PWM and are all rheostats created equal?

ChemMan · 03-25-2003, 10:59 PM

I'm building a PWM circuit based on Brian256's procooling article here. So far so good, I've got a prototype set up on a universal breadboard to test everything and adjust it to my liking. I have run into one problem though. The rheostats I am using are 5K Ohm linear radioshack specials

. I assumed that the resistance would increase relatively steadily as the shaft is turned, but these rheostats go from 10 ohms to 5K Ohms in about 1/2 a turn of the shaft. Once the resistance reaches 5K it stays at that value for last 1/2 turn. So I am not able to throttle the fan using the full turn of the rheostat shaft, only about 3/8's of a turn. I have checked all three rheostats that I bought and they all have the same crappy characteristic. So my two questions are: Are all rheostats like this, it is a limitation of the design? And if not, could someone point me to someplace that sells higher quality ones? Thanks, hope I made sense

.

mrmoustache · 03-26-2003, 02:50 AM

i don't really know much about electronics but i'm pretty sure they make rheostats that turn multiple times.

my guess is you can get them at most online electronics places like jameco or bgmirco if you know what to look for but i don't know if they have a special name or anything.

TerraMex · 03-26-2003, 04:03 AM

I've made a couple of PWM circuits, didnt have that problem, i think you just bought crappy rheostats. Althou almost all rheostats work with a single turn, well not even "a" turn, about 5/6 of one, and a decent one is pretty linear.

usually a multiturn one is different :

http://www.jprelec.co.uk/docs/order_...NONE&TEMP=NONE

Or with a similar design.

And another thing, the circuit on that page (Brian's circuit) uses a 10k Ohm variable resistance (R3), not a 5k, maybe thats part of the problem.

bigben2k · 03-26-2003, 08:41 AM

FWIW

You can add a 5k resistor before the pot. This will give you the total 10K you need, and allow the pot to vary the resistance between full and half, over the turn radius you have.

Yes, a lot of pots are like this.

murray13 · 03-26-2003, 09:05 AM

Sounds to me like Radio shlock screwed up and you got audio taper (log) pots not linear.

Brians256 · 03-26-2003, 12:48 PM

I believe you got some crappy pots. Sorry dude! What you want is multi-turn 10K rheostats. I prefer rheostats to pots for their size: pots tend to be small and thus a pain to adjust unless you love your miniature jewelers screwdrivers.

Also, some of what you see could be caused by dirty contacts. However that is probably not the case in your situation. If you scavenged an older rheostat, it could have dirt+grease buildup causing steady resistance that drops suddenly when the contact passes over the grease and then gets good (low resistance) contact.

ChemMan · 03-26-2003, 04:05 PM

Thanks for the replies. If you look a little further down in Brian256's article he mentions adding a second voltage divider circuit so that the potential can go from 8 to 4 volts using a 5k rheostat. I want to have the rheostats mounted in an unused face plate so that I can adjust the fan speed from the front of my case. My first thought was the these were audio taper not linear, but after repeated measuring I'm pretty sure they just suck. Not to mention what are the chances of buying 3 linear taper rheostats and having all three actually be audio taper.

I took the bottom cover off of one of the rheostats and though I can't describe it they look poorly made to me. I'll try finding some better ones elsewhere.

Brian256: Is it possible for 1 timer to drive 3 independant op-amps/transistor/mosfet combos? Or will the signal be too weak after splitting it 3 ways.

Thanks again.

Brians256 · 03-26-2003, 05:10 PM

I believe it will work, because the impedance (another name for the resistance to electricity) of the op-amp inputs is fairly high. High resistance means low flow, meaning that the op-amps won't suck down the voltage level seen on the high-side of the capacitor.

So, having three op-amps absorbing some the output of the triangle wave *should* work OK.

ChemMan · 03-27-2003, 04:12 PM

I can always use another transistor to amplify the signal if it doesn't work.

Brians256 · 03-28-2003, 02:59 PM

Exactly. You could use a transistor as a buffer to reduce the impedance seen by the sawtooth waveform. Realistically, though, you are probably not going to see any problems. Normal low-current opamps are high impedance on their inputs. I wouldn't want to connect an op-amp to a resonating crystal circuit, but this should work just fine.

ChemMan · 03-30-2003, 05:21 PM

Quote:

I wouldn't want to connect an op-amp to a resonating crystal circuit, but this should work just fine.

Huh? By resonating crystal circuit I believe you mean the transistor. I'm interested in what you are saying here, could you elaborate for me please.

I found 10 turn wirewound rheostat for $8.10 at Mouser Electronics here. This is the cheapest I could find them. If anyone can suggest something better or less expensive(assuming its not crap), fell free to mention it. Looks like I'll have to incorporate a LED into my pwm too.

ChemMan · 03-30-2003, 09:32 PM

And one more thing before I forget.

In Brian256's pwm circuit, if I wanted to add a capacitor to the circuit so that the fans are given full power at start-up(I believe its called power-on-reset), what would be the best place. I was thinking about placing it between Vcc and the MOSFET. I haven't had a chance to try it yet, so I'll ask now.

Thanks for all the input.

Brians256 · 03-31-2003, 02:23 PM

I didn't mean to send you off on a red herring with the comment about the crystal.

Basically, I was referencing the "chip" used by computers (and many other electronic devices) to keep time. A crystal resonates at a given frequency based upon how it is cut, but it depends upon a precise pair of capacitors on each side to keep the electronic ticks happening. So, adding anything to a crystal circuit can make it not work. Not that it matters to what you are doing!

Now, as for adding a capacitor to make it start up... that really shouldn't be necessary unless you are really downrating the fan speeds (i.e. less than 30% duty cycle). My circuit worked well down to about 30% speed on my fans, which was far below the (to my ears) silence threshold.

Good luck!

03-25-2003, 10:59 PM	#1
ChemMan Cooling Neophyte Join Date: Jan 2003 Location: New Jersey, USA Posts: 15	PWM and are all rheostats created equal? I'm building a PWM circuit based on Brian256's procooling article here. So far so good, I've got a prototype set up on a universal breadboard to test everything and adjust it to my liking. I have run into one problem though. The rheostats I am using are 5K Ohm linear radioshack specials . I assumed that the resistance would increase relatively steadily as the shaft is turned, but these rheostats go from 10 ohms to 5K Ohms in about 1/2 a turn of the shaft. Once the resistance reaches 5K it stays at that value for last 1/2 turn. So I am not able to throttle the fan using the full turn of the rheostat shaft, only about 3/8's of a turn. I have checked all three rheostats that I bought and they all have the same crappy characteristic. So my two questions are: Are all rheostats like this, it is a limitation of the design? And if not, could someone point me to someplace that sells higher quality ones? Thanks, hope I made sense . __________________ AthlonXP 1600+ Soyo KT400 dragon ultra 512 megs Samsung DDR333 PowerColor 9700 Antec SOHO case other junk

03-26-2003, 04:03 AM	#3
TerraMex Cooling Savant Join Date: Nov 2002 Location: Portugal, Europe Posts: 870	I've made a couple of PWM circuits, didnt have that problem, i think you just bought crappy rheostats. Althou almost all rheostats work with a single turn, well not even "a" turn, about 5/6 of one, and a decent one is pretty linear. usually a multiturn one is different : http://www.jprelec.co.uk/docs/order_...NONE&TEMP=NONE Or with a similar design. And another thing, the circuit on that page (Brian's circuit) uses a 10k Ohm variable resistance (R3), not a 5k, maybe thats part of the problem. __________________ "we need more cowbell."

03-26-2003, 08:41 AM	#4
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	FWIW You can add a 5k resistor before the pot. This will give you the total 10K you need, and allow the pot to vary the resistance between full and half, over the turn radius you have. Yes, a lot of pots are like this. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-26-2003, 04:05 PM	#7
ChemMan Cooling Neophyte Join Date: Jan 2003 Location: New Jersey, USA Posts: 15	Thanks for the replies. If you look a little further down in Brian256's article he mentions adding a second voltage divider circuit so that the potential can go from 8 to 4 volts using a 5k rheostat. I want to have the rheostats mounted in an unused face plate so that I can adjust the fan speed from the front of my case. My first thought was the these were audio taper not linear, but after repeated measuring I'm pretty sure they just suck. Not to mention what are the chances of buying 3 linear taper rheostats and having all three actually be audio taper. I took the bottom cover off of one of the rheostats and though I can't describe it they look poorly made to me. I'll try finding some better ones elsewhere. Brian256: Is it possible for 1 timer to drive 3 independant op-amps/transistor/mosfet combos? Or will the signal be too weak after splitting it 3 ways. Thanks again. __________________ AthlonXP 1600+ Soyo KT400 dragon ultra 512 megs Samsung DDR333 PowerColor 9700 Antec SOHO case other junk

03-27-2003, 04:12 PM	#9
ChemMan Cooling Neophyte Join Date: Jan 2003 Location: New Jersey, USA Posts: 15	I can always use another transistor to amplify the signal if it doesn't work. __________________ AthlonXP 1600+ Soyo KT400 dragon ultra 512 megs Samsung DDR333 PowerColor 9700 Antec SOHO case other junk

03-26-2003, 02:50 AM	#2
mrmoustache Cooling Neophyte Join Date: Oct 2002 Location: columbus, ohio Posts: 44	i don't really know much about electronics but i'm pretty sure they make rheostats that turn multiple times. my guess is you can get them at most online electronics places like jameco or bgmirco if you know what to look for but i don't know if they have a special name or anything.

03-26-2003, 09:05 AM	#5
murray13 Cooling Savant Join Date: Sep 2002 Location: Omaha, NE USA Posts: 216	Sounds to me like Radio shlock screwed up and you got audio taper (log) pots not linear.

03-26-2003, 12:48 PM	#6
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	I believe you got some crappy pots. Sorry dude! What you want is multi-turn 10K rheostats. I prefer rheostats to pots for their size: pots tend to be small and thus a pain to adjust unless you love your miniature jewelers screwdrivers. Also, some of what you see could be caused by dirty contacts. However that is probably not the case in your situation. If you scavenged an older rheostat, it could have dirt+grease buildup causing steady resistance that drops suddenly when the contact passes over the grease and then gets good (low resistance) contact.

03-26-2003, 05:10 PM	#8
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	I believe it will work, because the impedance (another name for the resistance to electricity) of the op-amp inputs is fairly high. High resistance means low flow, meaning that the op-amps won't suck down the voltage level seen on the high-side of the capacitor. So, having three op-amps absorbing some the output of the triangle wave should work OK.

03-28-2003, 02:59 PM	#10
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	Exactly. You could use a transistor as a buffer to reduce the impedance seen by the sawtooth waveform. Realistically, though, you are probably not going to see any problems. Normal low-current opamps are high impedance on their inputs. I wouldn't want to connect an op-amp to a resonating crystal circuit, but this should work just fine.

03-30-2003, 09:32 PM	#12
ChemMan Cooling Neophyte Join Date: Jan 2003 Location: New Jersey, USA Posts: 15	And one more thing before I forget. In Brian256's pwm circuit, if I wanted to add a capacitor to the circuit so that the fans are given full power at start-up(I believe its called power-on-reset), what would be the best place. I was thinking about placing it between Vcc and the MOSFET. I haven't had a chance to try it yet, so I'll ask now. Thanks for all the input. __________________ AthlonXP 1600+ Soyo KT400 dragon ultra 512 megs Samsung DDR333 PowerColor 9700 Antec SOHO case other junk

03-31-2003, 02:23 PM	#13
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	I didn't mean to send you off on a red herring with the comment about the crystal. Basically, I was referencing the "chip" used by computers (and many other electronic devices) to keep time. A crystal resonates at a given frequency based upon how it is cut, but it depends upon a precise pair of capacitors on each side to keep the electronic ticks happening. So, adding anything to a crystal circuit can make it not work. Not that it matters to what you are doing! Now, as for adding a capacitor to make it start up... that really shouldn't be necessary unless you are really downrating the fan speeds (i.e. less than 30% duty cycle). My circuit worked well down to about 30% speed on my fans, which was far below the (to my ears) silence threshold. Good luck!

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)