Designing a PWM to Analog mini board for fans

Egon.Net

@zapta said in Designing a PWM to Analog mini board for fans:

@Egon-Net, your analysis of the circuit is correct. A couple of notes:

The only rail to rail requirement in this design is the output of the first amplifier. It's easy to achieve that with a comparator since they don't need to be linear and can just saturate the output transistor.

For the off condition of the power transistor, instead of a diode you can use a resistor between base and emitter such that VBE is below the cutoff voltage when the op amp is at its lowest output value.

Stability and dynamic behavior is above my pay grade. When something oscillates I just stick capacitors. (I am not an EE).

I don't think we can about the magnitude of the input signal, just about its duty cycle and the comparator preserves that.

I would also add a resistor in parallel to the fan such that the circuit provides the correct output voltage also without load.

What simulation software do you use? Looks nicer than the freebies I could find.

My notes to your notes

The opamp I was going to use (TLV9352) is even more rail to rail (50mv from rail at 10K IIRC) than that comparator, so I can go with the comparator 1st stage without adding a second IC
I don't think so. I will simulate it, but I think it will only force the opamp to source more current, and not improve the off state of the BJT, but I may be wrong. A diode forces 0.6V voltage drop irregardless of the current. Any suggested resistor value to try?
I am myself an EE, but not working as it, so I'm quite rusty. But I've been doing a lot of DIY audio stuff design, and I'm very used to opamps oscillating because of the feedback loop/load capacitance/whatever. Analog electronics is "fun" XD
Yes, I agree losing magnitude is a good deal for going rail to rail. U can always chose a external supply to scale to your heart's desire.
I agree. I already thought of that, and it will be present in the final version.
I use National Instruments' Multisim since my early years as a Comm/EE student

As soon as I can check stability is not a problem, I will design the final schematic and go to the PCB phase

marcelorider

I'm also in, I've used a small normal rectifier I've found on the internet with a capacitor and whatnot, but I'm still interested.
Portugal based here btw.

Also one note on the fans, they work from 8v to 24v most of them so, would be nice to develop the board to have that range?

Regards!

zapta

@Egon-Net,

1 The TLV9352 looks good indeed.

2 Since the TLV9352 can go to 50mv, a VBE resistor may not be needed. If needed, you can make it, let's say, 1/2 of the output resistor, this way once the op amp output gets to 3x0.7V, the current through the VBE resistor will be fixed and any current increase will go to the base.

3 My formal EE education is from a vocational high school so not that deep. I watched though some stability videos on youtube recently and start to grasp the concept. https://www.youtube.com/watch?v=-CS5th2zKWY

6 I tried multisim two days ago, it was very easy to use but I didn't see much of a part library, maybe because I used the free version. I then installed LTSPICE which is more rough but very popular.

Egon.Net

Hi again!

I got the Tina-TI model of TLV9352 opamp and simulated the circuit as faithfully as I could. Depending on the input, I can get it to oscillate, but a capacitor in the feedback loop seems to get rid of the oscillation. Now, worst case just rings (when I apply a fast rising clock direct to opamp input, not real use), so I'll put the capacitor in the PCB and test with and without the capacitor.

Using the opamp model, there's no need for a VBE resistor (you where right @zapta) and max voltage is 23.82V (which is a beliebable output). Tina seems to simulate better than Multisim.

I just ordered the components for a 10 boards test batch (since I needed to order some components for other project). I'll order a PCB batch too, and report, but will take 15 day the less for the PCB to arrive.

I think is time to count how many board you all will need:

PWM signal support:

Open drain active low (Duet), 20k impedance
VCC level PWM (0.05·VCC to VCC with 10k impedance or 0.1·VCC to VCC 20k impedance)

Output

Scaled %PWM from VCC
VCC range: 4.5V - 40V
Current depends on power transistor, planned BD139

Are you interested in single PWM board or dual(more?) PWM board?

Egon.Net

Now that I read my own post... When I talk about VCC level PWM, 0.05·VCC, I mean that "high" level can be as low as 0.05xVCC, thats 1.2V for 24V VCC, and as high as VCC itself. Low level shoud be 0V.

ZipZap

@Egon-Net said in Designing a PWM to Analog mini board for fans:

Output

Current depends on power transistor, planned BD139

Does that mean, the user can decide on his own power transistor/fet (NPN) as long as you don't exceed the maximum Current suitable for the PCB-traces (or maybe even strengthen them with solder if necessary)?

Maybe a single board may be sufficent for most users. Is it possible, do design a dual board, that can be "cut" in two indipendent halves by the user?
If thats not possible i think it would be good to have dual boards (only).

Im willing to be a beta-tester for the boards (having a watercooled toolhead with oversized loud 120mm 3-wire-fans on the radiator).
I would also be interested in up to four (4) dual boards or five to six (6) single boards of the final version.

/Julien

Egon.Net

@ZipZap said in Designing a PWM to Analog mini board for fans:

@Egon-Net said in Designing a PWM to Analog mini board for fans:

Output

Current depends on power transistor, planned BD139

Does that mean, the user can decide on his own power transistor/fet (NPN) as long as you don't exceed the maximum Current suitable for the PCB-traces (or maybe even strengthen them with solder if necessary)?

Maybe a single board may be sufficent for most users. Is it possible, do design a dual board, that can be "cut" in two indipendent halves by the user?
If thats not possible i think it would be good to have dual boards (only).

Im willing to be a beta-tester for the boards (having a watercooled toolhead with oversized loud 120mm 3-wire-fans on the radiator).
I would also be interested in up to four (4) dual boards or five to six (6) single boards of the final version.

/Julien

Well, the board is not designed for high currents, but the transistor will limit the current well before the pcb traces would. BD139 IIRC has a thermal resistance of 100ºC/W, and a maxtemp of 150ºC, so it would be ok for currents about... 200mA at 24V with light heatsinking (by heart). And the limit is about disipated power, so if you can lower VCC, you would be able to use higher currents. And sice the transistor footprint is just 3 holes, you could always try different transistors if you want

zapta

@Egon-Net said in Designing a PWM to Analog mini board for fans:

I'll order a PCB batch too, and report,

Any thoughts regarding the board dimensions, mounting arrangement, choice of connectors and pin assignment, indicator LEDS, etc?

KevinSant

@Egon-Net im intrested also in your custoim changer as well

Egon.Net

@zapta said in Designing a PWM to Analog mini board for fans:

@Egon-Net said in Designing a PWM to Analog mini board for fans:

I'll order a PCB batch too, and report,

Any thoughts regarding the board dimensions, mounting arrangement, choice of connectors and pin assignment, indicator LEDS, etc?

This is what I thought:

Board dimensions: as small as I manage to do it
Mounting arrangment: I though about for 3.5mm holes, one in each corner
Connectors: open to suggestions, but I thought screw or pin
Leds: I'd prefer not to include any given the VCC variability
Arrangement of connectors: VCC/GND/PWM in one side GND/Vanalog in the opposite side.

Any suggestion?

Egon.Net

@bks31 said in Designing a PWM to Analog mini board for fans:

@Egon-Net im intrested also in your custoim changer as well

Do you mean my tool changer printer?

KevinSant

@bks31 said in Designing a PWM to Analog mini board for fans:

@Egon-Net im intrested also in your custoim changer as well

yes wasnt sure if you were making like a attachment for the carraige

Egon.Net

@bks31 said in Designing a PWM to Analog mini board for fans:

@bks31 said in Designing a PWM to Analog mini board for fans:
@Egon-Net im intrested also in your custoim changer as well
yes wasnt sure if you were making like a attachment for the carraige

It's a custom motion system, 300mmx300mmx300mm CoreXY, with e3d toolchanger plates...

zapta

@Egon-Net said in Designing a PWM to Analog mini board for fans:

Any suggestion?

All sounds reasonable to me. As for the LEDS, since they are for indication only, you can easily have a 1:4 voltage range (e.g. 2 to 8ma) without going out of spec.

They are nice for diagnostics, but not absolutely required IMO.

marcelorider

Well, I have one fan that is pwm regulated so, at least for now I’m in to one PCB also
My printer is a Longer LK4, but if that’s not a concern then perfect.

Do you think is easy to implement something like a 8v-24v range?

Regards

Egon.Net

@marcelorider said in Designing a PWM to Analog mini board for fans:

Well, I have one fan that is pwm regulated so, at least for now I’m in to one PCB also
My printer is a Longer LK4, but if that’s not a concern then perfect.

Do you think is easy to implement something like a 8v-24v range?

Regards

You can limit the PWM in M106 gcode so if it's less than a value, a minimum value is set (8V->33% equivalent for your case). The bad side is that from 0% to 33%, it would output the same 8V.

I'm afraid there's not an easy way to limit the lower side as you'd need.

dc42

Just a reminder that if you are not using a fan output on a DueX2 or DueX5 board, you can increase the PWM frequency to 65535 and then use a simple LC filter, i.e. capacitor in parallel with the fan, and an inductor between the Duet FAN- output and the negative fan wire + capacitor. Suitable component values would be 1mH and 1uF. For the inductor, part # 13R105C is widely available and rated up to 330mA, which should be sufficient for most fans used in 3D printers.

If using a heater output on a Duet 2 to drive a fan with LC filter, you also need to add a flyback diode across the output terminals.

DueX boards have a fixed PWM frequency, so this approach would need a much larger inductor, making is impractical unless you have some old transformers lying around.

Egon.Net

@dc42 said in Designing a PWM to Analog mini board for fans:

Just a reminder that if you are not using a fan output on a DueX2 or DueX5 board, you can increase the PWM frequency to 65535 and then use a simple LC filter, i.e. capacitor in parallel with the fan, and an inductor between the Duet FAN- output and the negative fan wire + capacitor. Suitable component values would be 1mH and 1uF. For the inductor, part # 13R105C is widely available and rated up to 330mA, which should be sufficient for most fans used in 3D printers.

If using a heater output on a Duet 2 to drive a fan with LC filter, you also need to add a flyback diode across the output terminals.

DueX boards have a fixed PWM frequency, so this approach would need a much larger inductor, making is impractical unless you have some old transformers lying around.

That's the very reason I'm doing all of this: I'm driving 6 fans from a Duex 5 board, and I already calculated that the inductor for a LC filter would be huge, and I'd need 6 of them. That's why I went with this route. But thanks for the advice!

marcelorider

@Egon-Net
Thanks for the reply.
Honestly I don’t use it below 50% so I don’t think that would be a problem.
Although as of now the fan doesn’t even spin correctly with any pwm state and just rotates above 50% but so slow, not controllable by any means.

fcwilt

@Egon-Net said in Designing a PWM to Analog mini board for fans:

That's the very reason I'm doing all of this: I'm driving 6 fans from a Duex 5 board, and I already calculated that the inductor for a LC filter would be huge, and I'd need 6 of them. That's why I went with this route. But thanks for the advice!

Have you determined what it is about your fans that cause them to not work with PWM?

I have dozens of different kinds of brushless fans from many different vendors and they all work fine with PWM.

Thanks.

Frederick