PDA

View Full Version : help on circuit



roboticmhd
2015-Apr-14, 07:19 PM
Basically a control circuit. There is a resistor whose resistance can be in two states: 240 Kohm and 24 Kohm. When it's resistance is 24Kohm I want a relay to be activated (relay properties are: about 65 ohm, current to activate it is 0.1A, so it consumes about 0.65watts ), when the resistace is high (240Kohm) I don't want the relay to be activated. I am not allowed to use a power source voltage above 25V, the total power consumption for such a circuit should not be above 10watts, I am not allowed to use amplifiers or microcontrollers or other semiconductor devices, just basic circuit elements like resistors, inductor... Can I build such a circuit ? I tried several options for myself but they didn't seem to work (like wheatstone bridge or simple serial conntion of the resistance with the relay) (note: a hypothetical ideal dc-to-dc transformer is not recommended but can be used also)

ShinAce
2015-Apr-14, 07:34 PM
If you can use transistors, that should work.

Let's call your resistor a 'state resistor'. Place that resistor between the base of a transistor and ground. Place a large resistor between the base and the 12 volt line (you shouldn't need 25 volts). This resistor will need to be something like 1 megaohm. Done!

When the state resistor is at 240k, then the voltage at the base is 2.3 volts and the transistor will saturate.

When the state resistor is at 24k, then the voltage at the base is 0.3 volts, which should be below the cutoff of the transistor. Perhaps you'll need a 2.2 megaohm instead of 1 megaohm.

Now you need to wire the relay coil, emitter and collector of the transistor. I can't remember exactly, but I'm thinking the relay coil should go between the emitter and ground. The collector would be connected to the 12V line. This is for an NPN transistor.


Otherwise, you'll need to get fancy with Zener diodes, but those are also semiconductors.

roboticmhd
2015-Apr-14, 08:13 PM
transistor is an amplifier. but thanks anyways.

danscope
2015-Apr-14, 08:34 PM
Perhaps a zener diode whose properties are such that when the voltage exceeds say 6 volts, the current is shunted to ground
inhibiting the relay, and when it is low (lower than required for purpose) the relay is closed. Just a thought.

pzkpfw
2015-Apr-14, 08:48 PM
Isn't a diode also a semiconductor device?

danscope
2015-Apr-15, 12:29 AM
Sorry, didn't read the " or other semiconductor devices " .

ShinAce
2015-Apr-15, 01:07 AM
Then no, I don't see how it can be done. If you could use a relay which has a coil resistance in the 20k range, you'd be set.

roboticmhd
2015-Apr-15, 05:58 AM
Actually after I posted this I found one method but I don't like it very much, although it satisfies all criteria. The thing that drives me crazy is, why isn't there a simple circuit for such a control problem, after all there is 10 times decrease in state resistor's resistance, from 240k to 24k, it is a significant change and yet I can't drive a simple relay with it, while I can drive an LED with it (note: notice that although LED is a semiconductor device, it is not a part of a control circuit, it is the driven thing so no violation of criteria, besides driving LED is another problem but for sake of the problem I mentioned it, I tried simple serial connection with it and it worked)

cjameshuff
2015-Apr-15, 01:53 PM
There are many simple circuits, they just use active devices based on semiconductors, which is really the right way to approach the problem. And you can't use semiconductors, but you can use a "hypothetical DC-DC transformer"? What precisely would this hypothetical DC transformer do? Does it trade voltage for current like a real AC transformer, or is it just a proxy for an amplifier? Are vacuum tubes ruled out? You mention a DC transformer, so presumably the power supply is DC. Is this a hard limitation?

Is this just a resistance measurement problem? You can connect the resistance anywhere you want, without concern about other connections?

One obvious possibility is to use a much more sensitive relay that can trigger with 1 mA of coil current, but not with 0.1 mA. It sounds like that may be ruled out as well, though you don't specifically state it. You state one specific relay, is that the only relay you can use? And only one of them? I can think of some workarounds using latching and oscillating relays to measure how quickly a capacitor is charging, but that would require several relays.

The only thing I can think of with the restrictions you seem to have is to put a resistance in parallel with the test resistance, both in series with the relay. Adjust the parallel resistance so that when the test resistance is 240 k, the current through the relay is just barely less than the triggering threshold. This is not a good way to do it (you'll get about 0.7% difference in the parallel resistance, and it'd be terribly unreliable in the real world), but it sounds like the good ways are all ruled out.

The simple way is to use a transistor to switch the relay. The good way is to use a comparator and another couple resistors to switch that transistor at a precise threshold. If the test resistance varies continuously, a better way would include feedback from the comparator output to add some hysteresis, to prevent chatter when the resistance is near the threshold. This involves active components, but forbidding them is requirement that is very hard to justify.

Grey
2015-Apr-15, 02:29 PM
The only thing I can think of with the restrictions you seem to have is to put a resistance in parallel with the test resistance, both in series with the relay. Adjust the parallel resistance so that when the test resistance is 240 k, the current through the relay is just barely less than the triggering threshold. This is not a good way to do it (you'll get about 0.7% difference in the parallel resistance, and it'd be terribly unreliable in the real world), but it sounds like the good ways are all ruled out.I thought about this last night, and came up with the same method (and agree that it's a terrible method, since it would be highly susceptible to slight fluctuations in the source voltage). I'm curious, roboticmhd, whether the method you came up with was something similar to this, or did you do something else?

profloater
2015-Apr-15, 03:40 PM
It makes no sense to specify no semiconductors and then allow a DC to DC converter unless it is a ward Leonard set (Motor and dynamo) The high resistance and limited voltage mean the resistor cannot power the relay. There are transistorised relayswhich would work but obviously contain semiconductors. There is a basic inconsistency in the question. It is easy to do this with a single transistor.

profloater
2015-Apr-15, 03:57 PM
Apologies the use of an adjustable parallel resistor might work but I think the hysteresis of the relay will mean it will not switch off so it is a one shot solution. A typical relay may go down to 20% current before unlatching because of the magnetic circuit improving as it closes. Similarly you could charge a capacitor slowly but you need say a Zener to allow the voltage to reach trigger point. Or a tiny spark gap! A 24 volt spark gap is very small!

molesworth
2015-Apr-15, 04:01 PM
Going by the "not allowed to" "must not" etc. it sounds a bit like a homework or assignment question, so that may be why it's so restrictive on the use of what would normally be practical solutions.

Like cjameshuff I first thought of using something like a reed relay with a much higher resistance coil (2K would be a good choice), or if that's not allowed, then maybe make one using "an inductor" and some other bits & pieces :-)

The other approach, making it a voltage divider with an extra resistor in parallel with the variable one, could be done, but it would need some very high tolerance resistors (and a very stable power supply). A back of the envelope calculation, if using the 25V supply, would need a second resistor between 185.15 and 186.43 ohms. I don't know where you'd get such a thing, but it's the only way I can think of to make it work using only inactive components.

Hopefully roboticmhd will be along soon to give us the "official answer" :-)

cjameshuff
2015-Apr-15, 04:33 PM
Yeah, you wouldn't be able to use a snap-action relay because it wouldn't release with the tiny reduction in coil current. A relay could be built without hysteresis, but it'd be an even more specialized component than a high-sensitivity relay. Just charging a capacitor is no good, you need to trigger the relay based on the charging rate...apart from being a semiconductor, a zener isn't useful for this. An avalanche diode might work, but again, semiconductor. There's that DC-DC converter with completely unspecified limits, though, so going by the stated requirements, you could use a spark gap operating at some kilovolts.

I came up with a way to do it with three relays, repeatedly charging a capacitor and checking whether it charged enough to pulse a relay, and then using the filtered output to drive a third relay, but it's again not clear if this would be allowed.

profloater
2015-Apr-15, 04:48 PM
allowed current is 1 mA so you could wind a special reed relay to operate at 1mA and hysteresis is less than ordinary relay so that may be an answer. Maybe this question is related to environments that kill semiconductors. How about a thermionic valve? There are low voltage triodes.

roboticmhd
2015-Apr-15, 05:38 PM
Thank you guys so much for replies, I wasn't expecting so much interest in this topic. I'll now write replies to some questions and issues below:
- hypothetical dc-to-dc transformer would trade voltage for current or vice versa, just like an ac-to-ac transformer.
- vacuum tubes are ruled out
- the power supply is 25V max dc
- if I were to actually build a circuit using hypothetical dc-to-dc transformer, I would actually use ac-to-ac and ac power. I said dc-to-dc just to simplify things. anyways using transformers is not a very good option, on which my idea is based on also. I should have never mentioned about transformers.
- are there any electromechanical relays that work with 1 mA ? I assume it would be a slow one, because it takes time for magnetic field to build up.
- the change of resistance from 240k to 24k should activate a dc machine that works with nominal voltage and currents, I said relay just as an example. by nominal voltage and current I mean about 1-25 V and 0.1-2.5 amps. (I am sorry if I confused you with the relay, again, I should nt have mentioned about it.)
- The idea of parallel resistance with the relay is not a good one, as you mentioned. I thought of it too previously.
- the test resistance does not vary continuously. two states only, 240k and 24k.
- are there any 25V spark gaps ?
- I totally forgot about capacitors, but right now don't know a solution using some.
- It is not a homework, who gives such homeworks :-) I have a phobia against transistors, amplifiers and any other semiconductor devices. I always try to find solutins not using them. especially when changes are not tiny but rather big, there's 10 fold change in resistance! This problem is completely my setup.
- I have a reed switch, it will require huge number of windings around it to bring resistance to k range. do you think it is worth trying ?
- I will think on oscillating relay option
- and relay-capacitor options as well

Hornblower
2015-Apr-15, 05:56 PM
I would suggest that you try to find ways to conquer that phobia about semiconductor devices. You are rejecting them for reasons th

Hornblower
2015-Apr-15, 05:58 PM
I would suggest that you try to find ways to conquer that phobia about semiconductor devices. You are rejecting them for reasons th

Sorry, I clicked on the Post Reply icon rather than the Cancel icon. I don't want to come across looking as if I am belittling your reasons for your criteria. My apology.

cjameshuff
2015-Apr-15, 06:29 PM
It may be a wide range of resistance, but combined with your maximum voltage it means you're limited to currents of at most 1 mA through the test resistance. You can achieve a wide voltage range by putting it in series with a similar resistance, but the current will be even smaller. The strength of an electromagnet is proportional to the number of turns and to the current, and very large numbers of turns are difficult to achieve, so electromechanical devices tend to inherently be high-current devices.

Relays with mA-range coil currents do exist, though they're expensive. Here's one with a 1.17 mA coil: http://www.digikey.com/product-search/en?pv671=71

With AC, you could use a magnetic amplifier to get a signal capable of driving a relay. If you're opposed to even this form of amplification, why do you allow relays?

ShinAce
2015-Apr-15, 06:53 PM
The current the coil consumes is not important. What you need is a coil resistance of at least 10k ohm and working voltage less than about 12V. Even higher resistance coils let you get away with lower power supply voltages. Then you just use the voltage divider already described so that the voltage across the coil is too low when the resistor is 240k, and just right when the resistor is 24k.

But you still haven't said if you must use your original relay(which won't work), or can use a different relay(which will work).

cjameshuff
2015-Apr-15, 08:38 PM
The current the coil consumes is not important.

The current the coil requires is quite important. Electromagnets function in the first place due to the current flowing through them. The maximum current that can be driven through the test resistance with 25 V is lower than the coil current for almost all relays. The additional coil resistance will only reduce the current further.



What you need is a coil resistance of at least 10k ohm and working voltage less than about 12V. Even higher resistance coils let you get away with lower power supply voltages. Then you just use the voltage divider already described so that the voltage across the coil is too low when the resistor is 240k, and just right when the resistor is 24k.

Higher resistance coils will drop a higher fraction of the power supply voltage in a divider, but they also tend to require higher voltages to begin with. The version of the relay I linked with the 1.17 mA coil is the 24 V version, with a 20500 ohm coil. The pull-in voltage is 18 V, and the most it'd see in series with a 24 kohm test resistance and a 25 V supply is 11.5 V.

roboticmhd
2015-Apr-16, 07:20 AM
Like I said before I mentioned relay just as an example, but I should not have done so.

The problem basically is, a test resistance is either 240k or 24k. whenever it is 24k a dc device must be activated, whose resistance is much lower compared to 24k and requires 0.1 A current. for example, this device has resistance of 50 ohm and requires 0.1 A. A control circuit must be built for such a purpose but using criteria I mentioned previously.

The allowed components of the control circuit are electromechanical components, resistors, inductors, capacitors but not semiconductors based components or vacuums. also max power supply voltage is 25V and max power cunsumption 10 watts.

ANyways thank you very much for your replies guys.

TrAI
2015-Apr-16, 10:12 AM
I don't see why you wouldn't want to use semiconductors, but before the invention of electronic amplifier components, engineers was faced with the issue of how to detect low currents, for example in long telegraph lines and for scientific instruments, and I believe they would often use a variant on the galvanometer.

Even today analog meters are essentially moving coil galvanometers, and they can be quite sensitive. Perhaps is would be possible to adapt such a meter to act as a relay.

cjameshuff
2015-Apr-16, 12:42 PM
Alright. Attach a mirror to a galvanometer in series with the test resistance. Aim a well-collimated incandescent light source at the mirror. Attach one end of a lead and harness to a switch, and the other end to a cat...wait, I see a flaw in this plan.

Use that relay I linked before: 20500 ohm coil, 18 V pull-in voltage, 3.5 V drop-out voltage. Just put it in series with the test resistance. Use that non-semiconductor DC-DC converter (a motor-generator? Relay-based chopper circuit?) to bump the voltage up to 40 V. When the resistance is low, the voltage across the relay will be 18.4 V, when it is high, the voltage across the relay will be 3.1 V. You may have to adjust the applied voltage due to variations in relay coil resistance, it isn't guaranteed to be exactly 20500 ohms.

Try not to think about the fact that your computer contains tens of billions of transistors.

Grey
2015-Apr-16, 01:53 PM
I'd kind of assumed that with the specific constraints that this was a class assignment as well, and just before roboticmhd commented that it was just that he didn't like semiconductors, I was going to suggest that maybe part of the point of the assignment was just to demonstrate how tricky it might be to do something like this without semiconductors, since it's so easy to do with a transistor.

ShinAce
2015-Apr-16, 04:45 PM
Still sounds like a homework assignment.

Why the need to supply 0.1A logic?

TrAI
2015-Apr-16, 10:48 PM
Alright. Attach a mirror to a galvanometer in series with the test resistance. Aim a well-collimated incandescent light source at the mirror. Attach one end of a lead and harness to a switch, and the other end to a cat...wait, I see a flaw in this plan.

As a current in the hundreds of microamps would be quite sufficient to cause visible deflection of a galvanometer movement, it is a bit excessive to use a mirror galvanometer, and detecting the light to switch a load would be kind of difficult with the restrictions.

Anyway, I was talking about a galvanometer relay, it would be similar to a common relay in that it switches contacts in some way, but it uses a galvanometer mechanism to actuate the switch rather than the electromagnet arrangement commonly used in relays.

I suppose it may be a bit tricky to make a contact that was light enough not to interfere with the meter movement, and still make a good contact. Perhaps it would be possible to mount a small magnet to the meter in some way, and have a reed switch as an output.

As a side note, if you wanted to go with an optical detection, it wouldn't be necessary to use a mirror, you could probably mount a light piece of foil or paper on the meter to act as a shutter for an opto-interruptor arrangement, the problem is that while you could use an incandescent lamp as a light source, I am not sure how you would detect it, even LDRs are actually a sort of semiconductor.

As for using a cat component, while able to detect the light and actuate movement that could potentially switch a circuit, they employ a wide range of chemical and ionic systems that would probably count as amplification, and so would probably not be allowed, also, there is perhaps an issue with the predictability and repeatability of a circuit employing such components.

danscope
2015-Apr-17, 03:41 AM
This problem smacks of a post apocalypse situation where one is trying to make circuits without the bennefit of access to
semiconductors . Is this an exercise for a sci-fi novel ?

roboticmhd
2015-Apr-17, 10:05 AM
Guys, not to mess things I will ask my another question here, which is not related to my previous question. We have a project at university, it is a vehicle which is driven by shape memory alloys. We have to make it controllable by a computer program. The program will be very simple just three buttons: go straight, turn right, turn left. Now how can activate a relay from signal coming from computer. How can computer communicate with outside world. On my laptop I have only vga connector port, usb port and a port where network cable is connected. I can use transistors for this, professors don't care. Only I need cheapest way.

I have to be able to cotrol three relay from my computer.

Extravoice
2015-Apr-17, 12:19 PM
This problem smacks of a post apocalypse situation where one is trying to make circuits without the bennefit of access to
semiconductors.

In that case, I'd recommend the use of stone knives and bearskins (http://www.imdb.com/title/tt0708455/quotes) :)

I've followed this thread with interest, but really don't have anything constructive to add at this point, other than to say that transistors are your friends.

danscope
2015-Apr-17, 03:09 PM
Yep. It reminded me of the Blind Hermit in "Young Frankenstein" whe he says " No....no... fire is good. Fire is your friend."
Only...with transistors. :)

TrAI
2015-Apr-17, 03:15 PM
Guys, not to mess things I will ask my another question here, which is not related to my previous question. We have a project at university, it is a vehicle which is driven by shape memory alloys. We have to make it controllable by a computer program. The program will be very simple just three buttons: go straight, turn right, turn left. Now how can activate a relay from signal coming from computer. How can computer communicate with outside world. On my laptop I have only vga connector port, usb port and a port where network cable is connected. I can use transistors for this, professors don't care. Only I need cheapest way.

I have to be able to cotrol three relay from my computer.

Hmmm, perhaps you should look into getting a USB relay board, either as a complete unit or a kit, they are not the cheapest option perhaps, but you wouldn't have to develop the hardware or the basic routines/libraries so you could concentrate on making the control software and the vehicle itself.

ShinAce
2015-Apr-17, 03:29 PM
How about using a transistor to amplify the computer's logic?

edit: You could build your own current amplifier with a transistor, or even use a packaged one, called a 'driver gate':
http://en.wikipedia.org/wiki/Gate_driver

74LS06 and 74LS07 are two of them. One inverts the logic, the other doesn't.


Why use a laptop when all you need is 'straight', 'left', 'right', and 'stop'? The controller for the car is so simple, there's no point in using a computer. I would use switches.

I get a weird feeling about this. You're in university. You're working on a simple car. But for some reason, you want to use a laptop to deliver 4(or 3) simple commands to motors/actuators. If you are the guy in charge of the electronics, you should be doing more homework than this. If you're in charge of mechanical, you should let the electronics people take care of that. I get the impression that you've not studied semiconductors and digital logic, yet you want to decypher how computer input/output works. A basic course in digital logic would get you 99% of the way there.

What part are you involved in? Do you not trust your teammates? How many people on the team?

roboticmhd
2015-Apr-17, 06:35 PM
How about using a transistor to amplify the computer's logic?

edit: You could build your own current amplifier with a transistor, or even use a packaged one, called a 'driver gate':
http://en.wikipedia.org/wiki/Gate_driver

74LS06 and 74LS07 are two of them. One inverts the logic, the other doesn't.


Why use a laptop when all you need is 'straight', 'left', 'right', and 'stop'? The controller for the car is so simple, there's no point in using a computer. I would use switches.

I get a weird feeling about this. You're in university. You're working on a simple car. But for some reason, you want to use a laptop to deliver 4(or 3) simple commands to motors/actuators. If you are the guy in charge of the electronics, you should be doing more homework than this. If you're in charge of mechanical, you should let the electronics people take care of that. I get the impression that you've not studied semiconductors and digital logic, yet you want to decypher how computer input/output works. A basic course in digital logic would get you 99% of the way there.

What part are you involved in? Do you not trust your teammates? How many people on the team?

I study in mechanical engineering department, 4th year, my last year project this is and it is such a stupid project. Damn professor wants the vehicle to be controllable from a computer. We are 7 in the team including me, all from same department. Somebody had to deal with the program and input/output.

What do you mean by amplifying computer logic? where is this computer logic? is it a port or what ? Why isn't there a simple port for whose pins I could set as high or low and thats it, like it is done in pic microcontrollers? I don't understand why such a simple thing requires special boards and stuff.

cjameshuff
2015-Apr-17, 06:45 PM
I study in mechanical engineering department, 4th year, my last year project this is and it is such a stupid project. Damn professor wants the vehicle to be controllable from a computer. We are 7 in the team including me, all from same department. Somebody had to deal with the program and input/output.

What do you mean by amplifying computer logic? where is this computer logic? is it a port or what ? Why isn't there a simple port for whose pins I could set as high or low and thats it, like it is done in pic microcontrollers? I don't understand why such a simple thing requires special boards and stuff.

http://en.wikipedia.org/wiki/Parallel_port
http://www.epanorama.net/circuits/parallel_output.html

roboticmhd
2015-Apr-17, 07:32 PM
http://en.wikipedia.org/wiki/Parallel_port
http://www.epanorama.net/circuits/parallel_output.html

But I don't have a paraller port, and many computers nowadays also don't have parallel port. In wikipedia link you suggested it also says that "Today, the parallel port interface is seeing decreasing use because of the rise of Universal Serial Bus (USB) devices, along with network printing using Ethernet.".

ShinAce
2015-Apr-17, 07:50 PM
I've never worked with USB communicating microcontrollers. You'll need that to translate USB into (go, left, and right) which can activate relays. You'll also have to write the computer program that displays on the laptop's screen.

When I was working with microcontrollers, the programming was low-level. So you had to tell the controller literally everything. Nowadays, the programming for Arduino's is fairly high level and simplified.

Try 'USB microcontrollers' in google. Better yet, try a microcontroller(PIC, Arduino) forum.

Even better, talk to the prof personally.

cjameshuff
2015-Apr-17, 08:43 PM
But I don't have a paraller port, and many computers nowadays also don't have parallel port. In wikipedia link you suggested it also says that "Today, the parallel port interface is seeing decreasing use because of the rise of Universal Serial Bus (USB) devices, along with network printing using Ethernet.".

There are USB parallel port dongles now.
You may be better off with a breakout board, though. That way, you're sure what chip you have, and can often access some extra features of the interface chip. For example: https://www.sparkfun.com/products/7841

roboticmhd
2015-Apr-18, 11:54 AM
I have a usb to rs-232 cable. Can that be used somehow? and how to send signals to rs-232 pins then ? on the cable written: "USB 232 converter" and one side has a usual usb the other one has something bigger with 9 male pins, one of which is shorter than others.

cjameshuff
2015-Apr-18, 01:50 PM
I have a usb to rs-232 cable. Can that be used somehow? and how to send signals to rs-232 pins then ? on the cable written: "USB 232 converter" and one side has a usual usb the other one has something bigger with 9 male pins, one of which is shorter than others.

If it has a FTDI chip, it may be able to be operated in "bit bang" mode, allowing you to control individual signals. It might use some other chip, however.
http://hackaday.com/2009/09/22/introduction-to-ftdi-bitbang-mode/

roboticmhd
2015-May-02, 09:28 AM
Hello Guys, Is it possible to electroplate carbon on copper ? Can it be done easily? And what would the electrical resistivity of electroplated carbon be ?

Jens
2015-May-02, 01:16 PM
I think the answer is no, because only metals can be electroplated, and carbon is not a metal.

profloater
2015-May-02, 11:49 PM
You can electroplate copper onto carbon so in effect you can create what you want. Resistive ty of carbon is low and drops with temperature but form of carbon obviously matters. Diamond is insulating but conducts heat really well, sintered carbon is a conductor but not so good as a heat conductor.
Can you refine your question?

roboticmhd
2015-May-03, 05:43 PM
You can electroplate copper onto carbon so in effect you can create what you want. Resistive ty of carbon is low and drops with temperature but form of carbon obviously matters. Diamond is insulating but conducts heat really well, sintered carbon is a conductor but not so good as a heat conductor.
Can you refine your question?

I was asking if it is possible to electroplate carbon on copper, not copper on carbon :)