Hello,

I'm building a replacement heater unit for my car, I'm controlling the bypass valve (controlling the amount of heat going through the radiator vs. bypassing the radiator) electronically with a PWM unit.

I originally hoped to use an off the shelf PWM controller from eBay, based on this: http://www.solorb.com/elect/pwm/pwm2/pwm3.jpg but with added voltage regulator.

It's my understanding that the frequency of the output is controlled by the 100 K resistor between pins 1 and 13, and the 10 nF capacitor between 13 and 14.

The catch here, is i'm targeting a frequency of 0.5 Hz (ish), which will involve replacing the resistor with a 500 K unit and capacitor with 1 uF, however this doesn't seem to change the frequency of the output.

Is my thinking correct here? Any suggestions, am I simply trying to get the unit to work outside of what it's designed for? If so can anyone recommend a different approach to generate a low frequency variable duty PWM output.

Thanks, Tom

 

This will likely get closed as a violation of the terms of service - no automotive topics allowed.

If I'm wrong, maybe you can explain why you want to use low frequency PWM to control a valve. Does the valve mechanism support that? Are you hoping to achieve proportional (not just on and off) control?

 

This will likely get closed as a violation of the terms of service - no automotive topics allowed.

If I'm wrong, maybe you can explain why you want to use low frequency PWM to control a valve. Does the valve mechanism support that? Are you hoping to achieve proportional (not just on and off) control?

I assumed the automotive issue was related to modifications which would impact the performance of the car and thus lead to members of the forum, or forum itself becoming a focus of lawsuits. Here there is no link to the performance or safety of the car i'd ask that any moderators who still take issue with this will edit the posts to remove mention of a car rather than simply closing it.

The valve mechanism is a solenoid and originally intended to be operated by a PWM - measurements of the original controller showed approx. 0.4 Hz, I am trying to achieve proportional control (hence variable duty cycle PWM)

 

I assumed...

Yeah, good luck with that. But we'll see.

The valve mechanism is a solenoid and originally intended to be operated by a PWM.

OK, that's great.

Someone here can better understand that oscillator, but first off I'd recommend against the LM324 op-amp and choose the LM393 comparator for this job. I believe you'll find some PWM motor control circuits in the Experiments section of this site, and those may be better starting points than the circuit you linked. Just look around here a bit. If I find stuff in the meanwhile I'll post back.

 

The LM324 will be more than OK for the low frequency involved, but is not rated for the automotive temp range. An LM2902 should be used instead, which is basically the same thing as an LM324, but rated for the automotive temp range.

Replacing the 100k resistor with 470k to 510k and the 10nF cap with one rated 1uF that is non-polarized should give a frequency of ~1Hz. You could use a 0.47uF cap instead to get ~2Hz. It must be non-polarized. If you must, you could use a pair of 1uf aluminum electrolytic caps back-to-back with 4.7MEG resistors in parallel with them to provide a current path to the junction between the caps.

 

The LM324 will be more than OK for the low frequency involved, but is not rated for the automotive temp range. An LM2902 should be used instead, which is basically the same thing as an LM324, but rated for the automotive temp range.

Replacing the 100k resistor with 470k to 510k and the 10nF cap with one rated 1uF that is non-polarized should give a frequency of ~1Hz. You could use a 0.47uF cap instead to get ~2Hz. It must be non-polarized. If you must, you could use a pair of 1uf aluminum electrolytic caps back-to-back with 4.7MEG resistors in parallel with them to provide a current path to the junction between the caps.

Thanks, I think I was going wrong using a polarised capacitor - looks like it's time to go shopping!

When you say automotive temp range, I take it you mean engine bay temperatures? In this application the controller is isolated from any temperature sources and should stay approximately ambient. Can I get away with the LM324 still?

 

You can buy an LM2902, automotive temp range equivalent of the LM324, for just about a dime apiece.
You can buy an LM124, a mil-spec version of the LM324 for under a quarter. Look on http://www.findchips.com
Why not simply use the right part for the job, instead of trying to cut corners for no reason?
You have to order non-polarized caps that are the automotive temp range anyway, so buying more stuff in the same order will spread your shipping costs around, decrease your net per-part cost.

When I make recommendations, it's for good reasons.

 

I'm trying to cut one big corner rather than a small one - I can buy pre-assembled boards using a LM324 chip (which is what i'm using now) and just need to change the resistor and capacitor. As the issue with my heater is universal across all of this type of car, I want to make it as easy as possible for others to do the same if they want to, so having a pre-manufactured part would be simpler. However I do plan to do some costing and work out if I can produce the same part at a similar cost (and am happy to consider using the LM2902)

I've just tested out the modified circuit with a 510 K resistor and 1 uF polyprop. cap and I'm still not seeing a 1 Hz frequency (i've got it attached to a 12V LED to test, a 1 Hz cycle should be very visible), any thoughts?

 

...I can buy pre-assembled boards using a LM324 chip ...

If that works for you and Sarge buys in, I retract my suggestion to avoid the LM324 in favor of the LM393. (I was concerned with clean switching without chatter, not the frequency, which is almost DC). I didn't realize you were just trying to hack an existing device.

Do you know that your valve cannot just use the existing frequency without modification?

 

If that works for you and Sarge buys in, I retract my suggestion to avoid the LM324 in favor of the LM393. (I was concerned with clean switching without chatter, not the frequency, which is almost DC). I didn't realize you were just trying to hack an existing device.

Do you know that your valve cannot just use the existing frequency without modification?

I'd like to hack the existing device as a proof of concept. I've gone through a few different designs up to now that couldn't deliver the full duty cycle (I need 0 - 100 % duty cycle to provide full temperature control) or dedicated PWM IC units that couldn't go to a low enough frequency.

I've burnt out a few of the solenoids trying to run them faster, I could probably get them into the tens of hertz but wouldn't want to push them much rather than that. The radiator has a decent thermal capacity so even at very low frequencies you won't experience heat pulsing and the lower frequency should in theory extend the life of the solenoid, so i'm keen to get it as low as possible.

Any other approaches to controlling this are welcome, I'm happy to build something from the ground up if there's a good reason to be doing so.

 

You need to find out what's wrong with the circuit.

 

Solved it; the resistors between 3 and 1 and 3 and 14 were the same value, changing it back to the specs on the diagram fixed it.

Cheers guys!

 

And with that, I'll close the thread. I just spotted it.