Which Transistor should I use?

Thread Starter

Robert Webb

Joined Dec 20, 2014
18
Hi
First the confession bit. The only thing I know about electronics is what I have picked up on the net. I do not have a technical background. I have been making clocks with electromagnetic coils controlled by a slot switch (the switch was designed by a friendly engineer at the Australian National University) for the last 8 years or so.
The problem is that these clocks require a pendulum and I am trying to build a clock that does not require a pendulum and I have been experimenting with a switch to drive clock mechanisms replacing the pendulum with a short shaft or rod.
I am now using a different switch which has a 12f683 chip (instead of the slot switch) to provide 30 impulses to the coil each minute. The resultant EMF attracts an armature on one end of a shaft. The other end of the shaft is used to drive a count wheel. I am experimenting with two coils. One is 16 ohms wound with .5mm wire and the other 30 ohms wound with .3mm wire. The input (regulated) can be varied between 4 to 5 volts but at the moment I am running at 5 volts.
The original transistor in the switch to power the coil is a 2n2222 but it lacked the power to drive anything very much at all. Initially I tried two 2n2222s as a Darlington pair and that certainly improved the output. I even considered a Darlington triple. Instead to tried a Mosfet J538 in another switch and it was an improvement and is driving the clock but I think it needs a bit more power because it does stop from time to time. I have tried to look at Mosfet data sheets to see if there is an alternative to the J538 that will increase the output but they might as well have been written in Swahili. So what I would like to know is if there is an alternative to the J538 to increase the output to the coil. Incidentally (or perhaps not) this switch was designed by a friend and we have been experimenting for a while. He has recently come to the conclusion that including a capacitor between the positive and negative rails may assist by my initial tests are inconclusive.
Sorry to be long winded.
Robert
 

takao21203

Joined Apr 28, 2012
3,702
150mA.

A digital MOSFET SOT23 can do the job easily its impulse only as well, and if you turn on high side or low side doesnt matter I guess.

BSS138 for instance.

Schematics? The 2n2222 can do it too, 680 Ohms base resistor, connected the emitter to ground.
 

tracecom

Joined Apr 16, 2010
3,944
Are you sure that the problem isn't the lack of base current to the 2N2222? If so, the solution might be as simple as changing the value of the resistor between the output of the 12F683 and the base of the 2N2222. How much current are you trying to switch with the 2N2222? A schematic would help.
 

Thread Starter

Robert Webb

Joined Dec 20, 2014
18
Hi Eric and Takao
Thank you for the prompt replies. I am looking for the simplest solution I can find because I am learning by doing. Perhaps I should also add that I have managed to spend 77 years on the planet in complete ignorance of matters electronic, but I really do want to get this system I am fiddling with to work so I need to persist even if I burn out a few parts along the way.

Originally I tried to read the description of the back of the Mosfet without taking it off the board it is soldered to but obviously part of the descriptive was corrupted. So I have now found the original circuit drawing which lists it as 2SJ538, sorry if I mislead you.

I have now looked at the list of SOT23s and BSS138s and am spoilt for choice. So if you can narrow it down a bit I would appreciate it.

Are you saying that the addition of the 680 resistor between the emitter and the ground will also do the job using a 2n2222?

Thanks for your time.
Robert
 

Thread Starter

Robert Webb

Joined Dec 20, 2014
18
Hi Tracecom
Thanks for your reply.

This is the schematic of the switch. I am using 5 volts and have not yet included the 4700uf cap but am experimenting with 2200uf which does not seem to do a lot. regards Robert
upload_2015-1-9_23-8-51.png
 

Attachments

tracecom

Joined Apr 16, 2010
3,944
The value of the 4700 uF capacitor is not critical; its purpose is to store power so that short interruptions of the voltage supply will not affect the circuit.

There are two things that are troublesome about the schematic. One is that there is no current limiting resistor between the output of the PIC (pin 7) and the base of the transistor. The maximum output of any one pin should be limited to 25mA (maximum.) The second thing is that I don't understand the purpose of the diode from the solenoid to ground.

The maximum collector current of the 2N2222 is 1A, but good design would limit it to .5A, or so. The collector current that is drawn is dependent upon the resistance of the solenoid coil. In doing a little math (which is not my strong suit,) I calculate that the 16 ohm coil should draw .268A, and the 30 ohm coil should draw .143A, either of which is well within the capability of the 2N2222. If you insert a 430 ohm resistor between pin 7 of the PIC and the base of the 2N2222, that should limit the current from the PIC to about 10mA, which should drive the 2N2222 into saturation.

ETA: I also note that there is not a .1uF cap from pin 1 to pin 8 of the PIC; physically, it should be as close to the PIC as practical. Absence of that cap often causes erratic operation of the PIC due to electrical noise. Adding that cap is the first step you should take.
 
Last edited:

ericgibbs

Joined Jan 29, 2010
18,766
hi Robert
If your posted circuit is correct.? and its using a transistor, say a 2N2222 then its missing the Base resistor.

If the transistor symbol represents a S2J538, the rest of the circuit is wrong as the S2J538 is a P type MOSFET which requires a negative supply.?
If so, a BSS138 would not be a direct replacement for the 2Sj538 as its an N type MOSFET.

The 2200/4700uF is only supply smoothing/reservoir cap, so you will not see a change.

Can you measure the resistance of the Solenoid coil, let us know the value, then we can suggest a transistor type.

E.
BTW: ref 77 years, I can top that with 6 more.!
 

takao21203

Joined Apr 28, 2012
3,702
I hope you can make some progress. 680 Ohms is more common eventually but 470 etc also work.

If theres no base resistor the base will hog current and prevent the PIC from working correctly. PICs dont like hard shorts on their IO. I've got some burned out bits even if it doesnt happen often, it does occur.

2200uF is more than enough, and if the PIC is at low frequency you dont need 100nF.
Indeed i've run them at 8 MHz with only a small electrolytic cap, no problem.
Its the high frequency controllers which need ceramic caps, and certain memory/logic ICs.

BSS138 is a N-channel digital MOSFET, it could directly replace the 2n2222, but that also should work almost equally. The diode across the transistor isnt needed I think.
 

Thread Starter

Robert Webb

Joined Dec 20, 2014
18
Hi All
Thank you for the input which it will take me a bit of time to absorb and I will need to read them all a few times so I don't miss anything.

I have two circuits running, one on the latest clock itself and the coil on that is 16 ohms.

The other is a test bed I have set up to play around and experiment with to see what happens and the resistance on that coil is 32 ohms.

I used the 2SJ538 because that is the transistor I use in the slot switches I build for my pendulum clocks and I had a spare one. It seems to work but, for want of a better adjective, is not very smooth.

Eric I am actually a woodturner and I never had any intention of getting involved in electronics. I took up clock making because I had run out of stuff to make that interested me and stumbled into electronics because it is the ideal way to drive a clock. I know some blokes my age and younger who made a living with electronics but have just switched off and have no interest any longer and the fact that you are still switched on is inspirational. For me this clock making lark gets me out of bed in the mornings. I call it my obsession that keeps me alive!

This is a wonderful forum for people like me I really appreciate the fact that you have all taken the time to respond.
regards
Robert
 

MaxHeadRoom

Joined Jul 18, 2013
28,618
. For me this clock making lark gets me out of bed in the mornings. I call it my obsession that keeps me alive!
Robert
I just saw the repeat of the documentary about the Englishman John Harrison ( which I am sure you are familiar with), incredible skill, too bad the Royal Society gave him short shrift.
Max.
 

tracecom

Joined Apr 16, 2010
3,944
If you decide to try a MOSFET, be sure that it is logic-level, i.e., that its gate will react to voltages of 5 volts or less.
 

Thread Starter

Robert Webb

Joined Dec 20, 2014
18
Hi Max
It is certainly true that Harrison suffered from the fact that he belonged to the wrong class - a mere carpenter. But his name became a household word unlike those who opposed him. I started with wooden geared clocks and as a result became friendly with a remarkably skilful fellow used to make replica wooden geared black forest cuckoo clocks (he actually won a gold medal for one of his clocks in the US) who was something of an authority on Harrison. His comment was that Harrison did suffer at the hands of the Royal Society, but he became a wealthy man whose priceless reward was to have his son follow in his footsteps and take his work to the next level. H1 is regarded by horologists throughout history as the time machine against which all others are measured.
regards
Robert
 

Thread Starter

Robert Webb

Joined Dec 20, 2014
18
Hi All

I will continue to use the 16 ohm coil wound with .5mm wire because I have quite a number of clocks using that size coil and experience tells me it has sufficient EMF for the clocks I build.

I hope I have understood what you have said and based on the combined input I have received I plan to do the following:

1. Add a 1uf Cap between pins 1 and 8
2. Add a 680 ohm base resistor between the output of pin 7 and the base of the transistor
3. Replace the existing transistors with a BSS138

What I am still a bit confused about:
I don't know what driving the 2n2222 to saturation means. I also don't know whether the BSS138 will be superior to the 2n2222.

Have I missed anything?

Thanks for your help
Robert
 

ronv

Joined Nov 12, 2008
3,770
The 2N2222 will work better in this case, but make the base resistor 120 ohms.
The BSS138 is a FET, but has a resistance of 3 ohms so a lot will be lost to the 16 ohm coil.
 

tracecom

Joined Apr 16, 2010
3,944
Hi All

I will continue to use the 16 ohm coil wound with .5mm wire because I have quite a number of clocks using that size coil and experience tells me it has sufficient EMF for the clocks I build.

I hope I have understood what you have said and based on the combined input I have received I plan to do the following:

1. Add a 1uf Cap between pins 1 and 8
2. Add a 680 ohm base resistor between the output of pin 7 and the base of the transistor
3. Replace the existing transistors with a BSS138

What I am still a bit confused about:
I don't know what driving the 2n2222 to saturation means. I also don't know whether the BSS138 will be superior to the 2n2222.

Have I missed anything?

Thanks for your help
Robert
1. No, the cap should be .1uF, not 1uF, and it should be physically close to the PIC.

Bipolar Junction Transistors (BJTs) such as the 2N2222 can function as amplifers or as switches. To function as a switch, it should be fully turned on, i.e., driven to saturation.
 
Top