What does a tach pulse look like?

[Gooserider]

I recently aquired a rotary type flowmeter switch, as described here: http://sales.goldmine-elec.com/prodinfo.asp?prodid=7052
I got it on sale for 9.95, but that sale is over. The description is not completely accurate, see this thread:
here for more details.

It has a rotor that feeds into a circuit board containing a relay that trips if the flow drops below a certain (user adjustable) minimum.

Just for coolness sake, I would like to be able to pick up a reading of what kind of flow I am getting when the relay isn't tripped. It needn't have a relation to any 'real' flow value, just a value to say how fast the rotor is spinning.

I am sure there must be a way to get a 'fan tach' type signal like that put out by a 3 wire fan which I could feed into a DigiDoc or otherwise unused mobo fan connector to tell me what it's doing. It might even be very easy, the circuit board has an undocumented, unused 2 pin connector sitting near the other outputs.

But before I go poking around with a scope, it would be nice to know what I'm looking for... I've seen something about the fan tach wire producing an intermittent connection to ground - is this right?

If so, does that mean the thing reading it is supplying a voltage? If so, what voltage and current?

Inquiring minds want to know...

Thanks,

Gooserider

[cyr]

The tach output is an open collector type output (pulled to ground or left floating).

With a pullup resistor to +5V or +12V you should see the pulses on a scope.

[Gooserider]

Appreciate the info...

As I reccollect the scope hookups, (It's been a few years)
then you're saying I should use something about like this?
I'll get a square wave pulse, correct?

(UGLY ASCII ART ALERT!!!)

..+12vdc----/\/~500Kohm res./\/\---|__probe____ _Test point_
----------- scope lead ------------|


-----------------------ground lead from scope--------

(END ASCII ART)

Gooserider

[Turbokeu]

I don't want to dissapoint you but I had a look at the 0100B110 datasheet on Proteus Industries's website:

1) It's a flow switch with user settable trip point, not a flowmeter, so no way to do RPM monitoring.
2) Output is a 120VAC-3A max SPDT relay contact.
3) Pipe connections are only 1/4" NPT, pressure drop is 6psi!
4) Flow speed is indicated visually by rotor speed.

There is another possibilty if you have some serious electronic skills:
Dismount the electronic circuitry to analyse the actual schematic.
I suspect the rotor to be a sort of a hall effect generator with magnets.
Create your own schematic using the hall generator's signal to output an open collector signal proportional to the flow speed and suitable for RPM monitoring.

CD

[Gooserider]

Turbokeu: I don't want to dissapoint you but I had a look at the 0100B110 datasheet on Proteus Industries's website: 1) It's a flow switch with user settable trip point, not a flowmeter, so no way to do RPM monitoring.

That is the official line. However, they do make a similiar model that DOES offer flow monitoring, as well as the switch. I know that it is very common as a cost saving measure to use the same circuit board for all models in a product line, and just not implement all the features.

The circuit board in this unit is almost fully populated, the only spaces left on it are for a resistor, a large transistor, and a relay (looks like they intended to have a second alarm for excess flow perhaps?)

There is also a two pin connector, which isn't documented, near the other connections to the unit. Makes one go Hmmmm....

2) Output is a 120VAC-3A max SPDT relay contact.

Looks about right. I figure I can use it to trip a larger relay that would trigger an alarm and/or shut off the PC and pump power supplies.

3) Pipe connections are only 1/4" NPT, pressure drop is 6psi!

True, but I don't think that will be a problem. I am going to be doing multiple parallel cooling loops and the loop this unit was going on won't need much flow, indeed it will help prevent the loop from getting excessive flow. My pump has a 14' head rating, so it should be able to handle it. I suspect that may also be a 'worst case' pressure loss value, the unit's block has 4 ports, two are used, and two are plugged, but which ones depends on the desired range of flow rates. I suspect the higher ranges will be less restrictive.

4) Flow speed is indicated visually by rotor speed.

True, but the only way the rotor can be sending info to trip the relay is if it's sending some sort of pulse to the circuit board. I have difficulty believing there wouldn't be some way to tap into that pulse and get a tach signal from it.

There is another possibilty if you have some serious electronic skills: Dismount the electronic circuitry to analyse the actual schematic. I suspect the rotor to be a sort of a hall effect generator with magnets. Create your own schematic using the hall generator's signal to output an open collector signal proportional to the flow speed and suitable for RPM monitoring.

Well, I want both functions, so I want to keep this board around. So far this is what my investigations have found:

The rotor is magnetic, there are magnets embedded in 3 of the 6 arms.

The block appears to be stainless, with the rotor chamber machined into it. The only openings in the champer are the front faceplate, and the 4 pipe entrances. The back side has a well machined into it that just misses the chamber, and contains the sensor.

The sensor appears to be an inductor, it's about 3/16" (5mm) dia. by about 3/4" (19mm) long, positioned such that the tips of the rotors probably sweep by it. It is the only component on a sensor board, which is secured in the well by two nylon screws. There are two wires from the sensor board to the main module, which can optionally be located a good way from the block.

The main board has only 3 active components on it, along with a small handful of diodes, caps, resistors, etc. One component is a Motorola 7812CT regulator. (marked VR1) There is also a 3904 transistor, and what appears to be the 'brains' of the unit; an 8 pin chip, I'm not sure of the make, labeled CA3140EX.

I haven't chased down the function of this chip yet, but I suspect it functions to integrate the pulses put out by the sensor and control the relay triggering.

Gooserider

[Turbokeu]

Originally posted by Gooserider
That is the official line. However, they do make a similiar model that DOES offer flow monitoring, as well as the switch. I know that it is very common as a cost saving measure to use the same circuit board for all models in a product line, and just not implement all the features.

You're right...

True, but the only way the rotor can be sending info to trip the relay is if it's sending some sort of pulse to the circuit board. I have difficulty believing there wouldn't be some way to tap into that pulse and get a tach signal from it.

Well, I want both functions, so I want to keep this board around. So far this is what my investigations have found:

The rotor is magnetic, there are magnets embedded in 3 of the 6 arms.

The sensor appears to be an inductor, it's about 3/16" (5mm) dia. by about 3/4" (19mm) long, positioned such that the tips of the rotors probably sweep by it. It is the only component on a sensor board, which is secured in the well by two nylon screws. There are two wires from the sensor board to the main module, which can optionally be located a good way from the block.

The main board has only 3 active components on it, along with a small handful of diodes, caps, resistors, etc. One component is a Motorola 7812CT regulator. (marked VR1) There is also a 3904 transistor, and what appears to be the 'brains' of the unit; an 8 pin chip, I'm not sure of the make, labeled CA3140EX.

I haven't chased down the function of this chip yet, but I suspect it functions to integrate the pulses put out by the sensor and control the relay triggering.

Gooserider

So it seems that RPM is done purely magnetical, not by hall effect.

The 7812CT is a TO220 12V-1A max voltage regulator.

The 2N3904 is a small signal NPN transistor, probably for buffering the opamp output signal and triggering the relay.

The CA3140 is a mosfet opamp and probably works as a comparator between a fixed (user settable) DC level and the integrated pulses from the sensor.

Best thing to do is to try to reconstitute the schematic following all the traces and components on the circuitry and drawing it down on a piece of paper, and then try to understand it (with the help of the datasheets of the components).

I suspect a usable RPM signal (AC pulses from the sensor) to be available just before the opamp, and before the integrator (which is probably just a (rectifier) diode, a serial resistor and a parallel cap).
All you should have to do with the AC pulses is rectify them with a diode, put them through a schmitt trigger to obtain clean square wave pulses, and then feed them to an open collector transistor.

Ofcourse a scope (even // port based or LCD based) would be usefull.

CD

[Gooserider]

So it seems that RPM is done purely magnetical, not by hall effect.

I agree, it seems like it might be easier that way.

The 7812CT is a TO220 12V-1A max voltage regulator. The 2N3904 is a small signal NPN transistor, probably for buffering the opamp output signal and triggering the relay.

That's what I thought, it's been a while since I did anything at the component level, but those parts looked real familiar

The CA3140 is a mosfet opamp and probably works as a comparator between a fixed (user settable) DC level and the integrated pulses from the sensor.

That was the part I didn't recognize, though that is about what I was guessing for a function - something had to be doing it and I didn't see any other suspects! Thanks for confirming my suspicion.

Best thing to do is to try to reconstitute the schematic following all the traces and components on the circuitry and drawing it down on a piece of paper, and then try to understand it (with the help of the datasheets of the components). I suspect a usable RPM signal (AC pulses from the sensor) to be available just before the opamp, and before the integrator (which is probably just a (rectifier) diode, a serial resistor and a parallel cap). All you should have to do with the AC pulses is rectify them with a diode, put them through a schmitt trigger to obtain clean square wave pulses, and then feed them to an open collector transistor.

Agreed, that is what I'm planning when I have a chance to sit down and play with it. I just hope the board isn't multi-layer. It doesn't look it, but when I hold it up to a light, I see shadows that I can't account for.

Ofcourse a scope (even // port based or LCD based) would be usefull.

Agreed. I have a cheapo single trace, no real calibration, low frequency scope that I got when I was in ET school, I just need to get it out of storage.

Gooserider