Heater fan low speed protection w/ LM2907/LM2917

[jpanhalt]

...prefer the retriggerable monostable PWM option so that different fan speeds give different power levels from the heater. But I've never designed any kind of PWM circuit before.

That's a new topic, it seems. We were discussing a way to shut the fans/heater off when the fan speed decreased below a certain threshold. Now it seems you want PWM power to the heater and different fan speeds according to that power. Is that correct?

John

 

[large_ghostman]

that seems a tight schedule before xmas! not sure how much time you get to work on this but i would be careful with the time limit, remember the project time constant PROJECT time T =4x what you thought it would take

seriously tho how much time in a day/week will you have to build this? and what equipment do you have access to? i only ask this because with the pwm route a scope would be one of my must have's just in case you hit a snag!
having said all that i have no idea what kind of time it would need as i dont have analogue experience. sounds a cool project tho , i would love to take a car like that apart see how it ticks!
if you had a micro think of all the bling you could add to an electric car!!! talk about pimp my ride
i would never get a electric car personally because by the time i had finished adding all the gadgets and Leds and stuff it would never have enough power left to move lol

 

[jelliott]

Now it seems you want PWM power to the heater and different fan speeds according to that power. Is that correct?

Not exactly. I'm referring to what Eric suggested (which I subsequently referred to as Option B) in response to my original query--if the fan quits, power is removed from the heater 330 microseconds later (or 3.3 ms--I still don't see how that corresponds to 1500 rpm). But it occurred to me that the same circuit would serve to modulate power to the heater if I intentionally ran the fan at a lower speed. (Unless I'm completely misunderstanding Eric's suggestion.) The idea is to modulate heater power in response to fan speed, not the other way around.

 

[jelliott]

that seems a tight schedule before xmas! not sure how much time you get to work on this but i would be careful with the time limit, remember the project time constant PROJECT time T =4x what you thought it would take

Well, I'm a professional engineer, so I only underestimate the time required to complete a project by a factor of 3.

seriously tho how much time in a day/week will you have to build this? and what equipment do you have access to? i only ask this because with the pwm route a scope would be one of my must have's just in case you hit a snag!

In theory I spend 6 hr/day sleeping, 8 hr/day at my day job, and 1 hr/day commuting, so that should leave me 9 hr/day for this if I never eat or bathe, right?

if you had a micro think of all the bling you could add to an electric car!!! talk about pimp my ride

And that's exactly why I resist the urge to take the "easy" route and use a microcontroller for every enhancement I come up with for my brother's car--once that line has been crossed, the 'To Do' list would be endless, as everything could be monitored, automated, displayed, data-logged, etc.

 

[large_ghostman]

Tx3!! wow i only dream of such short over runs lol. get up an hour early then you have 10 hours
i have to admit the anologue route takes some skill! yeah i guess with a micro its a bit like forbidden fruit.
the heater being done by fan speed is at least original, full marks there.
on average how far does he get on a charge and how long from almost empty to full charge? when mum got her new car (vauxhal astra) they did consider the ampeira or whatever they call it, really neat car but just not practical where we live and expensive, hopefully as technology moves on the cost of these type of cars will come down.

 

[jelliott]

Since I still don't know where to start (e.g. what IC to buy) on Eric's Option B, I went ahead and tried to work out the 'Figure 20' Option A approach on the breadboard this morning. As simple as it seemed, it's not working, i.e. the load isn't energized when the fan is running. (R and C were selected to specify a frequency of 119 Hz, corresponding to a 3570 rpm cutoff speed; normal operation of the fan is 3750 rpm.)

Anyone see any obvious errors? The one thing I did notice is that the internal resistance of the LM2097's input is low enough that it cut my 24 V square wave down to ~12 V. And I'm not sure how much current can be safely dumped into the tach wire on this fan, so rather than going to a smaller pull-up resistor on the fan, I just ran the breadboard at 12 V instead of 24 V. (It still didn't energize the load.)

 

[alec_t]

Can you post the schematic of what you've breadboarded?

 

[jelliott]

Can you post the schematic of what you've breadboarded?

 

[alec_t]

From the datasheet:

Two options (LM2907-8, LM2917-8) have one input internally grounded so that an input signal must swing above and below ground and exceed the input thresholds to produce an output.

Does your 24V square wave (and the fan tacho) go negative?

 

[jelliott]

From the datasheet:
Does your 24V square wave (and the fan tacho) go negative?

It does not. And sure enough, the IC I bought is an LM2907N-8. Is there a different part number that I should buy to make it work?

(Or I could just shelve this concept entirely and work on the aforementioned "Option B" if you have any suggestions to point me in the right direction on that approach.)

 

[alec_t]

You could try capacitively coupling the tacho/square-wave to the IC like this:-

 

[jelliott]

You could try capacitively coupling the tacho/square-wave to the IC like this:-
View attachment 82392

That did the trick. Thanks!

 

[alec_t]

Good to know you've got it working

 

[jelliott]

Of course, now that I"ve got that working perfectly on the bench, my brother has decided that he really likes Eric's suggestion from 15 November....

Is a CD4047 IC the right place to start for that 'Option B' approach? Something like **broken link removed** ?

That said, I did notice that the LM2907 'Option A' circuit exhibits a transition region wherein the load is neither on nor off, but in fact pulsewidth-modulated. If I can tune the circuit to be in this transition region when the fan is running at its low speed setting, I could theoretically achieve the 'Option B' functionality with the same basic circuit from 'Option A,' if I swap some kind of semiconductor switching device for the relay that's in the current plan. Can anyone recommend a transistor that's good for switching 5 A at 48 V? (Related question--when I see something advertised as a "solid state relay," is that basically a MOSFET [or ???] packaged to look like a relay? In other words, can I buy a "solid state relay" and switch it hundreds of times a second, PWM-style?)