage battery detection circuit one of my requirements for choosing those devices was low power consumption so that it chose the lowest power consumption possible when it was switched on and the other thing is when you've got a board as your front panel how do you mount an LED on there where you can't solder it on the top but then that's an extra lace assembly process because you've got to put the top instead of the bottom with all of the other parts so what I did is you got you can get these reverse LEDs which actually surface mount once I've actually instead of emitting here's the actual LED here so instead of actually emitting from the top light and coming out like that which would actually come out the bottom of my board I just drilled a hole in the board as part of the PCB fire I'll put a hole in there and emits from the bottom of the device and comes out panel and it works really well so the designs really starting to fall into place I've got ways to mount my connectors do a chip front panel which is integrated into the board it looks good get a red sole Damascus red looks sexy I've got a way to mount my LED now I've got a way to mount my battery with no wiring it's all looking quite good what's the last thing left well the switches now because I'm using the PCB as the front panel I can't use one of those traditional toggle switches because they would have stuck up about that far from the board and they look really ugly if I got like a through-hole version and I couldn't really mount it through the ball because I would have whirring on the bottom and that I didn't want to do whine it defeated the whole nice purpose of you know having everything surface mount really so I once again I did a parametric search for a double pole three throw switches our PCB mount um and a slide switches what I was looking for so I found a I found this cnk as it turns out there weren't too many actually that we're actually available so I found this see it nice seeing K sides the slide switch it was available in a vertical one and a right angle and the right angle is good because I can put the silkscreen on the board like that and just have the lever right next to it which points to the silkscreen really quite nice and it's only about 20 cents or something beautii so i decided to base my design around those now because i needed an on/off switch as well i had to buy this special cnk switch well it makes use to once again use common parts so i've decided to use the same switch for the on/off switch now because it's three pole double throw I thought aha I don't just want to go on off maybe I can have different modes so it has different modes you switch it off and then you switch it on with battery detect and the lead comes on now the LED actually draws excess current and if you're using this for a long time you don't want to drain your battery so you can add a third mode where it stays on but it switches the LED off and that's exactly what I did there was one limitation with the switch is that it only handle a limited current after several hundred milliamps and had a fairly high on resistance which will contribute to the burden voltage but it wasn't too much of a big deal so I'm especially considering the price and the suitability and it looks nice then I know I there weren't too many other choices and this was a clear winner and it was quite fortuitous that only one v8 or couple hundred milliamps max anyway so really it was a great choice now I find that often when I'm designing a product of designing the circuit or whatever they there are often these four childers circumstances that conspire to sort of move your project in a certain direction based on these certain lovely part you can get you fine they're an exact fit and it's lovely things like getting the switch I wanted firt cheaply and it just met the current range I wanted things like my maxim chip I wanted to use happen to just cut out at the two point seven volts right at the cutout voltage of a lithium coin cell battery things like that really nice for Childress design aspects which really help make a good product and a good designer will look for those things and take advantage of them and as you can see that's what I've got down here in the circuit by fortuitous use of the same rain switch up here using it for the battery detect I was able to switch between either by part turn it off or on the low voltage battery detective to say and the led to save power and that was a really neat little nice touch to make it a bit more professional product now to choose my op-amp for the voltage divider 0 volt reference once again I use the parametric search I go in a generic off op-amp I didn't care about the offset voltage or anything like that or the bandwidth because it's just a voltage follower and it doesn't matter if I was millivolts you know 10 20 50 millivolts off on the center voltage it's not really going to care that much so I'm really my main requirement for this was low power to to minimize the draw from the battery so in this case I did a parametric search by price and power consumption and pretty much the lnv three-to-one popped out and that's just a low-power version low power low voltage version of the classic gum lm35 one now the other thing you've got to consider as well is the accuracy of the device now the typical cheapo multimeter might be 0.5% basic DC volts I've seen a good meter might be our point one percent or noir point oh five percent so um but the current ranges this is the other thing why I wanted to do this project the current range is on most mid to low multimeters aren't very accurate at all they might be one percent one laughs two percent so they're pretty horrible so I wanted this thing to be pretty good you know I didn't want it to be one percent or half a percent so I decided to use point where I decide on this basic speck of roughly 0.1% because you can buy naught point one percent resistors very cheaply so the two current the micro amp micro amp and the nano amp ranges you just simply by 0.1% resistors and that sets a tolerance there the gain of the op amp these two resistors are important so you use 0.1% resistors there but the milliamp current Sharpe the ten milli ohm current shunt it's it's not impossible but it's very difficult expensive to get a naught point one percent ten milli ohm shock resistance so once again I did parametric search in Mouser and digi-key to see what I can get off the shelf and it turns out it popped out with a naught point five percent ten milli ohm current shunt resistor in the full terminal arrangement which we needed so I use that and that's the only limitation so the milli amp range is definitely point five percent ignoring that and the others are roughly point one plus point one you know 0.2 percent maximum over the temperature range so they're pretty good specs and that means that the micro current design is also going to improve the accuracy of measurements not just due to the burden voltage but because it uses your DC millivolt range instead your current range so I've decided all this sort of stuff the circuit how it's going to be constructed everything before I even prototype the circuit to see if it were cuz I pretty much knew the circuit was going to work it so simplistic the chip chips are going to meet their specs and and you're put in some bypass caps and they're not going to oscillate and it should work just fine so I jump straight into doing the actual design of the board and I pretty much got it first go I actually have to do another spin of the board because the old one actually I didn't get the fonts right and they were too small and stuff like that but dumb yeah pretty much it went first go and bingo at the end popped the fawn product and it was low-cost it was only about 17 dollars in parts all up for everything in this so it's it easily met the price I didn't really know what my price Taggart was but I know I didn't want to spend 50 bucks in parts because then when you try and sell it well you've got to sell it for a hundred plus dollars and you know that's just crazy no one's gonna pay that but um yeah even even in very small quantities the price was only you know fifteen seventeen sort of under that twenty dollar mark which was beautiful which means it can be sold for forty or fifty dollars once you build up your prototype the only thing left is to measure the performance of it now I just so happen to have a Keithley picot amp current source so I can generate our precision our currents down to the Pico ampere range so I could easily measure the nano amp range of this device now I added the Nano AF range on here cuz I thought it'd be real heavy but I knew it would be extremely sensitive so I'm a standard test for am just testing input um sensitivity to external fields is to get your mobile phone and put it near it and dial a number actually you know um actually make it transmit and put it near the airport terminals in this case it still worked pretty well because it's a nice tight our layout and it just seemed to tolerate um external electromagnetic fields pretty well so it worked it worked a treat really and I so at the time had access to a audio precision a very expensive audio precision audio analyzer so I was able to measure the the bandwidth of this thing are the noise and the total harmonic distortion and the noise and noise floor and the bandwidth over the entire range and if you take a look at my article for it you can actually see those plots so out the other end of all that design processes probably more things which go into and I've probably missed a couple of quite a few little subtle things which I put in there as well extra effort went into certain aspects and decisions and stuff like that but I couldn't do more but what popped out the other end was a really cool little product which um is reasonably recently low cost and there's nothing else on the market like it so very often a lot of my product are designs I want the circuits all right you know the circuit is almost totally irrelevant it's all about form factor meeting a price point usability things like that they're what's going to appear into the design they are what's going to they could make or break your product so really um you know you can't just slap something in in a jiffy box and whack some switches on it and stuff like that and expect it to be a winning product that's not necessarily the case so um I hope that next time you design something you'll actually put a lot of thought into a lot of little subtle aspects of it and it's electronics design is more than just circuit design now the other major thing you've got to think about during the whole design not just up front I should have probably mentioned this right at the start as well is that you've got to think about what the targets are what your target market or your target audience is now in this case it wasn't just for me otherwise I would have just you know slapped it in a box and and puts a letter set lettering on it or something like that real simple stuff um but no I wanted because I knew friends and colleagues would want one as well and I also thought hey it would be great project to share with share with the community as well so I thought it would make a real interesting a construction article in a magazine light silicon chip so I knew that I would have to my target market was the you know a small run of kits or something like that I knew people want kids already made ones you know we're only talking about you know a couple hundred we're not talking ten thousand or something like that so really those sort of choices that that market will determine a lot of you component choices which go into your design component and other aspects of the dozen design as well whether you know assemble his self whether you know test it yourself whether you going to get somebody else to do it or whatever there's lots of that factors there which can go into it as well you have to take into account from the start of the project now if you do a breakdown on how much time and effort actually goes into designing the actual circuit as opposed to the actual product your phone you probably spend most time searching for suitable parts to find to meet some design criteria every circuit design criteria and/or a visual usability functionality design criteria as well so a lot of electronics design is not just grabbing you know past from your jumping and slapping something together it's it's it's really making a lot you know sometimes hundreds of little individual choices and each one of those can involve hours sometimes hours of searching for each one of those decisions to come up with the best choice to go into your final product who so there you go that's how I designed a product from concept through to final design and as you can see there's more than just designing circuits there's all sorts of things even a simple design like this which is basically a shunt resistor and an op-amp in a box with an LED and a suite and a couple of switches there's there's more subtleties that go into designing a decent quality product like this so I hope that's been useful for you and you can make use of those techniques in your next design see ya you

Channel: EEVblog

Views: 175,148

Keywords: electronic, engineering, product, design, lecture, tutorial, how, to, multimeter, pcb, enclosure, battery, circuit

Id: g7b5YZENvjY

Length: 51min 3sec (3063 seconds)

Published: Mon Apr 05 2010