Dew Heater and Sensor in same cable

[MarkyD]

Maybe one for the Arduino/Programming build peeps?

My Dew Heater project has the Heater Power and the DS18B20 Temperature sensor cables (approx 2M) running from the same 5 pin plug/socket and enclosed in the same cable.  After testing with the live equipment last night as soon as the PWM starts for the heaters the Temperature sensor gets corrupted.  The cable I initially used was defence standard unshielded multi core.  I knocked up a cable with some UTP Cat5e I had and the corruption is not as bad ie the readings recover and then get corrupted.  With no PWM applied the readings are good and if I disconnect the power so just the sensor is using the cable all is good, so the length is not a problem

My question to all is - has anyone used a shielded cable with success, if so which one - STP Cat5e or some other Data power combined cable?  I need 5 cores or 4 with shield(use as sensor -ve) as I cannot common the grounds as the PWM is switched with N-Type MOSFETS.  I have a max diameter of approx  7-8mm to enter the plug restraint

TIA

--

Mark

[Dr_Ju_ju]

PWM switching is very 'noisy' due to the nature of the output signal, i.e. square-waves, therefore the only cable that may be suitable would be multicore with individual pairs having separate screening, the best would be those used to make microphone connection boxes.

[Gina]

You could try an LC filter as close to the drain of the MOSFET as possible.  That's what I do to cut down RFI.

[MarkyD]

I have never used or looked into lc filters tbh so I wouldn't know where to start

I have found this cable from rapid online which you can buy by the meter so I'm going to try this to see if it improves.  If not I may be picking your brains Gina

http://m.rapidonline.com/cables-connectors/kabeltronik-930402200-2x2x0-22mm-frnc-triple-shielded-pimf-49-0401

[MarkyD]

Quick Question - Would a RC low pass filter to smooth out the PWM signal work on the gate of the MOSFET or would this effect the firing of the MOSFET.  I'm using logic MOSFETS so they are fully on with 5V but if I smooth out the PWM if running at 50% this would give roughly 2.5V at the Gate so the MOSFET wouldn't fire - Am I correct in this thinking?

--
Mark

[Gina]

Pretty much correct.  But you need the filter on the output.  If you slow down the drive pulse edge it will have the effect of turning the MOSFET on slowly and produce heating.  The whole idea of using PWM is to switch the MOSFET so fast that it doesn't have time to heat up during the transition.  This means that the MOSFET is either OFF or ON.  In the OFF state no heat is produced and in the ON state the heating is due to the load current in the ON resistance (R_DSon) of the MOSFET.  HTH

[MarkyD]

Thanks for the clarification

After a week away in the Cotswolds (I took the scope but only had one good night) I have put my el cheepo ossiliscope on the PWM output of the MOSFET.  Below is what I am seeing.  TBH it doesn't mean alot to me as I have never worked with scopes so I'm not sure what I am looking at but I take it the switch on is fairly clean it's the switch off that looks dirty?  Below are screenshots of 0%, 20% and 50% PWM across the MOSFET Source

I have one got one of these which I am designing into the next PCB project (I'm on #4 atm) as a shield for the Mega.  I'm going to have the MOSFETS and Filters as a daughter board to keep it away from the Mega circuit just in case that gets disrupted by the 12V PWM

http://uk.rs-online.com/web/p/power-line-filters/0209083/?searchTerm=209083&relevancy-data=636F3D3126696E3D4931384E525353746F636B4E756D6265724D504E266C753D656E266D6D3D6D61746368616C6C26706D3D5E5C647B367D247C5E5C647B377D247C5E5C647B31307D2426706F3D313426736E3D592673743D52535F53544F434B5F4E554D4245522677633D4E4F4E45267573743D323039303833267374613D3032303930383326

The fully screened cable I have ordered is on back order for 3 - 5 days but belt and braces etc

--

Mark