PWM to Analog converter circuit

smsmasters · 6 Feb 2014 at 09:05

Tealc said:
I've got a spare circuit all built up here. I will heatshrink and then do the test after that and send it to you. I will get it to the post office Friday afternoon so you should get it Saturday with a bit of luck.

I'm not sure what has gone wrong with that one. This is the first one I've had a problem with out in the field, as it were, but it seems to be not receiving PWM properly. As I said it was tested but admittedly that was done before I fitted the heatshrink so maybe the extra pressure of the heatshrink has opened a connection somewhere.

I will do the testing before after heatshrink in future to be sure.

OK many thanks.

Shall I return the cable back to you?

Tealc · 6 Feb 2014 at 09:27

If you don't mind that would be awesome as I could then see what's wrong with the thing and make sure that it never happens again.

Thanks mate.

smsmasters · 6 Feb 2014 at 09:42

Tealc said:
If you don't mind that would be awesome as I could then see what's wrong with the thing and make sure that it never happens again.

Thanks mate.

OK please email me your address.

Thanks

p.s. should the PWM cable still function no matter which fans I use or are you adjusting the circuit specifically for my fans?

Tealc · 7 Feb 2014 at 10:26

smsmasters said:
p.s. should the PWM cable still function no matter which fans I use or are you adjusting the circuit specifically for my fans?

It is designed to work with a massive range of fans. The version I sent you should cope with fans ranging from 0.05A up to 1A, or combinations of 1-20 fans within that current range.

The potentiometer acts like a valve and allows more, or less energy into the fan output circuit. It is meant to be adjusted to suit an individual need. Originally I used resistors and a switch to select 3 settings, low med high but I feel the potentiometer is a much more elegant solution and allows any user to swap and change cooling fans without needing to change the circuit. In some respects I've harmed my repeat business, but meh.

The switches were expensive (comparatively so) and a pain to solder up neatly.

There are two main possibilities here.

1) The circuit has a fault and the PWM signal is not passing through the circuit.

2) The PWM signal from your motherboard is not actually changing when you adjust it in Speedfan and it is just sending 100% all the time.

I should be able to determine which of these two it is when I receive the board today/tomorrow.

Thanks for sending it back mate.

smsmasters · 7 Feb 2014 at 13:51

Tealc said:
It is designed to work with a massive range of fans. The version I sent you should cope with fans ranging from 0.05A up to 1A, or combinations of 1-20 fans within that current range.

The potentiometer acts like a valve and allows more, or less energy into the fan output circuit. It is meant to be adjusted to suit an individual need. Originally I used resistors and a switch to select 3 settings, low med high but I feel the potentiometer is a much more elegant solution and allows any user to swap and change cooling fans without needing to change the circuit. In some respects I've harmed my repeat business, but meh. The switches were expensive (comparatively so) and a pain to solder up neatly.

There are two main possibilities here.

1) The circuit has a fault and the PWM signal is not passing through the circuit.

2) The PWM signal from your motherboard is not actually changing when you adjust it in Speedfan and it is just sending 100% all the time.

I should be able to determine which of these two it is when I receive the board today/tomorrow.

Thanks for sending it back mate.

Did you receive it yet mate?

Thanks

Tealc · 7 Feb 2014 at 13:55

Yes mate, thanks. I will give it a test in a moment and update with results.

I've also posted the replacement off to you this afternoon.

Tealc · 7 Feb 2014 at 14:55

Right. Tested.

I don't think this broken ground wire helps much.

I also noted that a resistor was very close to the power rail so it may have been pushed down and shorted after heatshrinking, thereby popping the PNP. Once the PNP pops it goes short and you lose control. As a side effect when the PNP pops all PWM is cut and no current flows, protecting your PWM circuit and also running your fans at max. Always safety in mind you see

.

For the broken wire:-
I will need to devise some sort of strain relief for these wires. I was at one time potting them with hot glue but since using adhesive shrink I haven't thought I needed to. I will either go with a cable tie securing the wires in a loop so any force acts upon the cable tie, or use a blob of hot glue over the opening where they come out. I could also go up a wire size to 22AWG which will make the connection stronger.

For the resistor:-
I will need to make sure I don't cut the track too deep and that will allow me to use a SMD resistor, which won't have any risk of shorting. I will need to use a smaller drill to cut tracks, or devise a different method of track cutting.

I'd expect that with the PNP replaced the circuit would work just fine, once I replace that resistor.

smsmasters · 7 Feb 2014 at 17:34

Thanks for the test results.

Does the PWM cable you sent have the modifications you suggested?

Tealc · 7 Feb 2014 at 17:41

It uses the SMD resistor but has no strain relief on the wiring. They'll normally be fine unless you repeatedly pull on the wires. It was sent before I had yours back.

smsmasters · 8 Feb 2014 at 11:24

Hi mate received the new PWM cable and it works!

Can you please kindly explain the effects of turning the screw with regards to PWM control in speedfan.

So let's say it takes 10% PWM duty cycle for the fans to spin down to 1000 rpm, would adjusting the screw make the fans spin down to 500 rpm at 10%?

Testing with my fans:
10% = ~800 rpm
15% = ~1050 rpm
20% = ~1250 rpm
25% = ~1350 rpm
30% = ~1400 rpm
.....

Tealc · 8 Feb 2014 at 11:43

What I usually do is set the PWM at 20% or 30% and adjust until the fan speed is around 600rpm, or maybe higher if you run fast fans (greater than say 1500). Then it should ramp up speeds on a more or less linear fashion until it maxes out at 80-100% PWM.

You can also set max PWM and adjust until the fan just about reach their max speed. What'll happen is that the speed will increase gradually until it sort of hits a wall, this is the ideal point.

Remember that the fan closest to the cabling is the one that sends rpm. The others are just passengers.

You can choose to have your fans run a little faster at idle and have the max speed reached quicker if you choose. You can also have the fans never reach max speed by dialling down.

Look back to the first few pages of the thread and you'll see some graphs of fan speeds curves to get an idea of how adjustment affects the fan speed.

I may do a video on setting up eventually once I get my workbench set up.

pob9 · 8 Feb 2014 at 13:28

been working away from home so only just got a chance to have a go with mine. set up fans idle at 1000rpm idle and ramp right up to full when the load kicks in. does exactly what it says in the title. now i can browse the net in quiet mode xD. many thanks for sorting it, now just a bit of fine tuning to be done, then hide it away

Phunky · 8 Feb 2014 at 22:28

You should kickstart this to get this out on mass, shocks me that this does not exist. Drives me nuts finding out that a mobo doesn't have voltage control for my NH-D14.

smsmasters · 8 Feb 2014 at 23:36

Hi mate,

Would it be possible to make a cable that allows PWM control for a D5 pump? It gets power from a single molex and has a 3 pin tach wire.

pob9 · 8 Feb 2014 at 23:47

Phunky said:
You should kickstart this to get this out on mass, shocks me that this does not exist. Drives me nuts finding out that a mobo doesn't have voltage control for my NH-D14.

+1 ocuk should be stocking these. i have 2 case fans on atm to test then my cpu fans will join in

Tealc · 9 Feb 2014 at 07:27

smsmasters said:
Hi mate,

Would it be possible to make a cable that allows PWM control for a D5 pump? It gets power from a single molex and has a 3 pin tach wire.

Water cooling pumps are too power hungry for this particular design. I believe the D5 draws at least 2A at 12v. This circuit works up to 1A max. I also have no idea how slow the pumps are supposed to go, or how fast they go.

I have looked briefly into a high power version but I keep hitting the issue of the maximum voltage dropping to around 10v-10.5v at 2A.

I have also looked briefly at developing an IC based circuit that responds to PWM input and has a heavy duty FET on the output, that would drive a load of current but you'd lose tach output. You'd need to involve a micrcocontroller based solution to stretch the PWM or something to give a tach feedback.

Tealc · 9 Feb 2014 at 07:34

pob9 said:
+1 ocuk should be stocking these. i have 2 case fans on atm to test then my cpu fans will join in

You have to understand that this PWM converter is really quite simple, consisting of just 7 components, 5 of which are passives, a bit of wiring and some connectors. It takes time to put together of course.

A version that one would develop professionally would need to sense the current of the fans and determine it's own setting, rather than be adjusted by hand. Maybe if I were to develop a version with feedback and that level of control it would be worth considering. A professional board would be silkscreened so would not need the heatshrink.

tbyeah · 9 Feb 2014 at 23:26

Tealc said:
You have to understand that this PWM converter is really quite simple, consisting of just 7 components, 5 of which are passives, a bit of wiring and some connectors. It takes time to put together of course.

A version that one would develop professionally would need to sense the current of the fans and determine it's own setting, rather than be adjusted by hand. Maybe if I were to develop a version with feedback and that level of control it would be worth considering. A professional board would be silkscreened so would not need the heatshrink.

Could be worth the time, it's a useful tool.

Also, compacting it down a bit would be nice but hey you can't have your cake and eat it

Tealc · 10 Feb 2014 at 13:54

I can make it shorter by taking the fan connectors off board. That saves around 30mm length. Reducing the width is mostly pointless as the width of the heatsink governs the width of the device. I can remove it but it would get hot under load of several fans.

As for creating a smart device. That might be a nice project for the warmer months.

smsmasters · 12 Feb 2014 at 09:55

Hi mate, not sure what happened but I heard a "pop" and now the RPM reading has gone haywire.

Previously at 15% PWM duty cycle the RPM reading is ~800 RPM. Ever since that "pop" the RPM reading fluctuates between 1000 to 4800 RPM.

Setting it to ~40% PWM duty cycle resolves the issue. It's as if the minimum RPM has decreased for the given % of PWM duty cycle. So now 15% is now 40% for the same RPM.